Heeeeey!
I thought it was clear by now that these chips don't like high capacitance?!
They sound bad!!!
It's not a question of PSRR.
So, 1500~2000uf is as far as you can go with an unregulated PSU here, if you don't wanna kill the low-level detail, midband and treble putity. Bass improves, but kills the rest.
Soooo...
Regulated PSU gives much better results, and you can (must) have small caps between the regs and the power op-amp chip.
Only who has tested all this will know.
I'm soon going to make some more tests with a high capacitance PSU (with an old trick) and these chips, but for now it's secret, as I don't know if the results will be good.
I only post schematics that work.
(c) Elso
And if it does work, I will have an explanation of why do these chips sound bad with high capacitance.
Soon,
I thought it was clear by now that these chips don't like high capacitance?!
They sound bad!!!
It's not a question of PSRR.
So, 1500~2000uf is as far as you can go with an unregulated PSU here, if you don't wanna kill the low-level detail, midband and treble putity. Bass improves, but kills the rest.
Soooo...
Regulated PSU gives much better results, and you can (must) have small caps between the regs and the power op-amp chip.
Only who has tested all this will know.
I'm soon going to make some more tests with a high capacitance PSU (with an old trick) and these chips, but for now it's secret, as I don't know if the results will be good.
I only post schematics that work.
(c) ElsoAnd if it does work, I will have an explanation of why do these chips sound bad with high capacitance.
Soon,
re large xfrmer instead of regulation
Elso:
I must disagree with you that a large transformer will provide the regulation.
I had posted this some time back:
400VA Plitron 2 x 22V secondaries and BrianGT 3875 NIGC.
With heavy organ music or other hgih demand music I experienced over 10% variation (sag) on the +/- rails. This caused considerable volume compression of the music greatly reducing the dynamics.
At the same time there was no measurable (with DMM) sag of the AC voltage. Logically therefore any increase in xfrmer size would not have helped.
I then used a 100VA xfrmer 2 x25 V secondarieswith lm338 regs
T03 form. Please note that the output rail voltages were almost identical to the first case as there is a required voltage drop across the regs. Large caps were also used ahead of the regs and the caps on the chip +/- were reduced from 1500 ufd to 63 ufd
Now the sag on +/- rails was far less than 1 % actually close to 0.1 %
But the output from the AC secondaries varied by 10% or more.
The important thing is this DID not get reflected in the DC
rail voltages.
The sound was superior to the original setup and the dynamics were not compressed.
My conclusion is regulation makes up for smaller PS.
I have since regulated the 400VA amp with fine results
Elso:
I must disagree with you that a large transformer will provide the regulation.
I had posted this some time back:
400VA Plitron 2 x 22V secondaries and BrianGT 3875 NIGC.
With heavy organ music or other hgih demand music I experienced over 10% variation (sag) on the +/- rails. This caused considerable volume compression of the music greatly reducing the dynamics.
At the same time there was no measurable (with DMM) sag of the AC voltage. Logically therefore any increase in xfrmer size would not have helped.
I then used a 100VA xfrmer 2 x25 V secondarieswith lm338 regs
T03 form. Please note that the output rail voltages were almost identical to the first case as there is a required voltage drop across the regs. Large caps were also used ahead of the regs and the caps on the chip +/- were reduced from 1500 ufd to 63 ufd
Now the sag on +/- rails was far less than 1 % actually close to 0.1 %
But the output from the AC secondaries varied by 10% or more.
The important thing is this DID not get reflected in the DC
rail voltages.
The sound was superior to the original setup and the dynamics were not compressed.
My conclusion is regulation makes up for smaller PS.
I have since regulated the 400VA amp with fine results
Re: re large xfrmer instead of regulation
Or, regulation allows you to use large caps and "buffer" them from the chip.
That's why I recken a cap multiplier might work well with a GC. It allows very high capacitance without dropping voltage as much as a regulator. Is that what Carlos is testing?
http://sound.westhost.com/project15.htm
BTW: I haven't tried this so I could, and probably am, talking out my ....
SheldonD said:
Or, regulation allows you to use large caps and "buffer" them from the chip.
That's why I recken a cap multiplier might work well with a GC. It allows very high capacitance without dropping voltage as much as a regulator. Is that what Carlos is testing?
http://sound.westhost.com/project15.htm
BTW: I haven't tried this so I could, and probably am, talking out my ....
Re: re large xfrmer instead of regulation
Hi Sheldon,
This is what I wrote about the Plitrons:
http://www.diyaudio.com/forums/showthread.php?postid=492189#post492189
SheldonD said:
Hi Sheldon,
This is what I wrote about the Plitrons:
http://www.diyaudio.com/forums/showthread.php?postid=492189#post492189
Using opa´s as regulators
Hello,
I thouhgt that it is possible to use opa549 in the same way as lm338. Please correct me if I am wrong but I think that it is possible to connect up to +60V to positive rail and 0V to negative rail and set the desired output voltage with resistor as pictured here http://focus.ti.com/lit/ds/sbos093d/sbos093d.pdf in page 12.
And with opa541 you can use single supply (or the other secondary) up to +80V in positive rail.
Hello,
I thouhgt that it is possible to use opa549 in the same way as lm338. Please correct me if I am wrong but I think that it is possible to connect up to +60V to positive rail and 0V to negative rail and set the desired output voltage with resistor as pictured here http://focus.ti.com/lit/ds/sbos093d/sbos093d.pdf in page 12.
And with opa541 you can use single supply (or the other secondary) up to +80V in positive rail.
Power op amp that will regulate
Any body mention the LM12 power opamp? National has one Big momma of an app note on it. Here it is
http://www.national.com/an/AN/AN-446.pdf
Mark
Any body mention the LM12 power opamp? National has one Big momma of an app note on it. Here it is
http://www.national.com/an/AN/AN-446.pdf
Mark
Konnichiwa,
Well, the very best Amplifiers from DNM work in exactly that way, except they use discrete circuitry. My most extreme solid state linestage works that way too.
There are a few caveats.
1) Capacitive Load - basically don't. Make sure the PSU wiring from Regulator Amplifier to poweramp is low capacitance AND low inductance. You may have to play with things for a long time to get them stable.
In my linestage I ended up with a 3-D(imensional) layout with the Regulator IC's above the Linestage IC and rotated so that the PSU line was around 3mm long. The Linestage IC was LM6181, the regulator IC LM6171. It was a fun project and eventually stabel, but not for the fainthearted.
The other way is a way excessive capacitive load witha fairly high ESR, which will mostly swamp the "Amplifier as Regulator" (AAR) HF behaviour and will not move the Amp outside stability. Never played with that myself, as it struck me as counter productive.
2) Make sure the AC Gain of the "AAR" is within the stability margin.
3) Make sure that bandwidth (input and closed loop) of the actual Amplifier is narrower than that of "AAR".
As said, getting things to work reliable may be trickey and fun.
I personally would prefer a different approach to regulating a "gainclone".... I would suggest using a "current boosted" LM317/337 or higher performance equivalent.
A neat way is to feed the output from a generic LM317/337 circuit to the base of a Darlington Power Transistor with a maximum collector current several times that of the LM3875/3886 current limit. If you use a modest RC circuit (say 1K/100uF) between LM317/337 and Transistor base you reduce the noise no end.
The short circuit protection of the LM3875/3886 should protect the power darlington against excessive current. Place the LM317/337 on top of the Transistor in the hotspot (on the plastic surface above the actual chip) and the LM317/337 thermal protection will protect the Darlington transistor too.
The emitter of the Darlington transistor (actually a simple emitter follower) will offer a low output impedance and near unconditional stability into capacitive loads. If you choose a Darlington transistor with a seriously high Ft you can get a very nice sounding regulated supply this way....
Sayonara
Franz G said:
Well, the very best Amplifiers from DNM work in exactly that way, except they use discrete circuitry. My most extreme solid state linestage works that way too.
There are a few caveats.
1) Capacitive Load - basically don't. Make sure the PSU wiring from Regulator Amplifier to poweramp is low capacitance AND low inductance. You may have to play with things for a long time to get them stable.
In my linestage I ended up with a 3-D(imensional) layout with the Regulator IC's above the Linestage IC and rotated so that the PSU line was around 3mm long. The Linestage IC was LM6181, the regulator IC LM6171. It was a fun project and eventually stabel, but not for the fainthearted.
The other way is a way excessive capacitive load witha fairly high ESR, which will mostly swamp the "Amplifier as Regulator" (AAR) HF behaviour and will not move the Amp outside stability. Never played with that myself, as it struck me as counter productive.
2) Make sure the AC Gain of the "AAR" is within the stability margin.
3) Make sure that bandwidth (input and closed loop) of the actual Amplifier is narrower than that of "AAR".
As said, getting things to work reliable may be trickey and fun.
I personally would prefer a different approach to regulating a "gainclone".... I would suggest using a "current boosted" LM317/337 or higher performance equivalent.
A neat way is to feed the output from a generic LM317/337 circuit to the base of a Darlington Power Transistor with a maximum collector current several times that of the LM3875/3886 current limit. If you use a modest RC circuit (say 1K/100uF) between LM317/337 and Transistor base you reduce the noise no end.
The short circuit protection of the LM3875/3886 should protect the power darlington against excessive current. Place the LM317/337 on top of the Transistor in the hotspot (on the plastic surface above the actual chip) and the LM317/337 thermal protection will protect the Darlington transistor too.
The emitter of the Darlington transistor (actually a simple emitter follower) will offer a low output impedance and near unconditional stability into capacitive loads. If you choose a Darlington transistor with a seriously high Ft you can get a very nice sounding regulated supply this way....
Sayonara
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.