Searching \ for '[EE]: BooBu Converter - low cost buck boost - may' 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=boobu+converter
Search entire site for: 'BooBu Converter - low cost buck boost - may'.

Exact match. Not showing close matches.
PICList Thread
'[EE]: BooBu Converter - low cost buck boost - may '
2001\08\30@174524 by Russell McMahon

picon face
part 1 3489 bytes content-type:text/plain; (decoded 7bit)

Following on from the "buck converter challenge" I was thinking about how
one would implement an ultra low cost boost-buck converter - output can be
below or above input.

After brief thought the BooBu converter evolved. (Please note - not a BooBoo
converter although it may yet be :-).)

It may even work (although I haven't tried it yet).

I'm throwing this into the PICLIST melting pot early on so anyone else who
wants to can contribute to its evolution. By all means add critique but
before adding criticism please do note its "thought experiment" status (and
it's 12:30am here and I'm going to bed and .....

This is based on "my" prior 3 transistor buck circuit.
The circuit will definitely need work and the additions to the basic circuit
are certainly especially inspired (  :-)  ) but I feel it may be able to be
made to work quite well.

Components values where shown should be ignored!
I have taken some liberties in simplification which will probably not be
justified in practice eg Q2 and Q4 will quite possibly need separate base
drive "spreading" resistors.
Lines crossing don't join.
Tee intersections are joins.

I will refer to inductor as "L"

CORE GENERAL OPERATION

Most buck-boost converters works as follows.
Q3 and Q4 are both turned on.
Vin is applied across L1.
Current in L1 increases.

After a suitable time Q3 & Q4 are turned off.
L1 now 'rings" with LH end going to ground and rh end rising to supply load
via D1.

Cycle repeats.

Switching off may be controlled by time period or by current in L.
Switching on may be by time delay or Vout drop.


THIS CIRCUIT

Initial turnon is by RBUK2 biasing Q2 and Q4 on.
LH end of L will rise to Vin.
C2 will charge via R1.
When Vc2 exceeds Vzener_Z1 then  Q1 will be turned on as before.
Q1 on will turn Q2  & Q4 off and L will "ring" and deliver power to output.

The RH end of R1 will be at ground during ringing cycle and C2 will
discharge.
I have shown D2 allowing this discharge but on reflection a resistor in
series with D2 is probably needed or some other means to prevent immediate
discharge of C2 - needs thought. Regardless, when Vc2 reduces Q1 will turn
off initiating another turn on cycle.
However, if the output voltage has risen to the desired level on this cycle
then zener ZBUK1 will conduct and hold Q1 on and Q2/Q4 off until the output
voltage falls  enough to stop zener conduction. If this happens part way
through a discharge cycle L will deliver energy after switch off
contributing to Vout rising and off delay.

More thought is needed about the short cycling which could occur before
desired output voltage is reached. Possibly a capacitor C3 from the
collector of Q2 to Q1 base or R! C2 junction or some similar cheap solution
will become clear. Possibly a capacitor to ground from Z1/Zbuk1 anodes (as
per previous design).


An interesting thought:     If you want the ultimate in low parts count (as
opposed to practicality) it MAY be possible to eliminate Q4 and use Q2 to
serve both as driver (as now) AND as low side inductor switch. This may take
a diode or 2. We would then have a 3 transistor Boost-Buck circuit !!! At
approximately $nothing per transistor I don't see this as a major goal but
it could be fun.

And I bet they would never suggest you should build anything like these
circuits in EE school :-)





regards,


                   Russell McMahon













part 2 4683 bytes content-type:image/gif; (decode)


part 3 154 bytes
--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads


2001\08\30@202910 by Tom Messenger

flavicon
face
Re: [EE]: BooBu Converter - low cost buck boost - may even work

or may not.

At the very least you will need to add base resistors to Q3 and Q4;
otherwise, the one with the lowest Vbe will win.

Best regards,
Tom M.



>It may even work (although I haven't tried it yet).
>
>I'm throwing this into the PICLIST melting pot early on so anyone else who
>wants to can contribute to its evolution. By all means add critique but
>before adding criticism please do note its "thought experiment" status (and
>it's 12:30am here and I'm going to bed and .....

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads


2001\08\30@220616 by Russell McMahon

picon face
part 1 1896 bytes content-type:text/plain; (decoded 7bit)

Here's a revised circuit of the BooBu converter - addition of D3 is the only
change.
This should do exactly what is wanted to hold the converter in boost phase
until all inductor energy is expended.

Again, where shown compionent values should be ignored.

It works (hopefully) like this -

Afetr current input phase with Q3 and Q4 on the voltage on C2 rises, as
before, until Z1 conducts.
This causes regenerative turnoff action and L flips polarity so that the LH
end is at near ground potential and the RH end rises until current flows via
D1 to the load and CBUK2.
Q3 is off and the cathode of D1 has voltage present ONLY because of the
collapsing magnetic field in L.

D3 conducts and holds Q1 on thus keeping the converter in the "off" phase.
C2 discharges to zero to wait anothe rturn on cycle. Only when the energy in
L is dissipated and the rh end of L falls to near ground will D3 stop
conducting. At this stage Q1 again turns on and the cycle repeats EXCEPT
that if this last cycle has caused Vout to reach it's design level then
ZBUK1 wikll be conducting and will hold the conveter off until Vout again
falls.

D3 could instead be a zener and series diode to allow the inductor rh end to
not have to return fully to ground before the cycle restarts.

As before, under certain circumstances a design MAY be able to combine Q2
and Q4 but this would be more as a fun exercise than a useful cost saving.
3 of the 4 transistors are low current NPNs and could be in a single small
outline package leading to a very compact design. 4 x SOT-23 are also fairly
tiny. Even 4 x TO92 pkgs are rarther compact given the single inductor
winding and relatively few other components. The terminally enthusiastic may
be able to eliminate RBUK4 and PBUK1 but again this would be more for fun
than a sensible approach.


regards,



       Russell McMahon


part 2 4883 bytes content-type:image/gif; (decode)


part 3 154 bytes
--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads


2001\08\30@221516 by Tom Messenger

flavicon
face
What Tom meant to say was Q2 and Q4, not Q3 and Q4. Doh.


At 05:33 PM 8/30/01 -0700, you wrote:
{Quote hidden}

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads


2001\08\31@005310 by Russell McMahon

picon face
> At the very least you will need to add base resistors to Q3 and Q4;
> otherwise, the one with the lowest Vbe will win.

Yes.
You may note that at the top I said -

   > I have taken some liberties in simplification which will probably
   > not be justified in practice eg Q2 and Q4 will quite possibly need
   > separate base drive "spreading" resistors

Good to see that someone's actually looked at it closely enough to have
understood this point!
Any thoughts on aspects of the circuits (hoped for) operation which I
haven't covered?

Note my post of a few minutes ago adding D3 from rh end of L to rh end of
RBUK1 thereby (hopefully) adding proper cycle control.

regards

       Russell McMahon

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics


2001\08\31@093706 by Olin Lathrop

face picon face
> After brief thought the BooBu converter evolved.
> ...
> It may even work (although I haven't tried it yet).
> ...
> do note its "thought experiment" status (and
> it's 12:30am here and I'm going to bed and .....
> ...
> The circuit will definitely need work
> ...
> Components values where shown should be ignored!
> ...
> I have taken some liberties in simplification
> ...

Frankly Russel, this is getting a bit tedius.  I certainly don't want to
discourage the exchange of ideas here, but at least a modicum of brain in
gear before fingers in motion would be a small courtesy to everyone else.
As has become obvious over the last week or so, most people on this list are
not electronics experts and certainly not switching power supply experts.
They are trying to learn this stuff and mostly have no good way to tell for
themselves whether a circuit is a good design or some flaky hairbrain scheme
that plays fast an loose with the laws of physics cooked up in a sudden
flash of inspiration while watching TV, talking on the phone, and munching
on junk food all at the same time.

I am also dismayed to see this circuit is based on your earlier one that
several people pointed out had shortcomings.  Yes I know it works in your
particular instance, but I would never commit such a thing to production.  I
don't want to rehash all that, so I'll briefly comment on this new circuit.

> Initial turnon is by RBUK2 biasing Q2 and Q4 on.

And therefore Q3.

> LH end of L will rise to Vin.

I assume the "LH" end is the one connected to Q3?  The other end of LBUK1 is
held low by Q4 being on.

> C2 will charge via R1.

It would be nice if you used consistant and short names for parts.  Having
RBUK1, R1, C2, CBUK2 is a pain.

> When Vc2 exceeds Vzener_Z1 then  Q1 will be turned on as before.
> Q1 on will turn Q2  & Q4 off and L will "ring" and deliver power to
output.

Not for long (see below), and assuming it doesn't enter linear operation.
Once again I see no hysteresis.  You are counting on high gain and phase
shift to produce oscillation, but that is left wide open to part variations.
Can you prove (and I don't mean by example) this will always oscillate?

> The RH end of R1 will be at ground during ringing cycle and C2 will
> discharge.

Now the Q3 end of LBUK1 is called "RH"?!!  Please take a little care.  Make
up decent names, define them, and stick to them consistantly.

> I have shown D2 allowing this discharge but on reflection a resistor in
> series with D2 is probably needed or some other means to prevent immediate
> discharge of C2 - needs thought.

Not much.  A resistor dissapates power and isn't needed as long as D2 can
handle the maximum current of LBUK1 - which of course needs to be calculated
as part of the design.

> Regardless, when Vc2 reduces Q1 will turn
> off initiating another turn on cycle.

I'm going to stop here because going further is wasting time.  Too many
things are major screwed up here.

For one thing, LBUK1 is still fully charged at this point.  Remember, Q1
turned on because the voltage on C2 just reached a threshold.  Since there
is no hysterisis, Q1 will turn off shortly after C2 goes below this
threshold.  In boost mode, the collector of Q3 will be "regulated" to this
threshold level, DBUK2 never conducts, LBUK1 will keep charging until
something fries, ...

I could go on, but this is getting pointless.


********************************************************************
Olin Lathrop, embedded systems consultant in Littleton Massachusetts
(978) 742-9014, spam_OUTolinTakeThisOuTspamembedinc.com, http://www.embedinc.com

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics


2001\08\31@093714 by Olin Lathrop

face picon face
> Here's a revised circuit of the BooBu converter - addition of D3 is the
only
> change.

Aargh!  This one is so bad it hurts.

I give up.  Someone else want to explain this to him?


********************************************************************
Olin Lathrop, embedded systems consultant in Littleton Massachusetts
(978) 742-9014, .....olinKILLspamspam@spam@embedinc.com, http://www.embedinc.com

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics


2001\08\31@103844 by Russell McMahon

picon face
Olin said
> Frankly Russel, this is getting a bit tedius.
etc etc .....

Sorry about LH & RH confusion - I got one wrong at one point.
LH = left hand and RH = right hand - not necessarily obvious to people who
speak English as a second language.

As I said -

> > do note its "thought experiment" status

and

> > The circuit will definitely need work

That said, I believe that it's not as bad as you seem to think.
However, rather than traipsing through a detailed series of counterpoints
and rebuttals, I will build one at some stage and report on the working (or
not :-) ) circuit. A few brief comments only given below.

As I noted - there are various simplifications which may well prove
unjustified in practice.
This circuit is less likely to be useful than my original 3 transistor buck
only circuit.

As discussed repeatedly lately, hysteresis is present but not by normal
static means.
This can be increased (and established more formally) with a capacitor at
the junction of ZBUK1 and RBUK1.
You can certainly defeat the hysteresis mechanism by suitable "design
decisions".

The reason for the varying length but consistent naming scheme is that the
long names come from the original cicruit whereas the short names are more
recent changes.  In any future circuits I will rename all parts with the
same system.

The diagram with D3 added is simplified. I realise that there are back paths
through the zeners which can be corrected with extra diodes. I noted that a
hold-off mechanism would be needed. The point was that D3 provides the
hold-off mechanism need to discharge the inductor fully. Exact
implementation will obviously differ from this.

I'll let you know when / if I have a working version to demonstrate :-)
Not liable to be in the immediate future alas.




regards

           Russell McMahon

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics


2001\08\31@110146 by Scott Stephens

picon face
Sometimes its the question, not the answer.

How do you do a Monte Carlo yield analysis on nonlinear circuits with spice?
That would be a good course to take. One day I want to use spice to
simulating switch-mode circuits in the time domain. I have a feeling it will
require simulation in different modes. Where can I find a tutorial?

Having had to repair a seven transistor similar design, I would take the
no-brainer approach and run for da pill bottle - those darling 8-pin
switchmodes put out by NS and Linear, Maxim, et. with all those app notes
and recommended magnetics. I'm sure they make good money selling happy pills
to 'experts' who have proved their brilliance by designing expensive
abominations 8^)

Unless they sell hanger queens to the DOD. The predatory exploitation of
weakness and dependance 8^)

Scott

****************************************************************
Freedom is pursuing your carrot, not running from a stick.
The mob only rules what its members are allowed to achieve.
"Don't sweat the small things" - Dale Carnegie
"just bricks in the wall" - Pink Floyd
****************************************************************

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics



'[EE]: BooBu Converter - low cost buck boost - may '
2001\09\01@135821 by Peter L. Peres
picon face
> spice

There is nothing you can't do with spice. Some of the tricks involve using
slowly variable supplies added to a noise source (piecewise linear
approximation of noise using a voltage source f.ex.) to simulate for
example a ramping input voltage with noise on it, or a quickly switched
load. Just plan the campaign right and do it.

I think that there is no 'tutorial' on this because you can spend a
lifetime just explaining spice examples. There is a book however, and I
have not gotten to read it ;-).

Peter

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


2001\09\01@152549 by Dave Dilatush

picon face
Peter L. Peres wrote...

>There is nothing you can't do with spice.

As a long-time SPICE user and a devoted believer in its usefulness,
statements like this make me cringe.

SPICE is only a tool.  If used properly, it can save much time and
allow an engineer to significantly improve a design's chances of
ultimate success.  In the hands of an experienced engineer who works
in a disciplined fashion and who is respectful of SPICE's limitations,
it can be coaxed into yielding valuable information.

But in the hands of an inexperienced designer, SPICE is a disaster
waiting to happen: it's got a zillion ways to lull the unsuspecting
into complacency, sloppiness, laziness and terrible mistakes.  
SPICE doesn't tell you what your circuit is going to do- it tells you
what its mathematical model of your circuit does.  And your circuit,
and SPICE's model, are two different things.

A true connaisseur takes his SPICE with a grain of salt.

Dave

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


More... (looser matching)
- Last day of these posts
- In 2001 , 2002 only
- Today
- New search...