JCM800 clone tone problems

[MarcinD]

Hi guys, hmm I tried few things to improve stability and
1. up to 1nF ceramic cap over plate resistor or anode to cathode on V1 and V2 with no results in stability,
2. untwisted wires from main PCB to panel (V1 out, V2 in, V2 out and V3 in) and separate them with no effect
4. Removed anti pop 2m2 resistors from signal near relay (although it would rather make oscillation even bigger but I wanted to check whether there will be some change)
5. I swapped output tubes to other EL84 without effect
6. I swapped two first tubes in preamp without effect
7. I changed 100nF capacitor on PI in to 22nF to align it with original schematic
8. I swapped cheap foil caps 22nF and 680nF in preamp to Vishay mkt368/orange drop

Honestly I don’t have lots of other ideas and not a clue how to measure stability without vna, I have scope and audio analyzer (audio precision ats1 where I can measure gain and phase but not in the same time without computer connected and I don’t have gpib adapter)), next week I will be able to borrow bode 100 to measure phase and gain margin and I will try to swap components and observe margins. Do You have any other ideas guys?

[JMPGuitars]

You have any other ideas guys?

For us to help, we need to hear and see what you do. Take updated photos of your build, updated voltages using a voltage chart from the downloads section, and record a video showing us what the problems are so we can hear what you describe as instability.

Thanks,
Josh

[MarcinD]

Hi,
Thanks for answer and right away I'm giving what You need
1. Chart (screen in attachment)
2. Photos in attachement
3. Video in in link, volume is at about 40%, gain maxed out, i'm muting strings and just turning volume knob on guitar (https://drive.google.com/file/d/1-pyRhq ... sp=sharing)

Additionally I think I find a way to measure phase/gain margin without vna or at least to check if there is any improvement! I will let You know

[JMPGuitars]

I don't need to look at the video. You can't twist your tube wires like that. Only the heater wires get twisted. Anode, Cathode, and Grid wires need to be kept in certain positions to avoid noise issues. Anode and grid too close to each other will cause squealing. You also need to be able to chopstick the wires (move them around with a non-conductive wooden chopstick to get rid of oscillations and noise issues).

Thanks,
Josh

[Bieworm]

I don't need to look at the video. You can't twist your tube wires like that. Only the heater wires get twisted. Anode, Cathode, and Grid wires need to be kept in certain positions to avoid noise issues. Anode and grid too close to each other will cause squealing. You also need to be able to chopstick the wires (move them around with a non-conductive wooden chopstick to get rid of oscillations and noise issues).

Thanks,
Josh

My thoughts (from earlier) exactly! Why else are we always separating, right angle crossing them wires in the first place? I’m clueless why OP did this or where he got inspired to…

[MarcinD]

All right! I will check it tomorrow! I thought adding capacitance between anode and cathode will improve stability but now when I’m thinking about it it is also shifting with phase. In the meantime I measured entire amp phase and gain margin and I have +17dB gain margin at 3,2kHz xD

[MarcinD]

I resoldered sockets for preamp and PI, I think it’s a little bit better but it’s hard to tell, phase plot looks very similar 180degree shift at 3,4khz. I didn’t had time to do the same with output tubes. The one stage still twisted is FX loop but it is bypassed. I probably need to untwisted also ot cables right?

All best!

[JMPGuitars]

You can keep OT cables twisted. What we said was in reference only to the tube sockets.

Twisting wires does two things: 1. it keeps AC signals behaving; 2. it keeps things neat.

That's fine in most places, but with the tube sockets you need to be able to adjust those wires.

Now, ignore your phase shift and all that. Listen to the amp and see how it sounds. If there's noise, try to find out where it's coming from. You can tap things with the wooden chopstick, you can move wires with the chopstick, or you could use an oscilloscope moving along the signal path from the input jack. Once you have your signal path clean and stable, then you can make notes as a future reference for bode plots and phase shift data.

Thanks,
Josh

[MarcinD]

Yeah thanks, I will try to untwist also power tubes tomorrow and do some troubleshooting but I will be able to do only scope troubleshooting because I’m leaving in apartment with neighbours behind walls but I think it should be visible on step respond (overshoot and/or ringing). For real tests with speaker I will need to probably wait a little but either way thanks for hints, I will let You know what will be the outcome

Edit: one thought is it possible that I should connect speaker out of phase? I measured frequency respond and it looks like if I connected speaker in phase signal is in phase with input up to 3,4khz what is visible at step respond as weird spike/oscillation, over 3,4khz it looks that signal is quite ok. (Screens from scope in attachment)
Edit2: it was stupid I just realized that phase will depend of such things like distance from speaker :p

[MarcinD]

So,
Today I was able to perform some additional measurements and the conclusions are
1. I did measured and simulated entire preamp (without eq) and the results is quite similar in character to but it looks that bandwidth is slightly shifted (-3dB from bottom at 1,4kHz and top 19kHz) (1,6kHz and kHz in simulations 20kHz). Results in attachment.

V_meas_1.jpg

IMG_4407.JPEG

IMG_4406.JPEG

2. I measured tone stack and hmm... it looks like it is attenuating more than it should maximum what i get in entire spectrum was -6dB with treble at max... I checked values 2 times and those are ok, maybe I measured something wrong or maybe I have crappy ceramic cap. I will try to measure it impedance tomorrow:)
3. I measurent voltages on plates with cranked amp and it is still looks strange (screen in attachement)

IMG_4408.JPEG

4. Tomorrow I will recheck all the values in PI and power amp

[MarcinD]

Hi,
I untwisted power sockets today and rechecked once again all values in PI and Power stage and it looks ok. I did some tests today (with speaker and guitar ) and it looks that it works better now (less oscillations) but I will be able to really test it probably during the weekend. Thanks guys for Your help! I will probably spend some time to do measurements as I would like to understand how it works from egineering perspective, funny thing is that JCM800 amp looks like not stable at all from pure engineering perspective
Second thing is about this output power theoreticaly it should be about 18W but I'm measuring undistorted 12W and when amp cranked to maximum about 15W. I'm wondering if this is because of OT efficiency? I'm measuring with Audio precision ATS-1 (audio test system with power meter function).

[JMPGuitars]

Don't look at it like an engineer until it works correctly as a guitar amp. If you look at it like an engineer, you'll be focusing things into the hi-fi realm of things being clean and pristine. In reality, a tube guitar amp is ALWAYS distorted. Even a clean guitar sound has some level of distortion. If you have the gain cranked, then you'll see your sine wave start turning into a square wave.

Also, I'm guessing you're looking at the power on the output stage across a speaker load. Tube amp power is rated by the output tube dissipation, not the speaker load. 18 to 20W on the output tubes will look lower calculated across the speaker. You can calculate tube dissipation yourself , but I would start by looking at this calculator: https://robrobinette.com/Tube_Bias_Calculator.htm

Once the amp is biased where it should be, and you've chopsticked the wires and got rid of all the oscillations and unwanted noise, then you can turn back fully into an engineer to study what's happening with the amp during normal operation.

Thanks,
Josh

[MarcinD]

Ok, maybe I will have possibility to make real tests during the weekend. Now I found two additional things.
1. Wires to V1 are very sensitive, with tapping them (especially grid) with stick it is very loudly audible in speaker. Is it normal?
2. When I set gain to 0 and crank up master I can hear despite a little bit of hum and some noise a AM radio? I think it have more sense than FM have You ever encountered such problem?:)

[JMPGuitars]

Ok, maybe I will have possibility to make real tests during the weekend. Now I found two additional things.
1. Wires to V1 are very sensitive, with tapping them (especially grid) with stick it is very loudly audible in speaker. Is it normal?
2. When I set gain to 0 and crank up master I can hear despite a little bit of hum and some noise a AM radio? I think it have more sense than FM have You ever encountered such problem?:)

Input wires need to be shielded wire. Any long runs from the front to the back of the chassis should also be shielded. You probably already know this, but make sure the shield is only connected on one end.

"Grid stopper" resisters: it looks like you might have a grid stopper on 1 side of V1, but not the other. You may need to add grid stopper resistors to other stages as well. Pins 2 and 7 on 12AX7 tubes. If a stage has the grids in parallel (connected to each other), then only 1 grid stopper resistor is needed. Values can be from 10k to 33k.

[MarcinD]

Hi, Ok It looks that I had to add grid stopper directly on socket (I had them on PCB), now I would like to focus on noise/hiss/hum etc. After I untwisted wires a little bit of hum is audible I will try to find it wih scope but I'm guessing that it will from heaters but I have other question there is quite excessive noise when amp is cranked up and from time to time I can hear hmm... high frequency noise. I was able to record it (end of recording) https://drive.google.com/file/d/1UVuwdI ... sp=sharing. Do You have any idea what is the reason behind it?

All Best!

[JMPGuitars]

Yes, direct to socket is the most effective placement for grid stoppers.

Your recording is noisy, but not helpful. Without a comparison to guitar signal, we don't know what the signal to noise levels are. It could be a noisy room, noisy mic, etc.

The noise at the end is hard to place. I would suggest using a wooden chopstick as previously mentioned, and see if moving wires has any impact on the noise levels.

The noise didn't sound like heater wire noise to me.