Tube power supply to work with existing transformer

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

imo

Well-known member
Joined
Jan 19, 2006
Messages
397
Location
vashon, wa
Hello,
I am refurbishing an old Presto 3 channel mixer to use as a funky "character" pre for my friends studio. kind of a Christmas/bday kind of thing, so not wanting to spend any additional money. The topology is mixer/repro channel the mixer has: 26J7's, 6SL7GT, 6SN7GT. The repro channel mirrors this. Both sides have transformer ins and outs. There is an XLR after the mixer side that i could feed back into the repro side if i want to add another gain stage(using the repro volume to control it).
  I have cleaned it up and recapped it and now i need to make a power supply for it. I am trying to use some parts i already have laying around. I have a PT that i was using in a guitar amp(AC15) that is in the ballpark. Plenty of heater current(6a) and enough  current(150mA) for the HT. The B+ is a little high @ 330v. I am looking for the best ratio of elegance/frugality<g
If you look at the schematic you can see that there is already a ton of filtering within the circuit, so i was thinking a simple solid state rectified, into a choke, cap, resistor--Main thing is that i need to get it down from an anticipated 460v down to around 350 into circuit. The 6J7's are about 2.5mA class A, 6SL7 4.6mA, and 6SN7 9mA. I am assuming around 40mA total for both sections.
I have a couple of chokes laying around that will knock off a few volts, but still looking at 90 or so. With all the filtering downstream and it being a funky thing to begin with, would i be ok just kludging in a big 2-2.5k resistor in there, or would it be significantly better to split that up over a couple of filter sections(ie 2 5w 1k resistors)
Thanks for any help or ideas
Best
Ian
 

Attachments

  • Scanned Image 2.jpg
    Scanned Image 2.jpg
    718.9 KB · Views: 41
Dropping the excess volts via a resistor is fine. Splitting it into several equal RC stages will give a much lower ripple than the single equivalent RC stage. SO split the 2K5 into four shots of 620 ohms each with a filter cap. Eases the heat dissipation in each resistor and gets the ripple right down.

Cheers

Ian
 
A fellow Ian! Thanks for the tip. I have really enjoyed your writings about your tube mixer, etc. Very informative and helpful.
This piece is a bit of a mess. Someone worked on it a bit before, and short of doing a complete rebuild, which i don't have time for, I am stuck with lots of sloppy grounds and iffy routing. We'll see what kind of hum and buzz i get once i get the power supply up..
There will probably follow some hum chasing posts<g
Thanks
Ian
 
> I am assuming around 40mA total for both sections.

You can read it on the plan (maybe you did?). 350V comes in, goes through two 1.5K resistors, arrives as 320V. So 350V-320V is 30V drop, in 1.5K is 20mA, twice is 40mA.

> anticipated 460v down to around 350

462V no-load.

No-load happens at every switch-on, while the tubes are cold. Therefore "all" the B+ caps need to be 450V rated, at least down the string to C11 C23 (after the 27K, there's not enough current to burn-up a 300V cap in 11 seconds).

110V to drop at 40mA is 2800 ohms 4.5 Watts. Start with four 1K 5W in series (aim low-volt at first). If really too low, take out 1K.

The filtering is OK but not stunning. Start with 40uFd, go through several K dropper, another 40uFd, call that "+350V". (Ian's C-R-C-R-C-R-C-R-C is also fine but seems over-kill to me.)

As I read the squints, V4 V8 are run right AT rated dissipation. That's OK, but you do not want to go higher than 350V.

> to use as a funky "character" pre

Then you don't really need 350V. Indeed it may not have a lot of flavor as-drawn. Starve it. All of it would "work" (pass signal) at 150V on the "350V" supply. Now loud signal will be starved for swing, "funky". So I think you should have 8K or 10K of power resistors handy, and try/see what it sounds like.
 

Attachments

  • IMO-bw.gif
    IMO-bw.gif
    72.6 KB · Views: 15
you could go choke input, if you need to drop volts,

or go 5y3gt so the heat will be outside the chassis,

transformer with no load has to buck the pri emf with back emf,

since no current flows in sec, no current has to flow in pri,

so back emf keeps pri current at zero plus magnetizing current.

that is why flux density is maxed out with no load,

pri flux and sec flux are additive, small signal stuff excluded because pwr is so small,

now when the tubes heat up and draw B+ current, the transformer pri has to supply current, this means the back emf has to drop in order to let current flow in the pri so that current will flow in the sec. so overall flux drops, the transformer quits making noise, and the heat goes down, for a while, until the copper heats up,
 
PRR said:
to use as a funky "character" pre

Then you don't really need 350V. Indeed it may not have a lot of flavor as-drawn. Starve it. All of it would "work" (pass signal) at 150V on the "350V" supply. Now loud signal will be starved for swing, "funky". So I think you should have 8K or 10K of power resistors handy, and try/see what it sounds like.

I am not so sure about that. The first two 6J7 tubes are operated at plate voltages of 65V and 90V respectively. This is mainly for noise performance and the signal levels are very small. Reducing the HT to 150V might seriously compromise these stages. By all means reduce HT but it is probably better to arrange this to be confined to the later stages where signal levels are large enough for it to affect tube 'tone'.

Cheers

Ian
 
Wow! Great info, all of it. I found an old choke from a 50w Marshall that i am going to try. So far- rect(4007), 47u, choke, then 33u caps with 680 resistors between them. I'm gonna try plugging this in to the circuit and see what happens. If i am in the ballpark with proper load, I will start debugging. I guess the last time it was used it had a pretty good hum. I am hoping to have solved some of this with the recap, but the grounds are a mess, and I am anticipating that it is not going to be solved by the recap. Thanks for the article Ian. Will read shortly.
Best
 
well, phase one complete. That was almost 2 easy. I figured i would have to do some subbing with resistors, but the HT coming in is sitting almost right at 350. The repro channel is running a little low against schematic, but i was thinking of that being an extra gain stage accessed by connecting the mixer back through the repro, so I'm gonna go with that for now. Ripple seems pretty low, but haven't had a chance to put it on the scope. Passing signal with the generator, so i'm gonna run a mic through it and see what happens.
Will report back
 
OK, it does pass sound. There is a pretty nasty buzz, which i was suspecting.
A few things i noticed/tried:
Tried pulling out the tubes in the repro section, as well as subbing them into the mixer section-no change
-the grounds for each section are using the ground tabs from the old multi caps that were replaced with individual caps(this was done before i got it). It does seem, for the most part, to correspond with the proper section, but seems a bit messy
-C2A can't be 10u, can it? this must be a mistake. The other side has a .01, and it sitting on the cathode
-There seems to be an AC oscillation on the B+ line in circuit. It goes from a few mV to around 100mV or so in some sort of cycle. I have attached an image of the frequency plot.

Next step? Full disclosure: I am definitely not an EE. A touring guitar player/singer who does this for enjoyment. My aptitude is iffy and my skill-sets are questionable. I do have a scope and signal generator and would be willing to use them if necessary<g


 
update, more like a post to wire:
The mixer is actually relatively quiet. I'm sure i could pull some of the low level hum down, but not without a top down design. I will go back through the repro channel tomorrow and see what i can find. These are funky tubes, so i don't really have spares to experiment with.
I do have a question on my power supply. I just used 2 4007's to rectify. do you guys use snubbers on those to clean up the spikes or does it not matter with older designs like this?
The amp sounds like singing through a tube screamer! And this is within the correct specs for the circuit
 
you can try putting in a 62n7 in place of the 6sl7, this will lowwr the gain, the bias and plate resistor will be off a bit, but the pinout is the same,

if using choke input i would say forget the snubbers,

 
actually really digging the gain! Nice to have a crazy pre to get that sound. I think these Prestos were the predecessor to the Ampex, and it sounds like it.Reminds me of Little Richard vocal sound era '58
 
C2A bypasses the cathode resistor of V1, thus raising the gain of that stage.  Try eliminating that 10 uF cap to lower the gain, and connect the 'low" side of the .01 or whatever it is from the screen R to ground.

I really can't fully understand some of these ancient designs, since they seem to have been involved in some sort of "arms race" for insane amounts of gain.  Several years ago, I messed with a Magnecorder where the ORIGINAL spec sheet BRAGGED about having 90+ dB of gain via the mic input.

Dunno what Presto had in mind, however.

But, Maggie's design of 90 dB???????????  Why???  For recording farts of a flea at forty furlongs <G!!!> during a fortnight of full moon's ......with a ribbon mic??  (sorry, could not come up with a low output mic's name beginning with letter "F").

Maybe why the Presto sounds line a Tube Screamer.....cuz it has 597392380402658 dB of gain......just like a fuzz stompbox?

Bri
 
I think the 90dB gain thing only relates to 50 ohm mics via the input transformer which equates to less than 80dB for a 600 ohm input.

Cheers

Ian
 
Back
Top