“Blatty” type distortion problem

[andresound]

Compliments of the season to everyone and i hope we all have a good 2019.

I have just converted a small 1x 10 into a practice amp. I used a Hammond AO-66 reverb amp as a donor and just added a standard 100k/1.5k/10uf 12AX7 gain stage (with 1meg volume and 250k tone after) in front of the amp. Voltages are similar to Sluckeys conversion post of an AO-44. (Did not use his Fender preamp stage due to space restraints)

97C0CBCF-DF5B-44B2-80D7-AA7C1DF1C90A.png

The amp sounds good clean. The clipping has a strange “blatty/nasty” sound to it. I removed the first stage cathode cap, but this did not help. I then decided to scope the stage and noticed that the V1a - pin 9 was clean 1k sine wave, but v2a - pin 9 had a strange wave form when drivin hard. Here is a pic of the 2 triode outputs. Does anyone have an idea of what would cause this behavior?

V1a

2B70AC98-CCAD-4089-A5AA-949CDE3A545B.jpeg

V2a

19AE61B5-B056-4D4E-9FA2-CD3CD900D64C.jpeg

[Malcolm Irving]

Merlin shows that when a cathodyne PI is trying to overdrive the output stage, you can get a spike on the voltage waveform of the
cathodyne's anode output:

http://www.valvewizard.co.uk/cathodyne.html

Maybe that is what is happening here?

[andresound]

Thank you Malcolm. Interesting reading.. still a little above my pay grade.. but I’m getting there. So to be clear, do I add a 1meg grid stopper here? Or should I have a 470k/470k voltage divider here (reduce 1meg to 470k)?

FC877CBA-15D0-456F-849A-174D1D667274.jpeg

[Malcolm Irving]

Try a 1meg grid stopper, between pin1 of V2A and the junction point of the 0.02 cap
and the existing 1meg (i.e. on the grid pin of V2A in place of the existing direct connection back to
that junction point).

[Reeltarded]

Try a 1meg grid stopper, between pin1 of V2A and the junction point of the 0.02 cap
and the existing 1meg (i.e. on the grid pin of V2A in place of the existing direct connection back to
that junction point).

It is nice to have you around, Malcolm. Cheers!

[andresound]

Thank you. I will try this tomorrow morning and revert.

[andresound]

Ok.. progress. The addition of a 1meg grid stopper to v2a, slightly improved the sine wave, but did not make a big difference. However, placing grid stoppers (470k) on the grids of the Pentode sections, solved the problem. Thank you Malcolm for your input .
Now I have learned the cause/solution of a sine wave that looked like that. I will revisit the literature you referenced, and try to REALLY understand the physics behind it.

[Lynxtrap]

Great! While you're at it, add some screen resistors as well. 1K - 2.2K, at 3-5 Watts.
At least the tubes will thank you.

[andresound]

Will do. Thanks guys for the help. Herewith pics of the build so far. All of the original coupling caps were ceramic, so plan to change those eventually. Just need to play her a bit.

[pdf64]

...The amp sounds good clean. The clipping has a strange “blatty/nasty” sound to it. I removed the first stage cathode cap, but this did not help. I then decided to scope the stage and noticed that the V1a - pin 9 was clean 1k sine wave, but v2a - pin 9 had a strange wave form when drivin hard. Here is a pic of the 2 triode outputs. Does anyone have an idea of what would cause this behavior?...

Just to note that scoping waveforms within an NFB loop, especially if the circuit is being pushed to / beyond its linear region, will result in weird looking traces, due to the error component of the signal being significant.
It's generally best to open the NFB loop when undertaking such an exercise.

470k grid stoppers on the power tube control grids may have been excessive, 10k may have been plenty; it's often a case of anything is better than nothing

[Littlewyan]

470K is the maximum grid leak size you can have on ECL86 valves and at the moment you're double that (this takes into account grid stoppers). Try lowering them to 10K. If you start getting bad distortion again then raise them up to 100K. If you need to go any higher than I would lower the size of the grid leaks.

[Lynxtrap]

Sorry, I hadn't had my morning coffee so I thought those where EL84 tubes. The value for screen resistors I recommended may have been on the high side, I'm not familiar with these tubes. Maybe something like 470R could be a starting point.

[andresound]

Wow.. thanks for the feedback guys!!!
The “blatty distortion” still is there. However the strange waveform has been sorted.

470K is the maximum grid leak size you can have on ECL86 valves and at the moment you're double that (this takes into account grid stoppers). Try lowering them to 10K. If you start getting bad distortion again then raise them up to 100K. If you need to go any higher than I would lower the size of the grid leaks.

Please advise as to which grid leak/stoppers we are referring to. Remember ..this is a triode and pentode in the same envelope. To clarify.. grid leak or stopper on the triode section or pentode section?? I apologize for the stupid questions (I am still learning). Please remember my limited knowledge, tha is the reason I am a member on this forum.

[Littlewyan]

Pentode section. So going by your post you currently have 470K grid leak resistors and 470K grid stoppers. The limit on an ECL86 Pentode section is 500K so you're over the limit. This may not be an issue but it may also cause thermal runaway and a shorter valve life. Try lowering the grid stoppers right down to 10K, if this brings the problem back try going up to 100K. If that fixes it then leave it there. Going over the limit a bit probably won't hurt.

With regards to the screen grid resistors I couldn't find any info on these in the datasheet. Nor could I find anything on Google regarding it. Maybe they don't need them and the decoupling resistor (or choke resistor) is enough at 1K.

[Lynxtrap]

With regards to the screen grid resistors I couldn't find any info on these in the datasheet. Nor could I find anything on Google regarding it. Maybe they don't need them and the decoupling resistor (or choke resistor) is enough at 1K.

As I understand it, decoupling resistors or chokes don't really function as a screen resistors. That is, basically dropping voltage for the screen along with the plate when the tube is pushed. Some old or small/cheap amps sometimes lack these resistors.
But I guess the "need" for them might depend on the idle difference between plate and screen, and how much the plate actually drops. I've just kind of learned that screen resistors are good practice.