LM338 regulated snubberized PSU for audio amplifiers

[GeorgeBoles]

Dear Nuuk,

These things I understand. I was just hoping for some clues, that is all.

Thank-you for the information on your site, too, by the way.

With thanks,
George.

PS ... I still need guidance on what the the current LM3886 "working schematic" is, and the suitability of my transformers for the regulated power supply (regarding the voltage - possibly too low; and the single centre tap on the secondaries.) Sometimes my verbosity gets in the way of my enquiries.

[Nuuk]

OK George, let's try and answer a few questions.

To use the regulated PSU, you do (as you suspected) need to have separate windings on the transformer. Sometimes, it is possible to open up the transformer and saparate the windings that have been joined together to make the centre tap.

If you have a rail voltage of 26 volts you could drop that by 3 volts (I think that is the minimum drop for the LM338) to get 23 volts which would still be OK for the LM3875/LM3886.

Carlos posted the latest schematic recently but I haven't got time to look for it now (should be building a stud wall!) It could have been on this thread or one of his others!

[GeorgeBoles]

Thank-you for your assistance Nuuk.

Carlos' other thread on Improving the LM3886 chip has bubbled up to the surface again, and finally I can read it (been unable to do so from work or home for a few days, don't know why 'cause everything else is OK). Therein lies a link to "cfm ni lm3886.png" labelled version 3, which is identical to Bob's earlier in this thread, which, unless there is something else afoot, answers my question about the "current" schematic.

I guess it is time to bite the bullet, build something and lucky dip some capacitors, probably starting with the Panasonic FC series for Cs and Cm (a la Brian GT), perhaps bypassed with some Vishay MKP 1837's (recommended in several of the Pass Labs "capacitor" threads), and some larger Panasonic cans for the power supply.

I will study the performance of my transformers a bit and work out what to do in the power supply.

Lots of work to do given my old speakers AND amplifier have all passed on and I my son is listening to music through the 3" computer speakers. I don't get a look in.

Woo hooo! Off we go!
Regards,
George.

[carlosfm]

Originally posted by Nuuk
If you have a rail voltage of 26 volts you could drop that by 3 volts (I think that is the minimum drop for the LM338) to get 23 volts which would still be OK for the LM3875/LM3886.

Nuuk, George has a 2x26V trafo, so that should provide around 2x36V.
Could be regulated for 2x26~28V, but as the trafo is CT there's a problem.
Easily solved by a simple surgury on the trafo.

Sorry for the delay, George...
I've been busy in the weekend, and now my Benfica is champion...

[Nuuk]

That's typical George! We are trying to sort out your problem and Carlos is away watching football!

Well done Benfica - see you in the Champions League next season?

[carlosfm]

Originally posted by Nuuk
Well done Benfica - see you in the Champions League next season?

Yes, we'll meet on the Champions League.

[Cortez]

Hi Carlos !

First of all, i like your sharings, thx a lot !

May I have a few questions ? They are not attacks !
I'am just curious, before a try to build the LM338 regulated PS:
- foil caps are the best for 47nF - 100nF or maybe ceramic is also good ?
- why dont you use C and RC filters before the big caps, after the trafo ? Or did you maybe try these places too ?
(earlier Joseph_k in an other thread said, that the snubber here has theoretically also a great benefit)
- whats your theory what peranders askd many times: why to use 100nF paralell with the snubber (1R+100nF) ?
- there is a 47uF cap, direct at the LM338, there is no need for snubbering ?
- and a tip to try: allegedly an L improves the sound also, in a PS between 2 big caps after the graetz

Thanks again, and keep on !

[carlosfm]

Originally posted by Cortez
Hi Carlos !

First of all, i like your sharings, thx a lot !

May I have a few questions ? They are not attacks !
I'am just curious, before a try to build the LM338 regulated PS:
1. - foil caps are the best for 47nF - 100nF or maybe ceramic is also good ?
2. - why dont you use C and RC filters before the big caps, after the trafo ? Or did you maybe try these places too ?
3.- (earlier Joseph_k in an other thread said, that the snubber here has theoretically also a great benefit)
4.- whats your theory what peranders askd many times: why to use 100nF paralell with the snubber (1R+100nF) ?
5.- there is a 47uF cap, direct at the LM338, there is no need for snubbering ?
6.- and a tip to try: allegedly an L improves the sound also, in a PS between 2 big caps after the graetz

Thanks again, and keep on !

Hi Cortez,

I'll try to summarize this:
1. small 63~100V MKT caps are fine there.
2. I didn't try because I have no time... try it, 2x3,300uf with 1R resistor between them should be ok. I use CRC(CS) on my latest unreculated PSU, but I didn't find it necessary on the regulated.
3. Where? Before the regs? I don't think it makes much of a difference.
4. Now you are talking the other PSU (just to avoid confusions). They are bypassing the big caps, of course, and they make part for the trick (not enough, though) for a low impedance PSU. On the amp board the small caps must be located close to the chip's pins. I use them, and I have good results. Feel free to experiment the combinations you want.
5. That's the snubber
6. There are so many things we can do... send me a pair of good, big ($$$$) chokes and I'll report the results.

Guys, I don't have the time to try everything.
What I report is tested and works, and for now I'm happy with the results.
Until when I don't know, but my schedule is my work, my family and my like, I don't live for this...
I have many interests in audio (digital, analog, speakers, amps, etc), I'm always doing several things at the same time, and I enjoy it... sometimes I stop and think how can I do all this?

You guys go now!
Don't expect everything done for you...

[Cortez]

> 3. Where? Before the regs?
Yes.
> I don't think it makes much of a difference.
Why not ? After Joseph_k's measurements theoretically there is the best place for snubbering.
(Cause of the inductance of the trafo and capacitance of the diodes)

> 5. That's the snubber
I mean the 47uF "under" the LM338 (C5, C6 ?)
So, there is no need for snubbering ?

> Guys, I don't have the time to try everything.
Okok, just asking !
I just didnt know, what youve tested, and what is a principled theory from you.

> You guys go now!
> Don't expect everything done for you...
Ok, thats true, there are too many possibilitys to try, so lets work together !

[carlosfm]

Originally posted by Cortez
> 3. Where? Before the regs?
Yes.
> I don't think it makes much of a difference.
Why not ? After Joseph_k's measurements theoretically there is the best place for snubbering.
(Cause of the inductance of the trafo and capacitance of the diodes)

No, this is different.
My intention was to vastly reduce the (clumsy) output impedance of the LM338 regs.
The snubber is on the output of the regs.

Originally posted by Cortez
I mean the 47uF "under" the LM338 (C5, C6 ?)
So, there is no need for snubbering ?

Ah...
No, there you just need "big" caps, which also reduces output impedance of the reg.
You can go as high as 220uf.

Originally posted by Cortez
Okok, just asking !

Btw, tonight I've been testing and listening...
Go figure, these snubber values are spot-on for this PSU.
The calcs were spot-on for this reg.

This is high-end

Now I'm gonna upgrade the unregulated PSU schematic with my latest changes, things are now getting pretty serious.