High Voltage Regulators
- Thread starter analag
- Start date
Help Support GroupDIY Audio Forum:
Learner
Well-known member
so ahm..... got any full working HV PSU schematic you'd like to share????
Oh and uhm preferably minimal components with stable and reliable performance, if that's not too much to ask. :green:
Oh and uhm preferably minimal components with stable and reliable performance, if that's not too much to ask. :green:
[quote author="Learner"]any full working HV PSU schematic you'd like to share?[/quote]
I just finished laying out a PCB for a "3-in-one" PSU (HV, heaters and phantom power). Nothing groundbreaking - LM317 heaters, unregulated high tension, and TL783 phantom.
It's pretty big at 5"x6", but each section is independent of the other two, so you could cut the board to separate them, or just build the part you need... Also (like with my other PSUs), the IC's are mounted from the copper side, so you can bolt them to the bottom of the chassis for heatsinking. Let me move some things around and get it neater, and I'll share.
Again, it's no out-of-this-world power supply, it's simple and it works.
Peace,
Al.
I just finished laying out a PCB for a "3-in-one" PSU (HV, heaters and phantom power). Nothing groundbreaking - LM317 heaters, unregulated high tension, and TL783 phantom.
It's pretty big at 5"x6", but each section is independent of the other two, so you could cut the board to separate them, or just build the part you need... Also (like with my other PSUs), the IC's are mounted from the copper side, so you can bolt them to the bottom of the chassis for heatsinking. Let me move some things around and get it neater, and I'll share.
Again, it's no out-of-this-world power supply, it's simple and it works.
Peace,
Al.
Nope, you need the large in/out differential voltage of the TL783 in order to regulate safely on rather-high Zout supplies like HT's.
Jakob E.
VacuumVoodoo
Well-known member
A simple discrete HV regulator is described in detail in this PDF document
A lot of good tube material is on the mother site at http://www.next-power.net/next-tube/index.php3
Some documents are only in russian, some in both english and russian. Perhaps this link should be in a tube-meta..
Alex
A lot of good tube material is on the mother site at http://www.next-power.net/next-tube/index.php3
Some documents are only in russian, some in both english and russian. Perhaps this link should be in a tube-meta..
Alex
Ah...but I have floated the LM317, all it takes is a pre-regulator to hold the input voltage to the LM317 within a specified limit thus creating an even wider voltage input tolerance than the TL783CKC could ever hope to handle. Anyway I decided never to use the three terminal regulators for my B+ ever again. If I'm against IC's in my audio section, why use it in my P.S.
Analag
Analag
Learner
Well-known member
[quote author="alk509"]
I just finished laying out a PCB for a "3-in-one" PSU.....[/quote]
Thanks Al!
Just out of curiosity, how much more do you get by using a regulator anyone??
I just finished laying out a PCB for a "3-in-one" PSU.....[/quote]
Thanks Al!
Just out of curiosity, how much more do you get by using a regulator anyone??
What kinda question is that, the question should be "what is a regulator" then the most learned among us will go into details and lectures about shunt vs series and the benefits there of. I use it primarily for reducing P.S. noise and load induced voltage sag.
Analag
Analag
Learner
Well-known member
[quote author="analag"]What kinda question is that, the question should be "what is a regulator" then the most learned among us will go into details and lectures about shunt vs series and the benefits there of. I use it primarily for reducing P.S. noise and load induced voltage sag.
Analag[/quote]
I meant how how more benefits would you gain performancewise when you comapre a well filtered HV PUS vs a regulated HV PSU?Jah.
Analag[/quote]
I meant how how more benefits would you gain performancewise when you comapre a well filtered HV PUS vs a regulated HV PSU?Jah.
I think it depends on the circuit... I've put TL783's in some tube gear, and I could hear a difference for the better. Then again, I think a lot of what people call "tube sound" has to do with the unregulated supplies found in most tubed boxes... I guess it's really a matter of taste. I have yet to do any real A/B testing with different circuits, but as soon as time permits, I plan to get to it.
When I was building the PSU for my AKM78, Gus and Marik were adamant that I use an unregulated supply for sound reasons... I don't know about you, but I trust those two! :wink:
I like the TL783 better than the LM317HV, BTW.
Peace,
Al.
When I was building the PSU for my AKM78, Gus and Marik were adamant that I use an unregulated supply for sound reasons... I don't know about you, but I trust those two! :wink:
I like the TL783 better than the LM317HV, BTW.
Peace,
Al.
bcarso
Well-known member
[quote author="alk509"]I think it depends on the circuit... I've put TL783's in some tube gear, and I could hear a difference for the better. Then again, I think a lot of what people call "tube sound" has to do with the unregulated supplies found in most tubed boxes... I guess it's really a matter of taste. I have yet to do any real A/B testing with different circuits, but as soon as time permits, I plan to get to it.
When I was building the PSU for my AKM78, Gus and Marik were adamant that I use an unregulated supply for sound reasons... I don't know about you, but I trust those two! :wink:
Peace,
Al.[/quote]
When I did a tube signal path but sand-supported design for an audiophile friend I had final passive decoupling at the tube plate loads etc. The regulators were mainly used to further reduce hum and noise because the low frequency path had a huge amount of gain. The big box was an electronic crossover and the woofer channels compensated for what is practically a free-air big woofer. It does make the speaker system less placement-sensitive for bass though, but oh the power required.
Even at that the regulators were quite a lot of trouble, and I wound up with a pretty complicated discrete design before things had good transient response, were robust, and were adequately quiet.
It's instructive to play with three-terminal regulators and see just how bad the transient response is. Hook up a power FET and a pulse generator and a load to the output and look at the response of the regulator. Of course it's worst-case---no real audio signals will challenge the thing as much as that---but it still shows how slow and ringy most simple regs are.
With a discrete design you can fine-tune the response to be more or less what you want. Again, usually not necessary, but it can be done.
When I was building the PSU for my AKM78, Gus and Marik were adamant that I use an unregulated supply for sound reasons... I don't know about you, but I trust those two! :wink:
Peace,
Al.[/quote]
When I did a tube signal path but sand-supported design for an audiophile friend I had final passive decoupling at the tube plate loads etc. The regulators were mainly used to further reduce hum and noise because the low frequency path had a huge amount of gain. The big box was an electronic crossover and the woofer channels compensated for what is practically a free-air big woofer. It does make the speaker system less placement-sensitive for bass though, but oh the power required.
Even at that the regulators were quite a lot of trouble, and I wound up with a pretty complicated discrete design before things had good transient response, were robust, and were adequately quiet.
It's instructive to play with three-terminal regulators and see just how bad the transient response is. Hook up a power FET and a pulse generator and a load to the output and look at the response of the regulator. Of course it's worst-case---no real audio signals will challenge the thing as much as that---but it still shows how slow and ringy most simple regs are.
With a discrete design you can fine-tune the response to be more or less what you want. Again, usually not necessary, but it can be done.
Yes, when it comes to power circuits, PSU architecture has a lot of influence. Much of the "sound" of classic tube power amplifiers at loud levels has to do with the PSU getting on it's knees during transients. Still we like it very much 
But I feel that low-power tube circuits is a very different thing. Where power amplifiers are usually complementary and thus very insensitive to PSU noise, low-power circuits takes a lot more care with PSU. Here you really want good regulation (Solid-State e.g. '783) to get rid of noise - but you'd want R/C decoupling after it too, because you don't want the SS regulation transient dynamics to influence the rest of the circuit.
As a rule-of-thumb, I use a 470R resistor and a 220uF capacitor after the regulator will reduce regulator-influence to below what is noticeable.
Jakob E.
But I feel that low-power tube circuits is a very different thing. Where power amplifiers are usually complementary and thus very insensitive to PSU noise, low-power circuits takes a lot more care with PSU. Here you really want good regulation (Solid-State e.g. '783) to get rid of noise - but you'd want R/C decoupling after it too, because you don't want the SS regulation transient dynamics to influence the rest of the circuit.
As a rule-of-thumb, I use a 470R resistor and a 220uF capacitor after the regulator will reduce regulator-influence to below what is noticeable.
Jakob E.
SSLtech
Well-known member
[quote author="bcarso"]It's instructive to play with three-terminal regulators and see just how bad the transient response is. Hook up a power FET and a pulse generator and a load to the output and look at the response of the regulator. Of course it's worst-case---no real audio signals will challenge the thing as much as that---but it still shows how slow and ringy most simple regs are.[/quote]
Golden wisdom.
Keith
Golden wisdom.
Keith
One thing I like to do is to resistive load the reg for a good part of the circuits current load.
3 trems are series regs and only source current, So say the circuit uses 40ma I will often use a resistor from out to ground that "pulls" 20 to 40ma. The circuit is still a series reg but there is a path to ground if the voltage "tries" to rise on the rail. Also some 3 terms don't seem to reg good untill about 150 ma.
I think I frist read about doing this in a Jung or Marsh article.
3 trems are series regs and only source current, So say the circuit uses 40ma I will often use a resistor from out to ground that "pulls" 20 to 40ma. The circuit is still a series reg but there is a path to ground if the voltage "tries" to rise on the rail. Also some 3 terms don't seem to reg good untill about 150 ma.
I think I frist read about doing this in a Jung or Marsh article.
featherpillow
Well-known member
I've been wondering about some things along this line. In addition to regulator ringing and sloppy transient response, what about impedance issues? IIRC, there's potential for some type of impedance problem between transformer and diode? Does this cause problems?
bcarso
Well-known member
[quote author="Gus"]One thing I like to do is to resistive load the reg for a good part of the circuits current load.
3 trems are series regs and only source current, So say the circuit uses 40ma I will often use a resistor from out to ground that "pulls" 20 to 40ma. The circuit is still a series reg but there is a path to ground if the voltage "tries" to rise on the rail. Also some 3 terms don't seem to reg good untill about 150 ma.
I think I frist read about doing this in a Jung or Marsh article.[/quote]
That is a good idea, and somewhat analogous to the loading of op amps for enforcing class A operation.
For the highest performance independent of efficiency I like current-source-fed shunt regulators. If efficiency be considered having both sinking and sourcing capability in a regulator can get you close to the same performance.
In other complex systems I have gone to the extreme of having signal-driven compensating loads on power supply lines local to each module so that the current consumption is constant. With balanced design throughout this will tend to happen anyway, but there are ways with single-ended design to do the compensation with fewer parts. These arrangements allow you to use R-C or more complex decoupling networks to keep the circuits isolated at high frequencies, while not incurring the variations due to the higher local PS impedances.
3 trems are series regs and only source current, So say the circuit uses 40ma I will often use a resistor from out to ground that "pulls" 20 to 40ma. The circuit is still a series reg but there is a path to ground if the voltage "tries" to rise on the rail. Also some 3 terms don't seem to reg good untill about 150 ma.
I think I frist read about doing this in a Jung or Marsh article.[/quote]
That is a good idea, and somewhat analogous to the loading of op amps for enforcing class A operation.
For the highest performance independent of efficiency I like current-source-fed shunt regulators. If efficiency be considered having both sinking and sourcing capability in a regulator can get you close to the same performance.
In other complex systems I have gone to the extreme of having signal-driven compensating loads on power supply lines local to each module so that the current consumption is constant. With balanced design throughout this will tend to happen anyway, but there are ways with single-ended design to do the compensation with fewer parts. These arrangements allow you to use R-C or more complex decoupling networks to keep the circuits isolated at high frequencies, while not incurring the variations due to the higher local PS impedances.
