AC30 style build oscillation in bass and cut controls

[imo]

I am getting close to putting this into a box and wrapping it up, wanted to ask a couple of more 'AC30' type questions.
I currently have the amp sounding really good, it goes from really beautiful, classic Vox style chime to amazing blown up (think early Cream) distortion.
I am really happy with the sound, but i am concerned about the amount of current still in the output tubes.
I have talked to a lot of old timers, Vox techs, builders, etc, about my design and i have gotten a range of answers.
Basically what i have heard echoed again and again is that the higher plate voltage really makes for the ultimate tone. You can fudge a little on the cathode resistor, but again it quickly gets away from the things that make it great. So I am keeping the cathode close to the 50r in the original.
I have also heard that clamping the voltage from the choke (with resistance or zener) changes the nature of power supply and again takes it away from what makes the circuit great.
So, i have around 350-360 on the plates and around 11.5v on the cathode(2 tubes per 120r) Somewhere around 15V plate dissipation per tube.
I have JJ's in there. Can't really afford the sturdier old stock tubes that these guys say survive in there. Am i pushing myself into a corner with a touring amp that is bound to fail, or is this something reasonable that should be fine with regular tube changing?
Ian

[pdf64]

If you are running your el84s that hard and need to keep the amp reliable, ie carry on playing through a tube failure, I would have each cathode on a separate resistor (240ohms, cap bypassed) and put a fuse between the cathode resistor and ground (say 100mA quick blow).
That way, if the tube fails o/c (eg due to dead screen), the other tubes will maintain bias.
If a tube shorts, the fuse will pop - again the other tubes will maintain bias.
Your amp will carry on working even with 3 dead el84s.
What value are your screen resistors?
Also put silicon diodes in series with the plates of the gz34, to keep going in case that dies, see
http://www.geofex.com/tubeampfaq/TUBEFAQ.htm#mods
about 3/4 of the way down for other reliability mods.

[imo]

cool idea with the cathode resistor. hadn't thought of that.
The diodes over the rectifier i have read about, and thats a good idea as well. i had originally thought about putting them on a switch, but when i was dealing with the already high voltages, i moved on, and hadn't thought about it
Right now i have 100r screen resistors.
I have pretty much followed the AC30 schematic except for the 1st coupling cap and possibly the cathode resistor..

[pdf64]

Try measuring the Vac and Vdc across 1 of those 100 ohm screen resistors, with no signal, and big clean signal and when the amp is cranked. Work out the screen current in each condition. Then measure the screen to cathode voltage under the same 3 conditions.
Work out the power that those screens are dissipating - it will likely be many times the 2 watt rating.
That link I put in post #13 advised 2k5 screen resistors for JJ, but you might find that may reduce the output power a little. Whatever, 100ohms is way too low for current tubes. Try 470 and 1k in the above test to see if screen dissipation reduces without reducing output power.

[imo]

Sorry for the delay in getting back to you. I was rebuilding the amp in the finished chassis.
Right now i have 1.5k 5 watt screen resistors. I am finding around 7v dc rest, up to about 24v dc at full voume. I have not had a chance to measure AC current yet, but i'm assuming screen current would be simply measured by the DC+Ac voltage divided by the resistors. it seems that this would be within the 2 watt range.
The full current(at rest) coming through each tube is around 47mA. I am using the 130R(1 for each pair) giving me around 12.3 volts on the cathode. minus the resting screen current, that puts me a little over 42mA actually on the plates resting.
It does seem to me that with the larger screen resistors that the sound is a little less "clean" at higher volumes. I'm not necesarily concerned with losing power, but wondering if my ears are telling me the truth? What are your thoughts on the tonal differences when using larger screen resistors?

[pdf64]

The screen power dissipation is obtained by screen current x screen voltage, screen voltage being the voltage between screen and cathode.
So it would be a max of about 5 watts from your readings above, which will probably be an underestimate (going by my fluke 187, Vdc + Vac ~= 1.5 x Vdc).
However, it seems to be a standard thing with guitar amps (that have a high screen voltage) for the screens to run >>2 x the tube's max spec screen grid dissipation at max output.
I agree that larger than normal screen grid resistors (sufficient to limit the max screen grid dissipation even to <2 x max) do seem to thin the tone out a little.
As usual, it's a comprise between tube life and tone.

[imo]

I guess i am running up against the designs walls then. I always tend to the conservative side with design, really trying to understand the reasons for the way things are before i start making modifications. I have always loved AC30 tone, and as i've built this i've tried to build it to spec,using the higher plate voltage of the 'more desirable' mid 60's models as a template.
It seems that the modern tubes really limit this design more than in many other classic amps. The screens seem to be the 'X' factor in that the tone does get away from some of the fundamental sounds of the design as you raise the screen resistors; even more than raising the cathodes.
I really appreciate your time PDF. It really helps me understand design changes. I've done a lot of reading, more reading time than building time, and most of the info is really technical, which doesn't always help when you are trying to think of tone shaping as opposed to simply how to gain stage.
The amp sounds pretty cool. I think i may try to lessen the screen resistance a bit and put it in the trenches. We'll see how long they last at 470r. If the tubes go, i'll switch back to the higher screens
Again, thanks!
Ian

[pdf64]

I'm glad to have helped. Keep us posted on how you get on with it.
AC30s have got a rep for being about the hardest amp there is on their tubes - now you see why!
However, it's due to the 60s tubes being so robust, designers of many amps ran them over their max ratings without problems. Current production tubes are less forgiving of being abused.
To improve the reliability of the amp with modern tubes, consider the suggestions I made in post #17.

[ThermionicScott]

Sounds like it's getting somewhere! Since it's your personal build, consider improving the ventilation, too -- maybe add some fans to keep those hot EL84s happy.

And the idea of giving each EL84 its own 240-270 ohm cathode resistor and cap makes a bunch of sense to me as well. (Side note: when you double the resistor value, you can halve the capacitor value to maintain the same RC constant.)

- Scott

[pdf64]

Great call with the fan. I've been fitting 60 or 80mm 12V pc fans to heavy gigging amps for the few months, and mounted them on a bracket near the front panel, near the PT, to circulate the air around there.
Full wave bridge rectify the heaters supply with ~470uF reservoir to give about 9Vdc.
The fans run quiet but provide a significant benefit, with the PT running noticeably cooler.
Even with high gain amps, no additional noise / interference is introduced by doing this.
I got the idea from here -
Installing a fan in vintage Fender amps
I'm not sure where the best place would be to fit such a fan on an ac30 type build - maybe either end on, at the PT end of the power amp chassis blowing on to the PT, or mounted on the cab blowing out of the vent slot. Perhaps the best way would be to cut an extra hole in the cab for the fan at the side by the PT, and have the fan blowing out of it.

Re the cathode bypass cap, my understanding was the the cap should be sized according to the cathode source impedance, usually smaller than the bias resistor.
But there's a compliation if splitting a push/pull shared cathode resistor - the bypass cap should be larger than twice the shared value, because the cancellation effect from sharing the resistor has been lost.
Granted the cancellation effect when shared is lost at high signal levels when the tubes enter cut off, so with a shared cathode resistor/cap, the low end frequency response probably varies a little according to signal level, compared to having a resistor/cap per tube.
It's too complex to calculate, especially when the impedance of the OT at cut off is brought into consideration.
Therefore I make the bypass caps large enough so that the low frequency response, as measured at the speaker output, isn't limited by the cap value.
By that method, 470uF across 240 ohms (per el84) seems to be good down to 60Hz.

[imo]

I am looking to get a fan. I have to wrap the amp up and ship it down to Austin for the SXSW festival. I didn't have any 240r's handy so i have to stick with the dual 130r's for now. I am going to parallel some 1.5k's with the current 1.5's on the screens to try to grab back some of the 'bloom' that i have been missing.
I also tweaked the first stage a little. I like the sound of a smaller plate resistor in V1(around 150k) and also tried a 5751. That sounded good, unfortunately my 5751 was microphonic, but those 2 moves cleaned up the front end a bit.

My build is in a Marshall style chassis, with the tubes on top. It seemed like a much better design than the split chassis. The tubes do stay a bit cooler as is, and it would be easy to mount a fan on top.

I may experiment with the cap on the 84 cathodes, but i have no shortage of bass as is.
Unfortunately once the amp is in Austin its a bit harder to work on it as all of my tools are in Seattle, so any additional work will be only when really needed
Thanks fellas
Ian