Where to find accurate parts lists, circuit layouts, schematics for plexi replicas (like what metroamp used to have)

[itsneverloudenough]

Here is a link to the archive for the Plexi Palace forum for those who are interested:

Vintage Amps Bulletin Board • View forum - Marshall

Came here to say this, I didn't notice if it was already mentioned in the thread but if OP or anyone else ever come across a dead link somewhere that you think is important, try copy-pasting it to archive.org or archive.is or other such websites. It may still not have all the pics or functionality but it could really give you something you otherwise couldn't get. Hope it helps.

 

[valvestorm]

Id start with Valvestorm. They list all the components for a JTM45/100 build. And sell most of them, unless out of stock.

ValveStorm JTM45/100 Build Checklist

 

[William Payne]

ValveStorm JTM45/100 Build Checklist

Thanks for that. Sorry to be a bother regarding that missing turret. I really appreciated your help.

 

[arthur.lowery]

In a perfect world, you are correct, in that the capacitors will share the voltage because the charges are the same:
Charge Q1 = C1.V1, and this equals Q2 = C2.V2, so if C1 = C2, V1 = V2.
But is the capacitors have different values (electrolytics are notoriously sloppy on tolerance), then V1 will be different to V2. Also, if there is any leakage, there will be some imbalance.
And it's a bit of a run-away situation in that once the voltage on one cap goes down, the other will go up and that will fail.
So, I'm more of a Murphy's Law person when designing electronics!

The balancing resistors will also discharge the caps after you turn off (if the valves are not working or the HT fuse is blown) - so may save your life. Not bad for a few cents.

 

[William Payne]

I had my suspicions but the more I look the more it seems I may be right. The JTM45 circuit layout besides the secondary board appears the same as the layouts I have seen on JTM45/100's. I have not checked values yet but my inexperienced eyes sure see it that way.

 

[playloud]

In a perfect world, you are correct, in that the capacitors will share the voltage because the charges are the same:
Charge Q1 = C1.V1, and this equals Q2 = C2.V2, so if C1 = C2, V1 = V2.
But is the capacitors have different values (electrolytics are notoriously sloppy on tolerance), then V1 will be different to V2. Also, if there is any leakage, there will be some imbalance.
And it's a bit of a run-away situation in that once the voltage on one cap goes down, the other will go up and that will fail.
So, I'm more of a Murphy's Law person when designing electronics!

The balancing resistors will also discharge the caps after you turn off (if the valves are not working or the HT fuse is blown) - so may save your life. Not bad for a few cents.

Even still, I think there's sufficient margin of error if using 450V caps. Even if one is -20% and one is +20% of nominal capacitance (worst case), then by my calculations you get V1 = 1.5*V2. Rectified voltage is ~560V, so solving the simultaneous equations, you get V1=420V 336V and still have some plenty of breathing space remaining.

Edit: fixed some lazy math on my part.

Here is a link to the archive for the Plexi Palace forum for those who are interested:

Vintage Amps Bulletin Board • View forum - Marshall

These Wayback Machine links are bitter sweet. You just know this thread must have been a goldmine, for instance, but of course the link is dead:



"So close, yet so far away"

 

[arthur.lowery]

I was worried about leakage currents as well. If one cap measures 10 Meg and the other 5 Meg, and both are 50 uF, LTspice tells me that the mid point will be 133 V (e.g. like a potential divider). Does anyone see this?

 

[playloud]

I was worried about leakage currents as well. If one cap measures 10 Meg and the other 5 Meg, and both are 50 uF, LTspice tells me that the mid point will be 133 V (e.g. like a potential divider). Does anyone see this?

Technically you are probably correct. I recall finding this a while back (see p.4-5), which is specifically relevant to the RIFA caps some of us use in 45/100s: https://www.iea.lth.se/mek/Mekatronikkursen 2006/Energy Flow 06/electrolytic_appguide.pdf

Otoh, there is now plenty of empirical evidence of people running series caps like these in Marshalls - vintage or modern - without any such issue arising. It just doesn't seem to be a common point of failure.

Bizarrely, Marshall used balancing resistors on the screens (but not the mains) caps for some time.

Also, perhaps I am missing something, but wouldn't leakage current be less of an issue if the caps are being discharged (e.g. from playing)?

 

[arthur.lowery]

Thanks the interesting discussion.

I have a gut feeling that you may be correct in the situation where the caps are being recharged and discharged (have a much large a.c. component to the current), because the (reactive) impedance due the capacitance will be less than the (resistive) impedance due to the leakage.

Let me simulate it............ this is with a 400 v p-p transformer (half wave rectified for now, but at 100 Hz to simulate FW sort of), and a 250 Hz load that kicks in later. Yes, the caps seem to share the voltage in this case. But they would not for the DC case, so maybe that's why the resistors are just used on the screens! Fascinating.



And even if I add in the screen grid circuit, the caps seem to balance (even with 10:1 difference in leakage current. (Screen cap voltages are blue - top and cyan -middle ) (Main cap are red - top and green - middle)
]

 

[arthur.lowery]

Actually if I run a longer simulation, and lower the timeconstants by reducing the capacitance and increasing the leakage, the leakage does have an effect on the voltage sharing of the capacitors. (Green and Cyan for the B+ and screen supplies, respectively).

 

[playloud]

Interesting stuff!

Those latest capacitance/leakage values are quite extreme though

 

[arthur.lowery]

That was only to get the simulation to run fast. (And show actual waveforms.). Here's one with realistic values (I measured a 1970's cap out of a Marshall at 5 meg/50 uF). Actually modern computers are pretty fast, so here's 1000 seconds.
Thanks for the like!
The light green is the midpoint of the HT caps, the line in the middle is the screen caps. So the AC does not count in the end (we get a potential divider).

 

[William Payne]

I haven’t run away from this project. Just sorting life stuff out. I’ll get back to it soon.

 

[Marcomel79]

Welcome to the forum!

We're lucky to have @playloud

Agreed!