Next up: 1967 Vox AC30 tune-up

[Raoul Duke]

Yep, you guys were rIght (which I’m thrilled about).

Turned on with wall power and everything is working as it should, right on! Much quieter than it was, fuller sounding on all 3 channels, nice strong vibe/trem. Very pleasantly surprised I didn’t screw this up, lol.

Been playing with the AC voltage for a few hours to get more familiar with its effect on the operating voltages of the different nodes. Best I can do with the stock 47R bias resistor is at 108vac (-12vac from my bucking box) which yields the following:

Pre-amp tubes all stay within 10% of spec regardless of AC coming in
Plates: 312v
Plate - cathode: 8.2
Rectifier heater: 4.9
Filaments: 6.0
12.5w per tube @104%
I know part of the AC30 “mystique” is running it “just over” 100%, but I want to take it as easy as possibly on this amp. I’d like to see if it sounds good at about 90% just to be on the conservative side - unless I’m way off base with that thought?

I’ve got a new 10w Ohmite 56R on hand that I may try. From what I can find - the bias resistor spec is 12w; but I’ve seen a couple of videos where people have used 7w - which seems like tempting fate.

Is 10w too low? Is 104% too high? I’m researching; but always appreciate the advice I get here!

Thanks!

[martin manning]

The AC30 reissue schematic shows 2x 100R 5W, so I'm thinking go ahead and try the 56R 10W. In the end copying the reissue might be the way to go.

[sluckey]

Is 10w too low?

You have 8.2V across a 47Ω resistor. Do the math. Tell us what YOU think.

[martin manning]

Plates: 312v
Plate - cathode: 8.2

I couldn't make sense of this, but if Vk is 8.2, that's 13.25W per tube including Is. Deduct 10% and Pa would be right at 100%.

[Raoul Duke]

Plates: 312v
Plate - cathode: 8.2

I couldn't make sense of this, but if Vk is 8.2, that's 13.25W per tube including Is. Deduct 10% and Pa would be right at 100%.

Apologies - that 8.2 is voltage across the cathode resistor; so yes - “Vk” as Martin stated. Not sure why I typed what I typed.

Steve - Here’s what I come up with after doing some Googling:

8.2 / 47.1 = .1740
.1740 x .1740 = .0302
.0302 x 47.1 = 1.422
So… it’s dissipating 1.4w?

If that’s correct - then I would think 10w is good, but I feel like I’m missing something here. Please put me on track if I got this wrong. I’m eager to learn the basic calculations to determine these sorts of things.

Thanks all!

[martin manning]

Full marks!

[sluckey]

Your math is good. I agree with YOU that the 10 watt resistor is fine. It's fun fishing with your own pole, ain't it.

BTW, your three step math procedure was good, but there is a more direct one step method...

P = E²/R

Also, you may be interested in this formula wheel. Every formula in the wheel is derived from Ohm's Law (E=IR) and the Power formula (P=EI).

https://sengpielaudio.com/FormulaWheel- ... eering.htm

[Raoul Duke]

Thank you both! I continue to learn thanks to your patience and prompting. Yes - I definitely prefer fishing with my own pole for sure

Steve, I will definitely look into those references and try to apply them as much as possible as I move forward.
On a side note - I picked up a Crate V32 at a flea market the other day - super clean, hardly any use, and only $50. Everything works as it’s supposed to and I was going to flip it, but now I’m thinking of gutting it and using it as a platform for your Vox Double Shot design. My “dilemma” is whether to build it as a 30w (because of the transformers) or spend more and build it as the 15w version. I’m seriously thinking of trying the 30w route…

Thanks again guys!

[B Ingram]

... Best I can do with the stock 47R bias resistor is ...
I’ve got a new 10w Ohmite 56R on hand that I may try. From what I can find - the bias resistor spec is 12w ...

I've got an 1964 AC30 head pictured on this page.

It has what I believe is the factory EL84 cathode resistor: 50Ω at 5 watts. The resistor has a January 1964 date code on it.

Then again, with multiple places assembling chassis for JMI, maybe some used 47Ω while others used 50Ω.

[sluckey]

I’ve got a new 10w Ohmite 56R on hand that I may try.

Perfectly safe to try it. I bet you will be surprised at how little the dissipation changes.

[Raoul Duke]

Hey, that’s pretty interesting B!
I’m almost positive the 47R and 22k in mine are original. I’ll look for date codes and confirm.

I know mine is a Burndept build, I’ve dated the transformers, a few of the old caps, and the pots. My chassis number is in the 148XX range and there’s no evidence of a serial number plate ever being on the back panel - and the panel looks to be original based on the covering, plywood, chalk marks etc. Been over it with a big magnifying glass a few times looking for holes or adhesive traces; but have found nothing. Part of submitting its documentation to that web site is to try and pin-point a serial number range as closely as possible. I’m thinking it’s an early 228XX serial based on what I have here and been able to research on that web site.

So this morning I re-did the grid stoppers because I forgot to not route them through pin 1, then put the 56R in. Based on the Robinette bias calculator - at apx. 120vac and with minus 12vac on the buck box I’m at 94.2% using voltage drop and about one point lower using plate current. I notice his calculator doesn’t subtract 10% - is that figure the screen current subtraction?

Either way - I think where I am at minus 12vac gives me options. Sound wise; I couldn’t really tell the difference - but that could change depending on the day and room I think.

You’re right on Steve. It didn’t change much - but I feel like it gives me a little margin to operate in for this old critter. Oops - I better be careful how I word that; my wife is of the same vintage, lol.

[Raoul Duke]

B,
Both the resistors in the power amp section are green Welwyn with “XJ” date codes which I believe is September of ‘66 - so I’d say they’re probably original.

[martin manning]

...put the 56R in. Based on the Robinette bias calculator - at apx. 120vac and with minus 12vac on the buck box I’m at 94.2% using voltage drop and about one point lower using plate current. I notice his calculator doesn’t subtract 10% - is that figure the screen current subtraction?

10% is a reasonable guess at the screen current fraction of cathode current. You can check the data sheet for the tube type in question: The Mullard EL84 sheet has a 300V Va and Vs push-pull case where the quiescent plate and screen currents are 36 mA and 4 mA. That's 4/(36+4) = 0.1. For beam tubes the fraction is lower, 5% or less.

Here's how I would find average plate dissipation given Va, Vk, and Rk:

Pa = Va-k * Ia = Va-k * Ik * 0.9 / N, where N is the number of tubes drawing current through cathode resistor Rk.
Ik = Vk / Rk, so
Pa = Va-k * Vk / Rk * 0.9 / N
For the 47R Rk case, Pa = (312 - 8.2) * 8.2 / 47 * 0.9 / 4 = 11.926 W

[Raoul Duke]

Thanks Martin, understood.
I actually dug up an older thread here where you explained it in depth and very clearly for a feller who was cathode biasing a pair of 6V6s. Both your answers are very helpful, thanks!

[Raoul Duke]

So the wrap-up on this project has led me to more lessons and some small modifications that I’ll share for anyone interested or looking at a similar project.

Mods:
- grounded the first filter node over at the PT center tap/shield ground vs at the cathode resistor
- grounded the vibe/trem filter up at the pre-amp (where the filament ground used to be)
- used 1/2w 100r resistors for the heater artificial center tap - but fixed them to the power tube ground bus for small elevation
- changed the screen grids & stoppers to 200r and 2k respectively (from 100R and 1k5) and went with 2w rating
— read a lot about this and decided these values were right in the middle of stock and extreme
- slightly cooled the bias (56R) so that at 120vac into the minus 12vac receptacle on my buck box, I’m at 97% - which gives me minus 6vac and straight 120vac to scale up if needed

Regular stuff:
- replaced all the big filter caps and went back to original values (everything was x2) except for the first filter cap (dual 32uf)
- replaced every resistor and capacitor that exceeded 15% out of tolerance
— only one ceramic tube cap tested way out of spec - the 750pf, the rest went back in
- wired the power direct to 115v tap, capped and bundled the others, fused the hot (neutral was fused?)
- did a ton of cleaning to include what looked to be a Coke poured into it about 30 years ago
- re-wired the jack modification so that the speaker outs are switched
- because I focus on weird things sometimes; re-did the factory masking tape on the outer edges of the face plate
- spent a lot of time getting the lead dress neat and matching older pictures of AC30s (things were pretty tangled)
- re-flowed about two dozen dry or cracked solder joints
- even managed to preserve Cesar Diaz’s inscription inside the chassis (FWIW)

The amp is noticeably quieter, doesn’t crackle anymore (so far), and sounds fuller; really satisfied with the result.

Couldn’t have done it without the help I received here and I’m very appreciative for the lessons learned.
Thanks again guys!

IMG_0432.jpeg