Hello,
Does anyone have or know where I can find plate curves for the el34 with the screen at 350 volts. I've been monkeying with a pair and am really getting comfortable with drawing load lines and making power and distortion calculations, but I can't get more than 25 watts out of this thing to save my life (my life is literally being consumed by this amp).
475V on the plates, 360 on the screens, 1k shared screen resistor, 6.6k output transformer, gave exactly 50W last time I tried it. Are you using a 3.3k load?
"Enzo, I see that you replied parasitic oscillations. Is that a hypothesis? Or is that your amazing metal band I should check out?"
Perhaps the PSU voltages are sagging a lot at full power? Do you have enough reservoir capacitance to keep them up?
The amp is an old P.A. head I found on craigslist for $40. It is all octal and has a stancor power tranny and a bendix output with taps for 8, 100, and 500 ohms. The 8 ohm tap has a turns ration of 37:1 and the ratio from the top of the 8 ohm tap to the 100 ohm tap is 13.5:1. I cleaned and painted the chassis and transformers and had the steel top sandblasted and painted it gold with white racing stripes with acrylic urethane. I do a lot of work for a great banjo making company called Huber Banjos. They do all of their nickel plating in house, so I buffed the rusty steel handle and had it plated. I have not written a schematic, but the signal chain goes input-1/26sl7-tone stack-1/26sl7-6sj7 in triode-6sn7 ltp-power tubes. The rectifier is a GZ34 into a chokeless PI filter with 40uf at the output center tap, 40uf at the screen node and 33uf for the preamp tubes.
I can get the screens up to 28 volts rms each, and cleanly by cranking the bias up to 40 volts, but the output is very distorted.
The plates idle at 502v with the bias at 40v, and at 452 biased at 25v. There is a drop of about 20v between idle and maximum clean output of 16-20 watts rms. The mains voltage has varied between 119v and 126v over the last 12 hours.
The best I've been able to do was 30 watts rms with a pair of new JJ 6l6s with each tube dissipating 35.5 watts (screen and plate, measured across 1 ohm resistor and screen resistors, screens accounting for about 3.5 watts of that total) with the line voltage at 125v.
I've got it running on 6550s right now. This is what the output looks like with the bias at 47v and a totally clean 92v peak g-g. The clean output is 25w
I'm beginning to think that to get more juice out of it would need to go solid state for rectification with larger filtering and I'm think I can hear and see the core of the bendix saturating. I bought the thing to experiment and learn and probably sell at some point. I think that point is now, but do I order (both local electronics joints had to go out of business?) larger caps and a new output tranny or match a tube to the amp (maybe the 7591) to get some good sounding power section breakup before the core saturates?
Thanks for listening ampaholics.
I've got to run off and make some money from banjos.
Sounds like you might have a rubbish OT. A 37:1 turns ratio works out to roughly a 10k load. The section between the 8 and 100 ohm taps might have the correct ratio, but it could be wound with too-thin wire.
"Enzo, I see that you replied parasitic oscillations. Is that a hypothesis? Or is that your amazing metal band I should check out?"
I agree, a good conventional OT would be a good place to start.
Marshall 50W Output Transformer for Plexi, JMP and JCM800, 784-139
T
Keep Rockin!
Terry
I think I figured it out.
I don't think the tranny is junk, but just intended for P.A. use. I think the 100 and 500 ohm taps are for chaining a bunch of speakers in series throughout a school, factory, etc.. Does a hundred yards of speaker wire make you sound like Charlie Browns principal? The 8 ohm tap is right for 6v6s, but loaded with 4 ohms gives a 5.5k load, so is right for the usual suspects.
Overdriving the power section is squaring the wave completely and moving the amp into class B. The power section has to be designed as such, meaning the plate supply voltage has to be moved downward to accommodate the class B reflected impedance of 1.4k. This is why a Trainwreck has a plate supply voltage of 390v. It is designed as a class B amp. I threw a 5y3 in there and everything is right. I'm going to try a 5u4 and 5v4 also, as the have different characteristics, but all drop enough voltage to get this amp to work right.
I paid for this knowledge with much time. I always have to do things the hard way.
I'm totally failing to follow the reasoning here. You open the thread by complaining that the amp is making less output power than you want. You conclude that lowering the plate voltage is the way to fix it?
If the amp was designed for 6V6s, the OT could well be somewhat lossy, and the power supply saggy, under the heavier demands of the larger tubes. That's probably where your missing power went.
Yes, the 100 and 500 ohm taps are for driving a bunch of speakers, but probably all in parallel and each one with a little step-down transformer inside. A very common arrangement known nowadays as a 70 or 100 volt line.
"Enzo, I see that you replied parasitic oscillations. Is that a hypothesis? Or is that your amazing metal band I should check out?"
I'll try a new transformer
There are currently 1 users browsing this thread. (0 members and 1 guests)
Bookmarks