redplating 1959 w larmar

[raiken]

Okay, so if it only sags to 434 under continuous-duty heavy clip, then it is likely staying up in the 450V range you measured earlier with non-continuous guitar signals during normal playing, so that, in conjunction with the duty-cycle modulation, is the cause for your redplating, and likely why many original Marshall 100 watters don't exhibit the problem, because they sag a lot more.

RA

[jbzoso2002]

My Metro 68 100w 12xxx redplates at 480v.
Yesterday I put it back to the lv taps 407v
and no redplate.

Jimmy

[martin manning]

No, it actually means 53W each, 106W total peak power...

I thought this through again later in the day, and realized that I had to have gone the wrong way with the factor of two. It still didn't make sense to me because I wasn't thinking that the output power would be 106, even though I am aware that output of 2x "nominal" is possible. Thanks for straightening me out!

MPM

[raiken]

No, it actually means 53W each, 106W total peak power...

I thought this through again later in the day, and realized that I had to have gone the wrong way with the factor of two. It still didn't make sense to me because I wasn't thinking that the output power would be 106, even though I am aware that output of 2x "nominal" is possible. Thanks for straightening me out!

MPM

The big key here is the sag - not only in determining the redplating, but also the power the amp can put out during a sustained sine wave input.

The loadline has to be done at the full power sagged voltage to find the true output.

The problem is that it doesn't always operate way down there. For light loads and smaller signals, it operates closer to the unsagged load line (or somewhere in between max and min). Since maximum power dissipation does not occur at full output in a class AB or class B amp, rather at somewhere in the middle between idle and full power, even though the full power sag might get you out of the redplate zone at full clip, it may still redplate for smaller average loads, which is what we are seeing here.

RA

[martin manning]

Yea, I'm with you; fascinating! Thanks again.

[Roe]

No, it actually means 53W each, 106W total peak power...

I thought this through again later in the day, and realized that I had to have gone the wrong way with the factor of two. It still didn't make sense to me because I wasn't thinking that the output power would be 106, even though I am aware that output of 2x "nominal" is possible. Thanks for straightening me out!

MPM

The big key here is the sag - not only in determining the redplating, but also the power the amp can put out during a sustained sine wave input.

The loadline has to be done at the full power sagged voltage to find the true output.

The problem is that it doesn't always operate way down there. For light loads and smaller signals, it operates closer to the unsagged load line (or somewhere in between max and min). Since maximum power dissipation does not occur at full output in a class AB or class B amp, rather at somewhere in the middle between idle and full power, even though the full power sag might get you out of the redplate zone at full clip, it may still redplate for smaller average loads, which is what we are seeing here.

RA

my amp only redplates when running flat out for a minute or so. I will experiment more with tail resistor, PI voltages, tubes and bias to see if I can get rid of the last redplating

[flemingmras]

Almost sounds to me like Metro should redesign their 1203-80 PT to sag a lot more like the originals did.

[Roe]

Almost sounds to me like Metro should redesign their 1203-80 PT to sag a lot more like the originals did.

Randall's post suggests that 400-420v would be safe. however, this can still lead to redplating when amp isn't running at full. perhaps a 460v at idle (which sags to 400-420) would be safer?

[raiken]

I'm still wondering how much original Marshall 100 watters vary in the sag department. Some of them redplate, and some don't.

As for the Metro iron, I suspect George, being the stickler for accuracy and quality that he is, copied an original transformer design fairly closely...

RA

[novosibir]

Almost sounds to me like Metro should redesign their 1203-80 PT to sag a lot more like the originals did.

As for the Metro iron, I suspect George, being the stickler for accuracy and quality that he is, copied an original transformer design fairly closely...

The prob is, that i.e. the M6 sheet metal from the 70's or late 60's isn't 'the same' as the M6 from today and there's no posibility, to get it manufactured once again in amounts of less than probably 5 tons.

The M6 from today technically/physically regarded is much better than the old stuff, so a PT wound like in the old days and filled with the same thickness of M6 will saturate much later - resulting in less sag!

So what to do?

- You can overwind the PT both the primary and the secondary accordingly...
... but there isn't enough space in the sheet metal's windows! No place for more windings!

- You can make the lamination stack thickness thinner, to bring the PT earlier into saturation...
... but another size isn't vintage correct!

- You can fit in the sheet metal 2-2 instead of 1-1 into the bobbin...
... but this doesn't lower the core's permeability much, not enough!

- You can fit in the sheet metal 'butt-stack' like a Trainwreck OT...
... but again this isn't vintage correct!

- And you can try to excavate some NOS sheet metal from the outlying corners in good old UK...
... but you won't find that much anymore - and if still, what price does one have to pay for it

So who has "the answer", what to do?

I think, that George has done a lot of efforts, to come as close as possible to the vintage specs - and this should be appreciated accordingly! Unfortunately, despite this it isn't absolutely avoidable, that the one or other amp might have some remaining issues, where the owner/builder then has to try, to get rid of it by some minor design changes like a swamp R to the PI input, maybe a larger tail resistor and similar tweaks.

Randall Aiken has pointed out here extensively and very helpful, which way one has to try/to go, to get his amp 'straight'! And we owe him esteem and gratefulness for this!

Just my 2 washers

Larry


BTW: Randall Aiken is one of the very few guys, in front of whom I'd raise my hat anytime! And w/o painfulness I can confess, that as I've found his sites years ago, that even I have learned not less through this sites - things, I haven't read in any amp book before!
Randall! Thanks for your great and very helpful sites and for sharing your knowledge & experience with us!

[Roe]

yes, Randall and George have been doing a good job indeed. I'm just gonna tweak this amp a little and find some strong tubes for it. it sounds killer even with new production preamp tubes and a a little break in. I never liked much sag anyway (I guess I've listened too much to ac/dc )

[Midnite Zone]

Hi everyone, I have a similar problem with a 100w Laney amp (similar to a plexi with the two input stages in series).

It begins to redplate when I play the amp cranked for a minute or so, and if I stop playing the redplating subsides. I tried dropping the bias, all the way down to -70v, and it still redplated! So after finding this post (perfect timing), I tried doubling the load, and now the amp runs perfectly.

Here's some measurements taken with the doubled speaker impedance (presumably now around 3.4 k), with the tubes biased for 15w at idle:

At idle:
plates: 456v
screens: 452v

At full-blast:
plates: 416v
screens: 358v

...So I'm not getting a lot of sag on the plates, but my screens are coming way down. My high voltage overall is lower than the 480v you guys are talking about, but I'm still redplating with the 1.7k load. My question is, what is the best fix? Should I just use the amp with doubled speaker impedance? Or should I lower the high voltage?
...........................................ken

[Roe]

I still have problem with redplating (even though I reduced pi plate resistors to 56k and 82k). different tubes all redplate. v4-5 redplate clearly but there are indictions of redplating on v6-7 as well. even 2 kt88s (v4 and v7) started redplating when I turned up the bias to 60+ mA. At one point I even got an el34 to redplate a little on 400v.
the easiest way to get redplating is to run only one pair of tubes at 480v. I guess this gives more redplating because there is less sag in the power supply than when running all four tubes.

also, I reduced PI input cap to .0047uf. it didn't sound less bassy than the .022 that used to be there. but the sozo sounds much darker than the S&H cap I tried

I found some sweet sounding preamp tubes btw (v1: phillips; v2 siemens & halske; v3 telefunken). amp really sounds good. just need to minimize redplating

[VelvetGeorge]

Still tons of great info being added to this thread, thanks guys!

Larry has absolutely nailed the issue with replicating old PT's. I experimented with lower grades of steel, different lam thickness etc. The results were cool, but often ran too hot. And theres's no way I could send a PT overseas for 220/240V 50Hz operation with anything less than M6 grade steel.

I have to make PT's that run efficiently. Not doing so would be disastrous, especially in kits. So I try to mimic old PT performance in other ways, like the low original filtering values and choke specs.

Question: I agree 100% with Randall's load line assessments. But that begs the question: why am I able to build amps with these same parts that never have red plating issues? I'm asking this genuinely.

george

[Roe]

I'm gonna experiment more. gonna try running one MV instead of two. or removing the MV altogether.
I'm still not convinced this amp is absolutely free of oscillation. I need to learn to use the oscilloscope