Marshall TSL601 conversion into JCM 800 (of sorts)

[Willabe]

One interesting thing that happened was, after all the rewriting, I thought it prudent to fire it up on my light bulb limiter, just to be safe. Fired up fine. I then did a few things, and forgot it was on the limiter. Fired it up and started to do stability tests on the scope. It was oscillating in a minor (comparatively) way, before it started to flash the light globe. The bulb was blink to full brightness, then almost off, around twice a second, and I could hear a corresponding ‘click’ from the OT. Thoughts?

Seems odd.   

......and experiment with adding a shield on top of the OT, and one between the speaker jacks and preamp sockets. Will aluminium work ok, or does it need to be a ferrous metal?

If aluminum is all you have handy, you can use it. Steel is better.

[Willabe]

Is there a steel shield/bell end cover on the back side of that OT?

Was there ever a steel shield/bell end cover on the inside side of the OT?

I'm trying to ask if that OT ever had shield covers on both sides of it?

[Willabe]

I also did some rerouting of lead dress, and ran shielded sire from input to V2a, and from the gain knob to V2b. I will add a run from V2a coupling cap to the gain knob today. I put 68k grid stoppers on both V2a and b, right at the socket pins. It is much more stable than it was.

When you tested this, was it with or without the plate shunt caps?

[Fiat_cc]

Is there a steel shield/bell end cover on the back side of that OT?

Was there ever a steel shield/bell end cover on the inside side of the OT?

I'm trying to ask if that OT ever had shield covers on both sides of it?

Not to my knowledge. As far as o know the amp was stock when given to me, and there were no end covers on it, and no didn’t there ever were. Interestingly, the circuit that was originally in there was a PCB circuit, and incorporated shunt caps on almost every stage, plus the layout did have the preamp a little further away from the IT

[Fiat_cc]

I also did some rerouting of lead dress, and ran shielded sire from input to V2a, and from the gain knob to V2b. I will add a run from V2a coupling cap to the gain knob today. I put 68k grid stoppers on both V2a and b, right at the socket pins. It is much more stable than it was.

When you tested this, was it with or without the plate shunt caps?

I tested both ways. I have the caps on alligator clips so I can test in and out of circuit, and also see if there’s a difference shunting to ground or to B+. There wasn’t a huge difference with it without.

[Fiat_cc]

It’s getting there.
I ran shielded sure to the cathode follower grid, and played a little with lead dress. I can really see the sensitivity of that first grid wire around the OT now, where I couldn’t before.
More testing to do, and I think I’ll add shielded wire wherever it is suitable, but we’re getting there. Almost stable at idle... Haven’t injected signal yet, and haven’t listened to it, just looked at it idling on the scope.

[Fiat_cc]

I settled on 220pF snubbers on the tube socket pins from plate to cathode on V2a,b and V3a.

The first image (all the hash) is taken at the output jack, and is the amp at idle, all controls wide open, 510mV per division, so around 1.3V of noise at the speaker.  I calculated that to be around 0.02 watts or so into a 16ohm load. I can't really plug a speaker in, as it's getting late here, but to me, that doesn't seem excessive noise given the circumstances.  At least it is stable like that now.

Second picture is also taken at the speaker jack, and is a 200Hz 0.3V sine wave injected at the input jack, with all controls at full except the gain pot which is, wait for it, 1. SO it is making a full 55 watts of power before it starts clipping, which is good, but 0.3V is a warm humbucker.  It shouldn't be that nasty by 1 on the gain dial, surely?  That's not even starting to hit the clip in the cold clipper stage yet. So, I am thinking there's way too much gain somewhere, which is also contributing to my noise and oscillation issues.  Does this seem a fair assumption?  I've never built or played with a JCM 800 Marshall before, so maybe this is normal behaviour, but I wouldn't expect to start seeing power tube clipping until at least 4 or 5 on the gain control.

Still, I'm making progress towards a stable amp here.  Thanks for all the help and suggestions so far.

[Fiat_cc]

OK, so I've now tried using some aluminium on top of the OT, and also between the speaker jacks and the tube sockets, with seemingly no difference at all in either position.
I have also plugged it into a speaker, and with both gain and MV at full, and remaining controls at zero, it is pretty noisy, but it is also a JBL E120 speaker, so incredibly efficient.  Still, I'd like to reduce noise a bit further if I can, and see what I can do about the seemingly too much gain issue.

I swapped another known good tube into the cathode follower position, and no change there, so I don't think I've cooked that tube.

[Fiat_cc]

Latest voltage chart

V1 - disconnected from signal chain, so kind of irrelevant, but it will effect other voltages in the amp...
1 163.7
2 0.035
3 1.534
6 164.3
7 0
8 1.532

V2
1 261.5
2 0.007
3 2.309
6 313.7
7 0
8 3.218

V3
1 200.6
2 0
3 1.309
6 358.6
7 200.5
8 202.2

V4
1 260.8
2 31.22
3 46.9
6 268.3
7 29.25
8 46.9

V5
1 0
3 490
4 485
5 -43
8 0

V6
1 0
3 490
4 484
5 -43
8 0

B+
A 491
B 488
C 406
D 359
E 348.5

Wall voltage - 249.5VAC
Heaters - 6.65VAC

I'm of the opinion that my B+ at nodes C, D and E could still afford to come down another 30-40 volts.  I might try some different dropping resistor values in there and see what happens.

I also found that I'd put a 220K negative feedback resistor in, instead of the 150K I'd planned (the 220K was also cracked, but measured correctly on the meter out of the circuit, so I dunno).  I'm even wondering whether to try 100K there.  The original used 100K on the 4ohm tap.  I don't have a 4 ohm tap, so I opted for 150K off the 8 ohm tap.  Thoughts here?  Maybe part of my gain issue is a lack of NFB?

Edit: Idle voltage in the NFB circuit is pretty much perfect based on an annotated schematic I found.

[Fiat_cc]

One last thing... I played around on the scope a bit with the hash at idle with all amp controls at full. What I can see now after zooming out a bit, is an underlying 100Hz wave, with a whole lot of noise on top. Still some work to do to reduce that hum.

[shooter]

B+
A 491
B 488
C 406
D 359
E 348.5

got the AC component of the DC? 

[Fiat_cc]

B+
A 491
B 488
C 406
D 359
E 348.5

got the AC component of the DC?

I did look briefly last night, but it was late and I neglected to write it down.  I'll take those measurements again tonight.  What I do remember is that at node A it bounced around considerably, but was peaking around 11V on my DMM.  I guess I could put each node on the scope and measure peak to peak voltage of the wave form that way. DMM measures RMS though, yeah?

Can diodes introduce excessive noise?  The diode bridge I used wasn't chosen, I just had it in my parts bin.  It's big and waaaay over rated for the purpose, and possibly, it is not actually well suited to the purpose.

[shooter]

RMS

yeah
take ur scope PP *.707 .3535, and you'll be close enough, compare with meter
for me it's the "riding a bike thing", 97% of the time when I measure PS DC volts, I grab the AC at the same time.  I do SE, so I like <4vac at the A tap, and <20mV at last tap, Push-Pull you can be sloppier 

[Fiat_cc]

Ok, so I’m waiting on some 10K 3W resistors for the PSU, but figured I’d do some experimenting anyway.
AC measurements at each node. The DMM was utterly hopeless. It measured 2.2V on every mode for about 3 seconds, then slowly drained to zero.
The scope though, showed 100Hz partial saw tooth wave at 3.535v rms on A, 190mV B, 45mV C <20mV D and <10mV E.
Seems quiet enough in and of itself.

I was thinking about what I learned last night regarding amount of gain through the preamp so thought I’d take the following measurements.
All are approximate peak to peak voltages. Input at the jack was 200Hz sine at 300mV.

Preamp gain pot on full all others at zero- 200Hz at 300mV
Voltages AC peak to peak

V2a 0.3v in, 10v out
V2b approx 8v in, 50 out
V3a 11 in, 120 out
V3b 120v in, 160v out (cathode)

Master at full preamp at 1 all others at zero
Voltages AC peak to peak
V2a in 0.3v, 10v out (before gain control pot)
V2b in 26mV, 210mV
V3a 210mV in, 4.8V out
V3b 5.8V in, 5.2V out (cathode)
V4a 3.2V in, 59V out
V4b 1.45V in, 67.32V out
V5 58V in, 440V out
V6 69.3 in, 445V out (showing signs of clipping)
Speaker out 64.5V (showing slight signs of distortion)

Some of those numbers I just don’t believe (power tube plates at 440v peak to peak swing...?!?), but unless I totally screwed my calculations, that’s what I was seeing on the scope.
The thing is, it’s driving to the onset of clipping the power stage at only 1 on the gain knob. You can see from the measurements that the cathode follower can put out significantly more power than it is when power tube clipping starts.
Something seems really wrong. Either I’m not measuring right, or not understanding something, or this thing has waaaaay more voltage gain in the preamp than it should.

[shooter]

Speaker out 64.5V

using ohm, Vsq/8ohm 64*64 = 4096 then /8ohm = 512Wpp, that's a big #

Does you scope have a "calibrated" test point, typically it's a 5v square wave.
what do you measure when you scope your 120 house AC?

[Fiat_cc]

It’s a 16 ohm load, not 8 ohm.
So 256W.
That isn’t watts rms though (I know it’s voltage not watts measure in rms, but that’s what we all call it)

Yes, it has a calibrated 1V square wave test point, and it is spot on. I will try scoping my wall voltage tomorrow. 240vac here in Australia. (253+ I’ve seen some days).

[shooter]

64.5 *.3535 = 22.8^2 = 520Vrms /16 = 32.5Wrms very believable

[PRR]

> numbers I just don’t believe (power tube plates at 440v peak to peak swing...?!?)

Believe it or else.

The idea of the Power tube is to take ALL of the costly 480V DC you give it, and swing it into big AC/Audio waves.

The plates will typically swing (at FULL output) from 480V, to 80V, through 480V up to 880V, and back to 480V.

You really should not be poking a 'scope at this without an attenuator (something beefier than the usual 10X probe).

[shooter]

You really should not be poking a 'scope at this without an attenuator

I have a 100X probe, but 99% of the time I gave up on plate AC and just scope the speaker, since that's really the tell-all point anyway

good point sir

[Fiat_cc]

> numbers I just don’t believe (power tube plates at 440v peak to peak swing...?!?)

Believe it or else.

The idea of the Power tube is to take ALL of the costly 480V DC you give it, and swing it into big AC/Audio waves.

The plates will typically swing (at FULL output) from 480V, to 80V, through 480V up to 880V, and back to 480V.

You really should not be poking a 'scope at this without an attenuator (something beefier than the usual 10X probe).

Well there you go! I’ve learned something there! Now I know the output tubes are behaving as expected, I’ve no further need to probe them. Noted for the future though.

Now I guess my question is, is it normal for a 2204 JCM800 to be entering power tube overdrive at 2 on the preamp knob with a 0.3v sine wave? That seems like a significant amount too much gain. Does the rest of the gain structure look roughly correct?

[shooter]

is it normal

no clue, but I don't believe Marshall is known for "fender clean" 
If it's your amp and you wanna tinker, try a 12AU7a in V1

[Fiat_cc]

is it normal

no clue, but I don't believe Marshall is known for "fender clean" 
If it's your amp and you wanna tinker, try a 12AU7a in V1

No, not after clean, but I just did an experiment, and a 4mV input signal can drive this way into power tube overdrive. That seems ridiculous! It seems like there’s some interstage attenuation that I’ve either missed, or wired incorrectly. I just can’t believe that it could be ‘that’ sensitive.

I’m going to finish building my high power dummy load because the 16 ohm load I have is only 6 x 100 ohm 10 watt ceramic resistors in parallel. 60W. It’s marginal at best, (or being horribly abused at worst) when driving the full output of this amp into it!

New load will be rated at 400W, and will have a big heatsink, and a nice aluminium enclosure.

[PRR]

> is it normal for a 2204 JCM800 to be entering power tube overdrive at 2 on the preamp knob with a 0.3v sine wave?

Sight unseen, I would expect a guitar amp, full-up, to overload its final with 20mV input. Varies: 50mV on older student amps, <1mV on some "high-gain" head-shredder amps, but 10 or 20 mV is typical.

That's 60X down from 300mV. Depending on pot taper, "2" may be 60X down.

4mV is "hotter" than many but does not seem silly.

Counting on one thumb, for your post #8 plan I get 8mV sensitivity normal, 0.25mV boosted. That's a whole fork of a lot!

Counting on other thumb, I see 4 or 5 gain-stages before the output. 2 or 3 is more usual for "sane" gain.

[Fiat_cc]

> is it normal for a 2204 JCM800 to be entering power tube overdrive at 2 on the preamp knob with a 0.3v sine wave?

Sight unseen, I would expect a guitar amp, full-up, to overload its final with 20mV input. Varies: 50mV on older student amps, <1mV on some "high-gain" head-shredder amps, but 10 or 20 mV is typical.

That's 60X down from 300mV. Depending on pot taper, "2" may be 60X down.

4mV is "hotter" than many but does not seem silly.

Counting on one thumb, for your post #8 plan I get 8mV sensitivity normal, 0.25mV boosted. That's a whole fork of a lot!

Counting on other thumb, I see 4 or 5 gain-stages before the output. 2 or 3 is more usual for "sane" gain.

Thanks for the response.
Bare in mind that V1 is currently remove from the circuit. Input jack is going straight through a 68k gridstopper into V2a.

Edit: I reread your post and see what you’re saying about input sensitivity. I’m feeling like more attenuation between V2b and cathode follower is warranted, and also perhaps more between V1 and V2.

Still, it seems I’m not as far off as I though I was. It might be time to plug in a guitar and test results.
Could I replace R13 with a 220k or 330k to tame gain further between V2b and the cathode follower? Perhaps also add a 1M resistor between C2 and VR1?
It seems to me that, currently, you couldn’t really play this amp with both MV and preamp knobs turned to full. It would destroy itself.

The reason for the extra gain stage is the owner wanting something similar to the Jose Arrendondo mod. Basically a solo boost/overdrive built into the amp. I didn’t want to just straight copy something though, because I learn more by messing it up and picking the brains of the forums to help me figure out my mistakes.

[Fiat_cc]

Just found that I’d put a 1M grid stopper on V3a. Replaced with 100K and the oscillation is back. Not as severe as before, but it’s there.

*sigh

Gonna try upping V2a and be from 68k to 100k.

[Fiat_cc]

With the amp plugged in to a dummy load, putting 100mV input signal in, and running the amp until just before power stage clipping, I can clearly hear the rest tone emanating from the OT, at a not insignificant volume. Is that normal? Higher frequencies are clearer, and under about 500Hz I can’t hear it any more, but from 1KHz upwards, it’s quite loud.

[PRR]

Yes, transformers "sing". Being too small acoustically to sing bass, they sing the higher tones better.

[Fiat_cc]

Cool.  I won't worry about singing transformers then.   

So I got the new dropping resistors in. I'm happy with my B+ voltages now.  No need to post another voltage chart, but it brought V2,3 and 4 down 30 - 40 volts depending on the triode. Things seem much less on the edge in terms of voltages.

As mentioned, I realised I'd accidentally put a 1M grid stopper on V3a instead of a 100K. With the 100K resistor in place, the oscillation returned, so I played around with values.  I wound up with 100K each on V2a and V2b, and 220K on V3a. The oscillation is gone again at idle at least, but now there are noise artifacts on the sine wave when I inject signal. Could these be indications of an oscillation that is present when signal is present, or is it more likely distortion or noise? I don't really want to go above 220K as I think I'll start to cut to heavily into the usable audible range of the guitar.  I do see that in some amps, like the Soldano SLO, they use values as high as 470K, but I would think that would be rolling off quite a lot of actual usable high frequency.
I'm still toying with the idea of cutting my losses with this chassis and OT, and getting a proper Hammond JCM800 50W replacement transformer, and a custom chassis made, and designing the layout from scratch to be more optimal, but I would be gutted if I did that, only to find I was still having issues.

See pics attached for sine wave interference.

[shooter]

there are noise artifacts

glad it's clean at idle
where are we looking in the amp?
In my old life that would be a "ringing" artifact, usually caused by some burst interference, here, I'm leaning to loose connection, tube on the edge, did you verify input at jack was clean?

[Fiat_cc]

there are noise artifacts

glad it's clean at idle
where are we looking in the amp?
In my old life that would be a "ringing" artifact, usually caused by some burst interference, here, I'm leaning to loose connection, tube on the edge, did you verify input at jack was clean?

Sorry, we are looking here at the output jack.
This artifact only appeared after messing with the grid resistors. Nothing else changed . I didn’t even unplug my signal generator in between. Could weak solder joint cause this at all? I could re-examine the grid connections I messed with.

[Fiat_cc]

The plot thickens!
I was showing my 8 year old what I’d been doing in the workshop, and showed him the ‘ringing’ artefact on the scope. He said ‘could it be that thing?’ Pointing at the function generator. I said no, and that I’d show him, clipped the lead on to the outputs of the generator, and what would you know...? The artefacts were still there. At that point I figured I’d plug the guitar in and show my boy the guitar wave form etc.. And it sounded good. Clean, but with very conservative control settings. I figured I’d crank it up a bit, and then the scope blew its fuse. I don’t know why...
Still, the amp was sounding great. MV on 5, and preamp on 3 was giving great rock crunch, but then I noticed that on big chords the tubes were lighting up. Not like the plasma I’ve seen in the 6L6s in my old mans blackface Bandmaster, and not like red plating, but a red glow coming from inside the plate structure when I struck heavy chords. I’m guessing it’s the screens. It may also be normal, but I’ve not observed that in any of the other amps I’ve worked on. Did I mention it sounded fantastic? However at that point I decided that, with no scope, and something obviously getting very hot, I’d call it a night, and check in and ask for advice here.

[shooter]

did you verify input at jack was clean?

He said

been burnt more than once by that 

You need to get a Pdiss calculation for your PA section

[Fiat_cc]

did you verify input at jack was clean?

I have scoped the signal generator before and it was clean. I’m suspecting scope or probe maybe. Will replace fuse today and investigate.

He said

been burnt more than once by that 

You need to get a Pdiss calculation for your PA section

The hotter of the 2 power tubes (they’re quite close, but not identical) is at 33mA, which is 16.2W or near enough, and about 65% plate dissipation. After some reading last night some people say that’s easily safe in this circuit, others say because the tubes will likely spend so much time in overdrive (being a rock amp) that closer to 60 is better as screen current can become a problem. It was the warmer tube I noticed the glowing screen in, but it could have been happening in the other tube too.

[shooter]

It was the warmer tube I noticed the glowing screen

got G2 (screen) R's in, might verify values

you can swap tubes n see if it moves.  65% should be fine, assuming your current measurement is correct

[Fiat_cc]

It was the warmer tube I noticed the glowing screen

got G2 (screen) R's in, might verify values

you can swap tubes n see if it moves.  65% should be fine, assuming your current measurement is correct

Yeah, I can swap tubes and check.

I set idle by measuring OT resistance from CT to each plate, then measure voltage drop across same. Multiply plate voltage by mA to find wattage.

[Fiat_cc]

Dammit! My scope is dead! Or at least in a coma. Surgery may be required. 

[Fiat_cc]

Ok, so my scope blew one of the 35 year old line filtering safety caps. Have to wait to get another, but at least it’s not a tough fix.

I swapped tubes and made some observations playing guitar through it in a darkened room. The brighter tube is the brighter tube. It followed the tube when I swapped them.
Watching them though, they get brighter when taken off standby. You can see the heaters, and a blue glow inside the plate structure. Hitting chords etc dims or brightens the visible glow out the bottom of the plates (it’s hard to tell, but on further investigation I think they dim). Plates themselves aren’t glowing. Screens are very close to their max voltage rating, but I dunno. Maybe this is just an effect in some EL34 tubes and I shouldn’t worry about it too much.

[Fiat_cc]

I rebuilt this amp in a custom designed chassis, with a Hammond OT, and a much better layout than reusing the original chassis would allow. It seems stable now. The glowing tube issue seemed to be a dodgy tube. There was a small section of screen or suppressor grid that was glowing bright at idle and dimmed wit playing. I just bought another set of tubes, which is frustrating, but part of the risk these days I guess.

I still want to play with lead dress etc a little, as the boost circuit adds noise, but I’m going to let the client play it for a while, while I build a second one the same for myself, and try some tweaks. Once I have mine fully quiet etc, I’ll go back to this one and tweak it also. (reason being us the client lives 700km away over a large stretch of water, and I leave in 3 days...)

[Fiat_cc]

A couple of pics of the finished product, or at least the product as it stands.  I need to look at filament wiring, or any other possible noise sources.  It's not intrusive when playing, but it is quite loud at idle with the boost engaged.  I know there's a whole lot of gain going on there, but I feel like I should be able to reduce it further.
So far, reduction of the noise floor, and adding a little more bass response are the only requests from the client.  I'll play with different options in terms of coupling caps and cathode bypass caps and see what gives the best compromise between more low end, without farting out.

[Fiat_cc]

And one with the client...