Low watt amp power supply, design problems.

[rob_h]

    Normally on a higher watt amp, from what I've seen, the first tap off the rectifier, highest voltage, is the O/P transformer, then the cathode bias, the PI, and then preamp tubes.  In the amp I'm working on, a 6AK6 push-pull, the preamp voltages are 263v and 250v.  LTPI is 235v.  O/P transformer ct is 220v. Cathode bias is 182v. 
    In a normal higher wattage amp, with the first taps for the output section, and the preamp taps coming afterward, when the output tubes drive their load, voltage across at all the taps decrease with any sag across a tube rectifier, choke, series resistors, etc. Even though there is sag, it is a result of the increase in current to the output transformer.  With the voltages in reverse order, preamp first and output last, I don’t think there will be enough power available to the output tubes’ tap when driven hard, since it is at the end of the line.  I think it will be starved for current, and voltage, distorting to trash.
     If the power supply is split, two parallel lines, with the output taps on one section and the preamp on the other section, would it be best to put the PI with the output and cathode bias section or with the preamp section?  It seems to go with the output and PI, but it too is a preamp, is it not?  An added variable is the use of a PPIMV.  Then, there’s the preamp tube count, 4 dual triodes and a pentode.  Their bias current is greater than the PI and the output tubes.
Any clarity would be appreciated!
Sincerely,
rob_h

[PRR]

How does it sound now?

[jazbo8]

Could you post a schematic? 

[rob_h]

     First, I guess I should have said "design question".  I meant "design problem" as in looking for a soluiton in the design stage of the project. The BOM is complete except for the power supply and one triode bias decision.  I'll build the amp in sections, testing and tweeking each before doing the next.  PI and O/P with the 5E3 first, then the 5F6A/Little Wing, then the OD, then, after that the reverb section.  I'll post the schematic after it is finished, when I actually start building it, but here is the block diagram with power supply voltages.
     The power supply "design problem" is still what I am working on.  I already have the power transformer, a 270CAX, 250-0-250, so the voltages are limited to around 290vdc. 
     Any ideas, experiences, opinions, reasons, etc., about the PI, weather it should go with the preamp branch or the output branch of the power supply?
 sincerely, rob_h

[HotBluePlates]

     First, I guess I should have said "design question".  I meant "design problem" as in looking for a soluiton in the design stage of the project.

Then all I could tell you is I wouldn't design it the way you are.

I'd apply the maximum B+ voltage to the plates of the output tubes, same/slightly less to the output tubes screens, less to the phase inverter supply node, less still to preamp nodes working back towards the input.

I can't figure out where you got "182.5v cathode bias" for the 6AK6's, except maybe that the data sheet shows a condition for 180v plate and screen (but that condition also shows bias as -9v, so for cathode bias you'd need 189v total B+).

I would also suggest bread-boarding your circuit. I mean I'd be scared to attempt a circuit that complex without breadboarding, because something is bound to go wrong.

[rob_h]

Is this a test or something?
1) *   Ok, the 180v is the screen voltage, for the whole graph, but the curves are for Ip/Vp for conrtol grid voltages shown on each curve.  The text "Control Grid Volts =" just happens to be on the -9v curve.  With 220Vp (load line, class B), 300v max(class A load line) and a 22.5k primary, -13.5v is where the voltage is indicated, which gives 7ma for plate current at 0v signal in. 
2) *   The data sheet for the 6AK6 gives class a amp values where the grid voltage is -9v, but this is a class ab push-pull output.  At -9v, the tubes would be both biased for class A.  At -13.5v, they are biased way more toward their off state for class B. At -9Vg1 and 220v, the Ip would be about 16.7ma.  On the class a data, there is -9v, 180v and 15ma.   
3) X  Tell me it aint so!  I messed up the Vg2 voltage, it should be 193.5 for 180v Vg2-Vk.  With a 1k grid2 resistor, and a 2.5ma current, the supply voltage should be 193.5+2.5=196.0 for the cathode bias supply.  Have to say I went to 0v instead of Vk (-13.5) for that one. WRONG!!!!!!
     I really do appreciate your observations. You helped me fix my first of who knows how many mistakes.  I'm going to build the amp in small sections.  I agree that breadboarding would help.  My layout is tight, hopefully not too tight.  Each preamp and the output section will be on seperate boards, psuedo modular.  One piece at a time, it's not that complex.  And..., PI, with the output or the preamps?
Thanks, you made me think!

[jazbo8]

I see you are trying to do high-level design but with some lower-level calculation mixed in, but as HPB said - there is usually a natural decline in B+ voltages as you move from the PA back towards the input of the amp. Since you are already working with characteristic charts and setting bias, why not draw up a preliminary schematic (perhaps using SPICE's schematic capture if you want to sim it later), I think you will get better feedback from the group...

Jaz

[PRR]

6AK6 with comfortable load is probably cathode-bias.

Plate current will NOT vary much with level. "Sag" won't happen.

Build raw power supply. Two 1K resistors (estimate their wattage), 20+uFd filter cap, OT and 6AK6es. Fiddle the 1Ks and cathode resistor to get near your desired operating point. (This can even be done before you have the OT; a couple 200r resistors will approximate the OT's DC drop and also smoke if a 6AK6 sucks way too hard.)

You might drop the driver from here.

Then a second string from raw supply down to your preamp stages. And preamps pull much less current than a power stage (even 6AK6). But the first (guitar input) stage probably should have more than <180V to handle hot guitars. So come down from 290V with some 10K resistors, aiming at 250V-270V to the first stage and whatever in-between.

A reverb driver may suck significant current. (Not if it is 12AX7.)

I'd expect to design with a hot iron. I *could* pre-compute it, but resistor-tacking is faster and less brain-pain.

> My layout is tight

That, I have learned, is a mistake. The *second* build can be tight, but the first should sprawl all over a too-big chassis, even a bread-board.
___________________________

> there is usually a natural decline in B+ voltages as you move from the PA back towards the input of the amp

That makes perfect sense for normal amplifiers, which start with a very-small signal and build-up to a large output, with a free choice of tubes.

Geetar amps are funny.

The input *can* be very hot. We often don't like the sound of low-gain input tubes. So the first stage has to be high-strung.

Here he wants the output quite small, in a zone where few tubes aim for, also where OT impedances become uneconomically high.

We'd also like the B+ nodes in order of signal level, so a filter-chain runs from kinda-rippled down to super-clean, which implies drop-drop-drop. However today filter-caps are cheap enough we don't HAVE to do it that way.

It makes some sense to work the input over 250V, the output under 200V.

[DummyLoad]

use 6AR6 instead of 6AK6 - ARs are better than AKs... ;)

--pete

[HotBluePlates]

Is this a test or something?

No, I think I didn't understand your shorthand on the block diagram.

     If the power supply is split, two parallel lines, with the output taps on one section and the preamp on the other section ...

If you have to have the preamp voltage higher than the output tubes, this is probably your best approach.

I guess what I was getting at is, "do you have to have the output tubes at 180v for some good reason?" If no, then cheat the voltage up on the output tubes and calculate a good guess for cathode bias resistor. You can/should verify that piece by testing before you get too deep in the build anyway.

I already have the power transformer, a 270CAX, 250-0-250, so the voltages are limited to around 290vdc.

Well, it's limited to 250v * 1.414 = 353v and maybe higher if your line voltage drifts upward or if you have the older Hammonds with the 115vac primary. You'll lose a tiny bit across a tube rectifier if you want, but these output tubes (and amp as a whole) won't draw much current and so won't drop much in a rectifier. Might as well use solid state rectifiers and save a socket/wiring.

     First, I guess I should have said "design question".  I meant "design problem" as in looking for a soluiton in the design stage of the project. The BOM is complete except for the power supply and one triode bias decision. ...
     The power supply "design problem" is still what I am working on.

Suggestion:
All of the rest of the amp needs to be designed before designing the power supply. Every supply voltage and current draw must be a known.

See it works in a circle like this: You have to have an idea for what power transformers are available and what power supply configurations might work to set a basis for the available supply voltages for the rest of the amp. Then you have to design the output stage with the probable available power supply in mind, to determine how big a signal the phase inverter has to provide. And then you move to inverter design with a realistic drop in voltage from output stage (or with a concept of how a split-path power supply might be implemented) to set boundaries to work in the design of the inverter.

At the tail end, after every voltage and current draw is known, you circle back to the power supply. Now you know what the voltages are that you've outlined for each stage, and can calculate dropping resistors and see if these provide enough decoupling in conjunction with planned filter cap values. While it won't apply with a solid-state rectifier, you'd have to have a good idea of what this entire amp would draw if you wanted to figure likely rectifier voltage drop (normally, you'd assume knowledge of the output stage current would be enough, but your preamp will probably draw more current than the output stage in this case).

It seems you've got some good ideas going, I just wouldn't know how to recommend a power supply setup without knowing the rest of the amp, since I approach design as described above.

[rob_h]

These are the graphs used for the o/p and pi.  The o/p transformer is a 125D 10w, and the output wattage will be around 3watts.  I don't know if it will make much difference if I use the 22500 impedance or the 17600 since it's a 10 watt transformer.
This is the 3rd post tried with attachments. I finally saw the 800k max.  Schematic next.
rob_h

[rob_h]

Here is the Schematic " in progress".  The Deluxe is a copy, as is the Bassman preamp.  The Deluxe voltages are ok for what I have, but the Bassman voltage of 325v is a bit high with filtering for my power supply.  I'm leaning toward Geezer's Little Wing voltage of 163v.  The OD started as a Dumble OD section.  Then I read the Valve Wizard EF86 preamp guide, then I saw it used in the TOS.  I bought some, so I'll use one there.  I know the schematic is not finished, but you can get a better understanding of what I'm going to attempt.  

One thing I am trying to understand is how the addition of presence pot and cap effect the design of the phase inverter.  Also is the difference of PI Ra's and cap additions to bring the PI into a closer ballance?

PS- I do industrial electronics projects and machine control work, and electronic component level repair, so I'm accustomed to regrouping and troubleshooting my mistakes, and I have never quit or given up on a project.  I know this is a new world for me, but I love it!
Thanks for your input.

[rob_h]

I forgot.  The 180v for the Vg2 is the only one on any graph I could find on the 6AK6. I know raising it will increase the current, but I'm not confident enough to guess where the load line will be and I don't want to toast the tubes.

I'm using a 5Y3 rectifier, will the higher voltage, 325v, on the Bassman preamp be a benefit compared to the 293v or 263v used on the Little Wing?

On the power supply  program I have, the increase to 20-30ma shows a large sag.  I have no idea if it will be there in the real world.  I'm trying to emulate the qualities of the higher wattage amps.

Thanks again for your time and input!
Sincerely,
rob_h

[rob_h]

The "hum" is the 60hz ripple type, right?  The signal to noise ratio in the preamp is where the hum is introduced is it not?  That is, the amount of 60hz ripple from the power supply relative to the low voltage signal in the preamp, or 60hz noise; radiation, magnetic fields, etc. that add up to the "noise" relative to the "0v input signal", if large enough will be amplified through the preamp as hum, if I understand.  Virtual world again, but using PSU Designer, if I use a $16.97, 7H, 240 ohm choke and an extra pi filter in front of the preamp side of the power supply, my first current tap has 2.09mv of ripple on a 270vdc, the next has 335microvolts at 263vdc, the others have less.  I think if I don't introduce noise from wiring it should be ok. When I get it together, I'll use  a scope to look at the ripple and noise, and give it a listen.
Thanks, sincerely. ONCE AGAIN, one of you guys brought to my attention something that got me thinking, and I think I solved the problem before it was one.
rob_h