Have you tried disconnecting the phase inverter from the power stage to see if the power stage is loading the PI in an asymmetrical way to cause early clipping?
Tube Amp connections : 0V vs EARTH vs HEAT vs CURRENT
- Thread starter systemtruck
- Start date
Help Support GroupDIY Audio Forum:
Have you tried disconnecting the phase inverter from the power stage to see if the power stage is loading the PI in an asymmetrical way to cause early clipping?
systemtruck
Well-known member
I suppose just kill the connections after the C39 and C40 bypass caps and monitor the audio of either one of those leads? Do both those leads need to be dummy loaded down with some specific value so that the PI performs as it should? Or will the PI outputs perform just fine without any load other than an audio connection? And if so, what input impedance will i need for the audio input? My interface has 15k by default, but maybe i can add something more substantial like a series 200k so that I’m not loading down the PI output. It’ll create a voltage divider but it’s probably still plenty loud.
systemtruck
Well-known member
Ok i made a setup that seems to function pretty well for this test, and i think i can conclude that the EL84’s are somehow causing the PI to self clip FAR FAR FAR before it normally would self clip on its own.
I disconnected the PI’s bypass cap outputs and patched the anode’s and cathode’s respectively into a pair of 150K series resistors and connected those to + and - of a balanced line input to audio interface. I chose 150K so that it was a not heavy load, just in case i was going to load it down too much with a 15k line input. I connected the local 0V bus bar point to the ground of the audio input. Seems like great audio, like a solid DI signal.
There are two volume pots.. one within the preamp stages, and one just prior to the PI. Before, i had to keep both knobs at about 15% to avoid this particular clipping. Now, i can keep them both at about 50-60% before the clipping starts. They are linear pots too, so there is a lot of extra juice there between 20%-50%. Surely these levels would be blasting the great sounding overdrive of the amp, before this PI clipping sound would even occur.
But with this current setup, if i crank them enough, i will get a somewhat similar clipping. However i do have to say it is a lot less ugly somehow.
So is the PI getting badly loaded down by the EL84’s or something?
Should i maybe also measure my pair of 470K resistors at the EL84 inputs just to make sure somehow it’s the wrong resistor? I simply used the ones from the old amp.
What else can cause a PI to dramatically self clip only when connected to output tubes? Shouldn’t my 82k grid stoppers that i currently have in place help to stop this too?
Matador
Well-known member
I think you may end up chasing your tail quite a bit until you get your output tubes operating together with two known good (matched) tubes.
It's also possible your 'hot' tube may be getting driven into significant grid current. The cathodyne will behave strangely when one side is overdriven because the output impedances aren't equal between the two driven sides.
It's also possible your 'hot' tube may be getting driven into significant grid current. The cathodyne will behave strangely when one side is overdriven because the output impedances aren't equal between the two driven sides.
Or the imbalance in the output tubes is causing one tube to hog all the current, and the other tube is almost cut off. Again, I don't have much experience debugging power tube stages, but that is what would happen in a long tail pair of transistors, so I would definitely not trust any behavior that occurs with one of the power tubes almost red-plating. Getting the output stage working properly is the first job.
systemtruck
Well-known member
I did in post# 39, but here it is again with the new grid stoppers, as well as one correction to a rail cap leg that was missing.
If it doesn’t appear properly, I’ll re-up shortly with different file type.
But I believe if you click on this it will expand well enough.
View attachment Fisher R-20 Amp Feb 14 2025.jpeg
systemtruck
Well-known member
I ordered a pair of seemingly well matched EL84’s, pretty sure a close generation to the 1950’s British Mullards that my current pair is. Very modest price, so a low risk here.
I’ve thought about trying to keep these Mullards i have in there because buying a pair of these exact ones is very high priced, and I considered attempting to adjust each tube’s bias to see if i can get them to work better and not cause this clipping. But from what i am learning there are several factors in a well matched pair and i guess individual biasing might not help my situation at all if one tube is truly messed up. (Which seems likely at this point)
Also, i checked my first recordings of this Fisher amp when i first got it and performed a simple mod to play guitar with it, and sure enough this clipping is very much there. i just didn’t notice it because i was playing in a certain way. So i think this one tube has been faulty from the start, and it’s not something i caused.
I’ll report back when I receive and pop these replacement EL84’s in.
I’ve thought about trying to keep these Mullards i have in there because buying a pair of these exact ones is very high priced, and I considered attempting to adjust each tube’s bias to see if i can get them to work better and not cause this clipping. But from what i am learning there are several factors in a well matched pair and i guess individual biasing might not help my situation at all if one tube is truly messed up. (Which seems likely at this point)
Also, i checked my first recordings of this Fisher amp when i first got it and performed a simple mod to play guitar with it, and sure enough this clipping is very much there. i just didn’t notice it because i was playing in a certain way. So i think this one tube has been faulty from the start, and it’s not something i caused.
I’ll report back when I receive and pop these replacement EL84’s in.
Similar threads
