JTM45 rectifier question

[SPeter]

The 3H 250mA choke showed up the other day, I spent an hour going back and forth between the 20H/70mA and the new one. I could not hear any difference (I'm not saying there wasn't any but I could not hear it). As far as I could tell the new choke did not change neither the tone nor the ghost notes.

I could hear a slight difference in tone with the JJ KT66 tubes, I liked them better than my Gold Lions but they had no effect on the ghost notes either. I kept the JJ's in there for now.

So as it was expected choke and output tubes are not the root cause. I'm gonna order parts to build a line out next and also on the lookout for a cheaper attenuator. There is not much out there readily available.

[Tony Bones]

I have no direct experience with the ghosting problem or any solution. All I can offer is thoughts. But, if it's caused by insufficient filtering (presumably at the screens) then you could use both chokes to improve filtering.

If you feel like experimenting, put the 20H choke back in the original spot. Now we're going to add the new 3H, 250mA in the supply chain before everything, even before the plates. Besides the choke, you'll need an additional HV capacitor. I understand there was an unused 16uF section that you paralleled with the used section in a previous experiment. Disconnect them so that a 16uF section is available.

1. Find the wire that connects the standby switch to the first filter cap. Disconnect it from that cap and reconnect it to the free 16uF cap.

2. Connect the 3H, 250mA choke between where the wire is connected now and where it was connected. In other words, connect what was the first filter cap and the new 16uF cap using the 3H choke.

It shouldn't take too long to try.

[SPeter]

I have no direct experience with the ghosting problem or any solution. All I can offer is thoughts. But, if it's caused by insufficient filtering (presumably at the screens) then you could use both chokes to improve filtering.

If you feel like experimenting, put the 20H choke back in the original spot. Now we're going to add the new 3H, 250mA in the supply chain before everything, even before the plates. Besides the choke, you'll need an additional HV capacitor. I understand there was an unused 16uF section that you paralleled with the used section in a previous experiment. Disconnect them so that a 16uF section is available.

1. Find the wire that connects the standby switch to the first filter cap. Disconnect it from that cap and reconnect it to the free 16uF cap.

2. Connect the 3H, 250mA choke between where the wire is connected now and where it was connected. In other words, connect what was the first filter cap and the new 16uF cap using the 3H choke.

It shouldn't take too long to try.

Excellent idea! So it would look like this:

amp filterin mods.JPG

I also have a spare 32/32uF F&T cap which I can replace the 16/16 with next. So then the PI filtering would go up to 32uF (I tried that value before by jumping the two sides of the 16/16) and the other side would be as on the drawing but 32uF on both sides of the 3H choke.

[Tony Bones]

Excellent idea! So it would look like this:
amp filterin mods.JPG



I also have a spare 32/32uF F&T cap which I can replace the 16/16 with next. So then the PI filtering would go up to 32uF (I tried that value before by jumping the two sides of the 16/16) and the other side would be as on the drawing but 32uF on both sides of the 3H choke.

That's not exactly what I described, but that's the general idea and would accomplish a similar goal.

In most amps the plate supply comes directly off the first filter cap. Then there's a choke (or resistor) to the second filter cap that feeds the screens, and all points south. The reason that the plates don't get the benefit of choke filtering is that it takes a choke with more current capacity (bigger, more expensive.) That, and the fact that the plates are much less sensitive to PS ripple than the screens, especially with PP amps.

Your 3H choke is rated at 250mA, so it should be able to handle the current for the whole amp including the KT-66 plates. I described tacking it in front of the whole amp along with another cap (we need the rectifier to feed a cap first.) That configuration provides additional filtering for the power tube plates. The filtering for the rest of the amp would be improved as well. Putting the new choke after the first filter cap, in series with the existing choke with a cap to ground where they join together (as in your drawing) will provide additional filtering for the screens and preamp, but the filtering for the PA plates will remain basically unchanged. I would be interested to know if you hear/feel a difference.

EDIT: this is basically what I described in my first post. The cap values are not important to me, but they might be to you. It's where the OT center tap is connected that's different from what you drew.

[Roe]

it will work. I have tried a 3h choke like this (CLC filter) and didn't care for the sound personally, although the Fender 5e6/8 used a CLC filter

[Roe]

The 3H 250mA choke showed up the other day, I spent an hour going back and forth between the 20H/70mA and the new one. I could not hear any difference (I'm not saying there wasn't any but I could not hear it). As far as I could tell the new choke did not change neither the tone nor the ghost notes.

I could hear a slight difference in tone with the JJ KT66 tubes, I liked them better than my Gold Lions but they had no effect on the ghost notes either. I kept the JJ's in there for now.

So as it was expected choke and output tubes are not the root cause. I'm gonna order parts to build a line out next and also on the lookout for a cheaper attenuator. There is not much out there readily available.

have you checked the balance of the phase inverter? If you use the stock 27k negative feedback this is normally only an issue when you crank the amp. At least you have ruled out some potential causes, making progress in a Popperian fashion

[SPeter]

have you checked the balance of the phase inverter? If you use the stock 27k negative feedback this is normally only an issue when you crank the amp. At least you have ruled out some potential causes, making progress in a Popperian fashion

The NFB is the stock 27k.
I did change the 82K resistor to a 68K just because I had one on hand I could not really wrap my head around this PI balance thing. I typed up my values in the https://www.ampbooks.com/mobile/amplifi ... alculator/ calculator and got this:

PI balance.JPG

This looked balanced to me when it came to voltage gain...

With the 68k in place is is more out of balance:

PI balance2.JPG

I'm not sure I'm doing this right LOL, can I actually measure what is happening?

[Tony Bones]

That calculator can't be accurate as it doesn't include the bootstrapping of the negative feedback. The way NFB is applied to both the base of the tail and the unused side of the PI not only applies NFB as expected, but it also makes the tail resistor look bigger.

[SPeter]

Much to my surprise, the double choke method got rid of the ghost notes!!!!!! Completely gone.
I went all out and replaced the 16/16 cap to a 32/32.

So this raised the mains filtering to 64uF and the screen is now at 32uF. Plus the extra choke of course.
It is not all ok though as the change introduced a high pitch noise which changes with the volume (might be just lead dress issue but unlikely) and I'm worried about the Rectifier tube which is rated at 60uF vs 64 what I have. I looked awful bright LOL, honestly it might have looked like this always as I never really paid attention to that one.

Next up change back the 32/32 to 16/16 dropping back the mains to 48 and the screen to 16 plus the choke as Tony suggested.
I'm excited.

Edit: the high pitch noise came from a pedal with dying battery LOL

[SPeter]

Double choke with 16 cap still no ghost notes!!!

[Tony Bones]

Very cool! I'm glad that worked. Where do you have the OT center tap connected? Before the first choke or between the two? The sound and feel of the amp might be different.

As far as the rectifier is concerned, it's just the first cap that you need to worry about. Anything after the first choke doesn't count towards the 60uF max.

[SPeter]

I connected it between the two as you suggested.

Yes the amp feel changed but it is hard to describe. It is more modern sounding when it goes into overdrive maybe.. Dunno. I'm mostly deaf by listening to this amp in the last few weeks dimed testing the ghost notes Also there is not much clean headroom it seems. I always read that the JTM45 stays clean up to 4, mine starts to break up before that unless I dial down the guitar's volume around 5-6 and play soft (les paul).

Now onto the next decision, keep the double choke and find a place for it on the chassis or just screw it down somewhere in the cabinet. Or maybe I can replace the choke with a resistor and still have enough filtering left and give up some headroom due to the bigger voltage drop on the resistor vs the choke....

[martin manning]

So what are you running now? Rectifier-Cap1-3H-Cap2(plates)-20H-Cap3(screens)... What values are Cap1, 2, and 3?

Why is such extreme filtering required to make this amp behave when a 5F6-A, essentially the same circuit, works fine at 450V using Rectifier-40uF(plates)-4H-20uF(screens) ??

[SPeter]

So what are you running now? Rectifier-Cap1-3H-Cap2(plates)-20H-Cap3(screens)... What are Cap1, 2, and 3?

Why is such extreme filtering required to make this amp behave when a 5F6-A, essentially the same circuit, works fine at 450V using Rectifier-40uF(plates)-4H-20uF(screens) ??

Cap 1 is 16uF then 3H choke then Cap 2 is 32uF then 20H choke then 32uF just as Tony described it in his post see the marked up layout below. !!! THANK YOU THANK YOU THANK YOU Tony!!!

I don't know the answer to the 5F6-A question, even more interesting that ghost notes are not really common on the JTM45 yet there is quite a few builder who experienced the issue. I think there is variation between PT's, Rectifier tubes, Chokes. One guy claimed that when he did change to Celestion Cream back speakers he no longer could hear the ghost notes... go figure (I did no feel like dropping that kind of money to test his claim).
Also isn't the 5F6-A have a 12AY7 in V1? That tube has much less gain than the 12AX7 the JTM45 has in V1, this might be a contributor... Hmmm I should get one of those and see if it had any effect...

As I said I also tested that 16/16 cap I highlighted yellow on the layout replaced to a 32/32 cap. So that was Rectifier 32uF - 3H - 32uF - 20H - 32uF
Since then I played it more with the 16 front of the 20H choke and dimed on the neck pickup if I listen carefully and play real slow I can hear a hint of ghosting but nothing I would ever picked up on if it wasn't for the bad ghosting before. The 32 cap did not have even that small hint but I think the 16 is a good compromise.

final layout for now.JPG

[Roe]

the F&T 16+16uf caps only handle 450v but the 32+32uf caps handle 500v. I'd use 32+32uf on the mains personally