Distortion question

[riz]

I have been plugging away on this Harmony 415. Fixed the trem, and installed a new cathode resistor as the bias was way hot. Wired it up with two 8-ohm speakers in parallel, 4-ohm load, as it came to me stock. It has what I can only assume is blocking distortion (I know... I'm still learning!) at even moderate levels of drive. With one of the speakers unhooked (now an 8-ohm load), it sounds much better, with only the faintest hint of this distortion at high levels of drive (guitar volume all the way up, strumming hard). I know I need to remove the chassis and check voltages etc. again, but today my back hurts, so... I turn to this forum with questions, hopefully I will get some replies that give me something to think about, and look for when I get it back apart. This amp has a pair of el84 in cathode bias. I originally did all my measurements and calculated the new cathode resistor with the amp idling into an 8-ohm load (it was handy). Was this a really wrong thing to do? And, if this IS a case of blocking distortion, does it make sense that I would see these symptoms in an amp such as this, when I change the speaker load? BTW, it is NOT a bad speaker, I have eliminated that, and the results are the same also with a different pair of output tubes. Just wondering if my situation rings any bells with anyone.

EDIT: the amp sounded great into my test speaker, which is a Celestion vintage 10... otherwise I wouldn't have put it back together! And please forgive me for sounding like a totally disorganized newb, in many ways I am! But I spend a lot of time here reading, and I am slowly getting my bench together, and learning more organized testing procedures as I go, with each project. Just wanna say thanks to everyone here for making this forum the awesome place that it is! I am learning so much here. This amp is on the verge of being an absolutely awesome thing to plug a Telecaster into, just wish it were mine!

[riz]

Trying to learn!!

Hopefully I added this schematic right... Since making my original post earlier, I've had a chance to reflect upon what I wrote and have been studying the internet all day, and I got the chassis on the bench. My choice of test speaker was sort of hiding the nastiness, making it almost pleasing... Changing the 12ax7s and power tubes, and playing through different speakers has helped somewhat, but I realize those changes only mask or accentuate what's going on underneath... When I put the amp back together I wound up with a combination that really wasn't working. I'm glad I'm tearing back into this, it's a great opportunity to explore the following: Would I be on the right track to add 1k2 resistors from pin 2 of the EL84 to the junction of C9 and R23 (C6/R22 on the other EL84... what function do those 220K resistors perform anyway?)? These (1K2) would be my grid stoppers, correct, and would be there to prevent blocking distortion? And on pin 9 I could add screen resistors, say, 2K2? My intent here would be to protect the screens and promote tube life, right? I have arrived at these figures as a leaping off point by scouring forum posts, websites, schematics, etc. and realize there is some trial and error to taste involved here. BTW my plate and screen voltage is higher than the schematic, around 362v. Tubes are biased about 11w, not subtracting for screen dissipation, so that should be good. Like I said before, I'm trying to learn and any opinions are welcome! I'm in the process of swapping out a terminal strip for a larger one to make room for the additions.

[riz]

FWIW here's where I've gotten with this: borrowed a scope and set up the lappy using Audacity as a signal generator, works pretty well BTW. At the output of the PI the signal peaks had what looked like spiky mohawks- that was the distortion. A quick trip to the Valve Wizard site and a 470K grid blocker took care of that, now my wave form is exactly like the one shown there in the cathodyne PI section, nipple and all! I clipped in a 1meg pot to try a range of values and nothing was better than the 470K. Also added screen and grid resistors on the power tubes. The amp sounded clear, but lacked a little of the singing sweetness it had before all this started, and I remembered I had changed the cathode resistor to get the tubes down under 12w each at idle (el84). Reading about AC30s, and plate and screen dissipation ratings etc. got me thinking. Went back to the original value cathode resistor, and while it idles hot, almost 14w/tube dissipation at the cathode, the plate voltage dropped and it sounds sweet again. The tubes don't redplate. I dimed it in the dark for a while to make sure. What good is long tube life anyway if it's only because the amp isn't getting played? I checked the power output and it only looks like a little over 8 watts by my calculations. I'm not hung up on that, but it seems low and I'm wondering if that is normal for a smallish OT running a pair of cathode biased pair of el84, not the 15-18w I hear about... Also, I think I'm seeing crossover distortion while scoping the output as it goes into clipping and beyond... normal? It doesn't start developing until after the peaks start to flatten as I turn the amp up. What's left now is to chase out what I think is 120hz hum, probably due to Valco-induced rats nesting... one thing at a time!

[pdf64]

The schematic has a couple of issues - V4 lower lower el84 is shown triode connected, screen linked to plate? And V1a is shown with grid leak bias but no dc blocking cap; so when a guitar is plugged in, the pickup and 68k//68k resistors will shunt the 7M2 bias resistor, ie effectively no bias!
Additionally, the PSU hasn't got a screen node filter, so at high signal levels ripple is likely to modulate the output waveform.
Is the 8 watts into 4 or 8 ohms (sorry if I missed that)? Pete.

[riz]

Good catch on the screen supply filter, pdf, thanks! You are correct it's not there and I can fix that. V4 does not have the screen and plate connected it just looks that way on the schem. Ya gotta look close on this one, it's a bit messy! They used open dots to show connections, but the thing's been copied so many times they're hard to see. The 8 watts is into 4 ohms, clean at the onset of distortion on the scope. I'm still going to look into the V1a situation...