Stuck at power section

[Russ]

I'm in the middle of a build that has a Marshall jcm el34 50w power section/PI, and I've got the power supply, power amp and phase inveter wired up so far. I'm testing this by plugging into the input of PI, w/ a separate preamp and also w/o a pre, just guitar. Either way, I get sound out, but the problem is when I play softly, it sounds clean and normal, but if I hit the strings hard, especially low notes, I get this brief & loud 'splat' kindof distortion sound. Also, the presence control doesn't work. I've checked the wiring so many times I can't find anything wrong. I'm using a scrounged Stancor hifi OT for 6l6s with a 6600 ohm prim.

Any ideas about what could be causing these sort of symptoms?
why does it make this awful 'splat' sound?
and why would the presence control not do anything?

The voltages seem ok:
About 445v on plates, 438v screen, fixed bias at about 31ma/tube,
PI plates 265v & 270v, PI cathodes 39v, PI grids 25v & 26v,
I tried swapping tubes, there's no excess hum, heaters working fine.
Taking out the neg feedback by grounding the PI tail had no effect.

After 3 days of trying stuff, it's getting frustrating.
Thanks,
Russ

[loudthud]

The 6600 ohms is a little on the high side but you should be able to make it work. Does the amp make the splat sound with the feedback disconnected? What ohm tap is the feedback connected to? Have you tried reversing the OT leads to the plates?

[Russ]

To get 3300ohm pri impedance, I'm dividing the sec impedance in half to compensate for the imp mismatch. So when I connect the 8 ohm speaker, it's hooked up to the 16 ohm OT tap. It makes the splat sound either way, as I said disconnecting the NFB (grounding the PI tail) doesn't change anything. As it is now NFB is hooked up to the 16 ohm OT tap - the last output tap. The OT has 16, 8, 4 ohm taps, I've tried all the taps, same thing. Haven't tried reversing the leads, but I've made this mistake before and I don't get the loud howling, so I don't think that's the problem, but I'll try it - hate that howling though!

[loudthud]

OK, you should be good with that OT as long as it can tolerate the power this amp will generate.

When feedback is connected there should be a substantial reduction in voltage gain compared to operation with it disconnected. About as much as turning a guitar's volume pot from 10 down to 5 or 6. Are you seeing this?

Have you tried any changes to the power supply? Describe the components you are using. Sometimes a certain choke and cap combination will do strange things between notes or cords. Is the heater supply grounded (real or virtual center tap)?

Where are the grid stoppers for the power tubes located? Some of the Marshalls I've messed with were real sensitive to their location. They had to be right at the tube socket and anything less than 5.6K didn't work.

[Russ]

The OT's were originally for 6L6s, should be fine.

I'm hearing no volume change when I eliminate the NFB path - that's the weird thing!,
sounds the same either way.

PS is coming off of a Hammond 272HX 300vac, one 40uf atom cap for plates & one for grid, a smallish Hammond choke rated for 50ma, but I've used this same choke before - without prob.

Heaters are virtual CT raised to about +23v right now.

Grid stoppers are 5k6 & soldered into board, with about an inch oif wire to socket.

....Oh shit....I just tripped over my speaker wire and broke the cliff-style speaker jack
I'll be back in a bit, maybe this'll fix it.........

[Rob Mercure]

Loudthud: What load impedance would you consider optimal for the 6L6? While we're not dealing with the HiFi world the RCA tube manual suggests 6.6 K for 350 Vp and and 5.6 - 6.0K for 450V so Russ seems ball park to me. In fact the 3.3K seems to be on the low side IMHO but I don't tend to think this is his issue. Hope you doesn't insult you - no intention on my part - but I was just scanning through and the number jumped out at me.

Cheers,

Rob

[loudthud]

Most 50W Marshalls have a lower primary impedance than what seems normal in the US. I think 3.4K is pretty common. It may have something to do with EL34's. If the plate voltage swings too low, the screen current goes really high. With 6L6's this isn't as much of a problem but it takes more to drive them. Fender settled on about 4K for a 50W amp but B+ was a little higher. I can't do the math in my head but I think you need more than 450V B+ to get 50W out of a 6.6K primary. These things all have their effect on how an amp sounds so for a British flavor, one should lean towards the 3.4K. I think at 6.6K a 50W amp might sound on the other side of Fender, clean and sterile. 6.6K is a good choice for amps in the 25W to 30W range with appropriate B+. Just my own personal bias.

[Russ]

....well, I've got it back together, but it's still not working properly.
I've tried reversing the OT leads and that didn't make any difference.
It seems there's no NFB. That explains why the presence doesn't work and why when I try to eliminate NFB, there's no change.

I've got 100k and 4k7 for the NFB voltage divider. When I parallel the 100k with another 100k (total 50k) I get a squealing feedback when I strum hard, but nothing else.

I can't figure out why there's no NFB. It's all hooked up.
Should the NFB still be effective when just pluggin straight into the PI? Is the signal too low with just a guitar and no preamp?

[loudthud]

Feedback should be effective at all signal levels. If you amp is oscillating at some ultrasonic frequency, the DC levels should be low and there will be a large AC voltage at the output.

[Russ]

Here's a few more symptoms:

While in standby, I get a little hum through speakers, which changes in volume as I switch speaker imp taps. Hum goes up at lower imp settings.

When I switch the NFB (which doesn't seem to be there) in or out I get a loud thump, but no change in volume or a functioning presence control.

Could these be signs of a bad OT and if so, what could be bad about it? All the taps show continuity with a small amount of resitance: 241 ohm pri, .5, .7, .9 ohm sec taps.

[loudthud]

Hum through speaker while in standby is probably magnetic coupling between power and output transformers. Try rotating one carefully while power is on.

Loud thumps are normal when messing with PI tail resistors. You should even get a small one messing with the 100K to the OT tap.

The test for a bad OT can be done with a flashlight battery and a neon bulb or you can use the 6.3V heater supply and measure volts at the primary. Should be able to find threads about that here at Ampage, do a search. If you have continunity and no leakage (resistance to frame) then you could have a shorted turn. The bat-neon test checks if you can generate a high voltage spike with the battery, the heater volts test checks the turns ratio which will be way off if there is a shorted turn.

[loudthud]

Just saw a thread on another board where the builder had not grounded one end of the OT secondary. He used Cliff jacks and had no NFB.

[Russ]

LoudThud, thanks a bunch for all your input, I really appreciate it!!

Well, I've made some progress.
You were right about the hum coming from PT OT coupling. Getting the OT it away from the PT stopped the hum.

I didn't test the OT, but swapped it with a 4 x EL34 OT that was working fine the last time it was used, so I'm assuming it's still working fine.
Hooked it up & still got the same no NFB, no presence, awful 'splat' sound problem.

I've checked all the wiring, tried rewiring, replacing every part in the PI, I don't know what else to do.
Any thoughts on where a mis-wiring or some bad part could cause no NFB and the rest of the problems?

[Russ]

NFB

Also, when I change the NFB resistor from 100k to 47k, it becomes unstable and as soon as I hit a note I get a sqealing oscilation that doesn't stop.

What could cause all these symptoms?

[DirtyGeorge]

Your PI voltages look suspect to me. The voltage difference between the PI cathodes and grids (the PI bias voltage) should only be 1 or 2 volts if you are using a 12ax7. Check your cathode resistor. Are you using coupling caps on your PI input. If not then the input device could screw up the bias.

And I would leave the NF disconnected until the basic circuit is working.

DG