Marshall Super Lead build, sort of

[Helmholtz]

Right. The treble cap. But it sees significant DC. I had considered that regarding the PI plate coupler (low DC differential leg to leg) and the "bright" cap on the channel II volume pot. But the treble cap is a pretty big tone shaper. So it may be significant.

So only use class 1 (NPO/COG) ceramic or Mica types here (like Marshall did).
Both also have low THD (comparable to film caps).
Only class 2 ceramic types exhibit significant THD.

Looking at the dog bone (dogbone) caps I did notice that some makers marked one end. I guess this was more significant in this style cap because there's a definitive "outer foil".?. As far as I can tell the older red Lemco's have no marking. So I wonder if orientation for those caps could have been random, but made a difference no one was paying attention to.

EDIT: re: tubular ceramic polarity "but made a difference no one was paying attention to". I observe this regarding noise, not tone

The electrodes of a ceramic cap are formed by metallization layers on opposite sides of the ceramic.
So a tubular cap has an inner and an outer electrode.

[Helmholtz]

The value is probably the important thing. But maybe the tolerances are a possible factor in differences between Marshalls. Say, if the 500pF weren't really considered significantly different, but they used looser tolerance ceramics (compared to the 1 or 2% mica or 5% class 1), and maybe that made a slight difference.
.

If you really think a +/- 5% cap tolerance might make a tonal difference you should also consider resistor tolerance as corner frequencies depend on the RC product. The circuit resistance R is typically made up of several resistors and resistor tolerances may add up.
Add resistor drift.

You might use the Tone Stack Calculator to see the results.

[Helmholtz]

From the '69 Valvo shortform catalog:
(Notice difference between MKT and KT types)

img002.pdf

[dai h.]

Well your info about the vintage markings started me on my own venture to see just what's going on. As per your post above it seems nearly all the dog bone capacitors of the era were class 1 (black dot=NPO purple=N750). Almost none were marked with a poor temp coefficient.

Really? I'm pretty sure both class 1 and class 2 etc. were made. The dogbone Murata with purple mark does seem to be clearly class 1 (purple mark/N750 plus the relatively larger size), but (off the top of my head) there is a small red dogbone (seen in Marshalls) also which looks smaller (so most probably class 2). The very small values seem to be class 1 by default (not sure where the crossover takes place--100pF?). Also I can only think of the 47pF in the PI as one that was a NP0 in a Marshall.

suggesting higher voltage ratings.

Possibly (from the idea) higher rating = less change with applied voltage.

EDIT: This is purely speculation and some of it may sound silly, but...

I think it's probable that the tonal difference between the ceramics and micas that Marshall used was probably small enough to ignore. But enthusiasts, who are all about magic parts and their nebulous properties think that ceramics sound grainy because, well, they're made of a grainy material and feel grainy in the fingers. Likewise, micas sound smooth because they're physically smooth and shiny. For non stock replacement ceramics cheap ones are suggested because it's assumed that they are more like older technologies and higher voltage rating are suggested because of the more similar size to the originals. Cheap 1kV ceramic discs look more like the originals so they must be more like the originals.

Been a long time since I've experimented but the linear type caps seemed more clear (polystyrene as one example), so maybe if someone likes a mica, they will also prefer a class 1 ceramic and other 'cleaner' caps (and vice a versa)? (What I vaguely recall reading on forums long ago is) some recommending against the use of ceramics because they "sounded harsh". I do remember this being the case sometimes (trying out ceramics), but I suspect my amp was very far from the ideal state since I didn't have much of a clue about lead dress, grounding, layout, etc. back then so maybe there were oscillations and other problems going on (also maybe just being younger and having more sensitivity in the high frequency range). A lot of later PCB era Marshalls have small (so class 2) disc ceramics for the bright and treble cap spots (not all--I've spotted a red marking in some pics), so it's kind of hard for me to believe these all sounded harsh due to having class 2 ceramics. (Trying to think of guitarists who might've used later era Marshalls--Scorpions' Mattias Jabs (JMP 50W MV?)?

If you really think a +/- 5% cap tolerance might make a tonal difference you should also consider resistor tolerance as corner frequencies depend on the RC product. The circuit resistance R is typically made up of several resistors and resistor tolerances may add up.
Add resistor drift.

You might use the Tone Stack Calculator to see the results.

I don't know about the calculated aspect, but I was trying to say that I am grouping the 1, 2 &5% together, and the rest (10% and whatever else the other ones were) in the 'looser tolerance' group.

From the '69 Valvo shortform catalog:
(Notice difference between MKT and KT types)

img002.pdf

Thanks for that. I do think the mustards are KT/PET/polyester but (from your original quote plus the bit I described from the old Radiospares catalog--two sources of a "metallized" description for PET--or possibly just an error?) I was just wondering if there was some sort of technical explanation for that. Also, on www.vintage-radio.net/forum (which has been around for a while--there was a page describing how reliable some of the old caps were (which included some of the ones seen in old Marshalls)), and (I assume there are numerous technically literate and experienced people there) they do seem to know the mustards as polyester, so (seems to be additional evidence). FWIW I also have a couple of caps made by Matsushita (who was partly owned by Philips for a long time and made similar tubes, transistors, and caps, and possibly other parts) which look similar except they are a reddish brown on the outside (they are old and I couldn't find any data as I recall). I tried a used 22nF (it was used and was on a PCB which looked like part of a power supply) as a coupling cap (partly to see if it worked and partly to listen to it), and it *seemed* slightly fatter, but that wasn't really a perception I would put a huge amount of weight on. One thing I remember reading was film caps could have a bit more 2nd harmonic if the winding was loose, so wondered if there was something real perhaps from less than ideal manufacturing.

[Chuck H]

Really? I'm pretty sure both class 1 and class 2 etc. were made. The dogbone Murata with purple mark does seem to be clearly class 1 (purple mark/N750 plus the relatively larger size), but (off the top of my head) there is a small red dogbone (seen in Marshalls) also which looks smaller (so most probably class 2). The very small values seem to be class 1 by default (not sure where the crossover takes place--100pF?). Also I can only think of the 47pF in the PI as one that was a NP0 in a Marshall.

I admit I haven't seen many of these amps in person. The stock red dog bone caps I've seen in images do seem rather large to me and I haven't yet seen an image with a clear marker for thermal coefficient. It's very probable that some lower voltage, but still adequately rated dog bone caps have been used for restoration or private builds that are lower voltage or of a different thermal coefficient. In the stock amps even the RS ceramic caps in many images demonstrate the black cap indicating NPO classification. I'm not at all saying that Marshall may not have used lesser caps in a pinch. But it does seem that they were stringent on the criteria so I have to wonder why they would have had those lower rated caps on hand?. I do know that later amps did have lesser capacitors in the bright channel and treble cap circuits. I have one of those that I'm going to rebuild. But the vintage units do look to have low thermal coefficient caps in the bright channel and treble circuits as a rule. The only time I've seen small-ish red dog bone caps for these circuits were in restoration or personal build images. So not stock Marshall circuits. That doesn't mean it didn't happen though. And how could I know otherwise.?. I have a very small sample set to base my research on.

[dai h.]

In the stock amps even the RS ceramic caps in many images demonstrate the black cap indicating NPO classification.

Chuck, can you provide an example of this? This is confusing for me (FWIW, (sorry if you already know this) the flat rectangular ones are mica)). I just cannot think of many showing NP0 except possibly some 47pF in the PI or 100pF on the pot.

But it does seem that they were stringent on the criteria

(Possibly I am completely mis-reading things, but) in my mind it's the opposite, lol (i.e. "all over the place" inconsistency). If you can expound on your thinking here, I would appreciate it.

But the vintage units do look to have low thermal coefficient caps in the bright channel and treble circuits as a rule. The only time I've seen small-ish red dog bone caps for these circuits were in restoration or personal build images. So not stock Marshall circuits.

I don't think this is the case. The small red ones (which I presume to be 500pF or thereabouts) do seem to be stock in some amps. Some older amps also had a small disc ceramic for the treble cap. My red Lemco 100pF are 300V N750 2% tol. (I am 99.99% sure because I have the box) with no purple color dot (IIRC I have seen a red Lemco with color dot on the web somewhere but this is a foggy non-recent memory). (*Possibly* the red means Lemco manufactured--(I'm speculating but) maybe they did this to distinguish their parts from other makers.) So the red dogbones might not have a color dot indicating temp. coefficient, but in some amps that have red ones for multiple positions, like say a bass type circuit, there are ones where the mix R bypass (presuming 500pF) is (double value but) *smaller* than the 250pF treble cap (hinting the mix R bypass is not class 1), and I don't have the impression that they were all non-stock replacements (please check out the picture from a bass type circuit 100W). (Again, just a guess, but) if they went and bought 500pFs I think it might make more sense to just buy a higher voltage one (500V or whatever) so it's simpler and less risky (less chance of confusion if they chose a lower voltage part for the mix R bypass).




Also re: the EdVH amp, I found these in my files:


My memory is foggy but these may have been from ebay from a supposed "touring Marshall". Wires (presumably from the pot) are going somewhere near the PI (way too fuzzy to make things out clearly). Also interesting: PI output caps look conspicuously larger compared to the (presumed) 22nF elsewhere. Amp also shows the "fat cap".

[Chuck H]

The capacitor that would suffer most from DC bias would be the treble cap. I have seen the configuration in the first pic. That bypass cap for the mix circuit is doesn't have to be a high voltage part or have a stringent thermal coefficient. Many of the pics I see of Super Lead circuits do have the same mica cap for the mixer and treble. Some have the same, longer sized dog bone. Though more often when it's the longer dog bone on the treble circuit there's some other cap for the mix circuit. I've really only been focused on the treble cap since this is the one that needs a better thermal coefficient. I think Marshall must have recognized this too. I'll dig up a pic of the NPO discs and post it.

[Helmholtz]

Some nice pics here: https://www.marstran.com/Marshall%20...ial%20Page.htm

[Helmholtz]

I heated up some vintage Marshall caps to around 60°C (seems possible inside a closed tube amp chassis).
Results:
1) Rectangular RS 500p/220p: No C drop,
2) Red dogbone 500p: C drops by around 5%,
3) Lemco round brown 5nF ceramic disks : C drops by 30% to 40%.

[dai h.]

I heated up some vintage Marshall caps to around 60°C (not unlikely inside a closed tube amp chassis).
Results:
1) Rectangular RS 500p/220p: No C drop,
2) Red dogbone 500p: C drops by around 5%,
3) Lemco round brown 5nF ceramic disks : C drops by 30% to 40%.

Nice. Is the red 500pF the one that looks about the same size as 47pF showing up in the PI or a longer one (which IIRC there were)? I suppose the 5% drop strongly suggests a class 2 (i.e. if it was class 1 should've behaved like the rectangular mica)? Judging from the size (they look small), I suspected the 5nF discs was one with a relatively loose tolerance and more variance with voltage and temp.

Also, (re: the above chassis pic of the board with the red bones) notice the physical configuration for the B+ dropping resistors which are offset compared to the usual placing where they are in the middle (this offset configuration is seen in some amps). This moves one further away from the mustard film cap next to it in the tone stack. I've wondered if there was something there (change in frequency response, distortion or whatever) as the cap gets heated (in the usual configuration) and lowers(?) in value.

[Helmholtz]

Average length of the 500p red dogbones is 16.5mm.
Neither my '67 ST nor my '68 SL has any red dogbones.

The 5n discs vary in diameter. My smaller one measures 12.3mm dia/5.3nF, the larger one has 13.7mm dia/4.1nF.

An example can be seen here (treble bleed across vol. pot): https://www.marstran.com/1968%20100W.htm

The 47p cap between the PI plates only affects frequencies above 50kHz, so shouldn't be critical regarding tone.

[Chuck H]

Well I didn't find the black top ink caps (that I'm sure I saw in my previous and more extensive search) but I did find a bunch of purple tops (supposed to indicate N750 so still a class 1 marking though less good than zero like NPO). Four of them are in the image below. What I also saw is a few amps that had the smaller dog bone caps dai reported. I had not seen these before. These were JTM45's and early bass models so the caps may be 250pF (so physically smaller than the 500pF +/- spec for other models?) as are the caps in the images below. Class 2? Maybe. Still not a big deal as they should only exhibit a maximum 15% change in capacitance. Big kudos to you Helmholtz for taking the bull by the horns in the above test It seems Marshall was absolutely aware of where to use thermally stable caps to avoid the associated DC bias breakdown in capacitance because the higher shift on the caps in that test are in circuits that don't have any DC on them. This is good stuff and it's great to have some obviously inspired enthusiasm on the matter!

[dai h.]

Average length of the 500p red dogbones is 16.5mm.
Neither my '67 ST nor my '68 SL has any red dogbones.

The 5n discs vary in diameter. My smaller one measures 12.3mm dia/5.3nF, the larger one has 13.7mm dia/4.1nF.

An example can be seen here (treble bleed across vol. pot): https://www.marstran.com/1968%20100W.htm

The 47p cap between the PI plates only affects frequencies above 50kHz, so shouldn't be critical regarding tone.

Thanks for that.

Well I didn't find the black top ink caps (that I'm sure I saw in my previous and more extensive search) but I did find a bunch of purple tops (supposed to indicate N750 so still a class 1 marking though less good than zero like NPO). Four of them are in the image below. What I also saw is a few amps that had the smaller dog bone caps dai reported. I had not seen these before. These were JTM45's and early bass models so the caps may be 250pF (so physically smaller than the 500pF +/- spec for other models?) as are the caps in the images below. Class 2? Maybe. Still not a big deal as they should only exhibit a maximum 15% change in capacitance. Big kudos to you Helmholtz for taking the bull by the horns in the above test It seems Marshall was absolutely aware of where to use thermally stable caps to avoid the associated DC bias breakdown in capacitance because the higher shift on the caps in that test are in circuits that don't have any DC on them. This is good stuff and it's great to have some obviously inspired enthusiasm on the matter!

There's also the Murata 500pF dogone N750 (purple mark). The Murata 100pF N750 disc on the pot seems kinda rare.

re: Marshall awareness, possibly, but what I've read seems to indicate there wasn't a lot of awareness about the linearity of class 1 ceramics. For example someone who worked at Neve related that Rupert Neve considered ceramics "of the devil". Generally speaking many people seemed to think they all behave like class 2 (or other non-temp. compensating 'Hi-K') types. Also, does the non-ideal behavior mean they sound bad there (in a distorting guitar amp), or is it something complicated (some can't tell, some can and have a preference, or everybody prefers one)? Also, you've shown class 1s but the other ones do show up in Marshalls.


This is a schematic of some Jose Arredondo mods. EdVH (IIRC) said he played a Wurlitzer through (the one?) a Marshall. Would it make any sense to go into the power amp? (Just speculating about the mystery pot in the back.)

[Helmholtz]

The rectangular RS/Lemco mica caps can be found in all positions (except for 5nF).
So why don't you use mica caps throughout? Would take the guesswork out and still be "authentic".

[Chuck H]

The rectangular RS/Lemco mica caps can be found in all positions (except for 5nF).
So why don't you use mica caps throughout? Would take the guesswork out and still be "authentic".

I'm planning to use micas for the mixer bypass and treble circuits (and I already had them on hand too). One of the vendors for something else I purchased had the 47pF dog bone for the PI plate shunt for a good price so I bought one just for the "look". The tone snobs on the enthusiast forums seem to prefer ceramic for the "bright" cap. So that's what I'll use. No DC on that one anyway. All this ceramic thermal coefficient talk has just been for fun.