diyAudio (
-   Power Supplies (
-   -   Combined PFC AC/DC & DC/DC SMPS (

ChocoHolic 29th March 2005 08:28 PM

2 Attachment(s)
Hi SMPS fans:

For my subwoofer I will need a SMPS.

My dreams:
- AC Input voltage range: 120V – 277V , additional +/-10%
- Power Factor Correction: Yes, THD < 15% , at low loads 30%
- DC Input for car use.
- Power Rating: Suited for 1kW class D amp of subwoofer

Well, of course first I was thinking about a traditional push pull
car smps for the DC/DC part.
For a line isolated PFC AC/DC converter a flyback always looks charming. :clown:
No, not really… it is a :headshot:
After some short calculations a flyback
for 1kW clearly prooved to be a walk through hell ! Component loads will clearly exceed any reasonable limits.

My new idea seems to be much better.
The DC/DC part will still be a traditional car push pull smps.
The PFC AC/DC converter would now be a composite of:
- PFC boost converter (high efficiency, easy to handle, perfectly reliable)
- From the output rail of the PFC a half bridge is driving an additional primary
winding of the push pull transformer.

Of course you can only use AC input or car Battery. And the not used MosFets must be
turned off reliably….

Please refer to the attached schematic. It is just a principle drawing, but should help
to convey my idea. Only the components for the basic function are shown.
It is clear to me that I will need EMI filtering, may be some snubbers, etc….
Also it is clear to me that I will have to implement a safety isolation according
EN 60065 (170 pages full of safety requirements and lots of chances for misinterpretation…).
Don’t worry about that. I will manage that and in cases of unsolvable issues I will
simply post my future questions….
Also please don’t worry about the PFC boost. That’s the only section, where
I am 100% sure that I will get it working in a proper way (my everyday business).

My current question is, if you think that this combined topology can work in principle.
I do not see a flaw and it looks to promising to be true.

Looking forward to your comments!
Also additional proposals for the control ICs are welcome. The types which
I have proposed are just a first idea.


ChocoHolic 30th March 2005 06:48 AM

I feel lonely.. :sad:
So my idea must be perfect or complete rubbish.
To find out the truth: I could play with :dice: , or simply try it.

kASD 30th March 2005 09:24 AM

hi choco
hi chocolate man;)
try this link.
it features the qsc PL9.0 amp which incorporates the SMPS with PFC.
download the schematic.

check this out

ChocoHolic 30th March 2005 11:02 AM

Hi Amp man !

Nice insights. Interesting link. Thanks for this!
It will give me further inspiration.
But their SMPS is a push pull forward.
Can't follow your perception that it offers a PFC, that
draws sinusodial current from the mains.

Nevertheless: Really cool that QSC offers their schematics.
Probably they dare to do this, because they know, that this is not enough for copying their amps.

kASD 30th March 2005 11:38 AM


Originally posted by ChocoHolic
Hi Amp man !

Nice insights. Interesting link. Thanks for this!
It will give me further inspiration.
But their SMPS is a push pull forward.
Can't follow your perception that it offers a PFC, that
draws sinusodial current from the mains.

Nevertheless: Really cool that QSC offers their schematics.
Probably they dare to do this, because they know, that this is not enough for copying their amps.

hey buddy,

this schematic is of their discontinued model PL9.0 which is obsolete 2 years ago from now. THis amp is suffering from a problem thats why they have to discontinue it.
The Power factor correction is clearly evident on PFC LOOPS page of this document.
I have access to this document by some other means, qsc doesnt give the schematics of thier regular production models.

The qsc amps are very easy for copy indeed.
try this link for more discontinued models schematics and see how easy are their schematics.

Leolabs 30th March 2005 03:05 PM

Hi Choco,have u build the stuff???

ChocoHolic 30th March 2005 04:24 PM

Hi Kanwar!, I was less irritated by the control loops rather than by the
power electronics circuit...
At first galnce, it looked like there would be no possibility to transfer energy during that time when the mains sine wave is far away from
it's peak. But the clue seems to be the combination of L1 and the switched transformer T1. Quite interesting! Especially that you can
skip a lot of caps, compared to two separate stages.
But it does not look like beeing easy to handle.

@ Leolabs:
Up to know I did not build it. But I plan to do and because of that
I thought it might be a good idea to get some opinions first.
But here don't seem to be opinions.....


N-Channel 30th March 2005 05:01 PM


Sounds like an amp I wanted to do a while ago, minus the DC-DC section. I, too, wanted the output of the Active PFC to be a couple of series caps to provide a centerpoint for the half-bridge main power transformer. I have a pic of a 185W Active PFC I built about 5 years ago as my Senior Project for my undergraduate degree. I will post a pic of it when I get to the CPU where it is.

The control chip is OnSemi's MC33262 critical conduction PFC, run in the continuous-conduction mode, and the two output caps are 390uF, each rated at 250WVDC. The output is set for 385-390V.

The main MOSFET is an old Motorola MTW20N50E (20A, 500V), now discontinued. Any 500+ V N-Channel MOSFET from IR or others should do.

The input rectifiers are MUR8100s (8A, 1000V), ultrafast types. These should be ultrafast types as the current pulses being drawn through them are high frequency, modulated at the rectified line frequency of 120Hz (100Hz in Europe).

The flyback diode is also an MUR8100, although there is some ongoing discussion as to whether this diode should be just a fast or even standard-recovery type, depending on whether the circuit is run continuous- or discontinuous mode. I used an '8100, because that's what I had on hand.

I tested this circuit once, and promptly blew its input fuse upon power-up. Since I'm a little reluctant to go messing with AC mains voltages directly, it will be some time before I can get to powering it again, since I don't have a line-isolation transformer handy.

As for the PWM controller, I am using OnSemi’s MC33025. I have used this chip many times with good results. This is pin-for-pin compatible with Unitrode’s UC3825, and is the next-generation of the venerable SG3525, with higher speed oscillator (2MHZ), wider bandwidth error amp, voltage- or current-mode operation, ease of paralleling, and high-current totem-pole outputs(+/-2A each), for driving MOSFETs that have lots of gate capacitance. I'm not slamming the '3525, I still use them, but for those wanting to do an SMPS with all of the things they wish they had in the '3525, they're all in this chip.

If OnSemi was really on the ball, they would combine a dual-output PWM chip with an on-board PFC section to synchronize the two sections and avoid beat frequencies. In case there is such a chip today, I am not aware of it.

Anyway, I digress.

For the gate driver IC, I am using the old Motorola MPIC2113, their version of IR's IR2113, which is a high- and low-side MOSFET driver, rated at 600V, and is specifically designed for driving MOSFETs in half- or full-bridge configurations.

Since I have a few of the MTW20N50Es still laying around, I will use them for the PWM Section, as well.

Since I'm just starting out a new family, it may be some time before I actually get to do this, but I've been gathering the parts for all of this for some time now.

Hope this gives you some ideas that you're not the only one thinking along these lines.

Best of luck,


ChocoHolic 30th March 2005 06:08 PM

Hi Steve,
that's exactly the discussion I am seeking for.

Only point which seems strange to me that you were running your boost in continuos conduction mode. This causes full hard switching.
Especially when the MOSFET is turned on then it is switching into the conductive boost diode. This is giving quite impressive currentpeaks in the diode and in the MosFets you will see the high current peaks and still high voltage at the same time.

Usual you run these boost converters in the critical conduction mode. Exactly at border between continuos and discontinous mode.
Means: Right after demagnetization of the choke, the switch is turned on again. If you use nice timed zero sloping detection, then you can switch on lossless. L6561 and L6562 are providing this. Also there are several similar chips. Furtheron their criteria for MosFet turn off is the current in the MosFet. This makes the system very rugged against disturbances and also against saturating chokes.

I will have a closer look to the UC3825. If I remember right, then I have some of them somewhere in my mess IC assortment....


ChocoHolic 30th March 2005 06:17 PM

My high DC rail of 480 is a result from my 277V/AC wish....
Please note that your 390V are already the lowest value which you can handle at 230V mains. If the difference between mains maximum and the DC rail is coming to small then the zero detector of most
control chips start to make nonsense.... Usually no defects happen, but
control is a little bit "disorganized" :D

OK, we should stop the boost topic here, otherwise the moderators will close this thread..... :judge:

All times are GMT. The time now is 04:43 AM.

Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2019 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 17.65%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2019 DragonByte Technologies Ltd.
Copyright ©1999-2019 diyAudio