JTM45 rectifier question

[SPeter]

Done, man that was loud. I always run it in a 1x12 cabinet. 2x12 sounded much louder.
At 16 ohm amp settings the ghost notes still could be heard. The speakers become really farty though.
I switched to the 8 ohm settings and the fart was gone. Everythin cleaned up but the ghost notes were still present as expected maybe even more pronounced if anything.
It only happens when i play one note. As soon as I double it up picking any second note I can't hear it anymore.
There is no way i'm the first one this happened to. I quadruplechecked my wiring and i'm convinced it is not me it must be a faulty component.
The presence pot makes a lot of noise even when both volumes are all the way down...

About the filtering: i gathered somewhere that tube rectifiers can't really handle much more than 40uf. I have an additional 32 which i could replace the 16 with and increase the PI filtering to 64. Right now it is 16 and i did try bridging it for 32 in the past, bit i could not hear any improvements.

[martin manning]

Or some feedback path that resonates at that particular frequency. What note is it BTW?

I don't know how the presence could generate noise with the PPIMV down all the way. That seems odd, and may be revealing a feedback loop.

You can certainly add to the screen filter with no worries, but having doubled that up with no effect suggests that that won't do it. You can add to the reservoir safely if you don't use the standby switch (set it to play before you power up so the reservoir will charge slowly as the rectifier warms up). That would be more about proving that the cause isn't lack of decoupling in the power supply.

[SPeter]

PPIMV was on max the two loudness was all the way down, sorry for not being clear. If I turned down the PPIMV the speaker is dead silent.

I just replaced he PI Cap... Went from 16uF to 64uF (32/32 jumped) no effect on the problem.

It is most pronounced on G string around 11-15 frets, also on B but less, and I can hear it on D a little bit as well in the same area. It also depends how loud the amp is set but there were times when I could hear it all the way down to 5th fret G and B.

[martin manning]

PPIMV was on max the two loudness was all the way down, sorry for not being clear. If I turned down the PPIMV the speaker is dead silent.

Ok, that makes sense now.

I'm running out of ideas, but I can recall a difficult case like this where the solution turned out to be rerouting the power tube grid wires. In your case I would translate that into redoing the PPIMV leads with shielding from the board to the PPIMV pots, and from the pots to the tube sockets, with the shield carrying the bias voltage and connected at the board-end only.

[SPeter]

At one point i completely removed the ppimv and it had no effect on the ghost notes.
I'm going to bite the bullet and put in the ground rail Tony recommended.

[martin manning]

Sometimes a poor solder joint (and that could be somewhere in the preamp), or a poor ground connection can result in noise like you are describing, but I don't think a sub-optimal ground scheme would do it, as long as the connections are good. Above you said you reflowed a number of joints, but have you made a close inspection of all solder joints in the preamp?

[SPeter]

Sometimes a poor solder joint (and that could be somewhere in the preamp), or a poor ground connection can result in noise like you are describing, but I don't think a sub-optimal ground scheme would do it, as long as the connections are good. Above you said you reflowed a number of joints, but have you made a close inspection of all solder joints in the preamp?

I looked at every joint several times with a magnifying glass. I can't see any obvious problem. Now these turrets are a bitch if they are closed from the bottom, the air gets trapped inside when you solder from the top and it expands with heat. These can create issues under the surface of the solder I can't see.

I did the rail thingy last night:

20190414_221726.jpg

Something is wrong because now there is a crackling distortion every time I strum or pick a note. It is there even at every loudness level. The ghost notes are faded but maybe only because this new bad distortion is just louder.

[SPeter]

The distortion turned out to be the speaker rattling in the cabinet. I moved things around for the 16 - 8 ohm test and forgot to tighten the speaker down lol finally an easy fix.

Today I moved the ground of the speaker jack and the two output tubes to the buss. No change in ghosting,. The amp is dead quiet up till 7 and at 7 i think what i hear is the nos resistor hiss..

I read the metro forum for two hours. Lots of complaints about ghost notes. But as someone mentioned earlier in this thread the 100w marshall amps are prone to the problem, the jtm45 not so much. I still found a few people asking for help on the jtm45.

Based on info i found on the metro forum I replaced the 82k PI balance resistor to a 68k, it was a long shot and did not do anything.

Now I've seen people adding a cap between the 32uf main cap and one of the V5 pin. I could not figure out which pin and what cap, but that is something i could easily try.
EDIT: on this picture it looks like a cap between pin 3 (output transformer) and and the positive side of the 32uF filter cap:

Other metro jtm45 with 47nf cap.jpg

Any idea what this might be. I can't help but think that I need to increase filtering to get rid of the damn ghost notes...

[SPeter]

So how about this for a test. I have a unused 16uF side of the PI filter cap. Can I connect it to the 32/32 positive side where the Primary CT goes. This would create a 32+16uF filtering. The rectifier tube should be able to handle up to 60uF...
The Marshall JTM45 Reissue has a 50uF cap in that position... so it should be safe right?
As a longer shot I could replace the whole 32/32 to a 50/50 like in the reissue to see if that helped. The choke in between them decouples the second one from the rectifier correct?

[martin manning]

Should be ok as long as the voltage rating on that cap is adequate.

[SPeter]

Added the 16uF, did exactly nothing... this is frustrating.

I read almost every post on Metro forum about ghost notes LOL not very many people complain about their jtm45's... There are a lots of people experiencing ghost notes on the 100 watters though. The most often cited solution is add 100uF power filtering. I can't do that due to the limits of tube rectification. So as a last resort I might change to SS rectification and add the filtering. But that won't be a JTM45 anymore.

I'm thinking the high B+ voltage might contribute to the ghost note problem. The Metro amp kit instruction indicating a B+ of 400V mine is 465V.
For 115+shipping I can get this Power transformer with flexible B+ voltage: http://www.classictone.net/40-18054.pdf. This would provide 424V or 379V with the 5AR4 / GZ34

When looking at the info on the link I spotted that the same config gives different B+ voltages depending on rectification. The 5U4GB rectifier tube in the same configuration drops B+ by 29 volts vs the GZ34/5AR4. That got me thinking,... what if I replaced my GZ34 to a 5U4GB. If the same rules applied (and that I'm not sure of) my 465V B+ would drop 436V... much closer to the Metro spec. And very close to the flexible B+ PT above.
This is a relatively cheap easily reversible mod and if it improved the ghost notes I can justify the spending on a new power transformer...

The question is what else the change to 5U4GB would affect? Would I be still able to bias the tubes in the 40mV range for example?
Also how does a lower B+ affect tone? Would it increase the distortion at a given loudness level or decrease it? Does it improve clean headroom or there will be less of it?

[martin manning]

5U4GB needs another amp of heater current (3A vs. 2 for the 5AR4), which your PT doesn't have.

You can run higher reservoir capacitance if you don't use the standby switch, or you move the standby switch downstream, so that the reservoir will charge slowly as the rectifier heats up. I made that suggestion earlier. The issue is hot switching, so if you move the switch, you would only need to avoid that.

You could still try cleaning up the PPIMV wiring and shielding.

[SPeter]

Thanks for the info on the 5U4GB... it's a nogo then. Edit: I see that the 40-18054 do have 3A, damn I'm so mad that I did not investigate which power transformer to buy a little more. I just went with Triode Electronics set. For a little more I could have lower or higher B+ and more freedom on the rectifier choice too. Well... next time.

I went as high as 32+16 and used your suggestion of hot switching, just did not go higher yet. I pretty much stopped using the standby as I read everywhere that it is a bad idea with tube rectification. 60uF is safe per the GZ34 datasheet, so if I add another 32uF and use hot switching it won't stress the rectifier tube much. I can add another 32 to raise it to a total 96, which is now outside of the safe operating conditions of the tube and hot switching will be a must to avoid damage... makes me nervous For testing this I can temporarily connect the reservoir positive lug of the 32/32 to another 32/32 positive lug for additional 32uF, and I can jumper the two 32 lugs for an additional 64 (96uf total). Then I would just ground the added cap negative side to my star ground.

I don't believe the PPIMV has anything to do with this as I removed it at one point completely restoring the original non-PPIMV layout with no effect on the ghost notes. I'm running out of shielded cables so this is low on my list as I need to order more of that cable.

I found this in on the Metro forums discussing ghost notes, unfortunately the links are all messed up there so I could not trace it back to the original conversation. I have never seen this on any layouts, I wonder what this do, it is a 47pf cap between the reservoir cap/output CT + and the V5 Pin 3 where the output Primary lead goes:

47nf cap.jpg

[martin manning]

Hot switching, where the rectifier cathode is already hot when it is connected to the reservoir, is not a good thing. In the JTM45 schematic the rectifier is always hot switched because the standby is between the rectifier and the reservoir. The rectifier becomes hot when the main power is switched on, and then it sees a big current surge when the standby is switched to play. The reservoir therefore has to be kept small. If the standby is placed after the reservoir, the reservoir charges slowly as the rectifier warms, and it can be larger.

That cap is 47nF (not pF), and it looks like it’s across one side of the OT primary.

[xtian]

Or, you can remove the Standby switch entirely.