Searching \ for '[EE]: DC:DC regulators - what effects efficiency?' 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=regulators+what
Search entire site for: 'DC:DC regulators - what effects efficiency?'.

Exact match. Not showing close matches.
PICList Thread
'[EE]: DC:DC regulators - what effects efficiency?'
2004\05\31@033220 by Peter Mcalpine

flavicon
face
Hi All,
Trying to get my head around DC:DC buck regulators..
Have a circuit that generals runs from 12V input
and has 3.5V output at 150ma.
At present am using an ON Semi MC33063 chip. Although
cheap it uses transistor switches inside, and would like
to be able to have a bit more head room with a FET style
unit..

Anyway, at present with 12Vin, 3.5Vout@ 150ma I am
getting 69% efficiency.

What components most effect this efficiency number?

Also, it seems that the DC:DC solution likes a bit of
current going through it, as when I switch off things
my efficiency drops to 48% (13ma output current).
Should I do things differently?

Thanks!

Regards,
Peter Mcalpine

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

2004\05\31@082448 by Russell McMahon

face
flavicon
face
> Trying to get my head around DC:DC buck regulators..
> Have a circuit that generals runs from 12V input
> and has 3.5V output at 150ma.
> At present am using an ON Semi MC33063 chip. Although
> cheap it uses transistor switches inside, and would like
> to be able to have a bit more head room with a FET style
> unit..
>
> Anyway, at present with 12Vin, 3.5Vout@ 150ma I am
> getting 69% efficiency.
>
> What components most effect this efficiency number?

I like the 33063, but it is very old and clunky BUT very very flexible -
excellent building block if effiiciency is not the main aim.
The 33063 switch is usually a non saturating darlington (unles you take
special & annoying design care to make it saturate) so takes 1v plus across
it which is about 10% plus loss at 12v.
Use an external Schittky flyback diode for lower losses there. As this is in
the 3.5v path a diode drop of 0.5v loses ~~~ 0.5/(3.5+0.5) = 12%+
efficiency.
Inductor can be nasty - 150 mA isn't much BUT be CERTAIN that the inductor
is not saturating at your design current and frequency. Inductor needs to be
rated at about ~~ 400 mA in that application. A saturating inductor will eat
as much of your lunch as you will let it have. What inductor are youi
using? - type number and / or specs?

> Also, it seems that the DC:DC solution likes a bit of
> current going through it, as when I switch off things
> my efficiency drops to 48% (13ma output current).
> Should I do things differently?

Operating cuttent taken by IC becomes a larger % of total as load current
goes down.
If your application is well defined (as it seems to be) then there are any
number of modern ICs (Maxim, LT, natsemi, ...) that will give you 80% to
over 90% at the higher end of your range. Things like FET switches (rather
than non saturating bipolar), synchronous rectification (gold standard) or
schottly flyback diodes  (pretty good), higher frequencies and more help to
improve on 33063. It's an extremely useful chip, but can't compete in niche
applications nowadays.



       Russell McMahon

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

2004\05\31@110119 by Martin Klingensmith

face
flavicon
face
Peter Mcalpine wrote:
> Hi All,
> Trying to get my head around DC:DC buck regulators..
> Have a circuit that generals runs from 12V input
> and has 3.5V output at 150ma.
> At present am using an ON Semi MC33063 chip. Although
> cheap it uses transistor switches inside, and would like
> to be able to have a bit more head room with a FET style
> unit..
>
> Anyway, at present with 12Vin, 3.5Vout@ 150ma I am
> getting 69% efficiency.
>
> What components most effect this efficiency number?

Switching frequency: every time the transistor goes into a transition
period, you lose more energy than if it is just on or off.

Switch: a MOSFET is more efficient than a bipolar transistor most of the
time

Diode: The 'free-wheel' diode conducts when the switch is turned off, so
duty cycle can have some effect on losses.

Hope that helps. A low frequency super-low Rds(on) MOSFET and high speed
schottky diode would have the highest efficiency, but as the frequency
goes down, inductors and capacitors tend to get huge.


>
> Also, it seems that the DC:DC solution likes a bit of
> current going through it, as when I switch off things
> my efficiency drops to 48% (13ma output current).
> Should I do things differently?

Since the MC33063 is a fixed frequency device, it may force the switch
on every cycle, but since no current is being drawn, the duty cycle is
very short. Bipolar devices can be "slow", it may not have enough time
to fully turn on before the control is telling it to shut down. That's
my guess.

--
--
Martin Klingensmith
http://infoarchive.net/
http://nnytech.net/


>
> Thanks!
>
> Regards,
> Peter Mcalpine

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

2004\05\31@152132 by John N. Power

flavicon
face
{Quote hidden}

As the output current decreases, the switching transistor
stays on for shorter times. Eventually, the transition times
(rise and fall) become comparable to the nominal
conduction time. Since efficiency is lower while the switch
is in its linear region, the average efficiency drops. Some
regulators have what is called "burst mode" to deal with
this. If the on time falls below a certain minimum, the
regulator switches to a mode in which all pulses have that
minimum length. Regulation then consists in producing
bursts of those pulses with variable numbers of pulses in
each burst. The burst length controls the output voltage,
but each pulse by itself is long enough to keep efficiency up.
The pulse interval within the burst is the same as in
normal mode, so all magnetics calculations remain valid.

> Thanks!

> Regards,
> Peter Mcalpine

John Power

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


'[EE]: DC:DC regulators - what effects efficiency?'
2004\06\01@030829 by Peter Mcalpine
flavicon
face
Hi Russell,
thanks for the info (and to the others also :)

Inductor we are using is Tyco brand 220uh shielded ferrite core
style pn 3631b221k,
value 220uh 10%
max RDC 0.460ohms
max IDC 0.70A
LQ test frequency 1Khz

I have built another version using a LM2674, and on vero
board it performs similar to the MC33063, so guess with a
good layout and better selected components will get even
better.

I also have a Linear LT1934 sample on order - I think this
one will suit even better as its quiescent current is much
lower than the LM or MC parts, which means i should be able
to get much better eff at my low current levels.

Regards,
Peter

{Original Message removed}

2004\06\01@041412 by Russell McMahon

face
flavicon
face
> Inductor we are using is Tyco brand 220uh shielded ferrite core
> style pn 3631b221k,
> value 220uh 10%
> max RDC 0.460ohms
> max IDC 0.70A
> LQ test frequency 1Khz
>
> I have built another version using a LM2674, and on vero
> board it performs similar to the MC33063, so guess with a
> good layout and better selected components will get even
> better.

Yes. I think that's rather low for the LM2674.
If by LQ test frequency you mean the switching fequency of the converter,
then I suspect that the 1 KHz frequency is far too low for the inductance.
Lets see:
I =~ t.V/L
or t = LI/V
For 220 uH, 0.7A & 12-3.3 = 8v say.
t = 220E-6 x 0.7/8 = 20 usec !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
Your inductor is being well and truly saturated if you have a 1 kHz duty
cycle, unless you have a suitable cycle by cycle current trip operating
(which the 33063 supports).
Try a substantially higher frequency for the 33063.

*** You don't say what catch/flywheel diode you are using. It MUST be a
Schottky if you wish to minimise losses.

The 2764 looks after its own frequency. They claim around 87% efficiency at
500 mA , 3V3 out and 12v in so over 80% at 150 mA is expected.
If available I'd try another inductor of same or similar inductance but
rated at higher current just to see what effect it had on efficiency.

FWIW they recommend 100 uH in your application but 220 uH should be OK.




       Russell McMahon










>
> I also have a Linear LT1934 sample on order - I think this
> one will suit even better as its quiescent current is much
> lower than the LM or MC parts, which means i should be able
> to get much better eff at my low current levels.
>
> Regards,
> Peter
>
> {Original Message removed}

2004\06\02@201608 by Peter Mcalpine

flavicon
face
Hi Russell,
1Khz etc was from the data sheet of the inductor, we are running with
a 220pf cap which on a CRO gives 75khz frequency. circuit current is
around 150ma.. would I be right in then saying..
For 220 uH, 0.15A & 12-3.3 = 8v say.
t = 220E-6 x 0.15/8 = 4.125 usec

Does that sound better?

The catch diode is 10BQ030
Vf = 0.3V

The data sheet on the LT device talks about reverse leakage current
having a large effect on efficiency at lower currents, ie the diode
I am using say typical 0.5mA leakage, and the LT data sheet recommends
one with a uA leakage (MBR0540).
Do you think the diode I have selected is not that good?

Thank you very much for your help :)
Learning all the time.. as you do in this industry :)

Cheers,
Peter



{Original Message removed}

2004\06\02@202648 by Richard.Prosser

flavicon
face
Leakage will increase dramatically as the diode warms up. 0.5mA seems very
high - is this at ambient or max temperature?
Richard P





Hi Russell,
1Khz etc was from the data sheet of the inductor, we are running with
a 220pf cap which on a CRO gives 75khz frequency. circuit current is
around 150ma.. would I be right in then saying..
For 220 uH, 0.15A & 12-3.3 = 8v say.
t = 220E-6 x 0.15/8 = 4.125 usec

Does that sound better?

The catch diode is 10BQ030
Vf = 0.3V

The data sheet on the LT device talks about reverse leakage current
having a large effect on efficiency at lower currents, ie the diode
I am using say typical 0.5mA leakage, and the LT data sheet recommends
one with a uA leakage (MBR0540).
Do you think the diode I have selected is not that good?

Thank you very much for your help :)
Learning all the time.. as you do in this industry :)

Cheers,
Peter



{Original Message removed}

2004\06\02@211501 by Peter Mcalpine

flavicon
face
Hi Richard,
that is at 25'c, at 100'c it states 5mA
..ah, those figures are at Vr=30V
ok, from the graph at vr=10V
25'c = .002ma
50'c = .02ma

That sounds better :)
Helps when you read all the data sheet!

Cheers,
Peter

{Original Message removed}

2004\06\02@220158 by Richard.Prosser

flavicon
face
OK, - Sounds about right.
I've been caught by Schotty leakage current before. - they are
significantly more leaky then "standard" diodes though. Something to do
with the law of conservation of waste !
RP





Hi Richard,
that is at 25'c, at 100'c it states 5mA
..ah, those figures are at Vr=30V
ok, from the graph at vr=10V
25'c = .002ma
50'c = .02ma

That sounds better :)
Helps when you read all the data sheet!

Cheers,
Peter

{Original Message removed}

2004\06\02@223606 by Russell McMahon

face
flavicon
face
> 1Khz etc was from the data sheet of the inductor, we are running with
> a 220pf cap which on a CRO gives 75khz frequency. circuit current is
> around 150ma.. would I be right in then saying..
> For 220 uH, 0.15A & 12-3.3 = 8v say.
> t = 220E-6 x 0.15/8 = 4.125 usec
>
> Does that sound better?

Yes.
75 kHz sets upper time limit on inductor charge time.

(Vchg* efficiency) /(Vout) ~=  Tout/Tchg

(8 * 0.6)/3.3 ~=  1.%:1 or 3:2
At 75 kHz ttotal = 13 uS so Tchg ~~~= 5 uS
Delta I ~= t.V/L = 5E-6 * 8 /220E-6 =~ 200 mA
= OK
Notice all the "~" s :-)

In continuous mode the inductor current is never zero and oscillates around
a mean level which is the average current.
You can expect peak inductor current to be say twice the eman (without
actually working it out).

> The catch diode is 10BQ030
> Vf = 0.3V

Specs look OK. 30V at 1A Schottky.

Efficiency is lower than I'd hope for.
As noted before, if possible try another physically larger inductor of about
the same inductance value "just to see". This would not be your final
component but would show you whether it had a major effect.

FLASHBACK -
You said
> I have built another version using a LM2674, and on vero
board it performs similar to the MC33063,

I assumed you meant that the efficiency was similar. If so, then there's
still something strange. If efficiency is higher then maybe all is OK.


   Russell McMahon

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

2004\06\03@031157 by hael Rigby-Jones

picon face
{Quote hidden}

Oddly enough I have designed/simulated some low power switchers using Linear
Technology's "Switcher Cad" that have actually shown somewhat better
efficiency using a 1N4148 in place of the equivalent Schottky.  Was never
sure if this was genuine or an artifact of the Spice simulator.

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.
=======================================================================
Any questions about Bookham's E-Mail service should be directed to
EraseMEpostmasterspam_OUTspamTakeThisOuTbookham.com.

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email listservspamspam_OUTmitvma.mit.edu with SET PICList DIGEST in the body

2004\06\03@192245 by Peter Mcalpine

flavicon
face
Hi Russel,
The efficiency was a little better with a high load (150ma) by about
2%, but at low loads was worse by about 5%.
but having said all that, there was not a lot of thought put into the
components and there was a fair bit of HF noise on the output.
Interesting thing between the 2 circuits was the MC33063 no load Iinput
was 2ma, where the LM2674 was 4.5ma.. so when I do go to a low load
state the efficiency is worse due to quiescent current.

On a side issue, a module on my board calls for low ESR electros
of 1000uf or higher, as it can draw some high currents when it
transmits (GSM module). so my overall capacitance on the output
of the DC:DC reg is 10uf + 470uf + 2x1000uf - I would assume this would
not effect the DC:DC operation as the calculation are more for
a minimum filter cap... (also has various small nf ceramic on it)

Regards,
Peter


FLASHBACK -
You said
> I have built another version using a LM2674, and on vero
board it performs similar to the MC33063,

I assumed you meant that the efficiency was similar. If so, then there's
still something strange. If efficiency is higher then maybe all is OK.


   Russell McMahon

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

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email @spam@listservKILLspamspammitvma.mit.edu with SET PICList DIGEST in the body

More... (looser matching)
- Last day of these posts
- In 2004 , 2005 only
- Today
- New search...