diyAudio (
-   Power Supplies (
-   -   soft switching (

mflorin 21st June 2005 06:11 PM

soft switching

I have a dc/dc converter which is build with UCC2803.The power level is about 15W(15V, 1A all outputs together).It runs at 270kHz, hard-switching, and EMI/RFI and efficiency isn't so great.
The converter is a flyback topology and the stress over the main switch is significant.

I want to change the hard switching technique with the soft one, but I don't want to use a dedicated active clamp or resonant or phase shift controller.

I intend to use the UCC2803 in conjunction with an auxiliary circuit that assure ZVS for the main switch.

Personal, I think that active clamp looks very attractive.

Does anyone build a converter that achieve ZVS using regular PWM controllers (including eventually the logic control circuit for the active clamp switch)?
Could anyone give me some informations or pdf files or links with interesting auxiliary circuits (not IC's) dedicated for single ended converters (flyback, forward) that assure soft switching?


Sch3mat1c 21st June 2005 07:50 PM

I don't think you can do ZVS on flyback? Without resonance (quasi-resonant that is) anyways.


Eva 22nd June 2005 03:45 PM

I suppose that a flyback running at such a high frequency is operating in continuous mode. So an easy way to dramatically reduce EMI and turn-on losses is just to reduce the operating frequency and/or the flyback inductor value, until discontinuous operatrion is achieved for the usual load. The converter may still enter continuous mode during load transients, but this doesn't represent a problem as long as it happens only during a small fraction of the operating time.

mflorin 24th June 2005 07:37 AM

Thanks TIM, EVA.

Of course that a resonant circuit is mandatory for achieving ZVS.
I can't reduce the switching frequency, but a discontinous mode of operation could be a solution.

I find a pdf file with an interesting "auxiliary circuit" made for achieving ZVS in DCM even under light load conditions.

Now I have to think for a logic circuit capable to generate the signal for the auxiliary switch in accordance with the waveforms presented in the pdf file.

Thanks for your thoughts.

Best regards,


I tried to attache the pdf file but is too big.
Search for 00703113-RCP.pdf

Eva 24th June 2005 02:47 PM

Sorry, I'm just curious:

How is that you can increase circuit complexity and component count by adding an additional switch, but you cannot perform such a simple adjustment as reducing switching frequency?

Why 270Khz? Such a high frequency is hardly going to provide optimum efficiency.

mflorin 27th June 2005 07:08 AM

Well, a lower switching frequency means a biger core, and I have space limitations.Furthermore the PCBs and cores are already ordered.

I try to save some money.

I hope that auxiliary circuit could be build on a small PCB in such way that fits in the volume.


Eva 27th June 2005 04:49 PM


Well, a lower switching frequency means a biger core, and I have space limitations.
This isn't necessarily true for flybacks. I was assuming to keep the same core when I proposed to use both a lower frequency and a lower inductance (less turns or longer air gap in the same core) to achieve the slightly higher saturation current required.

If you could tell us more about the design, the optimum operating frequency and primary inductance could be calculated (input voltage range, core energy storage capabilities for maximum practical air gap I^2L, maximum allowed flyback voltage at the primary side, etc..).

mflorin 30th June 2005 07:20 AM

The input voltage range is 5Vdc-20Vdc. In this range the converter should be capable to deliver the full power.

As a transistor is used SUB75N06 (75A, 60V).

I will post the magnetic specs. soon as I get all of them.

Eva 30th June 2005 03:01 PM

My experience with high current >50A MOSFETs in fast switching circuits is that they experiment slow turn-on and current tail at turn-off like bipolar transistors. It appears that the farthest portions of the gate structure are lagging from the ones near the gate terminal bonding wire, or something like that. Also, their input capacitance is not easy at all to manage at 270Khz. I recommend a lower current device.

All times are GMT. The time now is 10:08 PM.

vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2015 DragonByte Technologies Ltd.
Copyright 1999-2015 diyAudio

Content Relevant URLs by vBSEO 3.3.2