Carlos stated:
"Go regulated and 150~200VA is enough for two LM3886 chips.
Bass is TIGHT, voltage doesn't sag and the "juice" is always available"
I agree
Some time ago I posted my experience with regulated 100VA
2x25 V transformer with 3875 Amp
Running with loud heavy bass music (pipe organ) and using large caps 10,000 plus the regulation was phenomenal. Much less than 1 % IIRC, the DC seen at the AMPs was very steady. The poor transformer got quite warm (not hot). THe AC out varied considertably but the regulated DC at the chips was steady.
This was much better than a 600VA transformer using the BrianGT stock GC with small PS cap and large caps at Amp chip.
This was true even though the transformer AC output was rock stedy But the DC at the chips was not.... it sagged approaching 8%. Using DMM which do not have the best response.
So it is quited clear that regulation can make a silk purse out of a pigs ear of a transformer.
"Go regulated and 150~200VA is enough for two LM3886 chips.
Bass is TIGHT, voltage doesn't sag and the "juice" is always available"
I agree
Some time ago I posted my experience with regulated 100VA
2x25 V transformer with 3875 Amp
Running with loud heavy bass music (pipe organ) and using large caps 10,000 plus the regulation was phenomenal. Much less than 1 % IIRC, the DC seen at the AMPs was very steady. The poor transformer got quite warm (not hot). THe AC out varied considertably but the regulated DC at the chips was steady.
This was much better than a 600VA transformer using the BrianGT stock GC with small PS cap and large caps at Amp chip.
This was true even though the transformer AC output was rock stedy But the DC at the chips was not.... it sagged approaching 8%. Using DMM which do not have the best response.
So it is quited clear that regulation can make a silk purse out of a pigs ear of a transformer.
Nuuk said:Carlos, you seem to have lost a lot of weight!
I started increasing capacitance slowly.
For unregulated, I would certainly make it with a 200VA trafo and 20,000~30,000uf capacitance per rail than a 500VA trafo with 1000~1500uf.
It's just a question of making that high capacitance to work well with the amp.
And it does work.
Bricolo said:btw, what's the difference between a snubber and a zobel?
That's a good question.
I call a Zobel to an L around an R on the output of an amp.
But an RC to ground is usually called a zobel, when applied to speakers.
But some (like TI/BB) still call that a snubber.
Attachments
Zobel network -- Dr. Leach put this excerpt from his book on his Georgia Tech website:
http://users.ece.gatech.edu/~mleach/ece4445/downloads/zobel.pdf
http://users.ece.gatech.edu/~mleach/ece4445/downloads/zobel.pdf
Exchange.
Hello!
I would really like to try the regulated PSU, but here in Norway the LM338 reg's are very expensive..
Someone want to exchange with me? (sorry my bad englich)
I have many opa's, 549, 541, 637, 134, 2134 and more!
Mail me!!
pks123456789@hotmail.com
regards
PK
Hello!
I would really like to try the regulated PSU, but here in Norway the LM338 reg's are very expensive..
Someone want to exchange with me? (sorry my bad englich)
I have many opa's, 549, 541, 637, 134, 2134 and more!
Mail me!!
pks123456789@hotmail.com
regards
PK
Hi Carlos, the moment that you have been waiting for is here! I have finished my snubberised, regulated GC (SRGC)!
It took me longer than expected as the 'testbed' that I am using had quite small heatsinks and I had to modify it for something larger.
So, what is the verdict from the Dungeon?
Well, this is clearly the best GC PSU that I have tried - apart from the SMPS! I find there is a little more fine detail with the SMPS, particularly run at +/-12 volts. I felt that the timing was better with the SMPS too but in fairness, I have bad stomach ache today and that may have affected how I feel about the music! I will reserve final judgement until I have listened some more.
But I would thoroughly recommend the SRGC too without hesitation. Performance across the board is improved from the standard 1000uF per rail and it is a big improvement on the un-regulated snubberised PSU (at least in my system).
One surprise too. The chips run much cooler than they did the first time that I tried the LM338 regulated PSU! Is this down to the snubbers or the fact that I am using one supply for both channels instead of one each?
Anyway Carlos, a big thank you for this design and here is my humble version. I should say that I copied your design as closely as I could but I used 10,000uF Slitfoil caps and 1200uF Panasonic FC's on the chip pins. Oh, and output voltage is 27 volts (and the regs don't get too hot).
It took me longer than expected as the 'testbed' that I am using had quite small heatsinks and I had to modify it for something larger.
So, what is the verdict from the Dungeon?
Well, this is clearly the best GC PSU that I have tried - apart from the SMPS! I find there is a little more fine detail with the SMPS, particularly run at +/-12 volts. I felt that the timing was better with the SMPS too but in fairness, I have bad stomach ache today and that may have affected how I feel about the music! I will reserve final judgement until I have listened some more.
But I would thoroughly recommend the SRGC too without hesitation. Performance across the board is improved from the standard 1000uF per rail and it is a big improvement on the un-regulated snubberised PSU (at least in my system).
One surprise too. The chips run much cooler than they did the first time that I tried the LM338 regulated PSU! Is this down to the snubbers or the fact that I am using one supply for both channels instead of one each?
Anyway Carlos, a big thank you for this design and here is my humble version. I should say that I copied your design as closely as I could but I used 10,000uF Slitfoil caps and 1200uF Panasonic FC's on the chip pins. Oh, and output voltage is 27 volts (and the regs don't get too hot).
An externally hosted image should be here but it was not working when we last tested it.
Well Nuuk, yes, it is THAT good.
The difference here, against your SMPS, is that this regulated PSU, besides amazingly detailed, allows an LM3886 amp to drive almost every speaker you throw at it.
And this is with just a single PSU for two LM3886 chips (a stereo amp), more than enough.
I am still avoiding paralleling LM3886 chips, which I think may degrade the sound, and concentrating on the PSU, which is the most important thing.
Btw, the "new" unregulated snubberized PSU is at the same level of performance as this.
Very slow progression around here...
PS: please allow the new caps to
-in for some days and listen again.
The difference here, against your SMPS, is that this regulated PSU, besides amazingly detailed, allows an LM3886 amp to drive almost every speaker you throw at it.
And this is with just a single PSU for two LM3886 chips (a stereo amp), more than enough.
I am still avoiding paralleling LM3886 chips, which I think may degrade the sound, and concentrating on the PSU, which is the most important thing.
Btw, the "new" unregulated snubberized PSU is at the same level of performance as this.
Very slow progression around here...
PS: please allow the new caps to
-in for some days and listen again.
And all the solder joints!
I can't say how well the SMPS will drive more difficult loads than mine. What I can say, is what I heard when using it at +/-12 volts would make me go out and get more efficient speakers if I needed them. It was really that good!
New unregulated snubberized PSU? Watch out mate - there are going to be a few people after you if you keep this up!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.