TRex 18watt, gating/farting/blocking?

[Michael Allen]

Been posting on this on the 18watt forum but may need more opinions:

Built an Trex, which is an 18w lite channel and a Trainwreck style channel into an 18w power section. http://mhuss.com/18watt/schematics/18wTRex-S.pdf

I'm getting close to done, but nagging issue here. On the Trainwreck side there is a gating effect like a misbiased or misreferenced tube. When I roll the guitar vol down or let the note ring out and die, it cuts out in a farty manner.

If I roll the treble down and mid up the problem goes away. Any ideas?

This is a sound sample. Treble is turned down and mid turned up at about 48 seconds.

[bluesfreak]

Sounds like you have blocking distortion. Not necessarily out of the PowerAmp stage, it could be also related to the PreAmp stage. Me personnally I would lower the first cathode bypass cap from 22µ to 1µ to reduce the gain hitting the tube in the first stage. Also on the PowerAmp Section, the grid resistors of 8K2 are a bit high in my eyes, I would lower them to 5k6 or even further and change the coupling caps on the Pi down from 22 to 10µ.....
Maybe one of the 18W cracks can chime in and suggest more as this are the major point sticking out to me at the first glance...

just my 2c

bluesfreak

[Guitarist]

That will be an amazing amp very soon. It has a tone that rocks, nice mp3. I didn't know we could post those!
As bluesfreak said, reign in those cathode caps; even 2.2 uF will help but those traditional values are most of your problem.
The last cap prior to the mix section, try a .0047uF or .01uF.
optional: Swap a 470kΩ vs. 150K on the grid to V2.

That should do it!

[pdf64]

There's a lot of gain in that channel, it's likely to be oscillation.
Does the problem only occur above certain vol / treb settings?
V2 is run pretty much wide open, gain of maybe 1800 between pins 7 and 1, with no grid stoppers or HF roll off.
See
New Page 1
for good info on lead dress etc.
Peter.

[Michael Allen]

Thanks for the replys guys,

I had earlier thought it was blocking distortion so I made some changes. Initially I only had 5.1k grid stoppers on the EL84s so I increased those to the schem 8.2k.

I removed the cathode bypass cap on the Trex second stage, decreased the grid leak between stage two and three to 75k to bleed some signal.

Also decreased the PI Grid leaks to 220k. These didn't ever take care of the problem, but I'll keep trying.

And yes, if I turn down the gain a good bit the problem starts to disappear. It must be oscillation of some kind. This is not necessarily a pretty looking amp inside since I shoehorned it into a too-small organ chassis, but the lead dress doesn't look to be too awful. I've built two Trainwrecks so I'm familiar with the usual lead dress importance in these type of amps.

Thanks guys, I'll make some more suggested changes!

[bluesfreak]

Hi Michael,

maybe due to raising the grid stoppers you might have made the problem worse. From the current 2x EL84 amp I like the most (Blackheart BH-15) I learned that you can go down to even 1.5k. Regarding the cathode bypass cap: you modified stage 2 but I think the problem already originates in the first stage...
Let us know what you found out...

cu
bluesfreak

PS; you can find the BH-15 Schematic in the Schematics Section

[pdf64]

'maybe due to raising the grid stoppers you might have made the problem worse'
What's your rationale for that?
A sure test to diagnose the problem as oscillation or blocking would be to bypass the 100k plate resistor on V2 pin6 with a small cap, say 470pF - 1nF. If that stops it, (or moves it's onset to a much higher vol / treble setting) then it's oscillation, if not maybe blocking etc.
That said, I'll be amazed if it's not oscillation.
There's a lot of gain in that amp. If it's not there already, put the 68k input resistor right on the tube socket. Same with the 220k mixers on the phase splitter input V3 pin2. I'd fit grid stoppers to every other grid, 10k-22k right on the tube socket. Peter.

[Michael Allen]

Changed the first stage cathode cap to 2.2uF and increased the grid resistor between stage 2 and 3 to 470K. I also lowered the power tube grid stoppers to 2.2k. As long as I keep the gain turned down to about 1 or 2 o'clock (3 o'clock being max) the problem goes away.

Would like to be able to max it out though. Here's some pics. Any suggestions for dress? I know it looks a bit rough in there.

Need to address some hiss issues next...

[pdf64]

The input tube is very close to an output tube. You'll have a hard time keeping that layout stable. Best to put some metal screening inside the chassis, to seperate the pre and power amp tube circuits/wiring. And put metal tube screens on the pre amp tubes.
Does the problem occur if you disable that 3rd stage, and take the channel output from the 2nd stage? Peter.

[tubeswell]

I know this sounds too simple, but did you try swapping pre-amp tubes out? Sometimes dud tubes make weird distortion sounds

[isaac]

I have never had problems with the grid stoppers on the el84's being 8.2k. I built a very interesting fender preamp into a marshall single channel style clone, and it will put out a lot of gain, so much that it _"safely"_ pushed the el84's into doing 28watts (that is measured with a signal gen, scope, and dummy load, and voltage squared x impedance) Something else I wonder, it seems like teh guys here are telling you to mess with PI mix resistors, you could try just using the two separate inputs into the phase inverter and skip the mix resistors, that might help a little. I found if 12ax7's or similar start motorboating, you will want to increase cathode resistors, and maybe put some small voltage dividers between gain stages if its REALLY bad. at least grid stopper resistors and maybe some impedance resistors can help sometimes. Just my couple ideas.

[Jazz P Bass]

[QUOTE=isaac;302450] (that is measured with a signal gen, scope, and dummy load, and voltage squared x impedance) QUOTE]

Wattage= voltage squared Divided By resistance.
Has to be a typo, I am sure.

[isaac]

A speaker has more attributes than just resistance. It has inductance and capacitance also. So impedance describes it well, but in this matter both of our equations would be correct.