New Marshall Project (‘77 1987)

[Roe]

OT has two commons (orange wire) coming from the secondary side. Is this typical

this is a typical British 70s thing yes, used by marshall/drake and vox

[Raoul Duke]

Thanks as well Roe, understood.

After doing a bunch of research it seems that I may have a ‘90s JTM reproduction 784-128 here. I’ve found aftermarket reproduction 784-128 OTs that refer to a “buried 100v tap, just like the original” which seems to fit. I’m still going to test it and see what I get.

Martin, quick questions after reading your test and watching the Uncle Doug video on the same subject, then unsoldering the primaries and CT so that the OT is completely disconnected (but still mounted to the chassis):

1. Which orange neutral should I use? The solid has continuity with the “known as speaker” taps (yellow&green) - so that would be my pick. Am I correct?
2. Same for grey: solid has continuity with all above - so that’s the speaker tap?
3. Should I test with the like colors (oranges and greys) combined as it was hooked up or everything separately?
4. The outliers here are stranded orange and grey - those should be significantly different from the others, correct?
5. I plan to use my Variac with current limiter - what would you recommend for voltage fed in?
6. Is there anything I need to watch out for so I don’t fry the OT or me besides not touching exposed live wires?

Intuitively - I would think the solid wires would be the PA tap based on the gauge (larger) and insulation (heavier). Coming out of the bell, the two orange and green are on one side of the rear-facing bell and the two grey and yellow are on the other side of the same bell. Red, white, and brown come out of the forward-facing bell.
If it matters…

[martin manning]

1. Which orange neutral should I use? The solid has continuity with the “known as speaker” taps (yellow&green) - so that would be my pick. Am I correct?
2. Same for grey: solid has continuity with all above - so that’s the speaker tap?

I would test this winding (with four leads showing continuity) by itself to see if it could be common, 4, 8, 16 without the separate stranded winding leads paralleled. You can connect the Variac across the assumed 16 ohm lead and common and then measure it and all the others and then the primary. Then do the math.

3. Should I test with the like colors (oranges and greys) combined as it was hooked up or everything separately?

I would then test it as it was wired and see what that gives.

4. The outliers here are stranded orange and grey - those should be significantly different from the others, correct?
5. I plan to use my Variac with current limiter - what would you recommend for voltage fed in?
6. Is there anything I need to watch out for so I don’t fry the OT or me besides not touching exposed live wires?

I would put a few volts AC on the secondary. The voltage measured on the primary will be much higher.

Intuitively - I would think the solid wires would be the PA tap based on the gauge (larger) and insulation (heavier). Coming out of the bell, the two orange and green are on one side of the rear-facing bell and the two grey and yellow are on the other side of the same bell. Red, white, and brown come out of the forward-facing bell.
If it matters…

I would guess that the separate winding is for PA speakers.

[Raoul Duke]

Understood and thank you for the breakdown - it helps more than you can imagine.

I will come back with my findings and see where we are.

Thanks again Martin!

[Raoul Duke]

Test completed, here’s the results:

IMG_0307.jpeg

I used the Ei and Svetlana datasheets (4k effective load resistance) as well as Mullard (3.5k ELR) and estimated 350v on the plates (although there wasn’t much difference going up to 375v from what I saw). Looks like either way will work; but paralleled orange and paralleled grey seems more favorable to my inexperienced eye.

Going the other way and trying to figure out primary impedance by applying known voltage to the primaries and measuring what comes out the secondaries, then dividing the “in” by the “out”, squaring the “out”, then multiplying by the load/tap - then averaging the three - I got 2909.6 with the wound orange and grey out of circuit and 2899.3 with the like colors paralleled.

I realize either way it’s effectively a 3k primary; but is there a “best” way to deal with the orange and grey from this perspective? Seems like a small difference - but it’s a difference none the less.

Thoughts or opinions always welcomed.
Thanks,

[martin manning]

I would say a 3k5 primary, since impedance ratio times assumed impedance gives
896.7 x 4 = 3,587
446.9 x 8 = 3,575
222.6 x 16 = 3,562

With the separate winding paralleled across the 16
221.7 x 16 = 3,547

Did you measure the impedance ratio of the separate winding by itself?

[pdf64]

... Did you measure the impedance ratio of the separate winding by itself?

That would be interesting.

[Raoul Duke]

Oops, meant to put that in, it’s pretty weird.
The separate winding (stranded orange to stranded grey) gives an IR of 183.3; nearly identical to 16 ohm tap. However…

Solid orange to stranded grey IR = 12.18

Solid grey to stranded orange IR = 7.56

Not even sure what the load would be there. I’ve read some old PAs had a 64 ohm load, but that doesn’t work out here; which leads me to two questions:

1. What do I have here (and I would think these characteristics would make OT identification easier).
2. What’s my best bet for reliable operation - parallel the colors as it was?

[martin manning]

Assuming Zpri from 8R measurement (Avg), this is what I get, if I'm understanding correctly.

[Raoul Duke]

You transposed the grays and yellow and I’m still googling the Zpri part - so I can’t really comment there - but other than that I think you have it mapped out just as it sits here.

Really interesting and hopefully the Marshall experts might have some ideas about this OT. All things considered - I feel that it’s a “well matched to this circuit” replacement unless I’m missing something.

Thanks Martin!

[martin manning]

Indeed I did; fixed now. I don't know why the stranded leads at 19.5R would be connected across the 16R. I think I would leave those disconnected and shrink-wrapped.

edit: swapped both instances of gray and yellow in the diagram, so the separate winding was connected across the 16R

[Raoul Duke]

So the gray (19.5) and orange stranded leads that do not have continuity with the yellow and green - cap them off?

Not trying to beat a dead horse, just making sure I have it correct.
Thanks!

[martin manning]

Correct.you have solid evidence that the other winding has the correct impedance ratios on its own.

[Raoul Duke]

Gotcha and done. Making good progress now that my AES box has showed up.

Also pretty convinced that this OT is a 783-128 repro - likely out of a Bluesbreaker reissue. All the specs I can find seem to match everything we’ve discovered through the testing. Sounded pretty good to me - far as I could tell.

So I’m contemplating a Lar-Mar MV in this amp and understand where everything needs to go with one small exception: the bias lead from the 220k “junction” on the board. Can I tap either 220k trace since they seem to be connected or do I need to jumper those and tap from that? Been researching and one method has top soldering from the leg of the bias cap - which seems kind of lazy to me. I thought for sure I’d find something conclusive since this is a practical mod for these amps - but the implementations are all over the place.

What’s the best practice here?

Thanks!

[martin manning]

Here is a JTM45 version: https://ampgarage.com/forum/viewtopic.p ... 60#p296260
You have access to the orange and green grid leads from the top of the board so you could break those and run the outputs from the dual pot back to the board. This would leave the existing grid leads running under the board in place. Lifting the board would be quite a bit of work, so no surprise people want to avoid that. Personally, I'd be reluctant to modify this amp.

...the bias lead from the 220k “junction” on the board. Can I tap either 220k trace since they seem to be connected or do I need to jumper those and tap from that? Been researching and one method has top soldering from the leg of the bias cap - which seems kind of lazy to me.

The inner end of the 220k's (nearest the middle of the PCB) and the "output" of the bias supply are all one node, so there are several options.