PSU Design of Chevin...

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.

Does anyone here uses the very simple PSU design of Chevin
(modified lamp balast switching psu)in their projects?
I saw the schematics, and it's really truly simple using Mosfets and a Ferroxcube transformer in a simple self oscillation mode.
No IC's, just a small bank of mosfets, used in a modified lamp ballast application.
(And they ask a lot for it..:))

[Edited by HaaiFaai on 11-22-2001 at 08:07 AM]
Self oscillating psu,see appl. Motorola


It is in original a design I saw at the application books of Motorola and International Rectifiers.
I can recommend you the Power Applications Manual DL410/D REV1 from Motorola.This is real a great book, not only you can find there lineair, also switching and even a class D amplifier.
If you want to know where the Leach amplifier is based on, just browse in this book.
PSU Design

BTW: the supply consists of (one for each channel)

1 ETD 49 Ferroxcube bobine (Philips)
12 times IRF 740 mosfet (hexfet) International Rectifier
small ferrit signal transformer 3C85
Capacitorbank approx. 10000-15000 Uf/450 volts
High speed rectifier diodes
Inrush current limiter
Those are the big parts, not so expensive and definitive lighter then any normal transformer, anyway they use it for their professional amplifiers rated 600-1000 Watt pro channel..
(I can specify details if needed)
I have some schematics of these amps. The power supply uses some IRF740 Fets (6 or 12 depending on model) and an output transformer (ETD49 core) with 3 secondary terminals(+V, -V and a central terminal). But the only thing I do not understand is why they connect this central terminal to the + terminal of loudspeaker instead of the ground point of the amp output stage.¿Can someone explain the reason of this?
The design Chevin uses is a patent from Hafler audio and can be viewed on their website,its the so called grounded
emitter design(with bipolair transistors)
For details on the psu I can scan some articles if you want.
The original design is a modified lamp ballast schematics with no IC's, just a self oscillating mosfet set up.
The trick is done in the ETD 49 transformer which has a small airflow channel to get more output while maintaining
good power characteristics.
The psu is real simple and light weighted.
Btw the advantage of the grounded emitter design is a very simple low voltage driver because the needed voltage swing is done by the output stage itself.
Chevin uses mosfets instead of bipolair in their output stage (I think they are using lateral ones)

[Edited by HaaiFaai on 12-09-2001 at 09:43 AM]
Who cares, as long as it is designed well, can you hear the difference between a LAB Gruppen amplifier/QSC powerlight series and a conventional design?
The usage of switching power supplies in the future will be more and more, and when careful designed it will reduce the rf output.
But I don't want to start a discussion about the usage of switching/high frequency power supplies and conventional psu's because that will be a never ending story about pros and cons...
Thanks for your info HaaiFaai.
Well, I,ve just finished a prototype of a 1000 w. power supply using the Chevin PSU design as starting point. I did some mods on the secondary of the output transformer to obtain +110 volts and an aux secondary to get +-15 volts because my power amp is a Powerhouse IcePower 1000A digital amp sample that needs these voltages.
It works, I connected two identical ETD49 transformers in parallel to obtain the current needed and the oscillator has 12 IRF740, by now everything works, so I,m thinking about changes to get more power ,maybe 2000 watts or so to supply 2 amps. Maybe it is possible to use different Mosfets and a bigger transformer or maybe IGBT transistors. Very clever and simple design and very powerfull.
¿It would be possible to change the small oscillator inductor by a oscillator IC to avoid the manual work?
Good work you get it running! I think the usage of the small signal transformer is just the great simplicity of the design,no IC's but a real high frequency (about 100 kHz)psu
which is light/small because of the ETD 49 transformer.
In the original idea is to use two steering units (two PCB units) one for each ETD 49 transformer. So that makes 12 times IRF 740 for each transformer(about 1 KiloWatt per channel) .If you want I can send you my drawing of the psu.
chevin psu!!

hi folks

i sold those amps in my area long ago!
if someone can e-mail me the psu schematic i will
appreciate it!!

or i can trade info for a new amp design for djs & clubs.
that i am designing using tube input/phase inverter & driver,on the output solid state!

i am trying to make a bulletproof amp that never blows
(djs) & the power of the solid state output

thank you
Disabled Account
Joined 2010
The original idea of self-oscillating power supply aimed at 12V light bulbs was brilliant. A simple solution with excellent efficiency.
Modding this concept for any audio supply ignores the restrictions of that concept: No secondary rectification, designed for pure resistive load, i.e. secondary current is a sinoid.
Rectification with bulk caps of several 10.000uF distort this current from sinoidal to heavy current pulses, increasing stress on primary power MOSFETs significantly .
All in all this this "simple mod" ends in a cheap and ugly design.
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.