Searching \ for '[EE] Chopper in front of LDO?' 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=chopper+front+ldo
Search entire site for: 'Chopper in front of LDO?'.

Exact match. Not showing close matches.
PICList Thread
'[EE] Chopper in front of LDO?'
2007\10\24@145622 by PicDude

flavicon
face
I have been searching for a very-low-cost switching replacement regulator for
a PIC-based application (automotive environment), but not coming up with
much, due to cost and space.  The goal is to find something smaller but not
be too much more in cost.  Efficiency is NOT a priority, except that too much
heat loss results in a heatsinking problem.

The app requires 5V and uses about 200mA on average, up to 250mA
occassionally.

Existing setup = SMA rectifier into a 7805 (TO-220) with a stud and standoff
to hold and heatsink the 7805 onto an (alum.) enclosure.  At 1.6W (average)
lost to heat, the enclosure does get warm, but not hot.  Total parts cost =~
$0.75 in 1k quantities.

For switchers, I came up with many arrangements, including this version that I
prototyped quite some time back...
  www.narwani.net/neil/electronics/78SWxx/index.html
With the parts count, I would definitely need to make this on a separate PCB.  
Total parts cost =~ $2.80 in 1k quantities, but labor is higher.  Other
lower-cost switchers require more expensive and larger passives.

So I came up with another idea recently and wondering if I could pull this
off...  input voltage (11-14V typical) to the emitter of a PNP transistor,
collector to the input of an LDO 5V regulator (SOT-223 or similar).  The
collector would also have a filter capacitor to ground.  The base of the
transistor will be driven by a spare PIC pin and that line would also have a
high-valued resistor to ground..  The idea is this... On power up, the base
resistor ensures that the PNP transistor is on, supplies power to the LDO,
and the PIC powers up.  After setting up ports etc, the PIC sets up the pin
that drives the PNP transistor to pulse on and off continuously, say 50%
duty-cyle for now.  That effectively chops the input voltage in half, with
the filter cap selected to ensure that the voltage at the input of the LDO
does not drop below 5V+dropout.  With a 0.5V dropout, that's 5.5V required,
and 50% duty-cycle means I need at least 11V at the input, which I have.  
With the PNP in front, I should be able to eliminate the input rectifier,
which is there for reverse-voltage protection.  And at 14V max continuous,
the LDO would dissipate 0.4W average, and 0.5W peak.  BTW, I'm proposing that
the PNP would chop at 50% all the time, with no feedback into the PIC to
regulate it.  The LDO would just dissipate the difference as heat.  Also, the
duty-cycle could be (and would probably need to be) set to say 55%.

Here, parts cost is about $1.25 to $1.50 (depending on filter cap required),
which I can live with.  But will it work?  Or is there something else I need
to be concerned with?  Is there a simple formula to calculate the filter cap
value for this current requirement?  Also, off the top of my head, I'm sure I
can get at least a few khz out of the PIC -- is this enough?

Thanks,
-Neil.

2007\10\24@152605 by Mark Rages

face picon face
On 10/24/07, PicDude <spam_OUTpicdude2TakeThisOuTspamavn-tech.com> wrote:
{Quote hidden}

This will not work as described.  The PIC cannot turn the transistor
off.  To turn it off its base voltage needs to go up to the 11-14V
input voltage, but PICs operate on 5V.  You'll need another
common-emittor NPN in there to do the level shifting.

That effectively chops the input voltage in half, with
> the filter cap selected to ensure that the voltage at the input of the LDO
> does not drop below 5V+dropout.

This will not work either.  The voltage isn't chopped in half.  When
the transistor is on, it will charge the capacitor to 12V.  When it is
off, its voltage will decline linearly (since you specified a constant
200mA current.)  Basically, you need to add an inductor in the
circuit.  And a feedback path.   Then you'll have a switching
regulator.

You didn't specify your output voltage tolerance, but the output of a
switching regulator will be fine to run digital circuitry like a PIC.
No LDO required.

In electronics, "chopper" is used to refer to a kind of low-offset
amplifier.   So that may be a confusing term to use here.

Regards,
Mark
markrages@gmail
--
Mark Rages, Engineer
Midwest Telecine LLC
.....markragesKILLspamspam@spam@midwesttelecine.com

2007\10\24@152634 by Martin

face
flavicon
face
Hello Neil,
You almost certainly need an inductor as well as capacitor in front of
the LDO. Adding an 1812 size inductor that can handle 250mA (largest
inductance I see on newegg being 68uH) would imply to me that you'd also
need to switch at at least perhaps 200kHz, to pull a number out of the
air. I picked 1812 size because I've seen your boards and I know how
much space you (don't) have. The inductor adds about $0.50
The MC34063 is about $.30 in high quantity and you still need an inductor.

I don't know though, you could try it without an inductor it just seems
like it would be pretty hard on the filter capacitor and LDO.
-
Martin K

PicDude wrote:
{Quote hidden}

2007\10\24@160138 by Harold Hallikainen

face
flavicon
face
I remember an application note from long ago (National, I think) where
they made a switching regulator out of a linear regulator. If you take a
regulator that has a sense input that has a ground referenced sense input
(not something like the LM317 where the output is always 1.2V above the
sense input), then put an inductor between the output and the filter
capacitor, and connect the sense to the top of the filter capacitor, and
add a catch diode between the regulator output and ground, you have a
switching regulator. When the sense voltage is low, the chip output swings
to the incoming rail. When it's high, it swings to ground.

Anyway, that's what I remember from maybe 30 years ago...

Harold


--
FCC Rules Updated Daily at http://www.hallikainen.com - Advertising
opportunities available!

2007\10\24@195945 by Xiaofan Chen
face picon face
On 10/24/07, Harold Hallikainen <haroldspamKILLspamhallikainen.org> wrote:
> I remember an application note from long ago (National, I think) where
> they made a switching regulator out of a linear regulator. If you take a
> regulator that has a sense input that has a ground referenced sense input
> (not something like the LM317 where the output is always 1.2V above the
> sense input), then put an inductor between the output and the filter
> capacitor, and connect the sense to the top of the filter capacitor, and
> add a catch diode between the regulator output and ground, you have a
> switching regulator. When the sense voltage is low, the chip output swings
> to the incoming rail. When it's high, it swings to ground.
>
> Anyway, that's what I remember from maybe 30 years ago...
>

You can use the LM317 as the PWM controller for
bulk converter. I have seen this in the previous job.

Xiaofan

2007\10\24@201757 by enkitec

picon face

       There is this discrete alternative:

       http://sound.westhost.com/appnotes/an006.htm

       Mark Jordan



On 24 Oct 2007 at 13:01, Harold Hallikainen wrote:

{Quote hidden}

> --

2007\10\24@224914 by Harold Hallikainen

face
flavicon
face

{Quote hidden}

You're right! It's in the application circuits in the datasheet at
http://www.national.com/ds/LM/LM117.pdf .

The discrete circuit posted by another listmember is a little closer to
what I was thinking of, though.

Harold




--
FCC Rules Updated Daily at http://www.hallikainen.com - Advertising
opportunities available!

2007\10\24@234234 by Marcel Duchamp

picon face
I have seen a couple of variants of that '317 circuit that were sold in
the millions.  They both used LM7805 regulators.  One was in a DEC
PDP-11 power supply; the other was in the supply of the Epson FX-80 dot
matrix printer that was ubiquitous in the 80's.

If designed right, it makes a very good switcher.  Not necessarily the
most efficient but ok.

Harold Hallikainen wrote:
{Quote hidden}

2007\10\26@204056 by PicDude

flavicon
face
Hey Martin,

I was thinking there would be some way to filter like a linear AC power
supply, because any inductor for these frequencies will be huge.  But doing
some math now, I realize that the cap will also be huge at these frequencies.  
I had prototyped an MC34063 in the past, but the low cost gets blown away by
the large and relatively expensive passives.  I don't mind inductors, just
large and expensive ones :-)

Cheers,
-Neil.


On Wednesday 24 October 2007 14:26, Martin wrote:
{Quote hidden}

2007\10\26@204339 by PicDude

flavicon
face
On Wednesday 24 October 2007 14:26, Mark Rages wrote:
> This will not work as described.  The PIC cannot turn the transistor
> off.  To turn it off its base voltage needs to go up to the 11-14V
> input voltage, but PICs operate on 5V.  You'll need another
> common-emittor NPN in there to do the level shifting.

Ack! Yes... you're right.


> ...
> Basically, you need to add an inductor in the
> circuit.  And a feedback path.   Then you'll have a switching
> regulator.

Which is not really what I intended, but I see your point.


>
> You didn't specify your output voltage tolerance, but the output of a
> switching regulator will be fine to run digital circuitry like a PIC.

A switcher is where I started, but the cost/space has not worked out so far.  
I do have hope for a part I found today though -- a semtech switcher.  I'll
report when I have more details.

Thanks,
-Neil.

2007\10\26@204442 by PicDude

flavicon
face
I believe I've seen that it was far more complicated (parts count, PCB
real-estate) than many of the switchers on the market today.


On Wednesday 24 October 2007 15:01, Harold Hallikainen wrote:
{Quote hidden}

2007\10\26@205008 by PicDude

flavicon
face
What I'm running into with a lot of these, is that the caps and inductors are
relatively large, and that adds cost and precious PCB space.  In this app,
the 100uh and 100uf's are the deal-killer for my app.

Cheers,
-Neil.


On Wednesday 24 October 2007 19:17, EraseMEenkitecspam_OUTspamTakeThisOuTgmail.com wrote:
>        There is this discrete alternative:
>
>        http://sound.westhost.com/appnotes/an006.htm
>
>        Mark Jordan

2007\10\26@213025 by Xiaofan Chen

face picon face
On 10/27/07, PicDude <picdude2spamspam_OUTavn-tech.com> wrote:
>
> I was thinking there would be some way to filter like a linear AC power
> supply, because any inductor for these frequencies will be huge.  But doing
> some math now, I realize that the cap will also be huge at these frequencies.
> I had prototyped an MC34063 in the past, but the low cost gets blown away by
> the large and relatively expensive passives.  I don't mind inductors, just
> large and expensive ones :-)
>

The main design constraint for DC/DC converters are space constraint,
cost and performance. If you want smaller passive parts (higher switching
frequency) and smaller PWM controller ICs, often the price will go up.
If you use lower cost PWM controller, often you need more spaces.
Often cost and space constraints are the killer.

If the space is limited and efficiency is paramount, then the problem
comes in. I have searched extensively on the high performance
Synchronous Buck Converter IC from Linear Tech/Maxim/TI/National
and finally I have not much choices because of space constraint. It
is amazing that I can not found good DC/DC converter for such a
"simple" requirement (10-32V input, 5V/1.8A output, synchrouns buck)
once space constraint comes into play.

It is rather frustrating to see I still need to use a 10 year old
24bit Sigma-Delta ADC for the current design and I could not
find decent +/-15V OpAMPs with rail-to-rail input/output.
Life as an high performance analog designer is still not easy.

Luckily TI and National are catching up and ADI/Linear Tech/Maxim
are again paying more attention to the situations again (high voltage
applications). In fact, I've getting more emails from ADI/Maxim
regarding the design specifications on ADC/DAC and other things
and ADI will pay us a 2nd visit in Singapore.

Xiaofan

2007\10\26@214223 by Xiaofan Chen

face picon face
On 10/27/07, PicDude <@spam@picdude2KILLspamspamavn-tech.com> wrote:
> On Wednesday 24 October 2007 19:17, KILLspamenkitecKILLspamspamgmail.com wrote:
> >       There is this discrete alternative:
> >
> >       http://sound.westhost.com/appnotes/an006.htm
> >
> What I'm running into with a lot of these, is that the caps and inductors are
> relatively large, and that adds cost and precious PCB space.  In this app,
> the 100uh and 100uf's are the deal-killer for my app.

You may not need 100uH and 100uF. You need to do some calculations
on the design. Maybe you can use the free LTSpice to simulate the design.
They are cheap but the emission is not good.

LTSpice: http://www.linear.com/software/

I will only use those solutions when there are no PWM IC solutions
available. For example, I use similar circuits for universal input
AC/DC sensors (20V-250V AC or DC input) which needs not much
power (often less than 1W from the converter output). There are no
IC solutions as far as I know. I was talking to some IR designers
last time but no results.

Xiaofan

2007\10\27@041417 by Dario Greggio

face picon face
Xiaofan Chen wrote:

> If the space is limited and efficiency is paramount, then the problem
> comes in. I have searched extensively on the high performance
> Synchronous Buck Converter IC from Linear Tech/Maxim/TI/National
> and finally I have not much choices because of space constraint. It
> is amazing that I can not found good DC/DC converter for such a
> "simple" requirement (10-32V input, 5V/1.8A output, synchrouns buck)
> once space constraint comes into play.

Hi Xiaofan, just wanted to point out this one (we already talked about
it on the forum)
http://www.recom-international.com/switching_regulator_R-78xx.html

(I've not tested it yet though)

--
Ciao, Dario

2007\10\27@043037 by Xiaofan Chen

face picon face
On 10/27/07, Dario Greggio <RemoveMEadpm.toTakeThisOuTspaminwind.it> wrote:
> > If the space is limited and efficiency is paramount, then the problem
> > comes in. I have searched extensively on the high performance
> > Synchronous Buck Converter IC from Linear Tech/Maxim/TI/National
> > and finally I have not much choices because of space constraint. It
> > is amazing that I can not found good DC/DC converter for such a
> > "simple" requirement (10-32V input, 5V/1.8A output, synchrouns buck)
> > once space constraint comes into play.
>
> Hi Xiaofan, just wanted to point out this one (we already talked about
> it on the forum)
> http://www.recom-international.com/switching_regulator_R-78xx.html
>
> (I've not tested it yet though)

Looks not bad to me but too big (too tall for me). The SMD version
does not handle 32V. The efficiency at 5V output is inline with
my first design using TI TPS5420 integrated buck. I got 87% at 32V
but that it not good enough for this particular design.

Xiaofan

2007\10\27@065155 by Morgan Olsson

flavicon
face
Den 2007-10-27 10:30:36 skrev Xiaofan Chen <spamBeGonexiaofancspamBeGonespamgmail.com>:

>> it on the forum)
>> www.recom-international.com/switching_regulator_R-78xx.html
>>
>> (I've not tested it yet though)
>Looks not bad to me but too big (too tall for me). The SMD version
> does not handle 32V.

I feel I am in an sililar gap; small cost-effective buck regulators usually can take max 6 to 10V input.
The ones for higher voltage is also bigger and/or requies more external components = to expensive.

In my app i need to convert 10 to 15.5 V (absolute max) down to 5V, 10 to 70mA with good regualton and low ripple.
-And not much EMI, so i believe a compact SMT design is best.

Suggestions?


--
Morgan Olsson

2007\10\27@083641 by Xiaofan Chen

face picon face
On 10/27/07, Morgan Olsson <TakeThisOuTost011EraseMEspamspam_OUTosterlen.tv> wrote:
> In my app i need to convert 10 to 15.5 V (absolute max) down to 5V,
> 10 to 70mA with good regualton and low ripple.
> -And not much EMI, so i believe a compact SMT design is best.
>

Try National LM5009 or similar. There are not many parts for higher
voltage but low current application. The current development
are more focused on low voltage higher current applications.

MAX1836 is another option.

Regards,
Xiaofan

2007\10\27@085002 by Xiaofan Chen

face picon face
On 10/27/07, Xiaofan Chen <RemoveMExiaofancspamTakeThisOuTgmail.com> wrote:
> On 10/27/07, Morgan Olsson <ost011EraseMEspam.....osterlen.tv> wrote:
> > In my app i need to convert 10 to 15.5 V (absolute max) down to 5V,
> > 10 to 70mA with good regualton and low ripple.
> > -And not much EMI, so i believe a compact SMT design is best.
> >
>
> Try National LM5009 or similar. There are not many parts for higher
> voltage but low current application. The current development
> are more focused on low voltage higher current applications.
>
> MAX1836 is another option.
>

LT3470 is another option, especially for your application.
10mA is quite low current.
http://www.linear.com/pc/productDetail.jsp?navId=H0,C1,C1003,C1042,P8810


Xiaofan

2007\10\27@113100 by PicDude

flavicon
face
On Friday 26 October 2007 20:30, Xiaofan Chen wrote:
> The main design constraint for DC/DC converters are space constraint,
> cost and performance. ...

Correct.  I want low cost and small footprint, but don't need the best
efficiency, as long as I don't have to heatsink any components.


> If the space is limited and efficiency is paramount, then the problem
> comes in.

I don't.  Give me 75% efficiency and I should be happy.  It's still much
better than my 36%-efficient linear regulator now.


> I have searched extensively on the high performance
> Synchronous Buck Converter IC from Linear Tech/Maxim/TI/National
> and finally I have not much choices because of space constraint. It
> is amazing that I can not found good DC/DC converter for such a
> "simple" requirement (10-32V input, 5V/1.8A output, synchrouns buck)
> once space constraint comes into play.
>
> ...
> Life as an high performance analog designer is still not easy.

This is why I'm a digital kinda guy. BTW, are your designs "high performance",
or are *you* "high performance"?


> Luckily TI and National are catching up and ADI/Linear Tech/Maxim
> are again paying more attention to the situations again (high voltage
> applications). In fact, I've getting more emails from ADI/Maxim
> regarding the design specifications on ADC/DAC and other things
> and ADI will pay us a 2nd visit in Singapore.
>
> Xiaofan

Agreed.  I've always had great results with National's parts, but I've started
searching the non-popular brands.  I think I found a solution now with a
Semtech part -- SC4517A -- 1.25Mhz, 1.5A internal switch (though I only need
200mA), and MSOP-8 package.  My calculated footprint for the whole circuit is
<0.4 sq in, which is much better than anything else I've found so far, and
total cost for the whole circuit is ~US$1.50 in 1k quantities.  I'll report
when I confirm these details and play with it a bit.

Since you're an analog designer, I would think it would be easy for you to
design a switcher with a PIC10F.

Cheers,
-Neil.





2007\10\27@113522 by PicDude

flavicon
face
I just posted this in another reply on this thread, but the Semtech SC4517A
looks real good right now.  Very small, low cost, and it will handle up to
16V continuously, with 24V max at the input.  Is your application automotive?

Cheers,
-Neil.



On Saturday 27 October 2007 05:50, Morgan Olsson wrote:
{Quote hidden}

2007\10\27@113737 by PicDude

flavicon
face
Actually, now that I think about it, isn't this spec (10-15.5V, 10-70ma) ideal
for a linear regulator?  You'd be dissipating 0.7W max.  Or is this a
battery-powered app that needs super efficiency?

Cheers,
-Neil.


On Saturday 27 October 2007 07:36, Xiaofan Chen wrote:
{Quote hidden}

2007\10\28@080809 by Morgan Olsson

flavicon
face
Den 2007-10-27 18:38:41 skrev PicDude <RemoveMEpicdude2EraseMEspamEraseMEavn-tech.com>:

> Actually, now that I think about it, isn't this spec (10-15.5V, 10-70ma) ideal
> for a linear regulator?  You'd be dissipating 0.7W max.  Or is this a
> battery-powered app that needs super efficiency?

For this card alone a linear is OK.
We use 78L05 today, the version with elongated TO92 body, mounted with short pins to large copper areas.
Works perfectly, but:

In near future the system will be loaded with more extra stuff, so to avoid changing the main power supply it will probably be cheaper to halve the consumption of my board.
"Changing main supply" would be expensive as it is now integrated "no cost" in another device.

Clarification: the load will *vary* dynamically about between 10 to 70mA, mainly as display LEDs go on and off.

I will check the suggested chips.
Thanks.


--
Morgan Olsson

2007\10\28@181914 by Russell McMahon

face
flavicon
face
{Quote hidden}

Yes, alas.
It won't work as expected.

Any system that does not use "energy transformation" must
instead use energy dissipation :-) :-).

In this case there is an implicit (and very real) resistance
in the feed to the filter cap via the chopper. You WILL get
I^2R losses in this resistance. If you make the traces etc
very solid the chopper switch (FET etc) will form the
resistance. There is no way of not dissipating the energy
resistively somewhere.

Energy converters (otherwise SMPS etc) act like an
electronic gearbox.
The norm is inductive storage (as in a classic buck
regulator).

An alternative less seen is to use capacitors charged in
series and discharged in parallel. if the ripple on the caps
is kept small then the energy losses can be low. Doing this
with discrete parts usually gets messy and it's normal to
use an IC based solution. In this case the Vin is > 2 x Vout
so you may have enough headroom for switch losses etc. At
lower current you could use eg a CD4066 but this demand is
in excess of that.

At the 75% efficiency that you've mentioned as acceptable my
GSR would be a possible solution depending on how precise
the 5V needed to be.  The "easiest" solution would be the
most economic commercial buck regulator IC available
operating at 1MHz or so. What has been mentioned looks
good - BUT and external switch FET will add little to size
and will probably be cheaper overall and allow more
flexibility.

It's easy to miss key points when design requirements are
conveyed in the informal manner that this sort of
interchange encourages. (It's an excellent way to explore
things but sooner or later a hard spec is needed.) Given all
design requirements I believe that a solution that met specs
and cost less than has been mentioned so far would be
possible.


       Russell






2007\10\28@191551 by Xiaofan Chen

face picon face
On 10/28/07, PicDude <RemoveMEpicdude2spam_OUTspamKILLspamavn-tech.com> wrote:
> > Life as an high performance analog designer is still not easy.
>
> This is why I'm a digital kinda guy. BTW, are your designs "high performance",
> or are *you* "high performance"?

My current design is supposed to be high performance and pushing
the performance limits within the space constraints. Cost is less an
issue.

> Since you're an analog designer, I would think it would be easy for you to
> design a switcher with a PIC10F.

As far as I think, PIC10F is not a good PWM controller in terms of
features. You need to add extra circuits like soft-start, protection,
driving and other functions to match the modern PWM controllers.
It can be used in certain applications but not in general recommended.

Xiaofan

2007\10\29@045627 by Alan B. Pearce

face picon face
>What I'm running into with a lot of these, is that the
>caps and inductors are relatively large, and that adds
>cost and precious PCB space.  In this app, the 100uh
>and 100uf's are the deal-killer for my app.

Hmm, I have used some switcher chips from Linear technology, and with tiny
low profile inductors from Coilcraft, would easily match that 7805
equivalent that you pictured. With A size case tantalums and very small SMD
resistors, you really cannot get much smaller with anything. Even using 1206
resistors it can be quite small.

2007\10\29@050817 by Alan B. Pearce

face picon face
>I feel I am in an sililar gap; small cost-effective buck regulators
>usually can take max 6 to 10V input. The ones for higher voltage
>is also bigger and/or requies more external components = to expensive.
>
>In my app i need to convert 10 to 15.5 V (absolute max) down to
>5V, 10 to 70mA with good regualton and low ripple. -And not much
>EMI, so i believe a compact SMT design is best.

Check out the Linear Technology range.
http://www.linear.com/pc/viewCategory.jsp?navId=H0,C1,C1003,C1042,C1032

They supply samples at N/C as well, just like Microchip.

2007\10\29@124601 by PicDude

flavicon
face
The 7805-equivalent switcher I pictured was an experiment a while back, and is
way to big to fit in my app.  I've not looked at Linear lately, but I
remember they were quite expensive relative to other competitors and required
a lot of external components, compared say, to National's Simple Switchers.

BUT -- the revelation now is that Semtech seems to have the right part for my
app -- SC4517A.  It actually fits on my app (at 0.36 sq in of total component
footprint using 0603 caps and resistors), and looks like it will come in
under $1.50 in 1k quantities (all other components included in that price).  
I've ordered samples and coincidentally the FAE will be down in Florida near
me tomorrow, so I'm meeting him for a bit.  I'll update once I know more.

Cheers,
-Neil.


On Monday 29 October 2007 03:56, Alan B. Pearce wrote:
> >What I'm running into with a lot of these, is that the
> >caps and inductors are relatively large, and that adds
> >cost and precious PCB space.  In this app, the 100uh
> >and 100uf's are the deal-killer for my app.
>
> Hmm, I have used some switcher chips from Linear technology, and with tiny
> low profile inductors from Coilcraft, would easily match that 7805
> equivalent that you pictured. With A size case tantalums and very small SMD
> resistors, you really cannot get much smaller with anything. Even using
> 1206 resistors it can be quite small.

2007\10\29@130807 by Alan B. Pearce

face picon face
>BUT -- the revelation now is that Semtech seems to have the right
>part for my app -- SC4517A.  It actually fits on my app (at 0.36 sq
>in of total component footprint using 0603 caps and resistors), and
>looks like it will come in under $1.50 in 1k quantities (all other
>components included in that price). I've ordered samples and
>coincidentally the FAE will be down in Florida near me tomorrow,
>so I'm meeting him for a bit.  I'll update once I know more.

Be interested to see how you get on. Don't have a website link handy? I am
on the lookout for small cheap switchers, and so far the LT ones appear to
be OK, space isn't so much of a problem, more the total cost.

2007\10\29@191858 by Xiaofan Chen

face picon face
On 10/30/07, PicDude <RemoveMEpicdude2TakeThisOuTspamspamavn-tech.com> wrote:
>  I've not looked at Linear lately, but I
> remember they were quite expensive relative to other competitors and required
> a lot of external components, compared say, to National's Simple Switchers.

That is true. In terms of power conversions, Linear Tech is still the
king. National is good -- relative small package, cheaper, and normally
good enough, but just not good enough sometimes. TI is also
catching up. ADI is very good in other part of analog (especially
data converter) but lag far behind in the power conversion section.
Maxim is quite all around but can not beat Linear Tech in power
conversion side. Intersil is very good at VRMs for PCs. Semtech,
Fairchild, OnSemi and some other companies have some good
parts as well.

Xiaofan

2007\10\29@235115 by Funny NYPD

picon face
Plus ST semi.

Funny



----- Original Message ----
From: Xiaofan Chen <EraseMExiaofancspamspamspamBeGonegmail.com>
To: Microcontroller discussion list - Public. <RemoveMEpiclistKILLspamspammit.edu>
Sent: Monday, October 29, 2007 7:08:54 PM
Subject: Re: [EE] Chopper in front of LDO?

On 10/30/07, PicDude <picdude2STOPspamspamspam_OUTavn-tech.com> wrote:
>  I've not looked at Linear lately, but I
> remember they were quite expensive relative to other competitors and required
> a lot of external components, compared say, to National's Simple Switchers.

That is true. In terms of power conversions, Linear Tech is still the
king. National is good -- relative small package, cheaper, and normally
good enough, but just not good enough sometimes. TI is also
catching up. ADI is very good in other part of analog (especially
data converter) but lag far behind in the power conversion section.
Maxim is quite all around but can not beat Linear Tech in power
conversion side. Intersil is very good at VRMs for PCs. Semtech,
Fairchild, OnSemi and some other companies have some good
parts as well.

Xiaofan

2007\10\30@001713 by PicDude

flavicon
face
On Monday 29 October 2007 12:07, Alan B. Pearce wrote:
> Be interested to see how you get on. Don't have a website link handy? I am
> on the lookout for small cheap switchers, and so far the LT ones appear to
> be OK, space isn't so much of a problem, more the total cost.


http://www.semtech.com/
The link is long, so search for "SC4517A" in the search box, then look at the
3.3V example in the datasheet.  The pull-up resistor on the enable pin can be
eliminated and EN tied straight to Vin.  Cin and Cout are 2.2uf and are very
low cost in ceramic.  Only bummer is that the compensation network seems to
be required, but with 0603's they don't take much space.

They've got a bunch of other nice parts too.

Cheers,
-Neil.

2007\10\30@002337 by PicDude

flavicon
face
Maxim... <shudder>!  I've slowly worked the Maxim parts out of my products,
except for the DS18B20, DS18S20.

TI's products don't fit *my* needs well, but their parts seem to be very
efficient.  I was playing with their "SWIFT" design tool recently, and it
claimed efficiencies of ~98% !  Not sure if it would actually get that in
practice, but even if it's close, it would still be great.

Cheers,
-Neil.



On Monday 29 October 2007 18:08, Xiaofan Chen wrote:
{Quote hidden}

2007\10\30@010330 by Russell McMahon

face
flavicon
face
part 1 667 bytes content-type:text/plain; format=flowed; charset="iso-8859-1"; (decoded 7bit)

> I am on the lookout for small cheap switchers

See attached "GSR" circuit.

I can see I am going to have to write this up properly.
See attached.
Can be very compact.
Many variants thereof.
Over 10,000 of mine in existence so far.
Efficiency can be OK to good.
Better FET drive gives better efficiency.
Can use a bipolar high side.
Vout is close to Zbuk1 value.

Inductor varies to suit your need - 100 uH here is not
usual. This was a 12-150V input example.

Spread spectrum waveform :-) gives low emissions.

Vout quite stable under load - far better than eg slightly
simpler Black regulator



       Russell



part 2 4235 bytes content-type:image/gif; (decode)


part 3 35 bytes content-type:text/plain; charset="us-ascii"
(decoded 7bit)

2007\10\30@011957 by Russell McMahon

face
flavicon
face
> TI's products don't fit *my* needs well, but their parts
> seem to be very
> efficient.  I was playing with their "SWIFT" design tool
> recently, and it
> claimed efficiencies of ~98% !  Not sure if it would
> actually get that in
> practice, but even if it's close, it would still be great.

It would have to be an extremely "special" case.
Synchronous switch, pure boost or pure buck with Vout very
close to Vin.
Not too much and not too little power (I^R and channel drop
versus standing losses.)

Consider that at say 5V out 2% is 0.1V drop in the output
circuit.
At say 100 mA out that's 0.1/0.1 = 1 ohm total resistance.
That's certainly easily doable, but care is needed. Includes
PCB tracks, synchronous switch loss, inductor loss and more.

Then all you need is a 100% input circuit to go with it :-)



       Russell



2007\10\30@021408 by Xiaofan Chen

face picon face
On 10/30/07, PicDude <KILLspampicdude2spamBeGonespamavn-tech.com> wrote:
> Maxim... <shudder>!  I've slowly worked the Maxim parts out of my products,
> except for the DS18B20, DS18S20.

Ah I was not using many Maxim/Atmel parts in my last job woking for a
medium size company (3000 people worldwide). Now I am working
for a bigger company (20000 people worldwide) so Maxim and Atmel
are again not bad suppliers.

> TI's products don't fit *my* needs well, but their parts seem to be very
> efficient.  I was playing with their "SWIFT" design tool recently, and it
> claimed efficiencies of ~98% !  Not sure if it would actually get that in
> practice, but even if it's close, it would still be great.

LTSpice is the best tool comes out of these vendors. I used the
SWIFT designer for my 1st deign with TPS5420. It is not bad.
I also used the National Simple Switcher designing software.

For a more complicated design, I will try to use PSpice or other
circuit simulator to design. For example, I am using SABER
(too expensive to justify the cost right now) to simulate two new
design ideas coming out of my design. SABER can do a lot of
cool things (system level simulation).

Xiaofan

2007\10\30@053356 by Alan B. Pearce

face picon face
>The link is long, so search for "SC4517A" in the search box,
...
>Only bummer is that the compensation network seems to be
>required, but with 0603's they don't take much space.

Hmm, that is one advantage of the LT3470 that someone else mentioned,
minimises these miscellaneous components.

Thanks anyway, certainly worth a look.

2007\10\30@131708 by Morgan Olsson

flavicon
face
Den 2007-10-30 06:03:44 skrev Russell McMahon <EraseMEapptechspamEraseMEparadise.net.nz>:

>> I am on the lookout for small cheap switchers
>
> See attached "GSR" circuit.

Nice :)

As for parts count, QBUK1 and the two base resistors may be exchanged for one package:
http://www.nxp.com/pip/PDTC123J_SERIES_5.html

I have looked before, but not found any wiht lower series resistance (making it turn off faster)

Theese kinds are one of my favourite components for low series manual production.

--
Morgan Olsson

2007\10\30@154737 by alan smith

picon face
Don't forget Micrel.  They have some great parts, and some new ones that have embedded inductors.

Xiaofan Chen <@spam@xiaofanc@spam@spamspam_OUTgmail.com> wrote:  On 10/30/07, PicDude
wrote:
> I've not looked at Linear lately, but I
> remember they were quite expensive relative to other competitors and required
> a lot of external components, compared say, to National's Simple Switchers.

That is true. In terms of power conversions, Linear Tech is still the
king. National is good -- relative small package, cheaper, and normally
good enough, but just not good enough sometimes. TI is also
catching up. ADI is very good in other part of analog (especially
data converter) but lag far behind in the power conversion section.
Maxim is quite all around but can not beat Linear Tech in power
conversion side. Intersil is very good at VRMs for PCs. Semtech,
Fairchild, OnSemi and some other companies have some good
parts as well.

Xiaofan

2007\10\30@191617 by Xiaofan Chen

face picon face
On 10/31/07, alan smith <spamBeGonemicro_eng2spamKILLspamyahoo.com> wrote:
> Don't forget Micrel.  They have some great parts, and some new
> ones that have embedded inductors.
>

We use Micrel LDOs. They are the rare ones who have a adjustable
nagative LDO. They do have some DC/DC parts but yet up to the
performance level I want now.

Xiaofan

2007\10\30@222323 by Russell McMahon

face
flavicon
face
>> Don't forget Micrel.  They have some great parts, and
>> some new
>> ones that have embedded inductors.

Embedded inductors are getting easier as frequencies climb.
Also, PCB based inductors start to be viable around 10 uH
(depending on other factors) and that's realisable in what
may often be an acceptable amount of PCB space.


       Russell


'[EE] Chopper in front of LDO?'
2007\11\01@073730 by PicDude
flavicon
face
AFAIR, for my apps, Micrel's parts did not have a high enough input voltage
range (10-15V), and were more in the 5V dropped tp 3.3V type apps.


On Tuesday 30 October 2007 14:47, alan smith wrote:
{Quote hidden}

> -

2007\11\01@074210 by PicDude

flavicon
face
3-4 years ago, I had seen PCB-based inductors (and resistors) at a
PCB-manufacturers trade show.  At that point though, they were quite
expensive, but wondering how they would fare today.  Also, I should look into
creating my own inductors on the PCB with bare traces.

Anyone have links to how the math is done for this, and recommendations for
layouts, etc?

Cheers,
-Neil.



On Tuesday 30 October 2007 21:23, Russell McMahon wrote:
> Embedded inductors are getting easier as frequencies climb.
> Also, PCB based inductors start to be viable around 10 uH
> (depending on other factors) and that's realisable in what
> may often be an acceptable amount of PCB space.
>
>
>         Russell

2007\11\01@081054 by Xiaofan Chen

face picon face
On 11/1/07, PicDude <TakeThisOuTpicdude2.....spamTakeThisOuTavn-tech.com> wrote:
> 3-4 years ago, I had seen PCB-based inductors (and resistors) at a
> PCB-manufacturers trade show.  At that point though, they were quite
> expensive, but wondering how they would fare today.  Also, I should
> look into creating my own inductors on the PCB with bare traces.
>
> Anyone have links to how the math is done for this, and recommendations for
> layouts, etc?
>

This is normally called planar inductor/transformer. My current design
uses two planar transformer but it is not designed by me.
I will tell you this is not easy and you might want to leave it for
experts. But this might be related to the fact magnetics design is
the weakest link for my understanding of DC/DC converters.

Xiaofan

2007\11\01@092226 by Russell McMahon

face
flavicon
face
>> 3-4 years ago, I had seen PCB-based inductors (and
>> resistors) at a
>> PCB-manufacturers trade show.  At that point though, they
>> were quite
>> expensive, but wondering how they would fare today.
>> Also, I should
>> look into creating my own inductors on the PCB with bare
>> traces.
>>
>> Anyone have links to how the math is done for this, and
>> recommendations for
>> layouts, etc?


Microchip AN710, which is targeted squarely at RFID
applications, has good information on PCB inductors.

>From long ago memory

       a^2.n^2/(9a+10b)

but An710 will make sense of that, and add a depth term.
I think it may actually be in error as they  seem to have
swapped two terms - one of which is in my above formula. A
web search will confirm this or not.


Gargoyles ..
Hmmm - this agress with Microchip.
Still may be wrong as this may be where they got it :-)


       http://en.wikipedia.org/wiki/Inductor


Formula 4 is the one I am referring to and 3. is the one I
have used for years. .
The change from 3. to 4. may or may not make sense.




       Russell


2007\11\01@094459 by Xiaofan Chen

face picon face
On 11/1/07, Russell McMahon <TakeThisOuTapptechKILLspamspamspamparadise.net.nz> wrote:
> >> Also, I should look into creating my own inductors on the
>>> PCB with bare traces.
> >>
> >> Anyone have links to how the math is done for this, and
> >> recommendations for
> >> layouts, etc?
>
>
> Microchip AN710, which is targeted squarely at RFID
> applications, has good information on PCB inductors.
>

Hmm, now I know where Neil comes from. However, I am not
so sure if pure PCB traces will be good enough for normal
DC/DC converter applications. I think a magnetic core will
be necessary (planar core).

Xiaofan

2007\11\01@094636 by Alan B. Pearce

face picon face
>        http://en.wikipedia.org/wiki/Inductor
>
>
>Formula 4 is the one I am referring to and 3. is the one I
>have used for years. .
>The change from 3. to 4. may or may not make sense.

I would have thought that 5 would be the one to use for a PCB coil - flat
spiral air core.

2007\11\01@150702 by Russell McMahon

face
flavicon
face
>> >> Also, I should look into creating my own inductors on
>> >> the
>>>> PCB with bare traces.

>> >> Anyone have links to how the math is done for this,
>> >> and
>> >> recommendations for
>> >> layouts, etc?

>> Microchip AN710, which is targeted squarely at RFID
>> applications, has good information on PCB inductors.


> Hmm, now I know where Neil comes from. However, I am not
> so sure if pure PCB traces will be good enough for normal
> DC/DC converter applications. I think a magnetic core will
> be necessary (planar core).

I suggested a few days ago that PCB coils up to about 10 uH
were feasible,
This depends on factors such as available PCB area, track
width, acceptable resistance, frequency and more.
If you want many watts then something else or added "planar"
cores will be needed for sensible coil sizes, but for mW up
to maybe a few Watts a PCB only inductor seems reasonable if
you can use a high enough frequency etc



2007\11\01@165016 by PicDude

flavicon
face
On Thursday 01 November 2007 13:55, Russell McMahon wrote:
> I suggested a few days ago that PCB coils up to about 10 uH
> were feasible,
> This depends on factors such as available PCB area, track
> width, acceptable resistance, frequency and more.
> If you want many watts then something else or added "planar"
> cores will be needed for sensible coil sizes, but for mW up
> to maybe a few Watts a PCB only inductor seems reasonable if
> you can use a high enough frequency etc


Correctamundo -- this is what I was looking into evaluating to see if it's
feasible for my apps, especially considering my minimal free PCB space.  The
Semtech solution I'm looking at requires a 4.7uh inductor.

Cheers,
-Neil.

2007\11\01@165254 by PicDude

flavicon
face
On Thursday 01 November 2007 07:37, Russell McMahon wrote:
> Microchip AN710, which is targeted squarely at RFID
> applications, has good information on PCB inductors.

I'll check it out, thanks!


> >From long ago memory
>
>         a^2.n^2/(9a+10b)
> ...

Gotta wonder what else is in your head :-)  Although I recently needed the
sine rule, cosine rule, and velocity/acceleration equations and was able to
whip those outta my head with no difficulty.

Cheers,
-Neil.,

2007\11\01@165338 by PicDude

flavicon
face
On Thursday 01 November 2007 08:44, Xiaofan Chen wrote:
> Hmm, now I know where Neil comes from. ...

Small Caribbean island. :-) :-) :-)

2007\11\01@195602 by Russell McMahon

face
flavicon
face
>> Hmm, now I know where Neil comes from. ...

> Small Caribbean island. :-) :-) :-)

Of ...


   Russell
   (medium+ size Pacific Island)

2007\11\01@223109 by Dave Lagzdin

picon face
On 01/11/2007, PicDude <.....picdude2spamRemoveMEavn-tech.com> wrote:
>
> On Thursday 01 November 2007 08:44, Xiaofan Chen wrote:
> > Hmm, now I know where Neil comes from. ...
>
> Small Caribbean island. :-) :-) :-)
>
>
A land full of Cobra's I thought
:)

2007\11\02@020218 by Russell McMahon

face
flavicon
face
> A land full of Cobra's I thought

But, Vietnam's not an island.
And the cobras are long gone.

       Russell


2007\11\02@134436 by Dave Lagzdin

picon face
I thought at least 427 of them remained...

On 02/11/2007, Russell McMahon <RemoveMEapptechspamspamBeGoneparadise.net.nz> wrote:
>
> > A land full of Cobra's I thought
>
> But, Vietnam's not an island.
> And the cobras are long gone.
>
>         Russell
>
>
> -

2007\11\02@150533 by PicDude

flavicon
face
On Thursday 01 November 2007 18:56, Russell McMahon wrote:
> >> Hmm, now I know where Neil comes from. ...
> >
> > Small Caribbean island. :-) :-) :-)
>
> Of ...
>

Trinidad (& Tobago)

2007\11\02@150624 by PicDude

flavicon
face
On Thursday 01 November 2007 21:31, Dave Lagzdin wrote:
> On 01/11/2007, PicDude <spamBeGonepicdude2@spam@spamspam_OUTavn-tech.com> wrote:
> > On Thursday 01 November 2007 08:44, Xiaofan Chen wrote:
> > > Hmm, now I know where Neil comes from. ...
> >
> > Small Caribbean island. :-) :-) :-)
>
> A land full of Cobra's I thought
>
> :)

Been thumbing through my website I take it? :-)


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