JCM800 clone tone problems

[MarcinD]

Hi There!
I'm new on forum and quite new on building tube amps. I decided to try to build PP EL84 based JCM800 clone with tube buffered FX loop and I did it, it looks like it is working but there is some tone problem especially at higher volume (it is not creamy crunch I was expecting but rather ear piercing harsh tone with some stability issues). I'm sure You will be able to give some hints where to look!
Main problems:
1. I think output power is to small (1kHz signal is giving max 9,59V undistorted signal over 8ohms resisitor (11,5W)
2. Master is usable up to 3 out of 10. Then tone starting to be really unpleasant and stability become a problem.
3. In general I would expect rather creamy tone and I'm getting shrilling and unpleasant especially on higher volume.

My guesses
1. Choke can be to small hammond 194A (however this replacement for 2x6V6 design)
2. Maybe I should try to connect common tap of OT before choke as in soldano?
3. PI is starting to have strange "spike" in waveform (on one side of wave form) at higher gain and in general maybe there is to much gain in PI for EL84?

Schematics is in attachment with measured voltages!

additionally below You may find all measurements I've done:
Stage / B+ / A / K // input AC (RMS) / out AC(RMS) / gain
V1 / 254V / 195V / 1,5V // 50mV / 2,72V / 34,7dB
V2 / 254V / 230V / 2,2V // 50mV / 325mV / 16,3dB
V3 / 261V / 147V / 0,83V // 49mV / 2,2V / 33,4dB
V4 / 260V / 260V / 148V // 2,26V / 2,35V / ~0dB
Inv1 / 280V / 148V / 36,8V // 50mV / 415mV / 18,3dB
Inv2 / 280V / 138V / 36,8V // 50mV / 382mV / 17,7dB

Output tubes - autobias (130ohms cathode resistor), Raa = 8400ohms
Tube / B+ / A / Sg / K
PV1 / 357V / 346V / 347V / 13,5V
PV2 / 357V / 346V / 348V / 13,5V

Caluclated output power on 8ohm resistor
Vo = 9,59V (measured just before clipping on output tubes)
Po ~ 11,5W which is rather low for 360V B+ (Am I right?)

I'm using OT (Hammond 1750PA)

I will really appreciate all Your hints!

[Bieworm]

Hi! Welcome to this community

I read your post yesterday and I decided to think on it somewhat before answering.
First off the name JCM800 is a wide term. There are 800’s with a smooth tone and there are these aggressive harsh metal beasts. So it’s hard to tell which version this amp is based on, since I’m not an 800 connoisseur. Don’t forget speakers can have a lot to do with it. A V30 will harshen the tone, while a G12M will soften the edges.
Anyway, could you post some detailed photos and maybe a sound clip so we can see and hear the problem child?

[MarcinD]

Hi,
Thanks for Your answear!
Unfortunately I will have opportunity to record something in friday (I did record sample (https://drive.google.com/file/d/1Lm4fai ... sp=sharing) on relatively low volume and bare speaker but i think the overall tone character is audible).and I'm not saying that the tone is very bad but it is a little harsh, not creamy that I'm able to get a lower volume. It is like the amp is choking a little, It is very hard to explain it:) Additionally I have stability issues especially when I'm boosting guitar signal with Tube Screamer. My biggest concern is why I'm getting only 11.5W out of this power amp? If I correctly understand I should be able to get at least 15W and maybe here laying the problem!
To answear Your questions
1. I have two cabinets, one with Creamback G12M and second with Jensen C12N.
2. I'm based pream and PI on JCM800 2203 Master volume (https://www.drtube.com/schematics/marshall/jcm800pr.gif)
3. Picture of amp in attachment:)

[JMPGuitars]

I see two big issues.

First, you calculated output wrong. Use this calculator: https://robrobinette.com/Tube_Bias_Calculator.htm

You have 2 EL84 power tubes, correct?
347 plate voltage, 13.5 cathode voltage, 130Ω cathode biased.
That's 16.3W PER TUBE. You're at 135.8% dissipation. We usually target 85%. You're way over, and you need to drop it. I'd start by using a 180Ω cathode resistor, but you might need to go higher, so if you're ordering, also get 200, 220, 240.

Second, your screen voltage should be lower than the plate voltage. Get a couple 3W 1K resistors and put them on the EL84s at the screen grids. If you already have 1K screen grid resistors, raise it higher.

Thanks,
Josh

[MarcinD]

Yeees… sorry cathode resistor is 180ohms not 130, but it still looks like there a little bit too much dissipation (12,5W). I will try to lower it additionally, do you think better will be to increase resistance or to apply additional negative voltage from psu? I have such capability. Additionally what will be sweet spot for biasing EL84.

[JMPGuitars]

Yeees… sorry cathode resistor is 180ohms not 130, but it still looks like there a little bit too much dissipation (12,5W). I will try to lower it additionally, do you think better will be to increase resistance or to apply additional negative voltage from psu? I have such capability. Additionally what will be sweet spot for biasing EL84.

Sweet spot for biasing EL84 (for us on 18w) is 85%, but that really puts you closer to 20W, which is fine.

With 347V plates, 13.5V cathode and 180Ω resistor, it's 11.8W or 98.3% dissipation. I would try 200 or 220Ω.

[MarcinD]

All right,
One more update,
1. I have currenly installed 1,8k screen resistors.
2. I Changed cathode resistor to exactly 203 ohms and I added additional negative bias voltage (in addition to catode voltage)
3. Installed 1ohm 5% resistors on anodes

Now I measure
High voltage (to GND) = 367V
anode to cathode = 343,9V/344V
Screen = 345V/346V
Cathode = 12,9V
Anode currents = 28,5mA/26,6mA

Screen is still higher than anode voltage and I have some unequal currents. Tomorrow I will try to replace tubes. Additionally do You think more that 2k ohm screen resistor still will be ok?

Additional question
Do you think in general this voltage is to high for EL84?

All Best
Marcin

[Bieworm]

You want that B+ near 345V to get in the obvious territory. Zener diodes might be a good help

[JMPGuitars]

All right,
One more update,
1. I have currenly installed 1,8k screen resistors.
2. I Changed cathode resistor to exactly 203 ohms and I added additional negative bias voltage (in addition to catode voltage)
3. Installed 1ohm 5% resistors on anodes

Now I measure
High voltage (to GND) = 367V
anode to cathode = 343,9V/344V
Screen = 345V/346V
Cathode = 12,9V
Anode currents = 28,5mA/26,6mA

Screen is still higher than anode voltage and I have some unequal currents. Tomorrow I will try to replace tubes. Additionally do You think more that 2k ohm screen resistor still will be ok?

Additional question
Do you think in general this voltage is to high for EL84?

All Best
Marcin

Your dissipation is good with those voltages. Yes, you can go much higher if needed on the screen grid resistor. I would go to 3k or 5.6k and see how the amp sounds.

Currents don't have to be perfectly even.

Yes, those voltages are fine. Look at the voltage charts in the download section, and you'll see what we typically go for.

345VDC (+/- 5V) is the target B+ for typical 18W amps. You can target that if you want, but if you like it where it is, it's fine.

Thanks,
Josh

[MarcinD]

And just to clarify by B+ you mean central tap of OT to cathode right? Without drop on cathode resistor? I will try to adjust screen current to have it about 10V below anode. Hopefully it will help for tone and output power my drummer have really loud rock cymbals :p

[Bieworm]

And just to clarify by B+ you mean central tap of OT to cathode right? Without drop on cathode resistor? I will try to adjust screen current to have it about 10V below anode. Hopefully it will help for tone and output power my drummer have really loud rock cymbals :p

That is one way of saying B+
Generally we call the B+ the highest DC voltage in the amp. It is the result of the rectified AC high voltage.
As to the volume question, an 18W amp is ideal to use with a moderately loud drummer. When you play with a softer sounding drummer like I do, then you don’t often get to extract the maximum out of an 18W. It’s just under the sweet spot and that’s a pity. I’ve found out that I have my sweet spot with 8 to 10 watt amps ran through a 95’ish dB speaker. I recently finished an amp that is spot on. It’s a tweed bassman circuit powered by 2 6AQ5 tubes and a B+ of 285V. The downside of playing at the sweet spot is that there is not a lot of volume left beyond that spot, so there’s no way up from there… only more saturation. But you can’t get THAT tone with a louder amp. This is the tone that keeps you wanting to play all day

[Bieworm]

About the instability… I notice all your plate/grid/cathode hookup wires from board to socket are twisted. I’m not a fan of that approach. You can’t separate the wires to tweak lead dress in case of coupling issues.

[MarcinD]

Ok,
So I changed screen resistor to 4,7k/2W and set bias to 29/27 mA.
Now I have
B+ 366 V
Va1 356 V
Va2 355 V
Vs1 345 V
Vs2 345 V
Vc 12,87 V
Ia1 29,0E-3 A
Ia2 27,0E-3 A
Rc 203 Ohm
Irc 0,063 A
Is 3,7E-3 A
Pv1 11,2 W 94%
Pv2 10,5 W 88%
Vspeaker 9,5 Vrms
Pout = U^2/R = 9,5V^2/8ohms = 11,2W (still low)
Unfortunately output wattage is exactly the same (around 11W of undistorted signal), I cannot tell anything about tone, as I have only bare speaker (rather poor one) without cabinet in house right now.
I started to look at PI this is my first measurement (sorry for photo but I lost somewhere my pendrive;))
I'm measuring by applying 50mV RMS to amp input through clean channel without any preamp distrotion and I'm rising master to the point of first visible clipping on power stage.
1. Yellow is on speaker
2. Blue and purple are inverter output (just behind coupling caps)

[JMPGuitars]

You're calculating the amp output wrong. If 1 tube is outputting 11W and the other 10W, you're outputting 21W. It's still slightly high, not low.

[MarcinD]

but I'm measuring on speaker (output of transformer) so rms is measured out of summarized (on OT) current from both tubes or I'm not understanding something very badly

[JMPGuitars]

but I'm measuring on speaker (output of transformer) so rms is measured out of summarized (on OT) current from both tubes or I'm not understanding something very badly

Ignore what you're trying to do with the speaker. Guitar tube amp power ratings are based on the power tube dissipation. You're still biased hot. Not terribly hot, but still hotter than what we typically target.

Measuring power from the speaker is more complicated than you think, and not relevant here when we're actually talking about the tube output power, not what the speaker load does to it. You can read up further on it, but it isn't beneficial to the conversation here about whether or not your amp is working correctly.

Thanks,
Josh

[MarcinD]

Sure I was curious, I will lower the bias tomorrow and finally I will have opportunity to check it on cabinet not on dummy load and scope I will let You know? Thanks!:)

[MarcinD]

Alright I adjusted a little bias to be more close to 85%, Today evening I will have opportunity to test it so I will for sure let You know!

Today measurements:
B+ 371 V
Va1 356,7 V
Va2 356,5 V
Vs 346 V
Vs 348 V
Vc 12 V
Ia1 27,0E-3 A
Ia2 24,8E-3 A
Rc 203 Ohm
Irc 0,0591133 A
Is 3,7E-3 A
Pv1 10,56734734 W 88%
Pv2 9,80331601 W 82%

[Bieworm]

Alright I adjusted a little bias to be more close to 85%, Today evening I will have opportunity to test it so I will for sure let You know!

Today measurements:
B+ 371 V
Va1 356,7 V
Va2 356,5 V
Vs 346 V
Vs 348 V
Vc 12 V
Ia1 27,0E-3 A
Ia2 24,8E-3 A
Rc 203 Ohm
Irc 0,0591133 A
Is 3,7E-3 A
Pv1 10,56734734 W 88%
Pv2 9,80331601 W 82%

Looking good

[MarcinD]

Hi,
All right so I did some tests and looks like it sound sweet, clean channel near breakup of power stage tone is awesome so thank You! By the way, place when I'm playing have some very old wireing in building and it looks like the voltage is a little lower than in my house so I this is what I measured:
B+ 352 V
Va1 341 V
Va2 341 V
Vs 335 V
Vs 337 V
Vc 11,25 V
Ia1 0,0252 A
Ia2 0,0228 A
Rc203 203 Ohm
Irc 0,05541871921 A
Is 0,007418719212 A
Pv1 10,75602266 W 89,63%
Pv2 9,96462266 W 83,04%

Now I still have few problems/questions I would like to ask You about.
1. Stability issues - on high volume, high gain especially with some boost (like tube screemer). It looks like this is problem in preamp especially that I recently changed front panel (on first PCB I did it a little bit too fast and potentiometer was on wrong side and I did some change in channel switching:D) those ares basically the biggest changes. In attachement You may find schematic comparison of new and old panel (on new tone stack LDR is not installed and R33 is 33k.
What I tried
1. I tried to put 100pF capacitor on first stage (between anode and cathode)
2. I untwisted grid from anode and cathode wires of each step.
3. I replace 68k V1 grid stopper with 33k and put it directly on tube socket
4. I added 10k grid stoppers resistor on V2 and V3
Nothing really help, the only way to play with higher gain is on relatively low volume.
2. How to properly connect input jacks (hi and low) - especially shields of those shielded input cables. I guess using shild as gnd connection is not the best option. Maybe I should use two shilded wires for GND and signal and both shields I should tied to GND on PCB? What do you think?
3. I noticed that stage 2 and 3 are particulary sensitive if I will put my finger close to elements for stage 2 and 3 there is very well audible hum in speaker, what is interesting there is no such effect near stage 1 where theoreticaly gain is the highest.
4. I think I have some ground loop problem in fx loop and in general it looks like recovery circuit in fx loop have to low gain (even with fx volume on 10 guitar is quiter than without fx loop)
5. In general there on higher volume there is quite a lot of noise (not hum), I assume this is mostly resistors thermal noise. Could You give a hint how to start investigation or where to look?:)

All Best guys and thank You very much for all of Your support!