• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

PP power supply design questions

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I plan to build a 6l6 PP (el34, kt66, kt 88 etc.) amp and need some help in want to consider for cap and choke values, etc. My B+ to the plate will be about 360v in ultra liner or triode strapped modes using a ST-70 reissue power transformer.

I do have Duncan amp software but I am seeking help on what issues to avoid in power supply design.

Thanks.........
 
Ex-Moderator
Joined 2004
Hi FD,

First things to decide, which will have a major bearing on the design and choice of components for the PS, are:

1. Will it be a guitar amp or a hi-fi amp?

2. If it's going to be a hi-fi stereo amp, do you intend to build a separate PS for each channel, or a shared PS that will feed both channels?

3. Will the rectification be done using a bridge arrangement (using an untapped power tranny secondary winding) or "full-wave" arrangement (using a center-tapped secondary)?

4. Is your preference for solid-state diodes or vacuum diodes?

5. Do you intend to use a choke-input filter or a cap-input filter?

6. What class of operation have you in mind - Class A or AB?
 
Not to thread-jack, but when designing in PSUDII, is it desirable to get near critical damping or is overdamped voltage response OK? I understand that underdamped/ringing is undesireable, since the voltage gets a little squirrely, but is there a reason not to want slow response?

Is it desireable to have the supply voltage change quickly or slowly? Is there an easily indentifiable sonic difference between the two?

I am going thru the same exercise designing a PS for a PP EL34/KT88 amp.
 
Hey Guys,

Let me add some detail to my request. I found a version of the “simple” HIFI stereo amp using KT77/6l6 tubes where the output section appears to act as a phase inverter. This has been discussed before using 6bq5 tubes. The KT77 article can be found here.


http://diyaudioprojects.com/Tubes/KT77-Push-Pull-Tube-Amp/


It operates in Class A with 350+/- on the output plates. The authors stereo version uses SS rectification. I plan to build a stereo version as well using tube rectification, a ST 70 power trans reissue rated at (360-0-360) 300 ma. My thinking is to use a cap input (to set the voltage)>5 H choke> 200 UF 2nd cap (supplying 350v) which connects to the PP OPTs. The balance of his power supply is one of my concerns. There is a big difference in the amount of UF in his design compared to historic schematics. Should I use more or less UF cap values in the R/C chain to the driver tubes?

My last issue is referencing the driver tube cathodes to the heater. I assume it is possible to use the center tap of the 6.3v heater or use two 100 ohm resistors across the heater wires as the author did. I would then tie this to a 250V reference if I am correct. The ST 70 reissue has two 6.3v heaters. Would I tie in both heaters?.

Thanks for your help.
 
There's so much wrong with that design, I don't know where to begin. First, the output stage as phase splitter throws away many of the advantages of push pull (hum rejection, distortion cancellation). The bible of tube electronics, RDH4, specifically slags this topology. Second, the 317 is a lousy CCS with limited bandwidth. Third, the overload recovery is dicey at best- 317s tend to oscillate under those conditions. Fourth the SRPP is fashionable but in this use has worse linearity than a simple grounded cathode voltage amp. Fifth, the lack of heater supply bypass to AC ground and the lack of HF differential bypass on the heater lines will negatively impact the noise performance.

You can do better.
 
Well... if it were me, I'd design from scratch using a 6SL7 diff amp cascaded to a 6SN7 diff amp, and then a set of either 6SN7 or MOSFET followers to drive the output stage grid. But doing that design might be a bit much for someone with less gray hair than me.:D IIRC, Ray Moth had a design done right along those lines, and Ray knows what he's doing. You might want to ping him or do a forum search for his schematics.
 
The 317 may be a lousy CCS for high impedance loading, but
here we only tail low impedance cathodes... I'm sure here at
this point, lousy as it might be, a lot higher Z than a cathode.
I do not think this topology has any problem rejecting ripple,
as long as output tube GM's are reasonably matched.

-------------------------------------------------------------------

Secondly, the input totem is not exactly wired as an SRPP.
I'd have described it as a Mu-Follower. And as such, might
intentionally behave single ended, rather than balanced...

Tap bottom end of the upper cathode resistor: SRPP
Tap top end of the upper cathode resistor: Mu Follower
Tap middle of the upper cathode resistor: Anti-Triode

These latter two topologies are not balanced, and do not
require the top device to be precisely matched. And PSRR
improves if the top device is deliberately selected to have
a much higher GM. I'd have subbed a MOSFET or IGBT here
at this location for sure. But definitely no LM317 as a load!
 
Fat Daddy said:
What would be some alternatives?

I have a stash of 6sn7s, 6sl7s and 6L6 tubes. My OPTs are Peerless units (10K) from Heathkit W5 units. I want to keep it simple in order to build and test with simple tools and meters.

Here's one alternative: Vixen Main Schemo

Screen Regulator

Main PS

6L6s are just a low voltage, audio amp version of the 807 made for higher voltages and mowatts as an RF final. Drop 'em in and they'll work just the same.

The only possible problem might be those 10K OPTs. You won't get the same output you'd get from the design specified 6K6 OPTs, but you might also have less of a distortion problem.

You might also want to go get yourself this: Rectifier Applications Handbook
 
SY said:
Ideally, the response time would be infinite, right? So, the longer, the better.


I'm not sure I agree with this philosophy for a class AB amplifier... class A, of course, a slow stiff supply is preferable. But if you are going to have sudden peaks of increased average current draw, it may be better to tune for a response time below audio frequency, but at critical damping so that it restores quickly to full voltage without overshoot after each trip into clipping. The amount of the drop is then determined by how much you can safely pull through the PT and diodes. As long as the amp can't wag the power supply with low frequency sine waves, then critical damping should be quite desirable for varying load currents.
 
Ex-Moderator
Joined 2004
I think SY is right about this. You do not want the voltage to drop quickly in response to increases in current draw, as the OP tubes go into cut-off (not clipping), i.e. when they start behaving on the B side of AB. (You do want quick recovery, though, if there is such a drop.)

If the power supply is too easily influenced by changes in the demand current, the tail is wagging the dog, so to speak, and that leads to distortion. This is known as Sag - beloved of guitar amp owners but not hi-fi amp buffs.
 
SY said:
Wouldn't you want to have the supply move as slowly as possible so that the clipping event disturbs it as little as possible? As a reductio ad absurdem, a perfectly regulated supply has an infinite response time.

But, a real power supply does sag slightly from idle to full power, so why not have this sag last for the smallest amount of time possible, without allowing the power supply to be modulated at LF, and with sufficient damping that it can't ring? With a big enough PT and solid state rectification, the total drop in B+ should be negligible anyways. Frankly, I can see the appeal of both sides, depending on what kinds of overload you expect.

ray_moth said:
...You do not want the voltage to drop quickly in response to increases in current draw, as the OP tubes go into cut-off (not clipping), i.e. when they start behaving on the B side of AB. (You do want quick recovery, though, if there is such a drop.)

If the power supply is too easily influenced by changes in the demand current, the tail is wagging the dog, so to speak, and that leads to distortion. This is known as Sag - beloved of guitar amp owners but not hi-fi amp buffs.

Oh I agree, but the sag allowed in a guitar amp is an extreme example. I'm not talking about a huge drop - for the power supply in my present project, this approach allows a ~1-2% drop from lean AB idle to clipping in the PSUD sims, smaller than typical line voltage variations. I don't think it would even budge at onset of AB transition. A fast power supply can fully recover before the next peak, rather than exhibiting a slow rise and fall - and a power supply that is slow to sag is also slow to recover.

Of course, a power supply can be too fast, which is especially bad with insufficient damping - the Citation II in stock form is a classic example, and a strong low frequency signal should be able to wag it. With much-bigger caps, recovery time can be moved further below audio range, step response near-critical, and with only a ~6-volts drop between 300 and 400mA in PSUD. Definitely not a guitar amp power supply :cool:
 
The point is the sag and response time are not independent. A really large filter cap stores a large amount of charge, which means it can provide a lot of charge when needed, but it also takes longer to recharge when depleted. Shorter time constants also mean worse overall regulation (higher 3db point in the filter), so I think you loose all around.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.