I am trying to regulate the output of an sg3525, here's the catch, I need the secondary isolated from the primary, no isolation, no problem.
I am taking 24V up to 75V, but I need the secondary floating so I can use an old crown trick of using a bridge amp with the output of one bridge grounded to main 0V to achieve double the voltage swing around 0V.
many people have mentioned opto isolation of the output from a primary 3525, in open loop all works fine, so are there any references people can point me to? or schematics?
all help gretafully received.
regards
I am taking 24V up to 75V, but I need the secondary floating so I can use an old crown trick of using a bridge amp with the output of one bridge grounded to main 0V to achieve double the voltage swing around 0V.
many people have mentioned opto isolation of the output from a primary 3525, in open loop all works fine, so are there any references people can point me to? or schematics?
all help gretafully received.
regards
i was thinking .....
maybe instrumentation amplifiers can solve this problem ....
i'm not sure though
maybe instrumentation amplifiers can solve this problem ....
i'm not sure though
The most popular way of doing this is with a TL431 on the output driving an optoisolator back toward the PWM.
Send me a PM and I'll attach a schem in a reply.
🙂
Send me a PM and I'll attach a schem in a reply.
🙂
Isolation
Anthony,
I have done several DC-DC converters (+12V to +/-37VDC), all with isolated input and output grounds. I AC couple the two via a 0.01 mF paralleled with a 0.1mF caps. As for getting the error signal back across to the primary side, use an optocoupler, preferably one with tight tolerances, and 100% CTR (Current transfer ratio) to simplify your calculations. Bias the opto's IR LED using a TL431 as poobah suggested, compensating it's error amp on the secondary side.
A good book for reference material on this subject would be Marty Brown's book:
Power Supply Cookbook (EDN Series for Design Engineers)
by Marty Brown
Newnes; 2nd edition May 2001
ISBN: 075067329X
The section entitled "Feedback" (I think that was the name of the chapter) gives very good design procedures with real-world examples which are readily adaptable to our bipolar audio needs.
Hope this helps,
Steve
Anthony,
I have done several DC-DC converters (+12V to +/-37VDC), all with isolated input and output grounds. I AC couple the two via a 0.01 mF paralleled with a 0.1mF caps. As for getting the error signal back across to the primary side, use an optocoupler, preferably one with tight tolerances, and 100% CTR (Current transfer ratio) to simplify your calculations. Bias the opto's IR LED using a TL431 as poobah suggested, compensating it's error amp on the secondary side.
A good book for reference material on this subject would be Marty Brown's book:
Power Supply Cookbook (EDN Series for Design Engineers)
by Marty Brown
Newnes; 2nd edition May 2001
ISBN: 075067329X
The section entitled "Feedback" (I think that was the name of the chapter) gives very good design procedures with real-world examples which are readily adaptable to our bipolar audio needs.
Hope this helps,
Steve
Thanks for the help-
I also got some answers on S.E.D, using the +in biased at Vref/2 and the -in fed from an opto emitter follower with 1K load, the LED of the opto being fed from the output.
I wish to post filter my output for emc reasons, I presume its best to take the feedback pre filter, but I think rain is wet- so who am I to judge?
any more tips appreciated.
regards
Anthony
I also got some answers on S.E.D, using the +in biased at Vref/2 and the -in fed from an opto emitter follower with 1K load, the LED of the opto being fed from the output.
I wish to post filter my output for emc reasons, I presume its best to take the feedback pre filter, but I think rain is wet- so who am I to judge?
any more tips appreciated.
regards
Anthony
poobah said:
just remember that the TL431 feedback loop has to be compensated if used in this manner -- it is an opamp with reference. there was an article last year in EDN which discussed this exact topic -- not bad for a device which costs about $0.25.
http://www.edn.com/contents/images/6255051f2.pdf
http://www.edn.com/contents/images/091505di.pdf
fwiw, the TL431 makes a very nice temp compensated comparator.
Anthony,
You have mail... you have the "whole enchilada"... keep it to your self. If other members have an interest, send me a PM and I will strip down the document to make it "public friendly".
Heed jackinnj's advice carefully. Some of the comppensation networks you see are merely for the benefit of the TL431.
You are quite right, you should NOT take your feedback after the filter. For every pole you add to the forward loop, you will have to add a cancelling zero to the feedback function... gets messy.
Also see APP NOTE 19 from Linear Technologies. It has a concise treatise on compensating the feedback in an SMPS empirically. I normally despise such methods, but they actually work much faster and better than "doing the math".
🙂
You have mail... you have the "whole enchilada"... keep it to your self. If other members have an interest, send me a PM and I will strip down the document to make it "public friendly".
Heed jackinnj's advice carefully. Some of the comppensation networks you see are merely for the benefit of the TL431.
You are quite right, you should NOT take your feedback after the filter. For every pole you add to the forward loop, you will have to add a cancelling zero to the feedback function... gets messy.
Also see APP NOTE 19 from Linear Technologies. It has a concise treatise on compensating the feedback in an SMPS empirically. I normally despise such methods, but they actually work much faster and better than "doing the math".
🙂
- Status
- Not open for further replies.