Searching \ for '[EE] Analysis of Olin's EUSB2 Circuit (sub circuit' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/index.htm?key=analysis+olins+eusb2
Search entire site for: 'Analysis of Olin's EUSB2 Circuit (sub circuit'.

Exact match. Not showing close matches.
PICList Thread
'[EE] Analysis of Olin's EUSB2 Circuit (sub circuit'
2007\01\31@071826 by Lindy Mayfield

flavicon
face
As an exercise Olin kindly gave me a circuit to try to analyze.  Here is what I can find in sub circuit R2, D3, C2, Q1.

I am deliberately not reading any of the previous posts on this yet until I can give this a try myself.

First the component list:
R2 2.7k resistor
D3 Zener Diode 5.6v
C2 1uF 10v capicitor
Q1 NPN Transistor

Since I had hints but nothing too specific about how and where to start, I'll have to fall back on my programming knowledge and give this a shot. If it doesn't work, I'll learn something and try a different approach.

Inputs and Outputs

Assumption: The emitter on Q1 feeds VDD on IC2.  A small current on the base of Q1, which looks like me to come from the 8.2V line after passing through R2. I'm not sure how to apply V=IR here, but from what I see the 8.2 volts passing through R2 drops to 5.6 volts which is used in both the base of Q1 and the output of the 5.6 volt power line from this power supply module.   But how is 8.2 volts dropping to 5.5?  Is because some of it is going to "leak" through R2.  

Resistors limit current, correct? So passing 5V through a resistor should yield 5V on the other side? Or did I get this wrong.

So here is one point that I'm stuck on.  The data sheet for the Pic10F204 says 0 - +6.5 volts.  But what I'm seeing is VDD gets 8.2V.

The zener at D3 and C2 have me stumped.  I've already been through a lot of Mims' books and other electronics books to get this far.  I'll go to E&E next.   I love that book!

Any corrections in my logic, just to make sure I'm on the right path, would be greatly appreciated Olin.  

I cannot tell you how grateful I am that you're taking this bit of time to help me get over my learning curve.

Kind regards,
Lindy



2007\01\31@090812 by olin piclist

face picon face
Lindy Mayfield wrote:
> Assumption: The emitter on Q1 feeds VDD on IC2.

Yes.  The purpose of this circuit is to supply somewhat regulated power to
IC2, which is a PIC 10F204.

> A small current on the
> base of Q1, which looks like me to come from the 8.2V line after
> passing through R2. I'm not sure how to apply V=IR here, but from what
> I see the 8.2 volts passing through R2 drops to

You are still sending paragraphs as single long lines, which are getting
truncated somewhere.  I already mentioned this once before, and I'm not
going to mention it again.  In the future I will likely just ignore your
posts until you set up your email client properly.

And to everyone else, let's not have this discussion of email standards
again.  Whether you agree or disagree it should be this way or what the
standard says is irrelevant.  I consider long lines silly and they are
certainly a hassle for me.  Whether anyone else likes it or not, I can
enforce sensible email practises when others want something from me.
Remember the golden rule: "He who has the gold makes the rules".

> Resistors limit current, correct? So passing 5V through a resistor
> should yield 5V on the other side? Or did I get this wrong.

You have some misconceptions here.  There is no such concept as voltage
passing thru something.  Voltage is like pressure applied accross something
or between two points.  Current flows thru things.

And resistors do limit current compared to a wire in the right
circumstances, but that way of thinking about them will cause confusion.
Think of resistors as allowing current to flow thru them proportional to the
voltage applied accross them, and conversely and equally important, they
cause a voltage to build up accross them proportional to the current flowing
thru them.  The voltage and current are directly proportional to each other,
and the purpose of different value resistors is to provide different
proportionality constants.  Ohms is the proportionality constant for the
second case (voltage accross builds up proportional to current thru, Volts =
Amps x Ohms).  That means that 1/Ohms is the proportionality constant for
the first case (current thru is proportional to voltage applied accross,
Amps = Volts x (1/Ohms)).

So let's say the voltage at the top end of R2 is 8.2V, and a the bottom end
5.6V.  What is the voltage accross R2?  What is the current therefore
passing thru R2?

> So here is one point that I'm stuck on.  The data sheet for the
> Pic10F204 says 0 - +6.5 volts.  But what I'm seeing is VDD gets 8.2V.

No, it gets about 5V, but we'll get back to that after you understand R2 and
D3.

> The zener at D3 and C2 have me stumped.

This is the heart of the circuit, so you have to figure this out first.
Look up the data sheet for the zener diode.  I'm using DigiKey part number
BZX84B5V6LT1GOSCT-ND, but just about any 5.6V zener will do.  Forget about
C2, Q1, and the fact that the signal on the top of the zener goes elsewhere.
Considering only R2 and D3 and by looking in the zener data sheet, what will
the voltage on the zener be when the top end of R2 is a 8.2V?

How sensitive is this to the top of R2 being exactly 8.2V?  What if it were
7V or 10V?  So overall, what does the circuit of just R2 and D3 accomplish
if the output is the voltage on the zener diode?


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000.


'[EE] Analysis of Olin's EUSB2 Circuit (sub circuit'
2007\02\01@013643 by Lindy Mayfield
flavicon
face
As an exercise Olin kindly gave me a circuit to try to analyze.  Here is what I can find in sub circuit R2, D3, C2, Q1.

I am deliberately not reading any of the previous posts on this yet until I can give this a try myself.

First the component list:
R2 2.7k resistor
D3 Zener Diode 5.6v
C2 1uF 10v capicitor
Q1 NPN Transistor

Since I had hints but nothing too specific about how and where to start, I'll have to fall back on my programming knowledge and give this a shot. If it doesn't work, I'll learn something and try a different approach.

Inputs and Outputs

Assumption: The emitter on Q1 feeds VDD on IC2.  A small current on the base of Q1, which looks like me to come from the 8.2V line after passing through R2. I'm not sure how to apply V=IR here, but from what I see the 8.2 volts passing through R2 drops to 5.6 volts which is used in both the base of Q1 and the output of the 5.6 volt power line from this power supply module.   But how is 8.2 volts dropping to 5.5?  Is because some of it is going to "leak" through R2.  

Resistors limit current, correct? So passing 5V through a resistor should yield 5V on the other side? Or did I get this wrong.

So here is one point that I'm stuck on.  The data sheet for the Pic10F204 says 0 - +6.5 volts.  But what I'm seeing is VDD gets 8.2V.

The zener at D3 and C2 have me stumped.  I've already been through a lot of Mims' books and other electronics books to get this far.  I'll go to E&E next.   I love that book!

Any corrections in my logic, just to make sure I'm on the right path, would be greatly appreciated Olin.  

I cannot tell you how grateful I am that you're taking this bit of time to help me get over my learning curve.

Kind regards,
Lindy

(Resent in plain text wrapping at 72 as was advised)

2007\02\01@063356 by Peter P.

picon face
Lindy Mayfield <lindy.mayfield <at> ssf.sas.com> writes:

> As an exercise Olin kindly gave me a circuit to try to analyze.  Here is what
> I can find in sub circuit R2, D3, C2, Q1.
>
> I am deliberately not reading any of the previous posts on this yet until I
> can give this a try myself.
>
> First the component list:
> R2 2.7k resistor
> D3 Zener Diode 5.6v
> C2 1uF 10v capicitor
> Q1 NPN Transistor

I'll try some hints:

1. The schematic does not indicate inputs and outputs. The wires that go
'offboard' to the right are not labeled with direction indicators. This can lead
to confusion when reading the schematic but it is not an error for a schematic
of this class (which is not a teaching schematic). Typically a teaching
schematic would include direction arrows for the wires, to show 'sources' and
'burdens'.

2. The circuit you focused on is a typical zener regulated linear power supply.
To analyze it you must see that the 'input' is the 8.2V voltage. The node where
the base of Q1 is connected must again be split into 'inputs' and 'outputs'. The
only input is R2, as it 'feeds' from a higher voltage (8.2V). All the other
branches are 'outputs'.

3. C1 is a filter capacitor. Since this is a DC circuit the capacitor will
charge at turn-on and thereafter cease playing a role in DC. So we can forget
about it for now.

4. The other branches are Q1 base and Q5 base (page3), and D3. Since transistor
bases cannot (usually) supply current these are clearly 'burdens'. Moreover they
are low current burdens, compared to what D3 is capable of (you need to know
part characteristics to be able to deduce this, D3 is able to handle about 5
times more current than Q1 and Q4 bases - but that sort of current is not
available in this circuit).

5. Thus the 'major' burden on the node seems to be D3. We know that D3 is a
zener, it will not allow the voltage across it to increase beyond the part's
spec (5.6V in this case). If the voltage increases the zener will absorb more
current and try to keep the voltage constant.

6. Now we can make sense of it: R2 feeds current from the 8.2V supply to the
C1/R2/D3/Q1B/Q5B node. The voltage on the node tries to raise. It does this
until the major burden on it limits it. That is D3. So the voltage rises to 5.6V
and stops there. C1 charges to 5.6V at turn-on and stays so, acting as a filter
(which counters voltage fluctuations).

7. Thereafter Q1 and Q5 are fed with 5.6V on the bases. They also receive 8.2V
on their collectors. Their emitters supply voltage to two chips (thus they are
burdens). Since a NPN silicon transistor has a 0.6V (nominal) Vbe drop the
voltage at the emitters will be 5.6V-0.6V = 5.0V (nominal). If the current
through D3 is a maximum of (8.2V-5.6V)/2000Ohms = 1.3mA when neither Q1 nor Q5
draw current from the node, then if the gain (beta) of Q1 and Q5 is 100 and we
accept that the current through D3 should not become less than half of what it
is without load, then the loads at the emitters of Q1 and Q5 can draw a maximum
of 0.65mA*100=65mA (sum) without any significant voltage change. If the loads
draw more, then the voltage will sag. Thus the constant voltage supply formed by
Q1 and Q5 is short circuit proof. This is called a 'distributed linear
regulator' and the 5.6V supply is often called a 'reference voltage' for this
supply.

8. In normal operation each of Q1 and Q5 will only dissipate about 90mW at most
(32.5mA*(8.2-5.6V)). This is within spec.

9. conclusions:

9.1 this circuit is a 'book move'. A standard circuit that anybody should
recognize directly at first sight. Even with the 5.6V wire going offboard.
Circuits like this are to be studied from intro books, not from real circuits.

9.2 the normal way to study such a circuit is to have the 'book moves' book open
and then look for circuits that appear in the 'book' in real schemes, like the
one you have decided to study.

9.3 simplified circuit analysis, like I did above, neglects A LOT of details and
it is not recommended that novices draw conclusions from such analysis. It is a
brain-teaser only. Simplified circuit analysis does not work well for low
current and AC circuits with reactive elements. In SPICE you can model such a
circuit with ease, and run a dc operating point analysis. SPICE will do the full
analysis (not simplified) and will let you see things like the current
distribution between the transistor bases and D3 and its effect on output
voltage.

9.4 imho the difference between doing circuit analysis in public on a mailing
list with 1000+ memebers on the scheme of a commercial device whole link was
published and viral marketing thereof is nearly negligible and could be a pretty
good idea.

hope this helps,
Peter P.


2007\02\01@081222 by olin piclist

face picon face
Peter P. wrote:
> 8. In normal operation each of Q1 and Q5 will only dissipate about 90mW
> at most (32.5mA*(8.2-5.6V)). This is within spec.

Actually in normal operation it is far less than that since the PIC 10F204s
don't draw anywhere near that much current.  The supplies appear to be
designed for higher current capability only because of the need to run at
least around 1mA thru the zener so that it gets close to its rated voltage.

> 9.1 this circuit is a 'book move'. A standard circuit that anybody
> should recognize directly at first sight.

Anyone that has gotten past a certain minimum level of electronics
knowledge.  Nobody was born being able to recognize such a circuit.

> Circuits like this are to be studied from intro books, not
> from real circuits.

I strongly disagree.  Circuits like this can be introduced in isolation, but
there is no substitute for analyzing and understanding them in context.
There are always issues that creep into the circuit from its context.  For
example, why is C2 there, and why was that particular value chosen?  The
answer only makes sense in the context of the whole circuit, and might be
instructive to someone that is not used to designing circuits for volume
production.

In addition of course you get to analyze not only the subcircuit, but also
how that fits into the bigger picture, why that subcircuit was chosen as
apposed to other ones that on paper perform roughly the same function, etc.
I think it is exactly this kind of analysis and understanding that is
lacking from most theoretical EE educations.  Don't get me wrong, I think
the theory is vitally important too.  Given a limited amount of curriculum
time, it makes sense to concentrate on the theory and expect students to
pick up practical knowledge on their own.  The best way to do that is to
design a circuit, build it, then debug it with a scope to figure out what it
does something unexpected.  Labs and projects are meant to do this to some
extent, but far too many of them are so poorly conceived, run, and taught,
that a great deal of the educational opportunities are not realized.

Other disciplines recognize this.  Architects study existing structures,
business student do lots of case studies, but EEs rarely do analysis of real
working shipping circuits with all their warts and tradeoffs.  I think this
is a mistake.

> 9.2 the normal way to study such a circuit is to have the 'book moves'
> book open and then look for circuits that appear in the 'book' in real
> schemes, like the one you have decided to study.

That doesn't make it a good way, though.  In fact I think that is a bad way.
It encourages getting a answer without real understanding.  Such a student
might be able to tell you it's a "regulated power supply using a zener
voltage reference driving a emitter follower" because that is what the
caption below the picture in the book said.  But what is forcing that
student to actually understand what that means and understand how the
circuit operates internally?  Loose the book.  After analyzing it and
figuring out what it does yourself, you will remember that configuration and
the tricks used, no book required anymore.

> 9.3 simplified circuit analysis, like I did above, neglects A LOT of
> details and it is not recommended that novices draw conclusions from
> such analysis. It is a brain-teaser only.

Nonsense!  This is exactly how real circuits are conceived.  The thought
process for the initial design of such a circuit isn't much more than "B-E
drop is around 600mV, need 5.6V zener, make R2 to get about 1mA at nominal
voltages, add C2 for transient response.".  Then you crank a few numbers
like you did to verify current capability is sufficient, regulation is good
enough, power dissipation is not exceeded, etc, although you're already
pretty sure they are all acceptable.  But the first step is the important
"design" part that requires some intuition of circuits.  The second step is
just cranking that any EE student should be able to do.

In fact it is my observations that good EEs use intuition and experience to
synthesize a design, then pick specific part values and verify with the
detailed analysis.  The not so good EEs that don't have the intuition and
can't seem to really "feel" the circuit and see the voltages and currents
and the components reacting to them over-analyize as sort of a compenstation
mechanism.

> Simplified circuit analysis
> does not work well for low current and AC circuits with reactive
> elements. In SPICE you can model such a circuit with ease, and run a dc
> operating point analysis. SPICE will do the full analysis (not
> simplified) and will let you see things like the current distribution
> between the transistor bases and D3 and its effect on output voltage.

Simulation tools have their place, but mostly for verification and detailed
numeric analysis, not often for design except for some specialty areas.
I'll confess right now I've never used Spice.  Something is wrong if you
don't already know how your circuit operates and can't work out the numbers
in a few minutes with a calculator.  There are exception, like antenna
design and analysis though where simulation is a critical part of the design
process.  But for basic circuits like this one, EE students shouldn't be
allowed to use a simulator except at the end to confirm what their thought
process and basic analysis should already have told them.  A simulator is
valuable as a virtual workbench because it allows trying out (within
limitations) circuits more quickly with more in depth information about
their operation.  This allows a student to try out more circuits more easily
and therefore this is a chance to learn more in the same amount of time.  Of
course a few need to be actually built too, because simulators only go so
far, and learning the limitations of simulators is important too.

> 9.4 imho the difference between doing circuit analysis in public on a
> mailing list with 1000+ memebers on the scheme of a commercial device
> whole link was published and viral marketing thereof is nearly
> negligible and could be a pretty good idea.

Say what?  Mr Kenny, my 7th grade english teacher, used to hand out awards
for "sentences" like this.  So what's the bottom line?  Do you think we
should be doing what we're doing on the list or not?


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000.

2007\02\01@085306 by Lindy Mayfield

flavicon
face
Olin said:

Say what?  Mr Kenny, my 7th grade english teacher, used to hand out awards
for "sentences" like this.  So what's the bottom line?  Do you think we
should be doing what we're doing on the list or not?

I answered:

I think there are a lot of people in the 1000's on this list that will gain some knowledge of the thought process that you are helping me with.  Many have already said so.  I'm stubborn.  I like to play Bach on the piano.  I won't give up so easily, though it may take some time to go through H&H and other books I have to get the answer.

P.S. I really don't get the big controversy over bottom vs. top posting.  For me to write at the top seems more logical.

2007\02\01@142647 by Peter P.

picon face
Olin Lathrop <olin_piclist <at> embedinc.com> writes:

> > 9.1 this circuit is a 'book move'. A standard circuit that anybody
> > should recognize directly at first sight.
>
> Anyone that has gotten past a certain minimum level of electronics
> knowledge.  Nobody was born being able to recognize such a circuit.

No, but imho it is better to know what one is looking for before looking, lest
one start seeing monsters.

> > Circuits like this are to be studied from intro books, not
> > from real circuits.
>
> I strongly disagree.  Circuits like this can be introduced in isolation, but
> there is no substitute for analyzing and understanding them in context.

I have to say that I do not have a degree in electronics (or anything else), and
that I never took an electronics course in my life (other than self
administered). Why this is so is beyond the scope of this discussion (if life
would have worked right I should have had a master in applied physics or such),
but with 20/20 hindsight all I can do is quote a saying as a fact here:
'experience is the worst possible teacher. It will show you all the mistakes
after you make them, and none of those you haven't made yet.'

> In addition of course you get to analyze not only the subcircuit, but also
> how that fits into the bigger picture, why that subcircuit was chosen as
> apposed to other ones that on paper perform roughly the same function, etc.

I think that that should also be a book move, just from another book. I.e. not
basic electronics but optimization by cost or for other constraints (like
manufacturability, testability, safe-fail modes etc etc). I agree that I am not
aware of any course that teaches these aspects in depth.

> pick up practical knowledge on their own.  The best way to do that is to
> design a circuit, build it, then debug it with a scope to figure out what it
> does something unexpected.  Labs and projects are meant to do this to some
> extent, but far too many of them are so poorly conceived, run, and taught,
> that a great deal of the educational opportunities are not realized.

I agree but still building 'book' circuits with SPICE will give people a running
start faster than any soldering iron. Of course they will not replace the
soldering iron but the bang/time gain of using SPICE instead of bunches of wires
is huge. It also teaches students early that there HAS GOT to be a ground wire
connecting ALL the parts early in the game (as opposed to certain questions
which crop up from time to time on some newsgroups, coming from freshmen).

> Other disciplines recognize this.  Architects study existing structures,
> business student do lots of case studies, but EEs rarely do analysis of real
> working shipping circuits with all their warts and tradeoffs.  I think this
> is a mistake.

You might find this to be not quite true. I have met many students of EE and all
those who had any talent showed a keen interest (and often a scary amount of
imagination) while analyzing any schematic they could lay hands on (like their
stereo, computer power supply, TV, etc). The scary parts made me ask myself some
interesting questions about what exactly they were taught sometimes. Eventually
they took their exams and started applying reality checks to their imaginations.

> > 9.2 the normal way to study such a circuit is to have the 'book moves'
> > book open and then look for circuits that appear in the 'book' in real
> > schemes, like the one you have decided to study.
>
> That doesn't make it a good way, though.  In fact I think that is a bad way.
> It encourages getting a answer without real understanding.  Such a student

It encourages knowing what to look for and then analyzing the differences from
the 'book move' imho. And that leads faster to the goal of understanding the
circuit and why it is different from the 'book move'.

> might be able to tell you it's a "regulated power supply using a zener
> voltage reference driving a emitter follower" because that is what the
> caption below the picture in the book said.  But what is forcing that
> student to actually understand what that means and understand how the
> circuit operates internally?  Loose the book.  After analyzing it and
> figuring out what it does yourself, you will remember that configuration and
> the tricks used, no book required anymore.

Imho nothing should be 'forcing' anything. Either the student's attention span
is sufficient to carry him from the book move to its implementation and to the
analysis of the differences (and to asking questions about it in a forum f.ex.),
or it isn't.

{Quote hidden}

The intuition is important but for optimized circuits it plays a surprizingly
small role ime. One could spend 3 weeks designing a knot of transistors,
resistors, caps and diodes that implements 6 different functions, is predictable
on power up and down, and can blink a led and beep if necessary, guaranteed over
temperature, safe fail etc. The other way is to use a chip or SPICE and do
it in 3 hours. Guess what gets used more. 20 years ago there was no SPICE and I
knew engineers who spent 2 weeks optimizing 3-transistor circuits using curve
tracers and log paper and such. Now one doesn't do that anymore. Anything with 3
parts or more is a candidate for an IC. F.ex. your regulators could be reduced
to just one LDO regulator I think, cutting the parts count by 5 (5 parts to 1).

> In fact it is my observations that good EEs use intuition and experience to
> synthesize a design, then pick specific part values and verify with the
> detailed analysis.  The not so good EEs that don't have the intuition and
> can't seem to really "feel" the circuit and see the voltages and currents
> and the components reacting to them over-analyize as sort of a compenstation
> mechanism.

Maybe but I was exposed to production/aftermarket/mods environments where one
cannot afford to write doctorates on circuits and getting things out the door on
time with minimum failure rate is a very high priority. In this context, 'good'
ideas are often discarded and 'medium' ideas are adopted which lead to lower
part count and/or cost (in fact these two are one and the same in production -
there is a coefficient for each). I also worked in research where spending 6
months on an IF stage is OK. So I know it both ways but I know what pays the
bills (unless one works for research, which changes the picture).

> > Simplified circuit analysis
> > does not work well for low current and AC circuits with reactive
> > elements. In SPICE you can model such a circuit with ease, and run a dc
> > operating point analysis. SPICE will do the full analysis (not
> > simplified) and will let you see things like the current distribution
> > between the transistor bases and D3 and its effect on output voltage.
>
> Simulation tools have their place, but mostly for verification and detailed
> numeric analysis, not often for design except for some specialty areas.

I beg to differ. The number of 'beer coaster' designs I doodled which were shot
down by a simple reality check with SPICE saved me months to years of soldering.
Not everything is obvious. Many standard circuits behave strangely when used in
certain combinations, or turn out to be too sensitive to something or other
(noise, temperature, supply at low or high end etc).

> I'll confess right now I've never used Spice.  Something is wrong if you
> don't already know how your circuit operates and can't work out the numbers
> in a few minutes with a calculator.  There are exception, like antenna
> design and analysis though where simulation is a critical part of the design
> process.  But for basic circuits like this one, EE students shouldn't be

I think that you are confusing SPICE with NEC. Anyway SPICE is a must use now,
NEC is a maybe (or RFSPICE or something like that).

> > 9.4 imho the difference between doing circuit analysis in public on a
> > mailing list with 1000+ memebers on the scheme of a commercial device
> > whole link was published and viral marketing thereof is nearly
> > negligible and could be a pretty good idea.
>
> Say what?  Mr Kenny, my 7th grade english teacher, used to hand out awards
> for "sentences" like this.  So what's the bottom line?  Do you think we
> should be doing what we're doing on the list or not?

There is no bottom line. It was a jocular observation. I forgot to add the ;-)
Keep it up. Your Mr Kenny used to hand out *what* kind of awards exactly ? ;-)
English is not my first language (by a long way) and I do not have a spell
checker here. I also economize commas (makes the keyboard last longer).

Peter P.


2007\02\01@143011 by Peter P.

picon face
> P.S. I really don't get the big controversy over bottom vs. top posting.  For
me > to write at the top seems more logical.

The idea is to split the original quotes and respond by points. The quote gives
the context and your answer goes under that. When you top post it is impossible
to guess what you are talking about after about 3 exchanges, especially if the
message is trimmed (as it should be).

Peter P.


2007\02\01@150319 by olin piclist

face picon face
From: "Peter P."
> your regulators could be reduced
> to just one LDO regulator I think, cutting the parts
> count by 5 (5 parts to 1).

You can get this opinion by looking at the regulator by itself.  You need to
look at the context in the circuit and other bigger issues to realize why it
was done this way.  Do you honestly think I'm unaware of LDOs and that using
them here never occurred to me?

In fact this is a good circuit challenge, although not beginner level.  Why
use R2, D3, C2, and Q1 instead of a 5V LDO?  What about the related supply
produced by Q5 on page 3?  Would you have done as I did, used one LDO for
both as Peter is suggesting, or two separate LDOs?


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000.

2007\02\01@150444 by Mike snyder

picon face
On 2/1/07, Peter P. <spam_OUTplpeter2006TakeThisOuTspamyahoo.com> wrote:
>
> > P.S. I really don't get the big controversy over bottom vs. top
> posting.  For
> me > to write at the top seems more logical.
>
> The idea is to split the original quotes and respond by points. The quote
> gives
> the context and your answer goes under that. When you top post it is
> impossible
> to guess what you are talking about after about 3 exchanges, especially if
> the
> message is trimmed (as it should be).
>
> Peter P.
>
>
> -

2007\02\01@182822 by Peter P.

picon face
Olin Lathrop <olin_piclist <at> embedinc.com> writes:

> From: "Peter P."
> > your regulators could be reduced
> > to just one LDO regulator I think, cutting the parts
> > count by 5 (5 parts to 1).
>
> You can get this opinion by looking at the regulator by itself.  You need to
> look at the context in the circuit and other bigger issues to realize why it
> was done this way.  Do you honestly think I'm unaware of LDOs and that using
> them here never occurred to me?

I do not make assumptions. All I can say is that if this project would be
'dropped' in my lap so to speak I'd probably shave off one third to one half of
the parts in various ways. I'm not saying it can be done, I'm saying that's what
I'd try to do.

I think that your distributed regulator scheme tries to use delayed power-on to
the MCUs which work as switchers such that they start working when the power
rail supplied by external USB power is high enough, and presumably after USB
power negotiation took place (I have no way to guess this it's a firmware
thing), but in any case after IC3 will have reset properly and initialized (and
they also isolate the programming Vdd when the 10Fs are ICSPd). I think that I
would have used a LDO with built in voltage monitor from Seiko, Microchip etc.
instead of IC1 and used the power good sense to release reset to IC3 and allowed
the switchers to be off (idle) at power on, and turned on by IC3 when needed.
I'd also have used a 1-way smpsu controller with 2 outputs instead of the two
10F's and used the 8.2V to make 17V. Also an active filter with gain for the PWM
filters (saves parts for the same slope). Also no -2.2V supply (324 outputs
pull down to 0 if loaded with 1k or active jfet or even bipolar current mirror
half).

<aside>10Fs are like aluminum paint. Once you start painting it's so much fun
you can't stop ;-)</aside>

But I'm just writing this email and it's your circuit, you tested it, and built
it, and are marketing it. That's a lot more than any email can do. So this is
just a guy's opinion.

Peter P.


2007\02\02@074855 by olin piclist

face picon face
Peter P. wrote:
> I think that your distributed regulator scheme tries to use delayed
> power-on to the MCUs which work as switchers such that they start
> working when the power rail supplied by external USB power is high
> enough, and presumably after USB power negotiation took place (I have
> no way to guess this it's a firmware thing),

Originally I had it the other way around where IC2 came on first, produced
the 8.2V and therefore the 5V supply to the main control PIC, which would be
held in reset until the 8.2V came up all the way.  I then had vague thoughts
of possibly being able to adhere to the USB spec for standby mode.  I
haven't really implemented this and am not sure the current consumption can
be made low enough anyway, but I flipped it around so that the main
controller starts running, then enables the switching supplies when it's
ready for them.  The current firmware just enables the switching supplies
right at the beginning and then waits a little while to make sure they are
stable before even turning on the USB peripheral.

> (and they also isolate the
> programming Vdd when the 10Fs are ICSPd).

Yes, this was one criteria.  The emitter outputs give inherent diode
isolation between the supplies.

> I think that I would have
> used a LDO with built in voltage monitor from Seiko, Microchip etc.
> instead of IC1

For others, IC1 is a 7805 regulator running from roughly a 7.5V supply,
making the 5V supply of the main controller PIC (18F2550).  There was no
need for a LDO here since something on the order of the 8.2V supply was
needed anyway.  Since the headroom was available, a 7805 was a good choice
because they are cheaper, widely available from multiple sources, and will
be the only things left after a nuclear blast along with cockroaches.

> and allowed the switchers to be off (idle) at power on, and turned on
> by IC3 when needed.

They are.  Take a look at the PWRDOWN line coming from IC3 at A3 on page 1.

> I'd also have used a 1-way smpsu controller with 2
> outputs instead of the two 10F's and used the 8.2V to make 17V.

The 10Fs also make the -2.2V supply.  Otherwise it's a close call to use a
programmable controller versus a dedicated chip.  Lately I've been using the
programmable controllers more because they get around some of the gray areas
with analog controller chips.  Startup, saturation of the inductor, and the
maximum current draw are all considerations, and these are often difficult
to guarantee with a analog controller chip.  The controller chips generally
do a better job once they get into the running regulation mode, but it's
often impossible to nail down exactly how they will react at startup,
overcurrent, or other exception conditions.  With digital control, I know
the switch will never be left on more than xx uS with minimum of yy uS
between pulses.  I know what the characteristics are at startup and the
worst case USB power current drawn.  I also know what will happen when too
much output curren is drawn, which in this case is that the supply voltage
will droop without anything getting stressed past its limit.  There is no
need for explicit over current protection.

Analog switching power supply control chips look good for normal operation,
but a 10F starts to look better when you start asking all the tough
questions around the fringes of operation that the data sheet won't answer.

> Also an
> active filter with gain for the PWM filters (saves parts for the same
> slope).

My filters are two resistors and two capacitors each.  And these are all
nice and cheap passive parts that don't introduce gain or offset errors.
You'd still have a R/C for each of the filter poles somewhere somehow even
for a active filter.  Besides, it's not about saving parts, but saving $$.
There is no way you can make a active filter for the less than $.04 each of
my passive filters cost.

> Also no -2.2V supply (324 outputs pull down to 0 if loaded with
> 1k or active jfet or even bipolar current mirror half).

First no opamp truly drives down to the rail with any kind of load on its
output.  However, a few 10s of mV output that should be zero could probably
have been tolerated.  The main reason is something you won't find in the
data sheets or you spice models.  Many opamps, definitely including the
LM324, act flaky when their inputs get near the negative supply, regardless
of what the data sheet says.  Some CMOS opamps could probably have worked
here with ground as their negative supply, but they generally have higher
offset voltages and higher cost.  And, I would have to test them for these
characteristics since the data sheets usually don't tell you this stuff,
then hope that the next lot works the same, the the same part from a
different manufacturer works the same, or specify a particular manufacturer.
No thanks.  The charge pump eliminates all these problems by providing over
2V headroom for the opamp at the low end.  Now I can use a well understood
dirt cheap opamp that I know will always do the job regarless of which
manufacturer or production lot it came from.

However all this dicussion aside, you didn't mention two of the three
advantages of the zener regulators that caused me to use them instead of a
single LDO as you suggested.  You did mention the inherent power supply
isolation, which is important during ICSP.  However the main reason was the
bootstrap operation.  Linear regulators are usually not specified for
operation with a input voltage below the regulated output voltage plus the
dropout voltage.  I've seen all kinds of strange behaviour, which included
oscillation and excessive output droop.

Think what happens before the switching power supplies come up.  At that
point you only have the USB power voltage, which can easily be a few 100mV
below the desired regulated voltage.  What would your 5V LDO do with only
4.5V input?  The problem is you don't know.  The zener regulator circuit on
the other hand is well behaved and clearly defined in that region.  It will
drop a bit under a volt at the few mA the 10F will draw, but otherwise
behave predictably and nicely.  Since a 10F is guaranteed to run with Vdd as
low a 2V, it is guaranteed to be operated well within spec.  The real issue
of voltage droop is not what the 10F can handle, but the higher on
resistance of Q3 with the lower gate drive.  The IRLML2502 is specified in
this region, although it will have somewhat higher Rdson at 4V than at 5V.
Since this is a programmable power supply controller, it can take that into
account during the startup phase.  During the startup phase the "8.2V"
supply will come up from one Schottky diode drop below the USB power voltage
to the nominal 8.2V regulation level.  Once it gets to a bit under 6V the
zener regulator will stabalize at 5V output and the gate of Q3 will be
driven from 0V to 5V.

The third advantage of the zener regulators over a LDO is cost.  I've got
two isolated 5V power supplies for the cost of 1 resistor, 1 zener diode, 1
cap, and two jellybean transistors, for a total of well under $.20 in parts
(and I'm being rather generous).


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000.

2007\02\03@055937 by Hector Martin [PIClist] n/a

flavicon
face
Olin Lathrop wrote:
> You are still sending paragraphs as single long lines, which are getting
> truncated somewhere.  I already mentioned this once before, and I'm not
> going to mention it again.  In the future I will likely just ignore your
> posts until you set up your email client properly.

It's something on your side (server or client), as I get the posts
correctly. Not trying to blame you here (I agree that manual wrapping of
lines is desirable), but you might want to look at getting that fixed. I
know I would, just to avoid the hassle of telling people to fix THEIR
e-mail clients instead :). Plus, although you can certainly request
people who need your services to use line-wrapping, sometimes you'll get
people that don't use line-wrapping from whom YOU need services.

Again, I agree with you on what the "right thing" is here - if I were
you though (which I'm not, so don't take this as blaming you or
otherwise), I'd get your side fixed.

--
Hector Martin (.....hectorKILLspamspam@spam@marcansoft.com)
Public Key: http://www.marcansoft.com/marcan.asc


2007\02\05@060212 by Gerhard Fiedler

picon face
Hector Martin [PIClist] wrote:

> Olin Lathrop wrote:
>> You are still sending paragraphs as single long lines, which are getting
>> truncated somewhere.  I already mentioned this once before, and I'm not
>> going to mention it again.  In the future I will likely just ignore
>> your posts until you set up your email client properly.

> ... but you might want to look at getting that fixed. I know I would,
> just to avoid the hassle of telling people to fix THEIR e-mail clients
> instead :).

We've been there before :)  For years Olin is telling people to "fix" their
clients, apparently because his POP3 server is broken (as in "not compliant
with Internet standards").

Not the first hint that could have lead to a solution (12/2004):
<http://tinyurl.com/2nt9e3>

The thread that seems to have cleared it up for Olin where the problem is
(7/2006): <http://tinyurl.com/3e37xl>

The last message of that thread: <http://tinyurl.com/2nx3z2>

Quoting Olin from that message:
> Given this new information, I may try to find where "256" was typed in
> and replace it with "1000" in the POP3 server.

Gerhard

2007\02\05@071458 by olin piclist

face picon face
Gerhard Fiedler wrote:
> Quoting Olin from that message:
>> Given this new information, I may try to find where "256" was typed in
>> and replace it with "1000" in the POP3 server.

Yeah, someday.  However this doesn't change the fact that sending long lines
is a stupid, irresponsible, and obnoxios thing to do and that I will try to
insist people send email properly whenever I have the power to do so.


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000.

2007\02\05@211553 by Gerhard Fiedler

picon face
Olin Lathrop wrote:

> Gerhard Fiedler wrote:
>> Quoting Olin from that message:
>>> Given this new information, I may try to find where "256" was typed in
>>> and replace it with "1000" in the POP3 server.
>
> Yeah, someday.  However this doesn't change the fact that sending long
> lines is a stupid, irresponsible, and obnoxios thing to do

Harsh words from someone running a broken mail server... :)

Gerhard

2007\02\06@114941 by Lindy Mayfield

flavicon
face

As an exercise Olin kindly gave me a circuit to try to analyze.  Here is what I can find in sub circuit R2, D3, C2, Q1.

http://www.embedinc.com/products/eusb2/eusb2.pdf

Please correct me if I am off track.  I thought that I should identify the inputs and outputs, or the positive to ground or flow of the current.  Someone mentioned that they aren't necessarily drawn that the flow goes off the page to the right.

I know from Q1 that the base and collector are input.  I know anything pointing to ground is flowing in that direction.

Therefore the 5.6V from the right side of the page is flowing towards Q1 base.  

And the 8.2V after SH1 is also supplying the current to the Q1 collector.  It also supplies the current from R2 to D3.  

So what is interesting to me is that Q1's base is getting a direct 5.6V as well as 8.2V passing through R2.  

The zener and capacitor have me stumped.  The H&H book is quite thick, but I'll delve into it next to try to unravel this mystery.

To be honest, I think this is a difficult challenge for me, but I won't give up. (-:

Lindy



2007\02\06@124714 by Vasile Surducan

face picon face
On 2/1/07, Lindy Mayfield <lindy.mayfieldspamKILLspamssf.sas.com> wrote:

> P.S. I really don't get the big controversy over bottom vs. top posting.  For me to > write at the top seems more logical.

Lindy, please answer to Olin's emails only at the bottom of his answers.
In that way you'll make him happy, there is a convention on this list,
posting at top means no respect, disagree etc, while posting on the
bottom means respect,
begging mercy, etc.
That's why some people are starting often a new subject with the same
topic name, they need to be always on top, it's just a human
(nonsense) behaviour.
There is one truth, if you've post at bottom it means you've read the
whole previous post...which could be good if hasn't 20 km lenght.

best wishes
Vasile

2007\02\06@131611 by olin piclist

face picon face
Lindy Mayfield wrote:
> Please correct me if I am off track.  I thought that I should identify
> the inputs and outputs, or the positive to ground or flow of the
> current.  Someone mentioned that they aren't necessarily drawn that the
> flow goes off the page to the right.

You are ignoring the things I told you last time.  As I said, forget about
the line going off the page, forget about C2 for now, and concentrate on
just what R2 and D3 do.  There is a reason I ask specific questions in a
specific order.  If you want me to continue with this you have to at least
answer the questions I asked before looking around trying to figure out
other things.  Since you aren't really replying to my last message, I don't
remember exactly where we left things.  You aren't making it easy for me to
help you.  I said that once before, so this time I'm not going to quit here.


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000.

2007\02\06@134255 by Lindy Mayfield

flavicon
face


-----Original Message-----
From: .....piclist-bouncesKILLspamspam.....mit.edu [EraseMEpiclist-bouncesspam_OUTspamTakeThisOuTmit.edu] On Behalf Of Olin Lathrop
Sent: Tuesday, February 06, 2007 8:18 PM
To: Microcontroller discussion list - Public.
Subject: Re: [EE] Analysis of Olin's EUSB2 Circuit (sub circuit R2, D3, C2, Q1)

Lindy Mayfield wrote:
> Please correct me if I am off track.  I thought that I should identify
> the inputs and outputs, or the positive to ground or flow of the
> current.  Someone mentioned that they aren't necessarily drawn that the
> flow goes off the page to the right.

You are ignoring the things I told you last time.  As I said, forget about
the line going off the page, forget about C2 for now, and concentrate on
just what R2 and D3 do.  There is a reason I ask specific questions in a
specific order.  If you want me to continue with this you have to at least
answer the questions I asked before looking around trying to figure out
other things.  Since you aren't really replying to my last message, I don't
remember exactly where we left things.  You aren't making it easy for me to
help you.  I said that once before, so this time I'm not going to quit here.

>> Lindy Replies:

I'm sorry Olin.  I'm really trying. Please don't give up on me as long as I try my best.  I think you are a great teacher, and forgive me for trying your patience.

Did you mean R2 and D2 or R2 and D3?

This is just new to me. I don't know which lines and components to follow. I was trying to be logical, but didn't succeed.

Am I on the right track to try to see in which direction the current flows?

Kindest regards,
Lindy




2007\02\06@141535 by olin piclist

face picon face
Lindy Mayfield wrote:
> I'm sorry Olin.  I'm really trying. Please don't give up on me as long
> as I try my best.

I don't care how hard you try.  The world looks for results, not effort.
Responding to the questions that I ask is a simple thing to do.  If you
think my patience is wearing thin, you're right.

>> As I said, forget about the line going off the page, forget
>> about C2 for now, and concentrate on just what R2 and D3 do.
>
> Did you mean R2 and D2 or R2 and D3?

I meant what I wrote.


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000.

2007\02\06@144553 by Lindy Mayfield

flavicon
face


Lindy Mayfield wrote:
> I'm sorry Olin.  I'm really trying. Please don't give up on me as long
> as I try my best.

I don't care how hard you try.  The world looks for results, not effort.
Responding to the questions that I ask is a simple thing to do.  If you
think my patience is wearing thin, you're right.

>> As I said, forget about the line going off the page, forget
>> about C2 for now, and concentrate on just what R2 and D3 do.
>
> Did you mean R2 and D2 or R2 and D3?

I meant what I wrote.
¨

Lindy Replys >> I'll try one more time as best I can.  But I don't want to wear your patience thin, which seems to be pretty easy to do, and I think if we knew each other personally you'd not be so harsh with me.  But I totally understand.  I get one more chance and if I fail it's over between your helping me.  I'm sorry about that.  I really respect your knowledge and experience, and ability as a teacher.  

I'm sorry it ended this way.  But I get one more try and that's the end of it. )-:

Very Kindest Regards,
Lindy

P.S. Results come from effort. And responding to your questions which you deemed simple weren't obvious or simple for me.  I've spent hours digging through my electronic books trying to understand this simple sub-circuit.

2007\02\06@145939 by Robert Young

picon face
>As an exercise Olin kindly gave me a circuit to try to analyze.  Here is
>what I can find in sub circuit R2, D3, C2, Q1.
>

--- stuff delteted ---

Lindy,

It seems Olin is just about to snap (and rightly so).  Stop thrashing around
with this.  Look only at the little bitty part he ask you to look at.  If it
helps, draw it on a separate sheet of paper.

The Art of Electronics is a thick book but you don't need to read all of it
at once.  Get through chapter one, understand the basics there.  Chapters
1-4 should be read in sequence.  After that you can jump around in the book.
 But chapter one should get you what you need to understand a voltage
source, a resistor and a zener diode.

If you can find a copy of the paper-back student's manual to accompany AOE,
get a copy of that too.  It has some good hands-on experiments and extra
material that had to be left out of the book.

Rob


2007\02\06@153108 by James Newton, Host

face picon face
> Lindy, please answer to Olin's emails only at the bottom of
> his answers.
> In that way you'll make him happy, there is a convention on
> this list, posting at top means no respect, disagree etc,
> while posting on the bottom means respect, begging mercy, etc.

If that is a convention on this list, it is one that I, as the longest
running list admin, am unaware of. For example, I'm disagreeing with you
here and I'm posting at the bottom rather than the top.

I would welcome the assistance if you desire to become an admin and help run
the list. You could then make that even more than a convention.

In the mean time, as far as I am aware: Posting on the top is best when your
thoughts stand on their own and do not need to be framed as a reply to
something else. Posting inline is best for replies that make sense only in
the context of the original post, need some setup to follow or are posted by
people who adhere to the "anti-top-posting" religion.

Members of the "anti-top-posting" religion should read the FAQ on
appropriate subjects for discussion and know that in the absence of logical
proof to the contrary, their religion is a religion.

Posting at all is a wonderful thing, and should be encouraged, top or bottom
as long as the content of the post is thoughtful, useful or humorous without
being rude, demeaning, or overstepping ones place.

Having said all that, top or bottom posters really should take the extra
couple seconds to trim what they do not need to send again. For top posts,
Ctrl-Shift-End / Del will trim from the end of your post to the end of the
page.

> That's why some people are starting often a new subject with
> the same topic name, they need to be always on top, it's just a human
> (nonsense) behavior.

It is also the default behavior of the most widely used email program in
existence. Never attribute to malice that which can be explained by
inexperience.

> There is one truth, if you've post at bottom it means you've
> read the whole previous post...which could be good if hasn't
> 20 km length.

And on that, I must agree.

Best wishes.

---
James Newton: PICList webmaster/Admin
jamesnewtonspamspam_OUTpiclist.com  1-619-652-0593 phone
http://www.piclist.com/member/JMN-EFP-786
PIC/PICList FAQ: http://www.piclist.com


2007\02\06@153223 by Gerhard Fiedler

picon face
Lindy Mayfield wrote:

> Lindy Mayfield wrote:
>>> As I said, forget about the line going off the page, forget
>>> about C2 for now, and concentrate on just what R2 and D3 do.
>>
>> Did you mean R2 and D2 or R2 and D3?
>
> I meant what I wrote.

> P.S. Results come from effort. And responding to your questions which
> you deemed simple weren't obvious or simple for me.  I've spent hours
> digging through my electronic books trying to understand this simple
> sub-circuit.

Maybe a few hints (I don't mean to disturb anything, just help keep the
ball rolling :)

For the purpose of this exercise (R2, D3), consider the 8.2V rail a given
(at first). Disconsider any other components. It then looks like this:

----+----8.2V
   |
  R2
   |
   +---- What's going on here? And why?
   |
  D3
   |
  Gnd

Another hint: the symbol Olin used for D3 usually means it's a zener diode.
Look up how these work; no fine details are needed for understanding this
subcircuit, only the basic principle.

Search at your favorite manufacturer's site or at Digikey for a 5.6V zener
diode datasheet (e.g. BZX84 is a common type). Wikipedia has an article on
zener diodes. Probably other electronics sites do, too, and in most
introductory electronics book you should find them explained, at least the
basics.

One difference between reading a traditional program and an electronic
schematic is that in a traditional program, there is a clear causality, a
clear sequence from input through processing to output. In an electronic
schematic that's not so simple: an input is always also a load that
influences the source, often there is no clear input and output (like with
a resistor), etc. It's more of a complex system where everything influences
everything else. But what probably most people do when reading such a
schematic is that they start out with some simplifying assumptions, and
later on verify whether these assumptions are consistent with the
schematic. So in this case, we just assume that whatever is connected to
the connection between R2 and D3 is not relevant for them to work as this
basic, common circuit is supposed to work. Later on we need to see whether
this is true, when exactly it is not true if it isn't always, and what the
consequences are of this. Same for the 8.2V rail.

Gerhard

2007\02\06@173507 by olin piclist

face picon face
Lindy Mayfield wrote:
> But I totally understand.  I get one more chance and if I
> fail it's over between your helping me.

Maybe I left the wrong impression.  This has nothing to do with your ability
to analyze circuits or how correctly you answer the questions.  That will
only be input into my attempts to guide the learning process.

What you are "failing" at is responding (again having nothing to do with
technical correctness) to the questions I ask.  I can see some
misconceptions you have, so I ask questions and that will cause you to
figure out the thing you need to get past your misconceptions.  The problem
is you won't do as I ask.  You go off and try to analyze other parts of the
circuit I specifically said to forget about for now.  Several times now I
have told you to concentrate on R2 and D3, and you have yet to respond to
that at all.  I specifically said to ignore the 5.6V line going off the
page, but you won't.

In other words, it feels to me like I'm wasting my time trying to guide you
thru the process of understanding, and that you don't trust me to have some
idea where you're at or what a good direction for you to proceed is to
understand the next hurdle.

> Results come from effort. And responding to your questions which
> you deemed simple weren't obvious or simple for me.

I never said the questions were simple, only that addressing *those
questions* as apposed to going off somewhere else was simple.  It's a
process issue, not a technical ability issue.

> I've spent hours
> digging through my electronic books trying to understand this simple
> sub-circuit.

Then come back and ask about details you don't understand, but stick to what
I'm asking if you want me to guide you.


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000.

2007\02\07@040434 by Lindy Mayfield

flavicon
face


-----Original Message-----
From: @spam@piclist-bouncesKILLspamspammit.edu [KILLspampiclist-bouncesKILLspamspammit.edu] On Behalf Of Olin Lathrop
Sent: Wednesday, February 07, 2007 12:36 AM
To: Microcontroller discussion list - Public.
Subject: Re: [EE] Analysis of Olin's EUSB2 Circuit (sub circuit R2, D3, C2, Q1)

What you are "failing" at is responding (again having nothing to do with
technical correctness) to the questions I ask.  I can see some
misconceptions you have, so I ask questions and that will cause you to
figure out the thing you need to get past your misconceptions.  The problem
is you won't do as I ask.  You go off and try to analyze other parts of the
circuit I specifically said to forget about for now.  Several times now I
have told you to concentrate on R2 and D3, and you have yet to respond to
that at all.  I specifically said to ignore the 5.6V line going off the
page, but you won't.

>>> Lindy Answers:
Ok, I get it now.  Actually, you've made my job simpler.  R2 and D3.  I'll keep banging at that nut until I crack it.

I included the 5.6 because it was connected to Q1.  But 5.5v line, what 5.6v line. I don't even see it.  All I see is R2 and D3.  Actually all the other components disappeared too. (-:

Warmest regards,
Lindy

2007\02\07@041811 by Lindy Mayfield

flavicon
face

As an exercise Olin kindly gave me a circuit to try to analyze.  Here is what I can find in sub circuit R2, D3, C2, Q1.  But first I'm to focus on R2 and D3.

http://www.embedinc.com/products/eusb2/eusb2.pdf

D3 is a zener diode with a reverse breakdown voltage of 5.6 volts, sort of like a voltage sensitive switch.  

So R2 and D3 in this configuration functions as a stable voltage regulator to the base of Q1.

Am I close?

Lindy

2007\02\07@055145 by Wouter van Ooijen

face picon face
> So R2 and D3 in this configuration functions as a stable voltage
regulator to the base of Q1.
> Am I close?

Right on target.

Next (small) step: that stable voltage feeds the basis of Q1, who in
turn feeds the 10F chip. Why this Q1, why not feed the chip directly
from D3?

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu


2007\02\07@070357 by olin piclist

face picon face
Lindy Mayfield wrote:
> D3 is a zener diode with a reverse breakdown voltage of 5.6 volts, sort
> of like a voltage sensitive switch.
>
> So R2 and D3 in this configuration functions as a stable voltage
> regulator to the base of Q1.

Right.  The point of R2 and D3 is to produce a 5.6V source that is
reasonably independent of the actual voltage on the 8.2V rail.  Do you see
how the top (cathode) of D3 will be close to 5.6V even if the "8.2V" rail
were 7V or 10V, for example?  I think you do, but this is important to
understand before proceeding.

Now figure out what Q1 is doing and how it works in that circuit.  You've
already figured out that R2 and D3 are producing 5.6V, so now just assume
the base of Q1 will be driven with 5.6V and that the collector will be at
somewhere around 8.2V.  Explain what happens as other parts of the circuit
try to draw current coming out the emitter of Q1.  What function is Q1
performing?


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000.

2007\02\07@072421 by Lindy Mayfield

flavicon
face
Thank you.  You've probably some idea of how much H&H digging plus other electronics books and Google I did to get that far.  

I'm happy we are on the same page and that I am following your advice in the right way.

Lindy

P.S. Thanks!

{Original Message removed}

2007\02\07@085809 by Lindy Mayfield

flavicon
face
The way I understand Q1 or transistors are as switches: a small current that controls a larger current from Current to Emitter.  I don't see that on/off part here in the circuit.

{Original Message removed}

2007\02\07@092617 by Timothy J. Weber

face picon face
Lindy Mayfield wrote:
> The way I understand Q1 or transistors are as switches: a small current that controls a larger current from Current to Emitter.  I don't see that on/off part here in the circuit.

You may have to expand your transistor model then - find out what
happens "between On and Off."
--
Timothy J. Weber
http://timothyweber.org

2007\02\07@095838 by Lindy Mayfield

flavicon
face
Please, kindly, elaborate. "Between on and off" sounds like a digital circuit, and I'm seeing everything in binary.

Lindy

{Original Message removed}

2007\02\07@101248 by Cristóvão Dalla Costa

picon face
It's not digital. This particular case is a common collector configuration.
Read everything here:

http://www.allaboutcircuits.com/vol_3/chpt_4/index.html

2007/2/7, Lindy Mayfield <RemoveMElindy.mayfieldTakeThisOuTspamssf.sas.com>:
>
> Please, kindly, elaborate. "Between on and off" sounds like a digital
> circuit, and I'm seeing everything in binary.
>
> Lindy
>
>

2007\02\07@101428 by Timothy J. Weber

face picon face
> Please, kindly, elaborate. "Between on and off" sounds like a digital circuit, and I'm seeing everything in binary.

I'd call the "in-between" stuff "analog"; the binary one-or-zero stuff
is "digital."

Let me ask you this: What, when you're looking at a transistor as a
current switch, does "off" mean?  And what does "on" mean?
--
Timothy J. Weber
http://timothyweber.org

2007\02\07@104053 by John Chung

picon face
Actually I do find that analog is the first step in
understanding Olin's circuit. Lindy, perhaps you
should reread  your chapters there .

John


--- "Timothy J. Weber" <spamBeGonetwspamBeGonespamtimothyweber.org> wrote:

{Quote hidden}

> --

2007\02\07@104802 by Lindy Mayfield

flavicon
face
If there is a current in the base, it allows a larger current to flow from Collector to Emittor.  


{Original Message removed}

2007\02\07@112535 by Timothy J. Weber

face picon face
Lindy Mayfield wrote:
> If there is a current in the base, it allows a larger current to flow from Collector to Emittor.  

Right.  So that's not an all-or-nothing definition.  That is, you could
have no current in the base, a very very small current, a bit more, on
up to a lot.

I don't know that this is going to give you a big a-ha, but think about
what happens when that base current varies, and what could make it vary.
--
Timothy J. Weber
http://timothyweber.org

2007\02\07@115749 by Peiserma

flavicon
face
piclist-bounces@mit.edu wrote:
> If there is a current in the base, it allows a larger current
> to flow from Collector to Emittor.

while that's true, in this case it is more appropriate to view it another way. You've got H&H, right? study sections 2.03 and 2.04.

(I personally don't like the way H&H approaches the various transistor models. But that may be due to having learned it a different way/order via Sedra & Smith. YMMV - everyone has a learning methodology that works best for themselves)

2007\02\07@120337 by olin piclist

face picon face
Lindy Mayfield wrote:
> The way I understand Q1 or transistors are as switches: a small current
> that controls a larger current from Current to Emitter.  I don't see
> that on/off part here in the circuit.

Right.  Transistors are inherently analog devices.  They can be slammed on
and off hard to act like switches in some applications, but that's not what
is happening here.

Go back to the basics of bipolar transistors.  It might help to try to get a
rough idea of the emitter voltage as 100uA are drawn from it, then 1mA, then
10mA.  Use 100 for the gain of this transistor and calculate the currents
for each case.


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000.

2007\02\07@120506 by olin piclist

face picon face
"Cristóvão Dalla Costa" wrote:
> This particular case is a common collector

No, it's not.  But the main point is that Q1 is being used in analog, not
switching mode.


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000.

2007\02\07@120644 by olin piclist

face picon face
Lindy Mayfield wrote:
> If there is a current in the base, it allows a larger current to flow
> from Collector to Emittor.

Basically right.  Now apply that in this circuit.

********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000.

2007\02\07@121121 by Steve Baldwin

flavicon
face
On 7 Feb 2007 at 13:24, Lindy Mayfield wrote:

> Thank you.  You've probably some idea of how much H&H digging plus
> other electronics books and Google I did to get that far.  

You may find that you learn better using experiments. Some do, some don't.
Maybe someone can recommend a book or source of simple experiments
that demonstrate fundamental concepts (voltage dividers, zener diodes,
transistors, etc). That way, Olin can direct you in the context of the circuit as
a system and you can take the fundamental concept of the component and
see the hows and whys of using it.

I suspect that if it took lots of digging to find the key points of a zener
amongst the not-so-key points, you are going to find transistors very
frustrating.

Steve.


2007\02\07@122011 by Michael Rigby-Jones

picon face

>>"Cristóvão Dalla Costa" wrote:
>> This particular case is a common collector

>Olin Lathrop replied
>No, it's not.  But the main point is that Q1 is being used in
>analog, not switching mode.
>

It looks exactly like an emitter follower (AKA common collector) to me too.  In the interests of slightly reducing my level of ignorance, what is the correct term for this configuration?

Regards

Mike

=======================================================================
This e-mail is intended for the person it is addressed to only. The
information contained in it may be confidential and/or protected by
law. If you are not the intended recipient of this message, you must
not make any use of this information, or copy or show it to any
person. Please contact us immediately to tell us that you have
received this e-mail, and return the original to us. Any use,
forwarding, printing or copying of this message is strictly prohibited.
No part of this message can be considered a request for goods or
services.
=======================================================================

2007\02\07@124219 by olin piclist

face picon face
Steve Baldwin wrote:
> You may find that you learn better using experiments. Some do, some
> don't.

That's a good point.  I think it would be quite instructional to build this
little circuit on a breadboard and make some measurements under different
conditions.

All you need is R2, D3, Q1, and a basic power supply in the 7-15V range or
so.  12V would do fine.  Use the power supply to provide the "8.2V" rail,
hook up R2, D3, and Q1 as shown.  Leave the emitter of Q1 open but allow for
connecting different resistors between there and ground.

Put around 50Kohms between the emitter and ground.  Measure the voltage on
the emitter.  Based on that, how much current is flowing thru the test
resistor.  Repeat the experiment for resistors around 5Kohms and 500ohms.
Roughly what was the ratio of highest to lowest test resistor current
accross the tests?  What wsa the ratio of highest to lowest emitter voltage
accross the tests?  What conclusion can you draw from that about the
stability of the emitter voltage as a function of the emitter current?  In
what application might this property be useful?

You don't have to use a MMTB4401 transistor.  Just about any jellybean NPN
will work here.  The zener voltage doesn't have to be exactly 5.6V either,
but should be at least a 2.5V or so less than your supply voltage.


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000.

2007\02\07@124509 by olin piclist

face picon face
Michael Rigby-Jones wrote:
> It looks exactly like an emitter follower (AKA common collector)

I agree about the emitter follower part.  I've never heard the term "common
collector" like that.  This may be due to different conventions in different
locations, although I don't see how the collector can be thought of a
"common" to both input and output.


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000.

2007\02\07@134905 by John La Rooy

flavicon
face
On 2/8/07, Olin Lathrop <TakeThisOuTolin_piclistEraseMEspamspam_OUTembedinc.com> wrote:
>
> Michael Rigby-Jones wrote:
> > It looks exactly like an emitter follower (AKA common collector)
>
> I agree about the emitter follower part.  I've never heard the term
> "common
> collector" like that.  This may be due to different conventions in
> different
> locations, although I don't see how the collector can be thought of a
> "common" to both input and output.


Maybe this helps
http://en.wikipedia.org/wiki/Common_collector

2007\02\07@222525 by Lindy Mayfield

flavicon
face


-----Original Message-----
From: RemoveMEpiclist-bouncesspamTakeThisOuTmit.edu [piclist-bouncesEraseMEspam.....mit.edu] On Behalf Of Olin Lathrop
Sent: Wednesday, February 07, 2007 7:45 PM
To: Microcontroller discussion list - Public.
Subject: Re: [EE] Analysis of Olin's EUSB2 Circuit (sub circuit R2, D3, C2, Q1)

Steve Baldwin wrote:
> You may find that you learn better using experiments. Some do, some
> don't.

Olin Wrote:
That's a good point.  I think it would be quite instructional to build this
little circuit on a breadboard and make some measurements under different
conditions.

Lindy Answers:

That is spot on.  My landlady promised I could put a floor and a desk in the attic in the summer and then I can start doing things like this again.  Breadboards, meters, ossiscopes, and all that.  I REALLY learn when I do things hands on.  Right now I have not place because of my cats, and the attic is too cold.

2007\02\08@012222 by Peter Bindels

picon face
Hello,

On 07/02/07, Olin Lathrop <EraseMEolin_piclistspamembedinc.com> wrote:
> You don't have to use a MMTB4401 transistor.  Just about any jellybean NPN
> will work here.  The zener voltage doesn't have to be exactly 5.6V either,
> but should be at least a 2.5V or so less than your supply voltage.

Does this have to do with it not switching on/off often enough to
provide a stable output voltage when you have less over-voltage?

Peter

2007\02\08@022402 by Marcel Birthelmer

picon face
Peter,
the transistor doesn't switch on/off at all. The overvoltage is
necessary to provide current through the resistor, both to keep the
zener in the proper region, and to keep current flowing through the
transistor (or so I think!).
- Marcel

On 2/7/07, Peter Bindels <RemoveMEdascandyEraseMEspamEraseMEgmail.com> wrote:
{Quote hidden}

> -

2007\02\08@033145 by Peter Bindels

picon face
On 08/02/07, Marcel Birthelmer <RemoveMEmarcelb.listsTakeThisOuTspamspamgmail.com> wrote:
> Peter,
> the transistor doesn't switch on/off at all. The overvoltage is
> necessary to provide current through the resistor, both to keep the
> zener in the proper region, and to keep current flowing through the
> transistor (or so I think!).
> - Marcel

Now that I read it back, "it" wasn't quite as clear as I'd imagined. I
was thinking the zener would have to switch between conducting (in
blocking direction) and not conducting quickly enough to keep the
voltage across it at 5.6V?

The transistor conducts with a voltage drop of .6-.7 V so at 4.9 -
5.0V. Current is from the collector (mostly), voltage is decided by
the zener (which I think is kept at 5.6 volts this way).

I was mostly wondering why Olin suggested the input voltage to be 2.5V
or more above the 5.6V the zener requires.

2007\02\08@064837 by olin piclist

face picon face
Peter Bindels wrote:
>> You don't have to use a MMTB4401 transistor.  Just about any jellybean
>> NPN will work here.  The zener voltage doesn't have to be exactly 5.6V
>> either, but should be at least a 2.5V or so less than your supply
>> voltage.
>
> Does this have to do with it not switching on/off often enough to
> provide a stable output voltage when you have less over-voltage?

I said several things, so I can't tell what "this" you are referring to.  In
any case the answer to your question is likely "no", since there is no
switching going on here.  This is definitely a analog, as apposed to
digital, circuit.


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000.

2007\02\08@065833 by Gerhard Fiedler

picon face
Lindy Mayfield wrote:

> Steve Baldwin wrote:
>> You may find that you learn better using experiments. Some do, some
>> don't.
>
> Olin Wrote:
> That's a good point.  I think it would be quite
> instructional to build this little circuit on a breadboard
> and make some measurements under different conditions.
>
> Lindy Answers:
> That is spot on.  My landlady promised I could put a floor and a desk in
> the attic in the summer and then I can start doing things like this
> again.  Breadboards, meters, ossiscopes, and all that.  I REALLY learn
> when I do things hands on.  Right now I have not place because of my
> cats, and the attic is too cold.

Until you can do that, you may want to look into simulation, as others have
said before. While simulation doesn't work well for the marginal cases and
"dirt effects" (because they are not simulated well or not at all in the
models), it works quite well for normal operation -- and to some degree
better than with breadboards, because it's easier to just look at the
current that flows in a wire or at the power that's dissipated in a part.
Or see dozens of voltage levels and currents change on the fly when you
change something.

There's even a not so expensive simulation package that allows you to add a
PIC to the simulation and run firmware on it (Proteus VSM
<http://www.labcenter.co.uk>).

Gerhard

2007\02\08@070412 by olin piclist

face picon face
Peter Bindels wrote:
> Now that I read it back, "it" wasn't quite as clear as I'd imagined. I
> was thinking the zener would have to switch between conducting (in
> blocking direction) and not conducting quickly enough to keep the
> voltage across it at 5.6V?

The zener doesn't switch at all.  Take a look at a zener data sheet and see
the plot of voltage as a function of current.  Note that it is a continuous
graph.  There are no sharp corners or discontinuities.  Look at just R2 and
D3 and you should be able to compute the steady state operating points at
various levels of the "8.2V" supply.

> The transistor conducts with a voltage drop of .6-.7 V

A statement like this is very bad form.  I'm going to object every time,
just like when someone leaves off units, etc.  Transistors have 3 leads, so
"with a voltage drop" is meaningless without specifying which of the 3
possible combinations you are talking about.

> so at 4.9 -> 5.0V.

Again, this is a meaningless statement.  It is only meaningful to talk about
voltages between two points, not 5V somewhere.  I can often be acceptable to
talk about voltage at a specific point when there is a clear implied
understanding of a 0V reference point, usually ground.  However, that
doesn't apply to your statement either.

You may think I'm nit picking, but getting these sort of details right is
essential to technical communication, and it is therefore pointless to
continue a discussion without them.  Often sloppiness with the details is a
indication of poor understanding of the fundamentals, which also has to be
addressed before continuing.


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000.

2007\02\08@071338 by Gerhard Fiedler

picon face
Peter Bindels wrote:

> Now that I read it back, "it" wasn't quite as clear as I'd imagined. I
> was thinking the zener would have to switch between conducting (in
> blocking direction) and not conducting quickly enough to keep the
> voltage across it at 5.6V?

No. It creates a very small noise voltage across it, but that can't really
be called "switching". Other than that, it settles at a value in its I/V
diagram (which has the typical "knee" form).

> I was mostly wondering why Olin suggested the input voltage to be 2.5V
> or more above the 5.6V the zener requires.

I started out writing an explanation, but in the spirit of this thread, I
deleted it :)

Look at the current through the zener and the resistor and what this does
to the voltage across the zener when the voltage across the resistor
becomes very small. Keep the I/V curve of a real-world zener diode in your
mind for this.

Gerhard

2007\02\08@072648 by Xiaofan Chen

face picon face
On 2/8/07, Gerhard Fiedler <EraseMElistsspamspamspamBeGoneconnectionbrazil.com> wrote:
> Until you can do that, you may want to look into simulation, as others have
> said before. While simulation doesn't work well for the marginal cases and
> "dirt effects" (because they are not simulated well or not at all in the
> models), it works quite well for normal operation -- and to some degree
> better than with breadboards, because it's easier to just look at the
> current that flows in a wire or at the power that's dissipated in a part.
> Or see dozens of voltage levels and currents change on the fly when you
> change something.
>

I second that. Even though there are some good electrical engineers
(eg: Olin) who do not use simulation now, simulations are actually very
effective in learning electronics and doing electronics design.

PSpice student version is good and very easy to use. I've been a long
time PSpice and Saber simulator user and I like them very much. They
are very good for analog electronics learning and practical design.

I was in different Power Electronics research labs before and all of
them use PSpice/Saber for circuit simulation. Design ideas often
started from paper and then simulation and lastly hardware testing.

Sometimes simulation can be tricky and can be misleading. However,
they have been increasing sophisticated and I think any new electronics
engineer should learn at least learn a bit on analog/digital circuit simulation.

2007\02\08@072703 by Peter Bindels

picon face
On 08/02/07, Olin Lathrop <RemoveMEolin_piclistKILLspamspamembedinc.com> wrote:
> Peter Bindels wrote:
> > Now that I read it back, "it" wasn't quite as clear as I'd imagined. I
> > was thinking the zener would have to switch between conducting (in
> > blocking direction) and not conducting quickly enough to keep the
> > voltage across it at 5.6V?
>
> The zener doesn't switch at all.  Take a look at a zener data sheet and see
> the plot of voltage as a function of current.  Note that it is a continuous
> graph.  There are no sharp corners or discontinuities.  Look at just R2 and
> D3 and you should be able to compute the steady state operating points at
> various levels of the "8.2V" supply.

You're right, I oversimplified the model of the zener diode in my mind
to either conduct in blocking direction above its specified voltage
and to block voltages below it.

> > The transistor conducts with a voltage drop of .6-.7 V
>
> A statement like this is very bad form.  I'm going to object every time,
> just like when someone leaves off units, etc.  Transistors have 3 leads, so
> "with a voltage drop" is meaningless without specifying which of the 3
> possible combinations you are talking about.

A voltage drop of 0.6-0.7V from the base to the emitter (is that sufficient?)

> > so at 4.9 -> 5.0V.
>
> Again, this is a meaningless statement.  It is only meaningful to talk about
> voltages between two points, not 5V somewhere.  I can often be acceptable to
> talk about voltage at a specific point when there is a clear implied
> understanding of a 0V reference point, usually ground.  However, that
> doesn't apply to your statement either.

"so a potential of 4.9-5.0 V between the emitter of the transistor and
the ground of the zener diode", is that more correct?

> You may think I'm nit picking, but getting these sort of details right is
> essential to technical communication, and it is therefore pointless to
> continue a discussion without them.  Often sloppiness with the details is a
> indication of poor understanding of the fundamentals, which also has to be
> addressed before continuing.

The reason I replied is because I think I don't know enough about the
fundamentals for any bit of analog work. I also "suffer" from the same
problem as Lindy, namely that I can look at a schematic and understand
nothing, I can follow (and occasionally prove wrong) possible ways the
schematic would work but I can't figure out the right one myself.
Thanks for your time already Olin.

2007\02\08@074857 by olin piclist

face picon face
Xiaofan Chen wrote:
> I second that. Even though there are some good electrical engineers
> (eg: Olin) who do not use simulation now, simulations are actually very
> effective in learning electronics and doing electronics design.

I don't think simulation is a good design tool, but it can be useful for
design verification.

I definitely agree that simulation can be useful for learning, especially in
cases like this where everything is used in its normally intended mode of
operation.  Even the dumbest of simulators should be able to show what
happens in this zener circuit well enough, and then allow for quick
experiments and show the resulting voltages and currents.


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000.

2007\02\08@083312 by Peiserma

flavicon
face
piclist-bounces@mit.edu wrote:
> PSpice student version is good and very easy to use. I've
> been a long time PSpice and Saber simulator user and I like
> them very much. They are very good for analog electronics
> learning and practical design.

I have played with SwitcherCAD from linear technology in the past. Free and lets you add your own parts. I'm not a big user of simulators in general, but they can be useful tools. I'd be cautious to depend on them for design, though.

2007\02\08@091625 by Lindy Mayfield

flavicon
face
Gerhard Fiedler wrote:


Until you can do that, you may want to look into simulation, as others have
said before. While simulation doesn't work well for the marginal cases and
"dirt effects" (because they are not simulated well or not at all in the
models), it works quite well for normal operation -- and to some degree
better than with breadboards, because it's easier to just look at the
current that flows in a wire or at the power that's dissipated in a part.
Or see dozens of voltage levels and currents change on the fly when you
change something.

There's even a not so expensive simulation package that allows you to add a
PIC to the simulation and run firmware on it (Proteus VSM
<http://www.labcenter.co.uk>).

Gerhard

Lindy answers:

Do you mean the Proteus VSM for Microchip PIC16 (Complete) for £295?

That's a tad steep, but affordable if it will help me in my learning.  And I mostly need to learn how the components react with each other and not anything super complex.          

2007\02\08@093217 by Robert Young

picon face
Olin said:
>
> I don't think simulation is a good design tool, but it can be
> useful for design verification.
>
> I definitely agree that simulation can be useful for
> learning, especially in cases like this where everything is
> used in its normally intended mode of operation.  Even the
> dumbest of simulators should be able to show what happens in
> this zener circuit well enough, and then allow for quick
> experiments and show the resulting voltages and currents.
>

A few "cheap" SPICE type simulators have been proposed.  Linear
Technology has a very good one with good outside support through a Yahoo
group.

Link to download:
http://www.linear.com/company/software.jsp

Link to Yahoo Group:
http://tech.groups.yahoo.com/group/LTspice/

But don't take what any SPICE program tells you to be gospel!  Try and
understand what is supposed to happen by understanding the basic
equations and behaviors first.  Then with SPICE you can experiment and
fiddle a bit.

Rob

2007\02\08@134610 by Steve Baldwin

flavicon
face

On 8 Feb 2007 at 7:50, Olin Lathrop wrote:

> I definitely agree that simulation can be useful for learning,
> especially in cases like this where everything is used in its normally
> intended mode of operation.  

I disagree quite strongly. A simulator at the stage where Lindy is at now
would serve to teach very little. Perhaps it would increase the skills of
dragging, dropping and clicking, but very little electronics learning. It would
be the equivalent of copying some DSP assembler and clicking run. It might
work, but it doesn't teach any DSP.

At the level being discussed, I don't think you need any more space than you
need to eat. An adjustable plugpack supply, a super cheapo multimeter
(maybe two), a breadboard and a notebook will teach far more than a
simulator.

You need to try and take measurements, to duplicate what the textbook says
will happen, ask questions and determine a means of answering them by
experiment and so on. You need to make mistakes and have the smoke
come out.

Steve.





Even the dumbest of simulators should be
> able to show what happens in this zener circuit well enough, and then
> allow for quick experiments and show the resulting voltages and
> currents.
>
>
> ********************************************************************
> Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
> (978) 742-9014.  Gold level PIC consultants since 2000. -

2007\02\08@154311 by Peiserma

flavicon
face
piclist-bounces@mit.edu wrote:
> On 8 Feb 2007 at 7:50, Olin Lathrop wrote:
>
>> I definitely agree that simulation can be useful for learning,
>> especially in cases like this where everything is used in its
>> normally intended mode of operation.
>
> I disagree quite strongly. A simulator at the stage where
> Lindy is at now would serve to teach very little. Perhaps it
> would increase the skills of dragging, dropping and clicking,
> but very little electronics learning. It would be the
> equivalent of copying some DSP assembler and clicking run. It
> might work, but it doesn't teach any DSP.

your argument can just as well apply to building a circuit on a breadboard: connect a 2k2 resistor in series with a 5.6V zener and hook up to a 10V power supply. read zener voltage with an ohm meter, just like your lab manual says. what have you learned that you couldn't get with a simulator?

Alternate viewpoint: for any learning activity, you tend to get out of it what you put into it. If you just blindly follow an example (be it the breadoarded circuit above, spice, dsp code, or whatever) without understanding what's going on, then I agree, you're wasting your time.

But on the other hand if you're trying to learn and understand, then I don't see how spice hinders you. work out the equations on paper and simulate. or work out the equations and build the circuit (assuming you have all those components on hand and dont have to go order them). Change the resistor, voltage, etc and see what happens to the zener voltage and current. Do they follow your equations?

The danger, in my opinion, is when students DO NOT know what a circuit is supposed to do, or how it operates, and they instead use spice to tell them the answer instead of trying to figure it out for themselves. But I didn't get that feeling in the discussion. The other danger is, as you get into more advanced simulation, is a failure to understand the simulation's limitations, but that's a another can of worms entirely.

> You need to try and take measurements, to duplicate what the
> textbook says will happen, ask questions and determine a
> means of answering them by experiment and so on. You need to
> make mistakes and have the smoke come out.

this I think the majority of us would agree with. There's just no substitute for hands-on learning, or letting the magic smoke out of the wires. A simulator can't take the place of a breadboard. But until Lindy can get his area set up, I don't see anything wrong with him using it as a virtual breadboard (keeping the caution above in mind - don't use it to give you the answers) for simple circuits like this.

2007\02\08@164916 by Herbert Graf

flavicon
face
On Thu, 2007-02-08 at 15:43 -0500, peisermaSTOPspamspamspam_OUTridgid.com wrote:
> spamBeGonepiclist-bouncesSTOPspamspamEraseMEmit.edu wrote:
> > On 8 Feb 2007 at 7:50, Olin Lathrop wrote:
> >
> >> I definitely agree that simulation can be useful for learning,
> >> especially in cases like this where everything is used in its
> >> normally intended mode of operation.
> >
> > I disagree quite strongly. A simulator at the stage where
> > Lindy is at now would serve to teach very little. Perhaps it
> > would increase the skills of dragging, dropping and clicking,
> > but very little electronics learning. It would be the
> > equivalent of copying some DSP assembler and clicking run. It
> > might work, but it doesn't teach any DSP.
>
> your argument can just as well apply to building a circuit on a
> breadboard: connect a 2k2 resistor in series with a 5.6V zener
> and hook up to a 10V power supply. read zener voltage with an
> ohm meter, just like your lab manual says. what have you
> learned that you couldn't get with a simulator?

Hehe, LOTS. To somebody COMPLETELY new to this stuff, even something as
simple as you describe can be a VERY effective learning experience.

For example, what happens if you hook up the power supply wrong? What
does that little plus mean? Which side of the zener connects where? What
are those pretty colours on my resistor.

After that, what happens if I reverse the zener, what happens if I
change the resistor (puff of smoke...), etc. You don't get "puff of
smoke" from a simulator.

On top of that, most people retain far more of an experience when it's
"real".

>
> Alternate viewpoint: for any learning activity, you tend to get out
> of it what you put into it. If you just blindly follow an example
> (be it the breadoarded circuit above, spice, dsp code, or whatever)
> without understanding what's going on, then I agree, you're wasting
> your time.

I wouldn't necessarily call it a waste of time. For a complete newbie,
even something like that can give valuable experience.

The first circuit I ever built was from a book (a touch switch for an
LED, used one or two BJTs). I had NO clue how it worked. But the memory
of following the directions and soldering together that circuit, and
having it work, live with me even today. It was a wonderful learning
experience and gave me the motivation to learn how those things all
worked.

> But on the other hand if you're trying to learn and understand,
> then I don't see how spice hinders you. work out the equations
> on paper and simulate. or work out the equations and build the
> circuit (assuming you have all those components on hand and dont
> have to go order them). Change the resistor, voltage, etc and see
> what happens to the zener voltage and current. Do they follow
> your equations?

Equations? Equations come later. If you try and learn electronics by
chugging through all the equations most people will become so annoyed
they'll just give up. Equations are very important, but when first
learning something there's nothing like doing.

There is a fallacy in our education system in that theory is taught
almost exclusively. I've known electrical engineering students who were
EXCELLENT students, great marks, knew the math to a tee. But when they
were presented with a simple BJT circuit, they had ZERO intuitive sense
on how it worked. They immediately went to look up their equations,
trying to figure out what the circuit did (common collector? Is the BJT
in saturation? What beta should I use?). I, OTOH, had experimented with
electronics for a long time, I could look at the circuit and get a
"feel" for what it was doing. Then I choose a couple equations, crunched
a little math and had the result.

My best example of this was in my fourth year of university (I've
mentioned this before). We were in a lab, putting together our final
projects, and I had several fellow students walk up to me for help.
Their questions? Which way does this resistor go in? What end of this
diode is the cathode? These were GOOD students, very good marks, but
something as simple as which side of a diode was which was never
mentioned to them. They knew the theory of a diode, they could quote the
equation relating voltage to current, but they had no sense of what a
diode actually did. These were 4th year electrical engineering students,
weeks away from graduation. That's just wrong.

> The danger, in my opinion, is when students DO NOT know what a
> circuit is supposed to do, or how it operates, and they instead
> use spice to tell them the answer instead of trying to figure it
> out for themselves. But I didn't get that feeling in the discussion.

I completely agree with you there. Spice can often give answers that are
either outright wrong, or simply not possible in the real world (i.e.
spice reporting 1.9kA going through a 1N4001 diode over a span of a few
seconds...)

> The other danger is, as you get into more advanced simulation,
> is a failure to understand the simulation's limitations, but
> that's a another can of worms entirely.

Absolutely.

TTYL

2007\02\08@172425 by Steve Baldwin

flavicon
face
> your argument can just as well apply to building a circuit on a
> breadboard: connect a 2k2 resistor in series with a 5.6V zener and
> hook up to a 10V power supply. read zener voltage with an ohm meter,
> just like your lab manual says. what have you learned that you
> couldn't get with a simulator?

That you should have used a voltmeter instead of an ohmeter.:-)

Steve.

2007\02\08@180656 by Gerhard Fiedler

picon face
Lindy Mayfield wrote:

> Gerhard Fiedler wrote:
>
>> There's even a not so expensive simulation package that allows you to
>> add a PIC to the simulation and run firmware on it (Proteus VSM
>> <http://www.labcenter.co.uk>).
>
> Lindy answers:
>
> Do you mean the Proteus VSM for Microchip PIC16 (Complete) for £295?
>
> That's a tad steep, but affordable if it will help me in my learning.
> And I mostly need to learn how the components react with each other and
> not anything super complex.

There are the various starter kits for £150. But there are also a number of
free simulation packages out there; the thing that makes Proteus
outstanding is that you can run firmware inside your simulation. (Somewhat
limited -- not all peripherals are always supported --, but still.)

So if you don't want to spend a dime, get one of the free simulators. There
have been quite a few discussions here over the past years; you should find
plenty in the archives. Or just try the LT simulator -- or the eating space
home experimenting setup Steve suggested. You don't need a whole fully
dedicated desk to start experimenting. The argument about the smoke coming
out has something going for it :)

But /something/ for you to build (manually or virtually) circuits you
envision and compare how you think they work with how they actually work is
IMO essential for understanding.

Gerhard

2007\02\08@191221 by Alexandre Guimarães

face picon face
Hi,

>Do you mean the Proteus VSM for Microchip PIC16 (Complete) for £295?

>That's a tad steep, but affordable if it will help me in my learning.  And
>I mostly need to learn how the components react with each other >and not
>anything super complex.

   Take a look at switchercad from LT ( http://www.linear.com ). It is free, easy
to use and has some really nice features... It would be perfect for analysis
of small circuits.

Best regards,
Alexandre Guimaraes

2007\02\19@141317 by Lindy Mayfield

flavicon
face
Olin,

I haven't given up.  It's just that the book recommended to me, _Introductory Circuit Analysis_ is over 1100 pages and it will take me a while to get through it properly.

Great book though.  One of the best I've come across so far.

Lindy

More... (looser matching)
- Last day of these posts
- In 2007 , 2008 only
- Today
- New search...