JTM45 rectifier question

[Roe]

THE 100w marshalls often ghost, but the 45s and 50s don't ghost that much if they are in good shape.

[martin manning]

Yes I do, should I try to replace the 470 ohmite ones to these 1k ceramic ones?

Increasing the screen resistors from 470Ω to 1k will have a similar effect as did adding the flying resistor, but the increase in effective resistance will be about half as much. I believe it's pretty common to increase screen resistors to 1k on EL34's to protect the screens, but KT66 are beam tubes and more robust in that respect. You could experiment a bit by installing them, and then try reducing the flying resistor by paralleling another 1k with it, and then shorting it.

[SPeter]

At he current state, I'm back to the metro layout so no flying resistor. I'll keep it that way for now.

I have a set of EL34 tubes , other than these 1K resistors over Pin 4 and 6 do I need anything else to be able to use them in my amp?

Also Diode question. I found that the TubeDepot JTM45 schematic has an R3000 diode on both V4 and V5 over Pin 3 and Ground. In the description they explain:
"This rectifier is commonly used as a protection diode in tube amps. It is installed with the cathode connected to the plate of the power tube and the anode connected to ground. There should be one diode per push-pull half. This diode keeps unwanted negative voltage spikes from damaging the power tubes or output transformer. This diode can be found in the Fender Hot Rod Series as well as many other amps."

This is the wiring layout:

r3000diode.JPG

I wonder if this was something I should consider adding...

[martin manning]

For EL34’s, you need to tie pins 1 and 8 together, which you may have already done. Check the idle current and adjust bias as needed, of course. I don’t like protection diodes on the plates. They’re like fuses more or less, but they can have soft failures that can be hard to find. Just avoid running without a load. The shorting speaker jack protects you if no cable is plugged into the amp. To protect against the case where a cable is plugged into the amp, but not connected to a load, I’d rather put a resistor across the output (250 ohm, 5W say).

[SPeter]

Pin 1 and 8 is connected through the 1 ohm resistor per the Metro Instructions then grounded to Star #5 lug. So I'm good there I guess.

[SPeter]

Yesterday I got so annoyed by the intermodulation distortion to the point that I decided I need to redo the grounding :

20190409_181751.jpg

Any pointers on grounding or links to information on the topic would be appreciated

[pompeiisneaks]

The best grounding scheme I've seen and follow is to split up the major sources of current flow and separate them accordingly.

these are:

1. main transformer grounds and first filter cap
2. power tube grounds, and all power stage filter caps
3. PI
4. preamp sections
5. inputs.

Basically if you keep each of these extremely isolated from one another by grounds being as physically separated from one another as possible, it creates a strong separation of the differing sections so they can't interact and create ground loops. Take a look at a dumble and the grounding scheme, it follows this pretty well. His amps, with the drive channel, needed to be as clean as possible, as a drive channel is overly sensitive to any added noise, no matter where it comes from.

~Phil

[SPeter]

Thanks, what I can't get my head wrapped around is the chassis connections. Are you saying that these groups should be connected to a ground lug each separately which is bolted on the chassis (is this what they refer as galactic ground scheme?), or these should be connected together with a buss wire of some sort then those wires connected to one star ground (star ground)?

I've seen recommendation of both and I lean toward the galactic just because it feels that is less work.

I've found this: http://el34world.com/charts/grounds.htm but I'm not sure how this would translate to the JTM45 layout

[Tony Bones]

Read what Merlin has to say about grounding. You can download a pdf from here: http://www.valvewizard.co.uk/Grounding.html

Pay close attention when he writes about ground currents. Once it becomes clear in your own mind that different stages share ground currents you'll realize that you can't really isolate grounds. All you can do is arrange things so that the different ground currents don't get in each other's way.

FWIW, I don't like using the chassis as a ground bus. I know that lots of awesome amps have been made that way, so whatever, but I don't do it. What I do is run a single large diameter (12AWG) solid copper wire from one end of the amp to the other. Circuit grounds are connected along the bus in a fairly natural order. Input jack and first stage are at one end, then the second stage. Volume and tone controls connect where's you expect them to be near whatever gain stages they are near in the signal chain. The far end of the bus connects to the first filter cap after the rectifier, but not to the chassis. The only place amp ground connects to the chassis is at the input jack. Output jacks are isolated and grounded near the PI.

Of course earth ground is connected to the chassis for safety. That's the ground wire from the power cord.

[SPeter]

...
What I do is run a single large diameter (12AWG) solid copper wire from one end of the amp to the other. Circuit grounds are connected along the bus in a fairly natural order. Input jack and first stage are at one end, then the second stage. Volume and tone controls connect where's you expect them to be near whatever gain stages they are near in the signal chain. The far end of the bus connects to the first filter cap after the rectifier, but not to the chassis. The only place amp ground connects to the chassis is at the input jack. Output jacks are isolated and grounded near the PI.
...

That is really an awesome idea Tony!

[Tony Bones]

I'm sure I picked it up from someone else. Read Merlin.

[SPeter]

Rebuild is done, reflowed almost every connection, even ripped off the power trans to put the mounting bracket on the inside. Replaced a bunch of resistors to Allen Bradley NOS. The only thing I did not redo is the Output transformer leads to the impedance switch. It sounds great BUT the stupid ghost notes are still present. It is so disheartening...

These are the random ideas I have:
- Try the EL34 tubes I have on hand (need to rebias, I already have the 1K resistors over the power tube sockets I'm not sure they would play nice with the Output Transformer)
- new set of KT66 tubes
- New rectifier tube
- New Output transformer
- New power transformer
- Replace the choke with a resistor (???)
- Pay a guy to try to find a fix

I can also try to take the preamp signal and feed it in the FX return of my other amp and see if the ghost notes are coming from the preamp. I'm just not sure how to do this safely.

This is how it looks now, I decided to run the leads from the top of the circuit board to help troubleshooting:

20190414_110652 (1).jpg

(high res: https://photos.app.goo.gl/b1QABEy36Szm1ZSo9)

[martin manning]

Looks much better!

Thoughts:

I wouldn't think replacing the transformers or choke will solve it, and replacing the power tubes is probably a long shot, too. OTOH, Metro is running at lower plate voltages, and they may be doing that for a reason- they want happy customers. Have you reached out to them for advice?

Have you tried miss-matching the speaker load (8-ohm speaker on 16 ohm tap) as I suggested above? That will get the load line up into a more normal location, and would be safe for KT66 or EL34. That will lower the transient screen current and plate voltages peaks.

Unless I'm just missing it, I can't see where the shield in the cable that runs from the board to the MV pot is connected. I would do it like the drawing I posted. Not promising anything, but it's worth a shot.

Using another amp to drive the power amp is a good idea, and would indeed help isolate the problem. You could do that by taking the wire off the treble pot wiper and using that as an input.

Some people have said that increasing the filtering will eliminate the ghost notes. You could try that, tacking or clipping another cap across the existing reservoir or screen filters may have some effect.

[SPeter]

So, i tested the el34's in the amp. Please check my math. Max plate dissipation of the el34 is 25w. My B+ is 465V. 25x0.7/465=37mV. The problem is the pot maxes out around 24mV I plaued it anywys. I can hear the same thing.

I did not try your suggestion yet. I have two 16omh speakers if i wire those paralel then they will represent an 8ohm load, correct? I did that but switched the amp to 8 ohm when i tested it... damn should have tried the 16 ohm setting. I'm gonna do this next.

[martin manning]

So, i tested the el34's in the amp. Please check my math. Max plate dissipation of the el34 is 25w. My B+ is 465V. 25x0.7/465=37mV. The problem is the pot maxes out around 24mV.

Correct. So you'll need to do some tweaking in the bias circuit, and it's probably easiest to increase the 220k at the end of the board. I'd try 240k or 270k.

I did not try your suggestion yet. I have two 16 ohm speakers if i wire those parallel then they will represent an 8ohm load, correct? I did that but switched the amp to 8 ohm when i tested it... damn should have tried the 16 ohm setting. I'm gonna do this next.

Correct.