Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › Best dual PS for highend hybrid buffer
New Posts  All Forums:Forum Nav:

Best dual PS for highend hybrid buffer

post #1 of 13
1/7/10 at 7:32am
Thread Starter 
'

I am using a dual tread to supply +/-18V to a high end hybrid amp SS buffer. Am I correct in thinking there is something better out there (I'm not a SS guy.)

I've seen Tangent has a new Jung type regulated supply and that Twisted Pair has a Shunt type powersupply. Then there is the Series regulated supply from AMB. I read a post on Diyaudio that a shunt reg is best for analog, but I don't know if that is true or not.

What would the better +/-18V 10W powersupply for me to build for this hybrid tube amp?


'
Reply
post #2 of 13
1/7/10 at 10:27am
Quote:
Originally Posted by regal View Post
'
I read a post on Diyaudio that a shunt reg is best for analog, but I don't know if that is true or not.
From my point of view it is true. You should skip all NFB driven PSUs as it is more suitable for application with more or less constant current draw and necessity of exact voltage. When you care for dynamic behavior rather than it's 12.0V instead of 11.6V, shunt regulators should be optimal.
Reply
post #3 of 13
1/7/10 at 10:36am
X2 go for shunt. you could even check out the salas simplistic shunt regulator on DIYAUDIO, you could build one up on perf and see if you like it.
Reply
post #4 of 13
1/7/10 at 11:31am
Quote:
Originally Posted by majkel View Post
From my point of view it is true. You should skip all NFB driven PSUs as it is more suitable for application with more or less constant current draw and necessity of exact voltage. When you care for dynamic behavior rather than it's 12.0V instead of 11.6V, shunt regulators should be optimal.
Sure, but how dynamic is an output stage deeply biased into class A?
Reply
post #5 of 13
1/7/10 at 11:43am
Thread Starter 
Quote:
Originally Posted by majkel View Post
From my point of view it is true. You should skip all NFB driven PSUs as it is more suitable for application with more or less constant current draw and necessity of exact voltage. When you care for dynamic behavior rather than it's 12.0V instead of 11.6V, shunt regulators should be optimal.
I would worry about DC offset problems with the DC servo if the voltage was swinging around. Headphones are different than speakers.
Reply
post #6 of 13
1/7/10 at 11:55am
I wasn't writing about swinging but the average or settled value.

Quote:
Sure, but how dynamic is an output stage deeply biased into class A?
Just evaluate the current swing thru the headphones. The PSU is always "seen" as a part of the signal chain due to its low AC impedance. When it's something of a band gap type, it acts static, with no delay due to closed loop response producing TIM, IMD and other garbage.
Reply
post #7 of 13
1/7/10 at 11:58am
Thread Starter 
Quote:
Originally Posted by majkel View Post
I wasn't writing about swinging but the average or settled value.



Just evaluate the current swing thru the headphones. The PSU is always "seen" as a part of the signal chain due to its low AC impedance. When it's something of a band gap type, it acts static, with no delay due to closed loop response producing TIM, IMD and other garbage.
Has anyone tried a shunt reg on a headphone amp? I'm afraid DC offset would be a deal killer the more I think about it, no way a DC servo sould react fast enough. I think thats why AMB made the decision to go with a series reg.
Reply
post #8 of 13
1/7/10 at 12:04pm
Quote:
Originally Posted by majkel View Post
Just evaluate the current swing thru the headphones. The PSU is always "seen" as a part of the signal chain due to its low AC impedance. When it's something of a band gap type, it acts static, with no delay due to closed loop response producing TIM, IMD and other garbage.
But a proper class A output stage that stays in class A should draw constant current from the PSU all the time. The current swing into the headphones just becomes part of this constant current.

Quote:
Originally Posted by regal View Post
Has anyone tried a shunt reg on a headphone amp? I'm afraid DC offset would be a deal killer the more I think about it, no way a DC servo sould react fast enough. I think thats why AMB made the decision to go with a series reg.
It depends on the PSRR of the whole circuit as well. I recall AMB simulating feeding the B22 with way lower voltage on one rail, and there were no problems.

And I think AMB went with a series reg because it can supply small or very large amounts of current quite easily. A similar shunt supply would probably get way too hot supplying large currents, and would need to be configured for individual circumstances. It is generally a lot more difficult to make a 'generic' shunt supply.
Reply
post #9 of 13
1/7/10 at 12:12pm
I've been using shunt regulator for months in my headphone amp. However, it uses single supply voltage so I don't care about the symmetry. It's done by the amplifier.
Quote:
But a proper class A output stage that stays in class A should draw constant current from the PSU all the time.
It's true only for bridged class A amplifiers.
Reply
post #10 of 13
1/7/10 at 12:22pm
Quote:
Originally Posted by majkel View Post
It's true only for bridged class A amplifiers.
Yeah, I'm re-reading a thread now that you yourself participated in......
http://www.head-fi.org/forums/f6/do-...ml#post3567005

There is a touch more than zero dynamic draw without bridging or active ground.

Still, I find it hard to believe that something like a S22 would cause significant audible 'problems' in a non-bridged amp compared to a shunt.
Reply
post #11 of 13
1/7/10 at 12:57pm
I wouldn't call it a problem. Negative feedback is nothing wrong per say. It becomes nasty when the loop delay is high because the circuit responds too late. It also limits its speed and slew rate which is the source of intermodulation. This is why it is recommended to use local feedbacks instead of global or just avoiding them. Sure the PSU is not the whole amplifier but the idea is similar, you just "amplify" a constant value in op-amp based closed loop circuits. Shunts "produce" the voltage without further buffering.
When you ask me what is better, I present my point of view. You can have another and I'll be glad to know it. People learn the whole life.
Reply
post #12 of 13
1/7/10 at 1:23pm
Quote:
Originally Posted by regal View Post
Has anyone tried a shunt reg on a headphone amp? I'm afraid DC offset would be a deal killer the more I think about it, no way a DC servo sould react fast enough. I think thats why AMB made the decision to go with a series reg.
not specifically until sometime in the next day or 2. I use the placid shunts ito supply the analogue section of my buffalo with good results. havent tested for DC though TBH as I trusted the TP guys to have dealt with that. I will be using a combination of shunts and series regs in my ackodac and also for a dac mod supplying some burson discrete opamps in the output stage. firing that up sometime in the next couple days and will report back if you like?
Reply
post #13 of 13
1/7/10 at 2:24pm
Quote:
Originally Posted by regal View Post
I'm afraid DC offset would be a deal killer the more I think about it, no way a DC servo sould react fast enough
Can you elaborate more on why you think this would be the case?
Reply
New Posts  All Forums:Forum Nav:
  Return Home
Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › Best dual PS for highend hybrid buffer