Schematic and parts list for DC heater supply?

[dehughes]

Mark, Paul, Tom.....THANK YOU. I'll get to your posts when I get back from work.

BTW, the PT I have is a Mercury Magentics Brown Deluxe clone tranny...model number FBDP, with a 5A heater capacity. So, with a pair of 6V6s and one 12AX7 (PI) on AC, and then running DC to the other 12AX7 and the 6SN7, would this put too much strain on the tranny, then? I know it's plenty for the tubes running on AC, but I don't know if running part DC would stress the heater winding too much...???

[Andy Le Blanc]

A full wave bridge rectifier with a capacitor input load makes I D.C. = 0.62 X Sec. I A.C.
5A heater winding will give you around 3A D.C. after rectification.
A choke input on the load will get you a I D.C. = 0.94 X Sec. I A.C.

[Structo]

Hey, I just love this stuff!
I learn something on every visit.

I'm getting a little long in the tooth and although I had rudimentary education in electronics, they failed to spend much time on tube theory since everybody knew that transistors were much better........

I even grew up in the 60's too.

Remember Dad pulling all the tubes from the TV and going to the store to test them and buy replacements.

I wonder what happened to all those old crappy tube testers?
More importantly, all the tubes that were stored below in the cabinet?

So I try to learn where I can.

[dehughes]

A full wave bridge rectifier with a capacitor input load makes I D.C. = 0.62 X Sec. I A.C.
5A heater winding will give you around 3A D.C. after rectification.
A choke input on the load will get you a I D.C. = 0.94 X Sec. I A.C.

Cool. So how does that work out in terms of making DC out of some of the AC, in order to run some tubes with one and some tubes with the other? Really, all I'm concerned with is 1) providing enough current to the heaters 2) without stressing the PT or the tubes.

[paulster]

You've got bucket-loads of current available. The 0.9A that the 6SN7 and 12AX7 would draw will probably end up at around 1.1-1.2A with rectifier losses.

Your total heater draw should be under 2.5A all things considered, so plenty of headroom left!

[dehughes]

You've got bucket-loads of current available. The 0.9A that the 6SN7 and 12AX7 would draw will probably end up at around 1.1-1.2A with rectifier losses.

Your total heater draw should be under 2.5A all things considered, so plenty of headroom left!

Sweet. Good to know. I'll go compile the parts list and order the DC stuff...

[drz400]

You've got bucket-loads of current available. The 0.9A that the 6SN7 and 12AX7 would draw will probably end up at around 1.1-1.2A with rectifier losses.

Your total heater draw should be under 2.5A all things considered, so plenty of headroom left!

Sweet. Good to know. I'll go compile the parts list and order the DC stuff...

You could do it easily, total cost $30
http://search.digikey.com/scripts/DkSea ... 02-1336-ND

[dehughes]

Dude, that part is made by my the company my brother works for! CUI. Too funny....I'll have to ask him if they have any rejects lying around....didn't even THINK of that....

So, would I be better off with the 5v or 7.5v model? I'm shooting for 6.3v, right?

[martin manning]

Here's another incarnation; this one is from the Marshall 4001 Studio 15. I have an '88 4001, so I can say that it works fine and has held up well.

MPM

[drz400]

Here's another incarnation; this one is from the Marshall 4001 Studio 15. I have an '88 4001, so I can say that it works fine and has held up well.

MPM

I say either float the center tap to DC potential (also relieves the cathode followers from excess voltage on the cathode)

or use a tight DC supply

If the DC supply is not tight or regulated enough it can actually cause more buzz than not using it at all. You can hear that sawtooth

[dehughes]

Here's another incarnation; this one is from the Marshall 4001 Studio 15. I have an '88 4001, so I can say that it works fine and has held up well.

MPM

I say either float the center tap to DC potential (also relieves the cathode followers from excess voltage on the cathode)

or use a tight DC supply

If the DC supply is not tight or regulated enough it can actually cause more buzz than not using it at all. You can hear that sawtooth

Right on. Please elaborate as to what a "tight DC supply" would look like. I'm a DC newbie.

[dehughes]

Here's another incarnation; this one is from the Marshall 4001 Studio 15. I have an '88 4001, so I can say that it works fine and has held up well.

MPM

Dude, Martin....that's nearly my situation. 3 tubes on AC, two on DC. I'm gonna have to look into that one for sure. THANKS!

So, the thing is, the AC side's CT isn't grounded, but the DC side is via the two 100 ohm resistors, correct? Could I just leave the AC side grounded and let the DC side be ungrounded?

Also, the cap looks like it is sitting between the two heater lines, whereas in my head I was thinking that the + from the cap would go to the heaters, and the - would head to ground. This is different than what I've seen...

Couldn't I just run a pair of diodes off the AC line, connected to the + of the 10000uf cap (the - of which would be grounded), and then run that + to the heaters of the 6SN7 and 12AX7, and then ground the other heater pin of the tubes? Wouldn't that do the trick? I believe that's what they did with the Constellation...albeit off a dedicated supply of AC.

[paulster]

Please elaborate as to what a "tight DC supply" would look like. I'm a DC newbie.

A tight DC supply is one where there is very little AC component to it, i.e. very little ripple.

[img:400:184]http://upload.wikimedia.org/wikipedia/e ... le.svg.png[/img]
This is the type of power supply you're looking at building. You'll see how the capacitor charges 120 times a second and then begins discharging again immediately because of the load. If this ripple is significant then you end up hearing it as you essentially have an AC voltage sat on top of a DC voltage. Fortunately your load will be constant so you can minimise the ripple by heavily filtering. That's tending towards a tight supply by brute force regulation.

Doing a quick calculation you're likely to have about 0.8V ripple with 10,000uF since the 6SN7 draws a reasonably substantial 0.6A, so that plus the first 12AX7 takes you up to just under 1A. If this is really sensitive to noise then I'd be thinking about using a pair of caps, to halve this ripple.

The other option, which is what I do, is to sacrifice a bit of peak voltage and use a regulator IC to force the output to remain at exactly 6.3V, whatever the point in the cycle. I do use massive amounts of filtering beforehand as well, but that's to give me plenty of headroom for dealing with low wall voltages because you have to ensure that your voltage under load never drops below 6.3V at any point in the cycle.

So, the thing is, the AC side's CT isn't grounded, but the DC side is via the two 100 ohm resistors, correct? Could I just leave the AC side grounded and let the DC side be ungrounded?

You can try leaving the AC side grounded and the DC side floating, the DC side grounded and the AC side floating, or the AC side grounded and the DC side grounded via two 100R resistors.

I've tried all of these and in my experience the quietest solution is with the DC side grounded and the AC side floating. It's a really quick and easy test though so I'd advise you to try it for yourself.

Couldn't I just run a pair of diodes off the AC line, connected to the + of the 10000uf cap (the - of which would be grounded), and then run that + to the heaters of the 6SN7 and 12AX7, and then ground the other heater pin of the tubes? Wouldn't that do the trick? I believe that's what they did with the Constellation...albeit off a dedicated supply of AC.

If we describe the heater winding as 3.15-0-3.15V you can probably see why this won't work. Using two diodes will only use half of the winding at a time, so you'll have rectified a 3.15V supply. I'm fairly sure the Constellation must have a 12V centre-tapped (6-0-6V) secondary.

[dehughes]

Man, thanks Paulster. MUCH appreciated.

So, what about a 12v filament transformer running into two diodes connected to the + of a 10000uf cap, the - grounded as well as pin 8 (6SN7) and pin 9 (12AX7)? Would that eliminate the ripple sufficiently, allow enough voltage, and set things up properly? I'd like to emulate the 6v lantern battery as much as possible, as that was pretty quiet....

[martin manning]

Also, the cap looks like it is sitting between the two heater lines, whereas in my head I was thinking that the + from the cap would go to the heaters, and the - would head to ground. This is different than what I've seen...

The cap is across the 6.3V on the heaters because you need a store of charge at 6.3V. The ground reference (junction of the two 100R) is half way between so the heaters are running on a +3.15 to -3.15V DC potential relative to chassis ground. BTW, that circuit supplies one 12AX7 (about 300 mA), so you might triple the size of the cap if you want to keep the ripple down to the same level or less at ~1A.

MPM