Tube Amps / Music Electronics
For current discussions, please visit Music Electronics Forum. The sunn still shines online!

ampage archive

Vintage threads from the first ten years

Search for:  Mode:  

couple of Marshall notes

4/19/2003 1:07 AM
Dai Hirokawa
couple of Marshall notes
Some notes:  
The amp: later period JCM800 2204 (pot-mounted PCB) converted to later 1986 bass-type cct. (tied cathodes, Bassman tone stack) with a gain boost via 250uF on V2a cathode. Stock Xfrmers. 27k fdbk. R off of 8 ohm tap (to 5k presence pot w/.68uF). Cathode biased (380 ohms/100uF). Filters (from rect. output) 50-20-16-8-8uF.  
The amp was configured basically the same as the 5F6-A before (one less filtering stage) and it sounded okay. After I changed it, it sounded bad. What I found was so far:  
a) 22nFs out of the PI sounds better than 100nF  
b) The .1uF in the PI (grounded grid side) affects gain. Might be interesting to have a variable cap here.  
c) Had nasty oscillation problems which I eventually traced (with Mike S.'s trick of small cap and connected wire heatshrunk on the end of a bamboo chopstick and a wire going to an insulated alligator clip which is connected to ground) to the V2 bright channel wire after the volume pot going to the grid (inducing squeal then using the probe here killed the oscillation). Used a shielded wire for this run, and consequently was able to now turn up TREBLE and PRESENCE all the way up. I do remember someone using the high gain Marshall mod over on the Hoffman amps site and having squeal problems with that which he cured with using a shielded wire on the same run, IIRC. Maybe the same with other amps(?).  
d) tried some rewiring, rerouting to try to lower noise. Somewhat successful. No major improvement, but a bit less background noise/buzz seems to be possible with better grounding  
Stuff that DIDN'T work:  
-twisting the PI output wires to the power tube grids  
-220pF going from plate to grid on the cathode follower stage  
-56k grid stoppers right on the tube pin at the input on V1 (instead of 33k)  
-moving the two 470k mix Rs close to the V2a grid pin  
-using thinner-insulated Teflon wire and moving plate, grid and cathode leads of respective triode stages close together.  
-trying to route the power tube cathode wires near the OT leads going to the EL34 plates  
-50pF to 100pF PI suppression cap  
-making sure unused PCB pads were connected to ground  
-using small 10ohm Rs connected on the plate pins of V1 (bright ch. side) and both sides' plates on the PI  
-trying to move wires around (hard to do with limited length  
The sound now is closer to what I expect a Marshall should sound like but not perfect. From the experience of trying .022uF in place of .1uF for the grid to ground (to fdbk.), it really impressed in my mind how true it is that a lot of gain comes from the PI stage. So I wondered how much the tone could be tailored by playing with the PI stage. The overall response seems to mush out some at higher volume settings. I suppose this is due to the inferior OT. Still needs to be smoother. Will have to keep working on it.  
4/21/2003 1:38 PM
Interesting findings. I've been wanting to experiment with that .1uF cap on the grounded PI grid but just haven't got around to it.  
I've been fine-tuning my homebrew 50-watt plexi and recently tried adding the 10k grid stoppers a la TUT3 (at least, that's where I saw it first). I only added one to the second gain stage, not the input stages. I also added a 220R grid stopper to the CF (replaced the jumper wire with a resistor). I did both at the same time and when I buttoned it up, it seemed to improve the slight "mushing out" when playing at really high volume settings. I know your amp is different but just thought I'd mention it.  
Also, it seems that the tone stack is much more "active" at high volume settings than before. Before the grid stoppers were added, the tone stack didn't do much at high volume settings.  
4/22/2003 1:39 PM
Dai Hirokawa
Spoke too soon...
Thanks Matt for the tips/ideas.  
Well I tried clipping off the 220pF, and the oscillation (at the upper extremes of the TREBLE and PRESENCE controls) came back. LOL. Pulled out the probe again, and found that shunting highs to ground from the green grid wire going to V3a, or pin 3 on V4 or V5 (the plates of EL34s), and putting the probe cap across pins 2 and 7 of V3 killed the oscillation. So went on to try a shielded wire for that V3a green, and a shielded wire for the purple fdbk. wire, then put the purple wire back but wrapped it around the OT wire bundle (just experimenting on a hunch that it would do something). Then got kind of confused about the sound (didn't seem to like it--or maybe it was just me) so decided to take a step back and put the purple wire back the way it was, the stock unshielded green wire back on V3a, and as an experiment put 5 33pF mica caps on V2 between pins 6 and 7 (where the 220pF was) so I could adjust downwards by clipping them off one by one. But it wasn't enough capacitance, and I've got it squealing again... LOL. (The cap there BTW, was an idea nicked from a picture of a JTM45 offset reissue. Not sure if the main idea is increasing interelectrode capacitance or shunting highs from the grid to ground, or both, but I thought perhaps that's a better spot to use a cap to cut highs because it's lower impedance. Interesing thing is, in another tag board pic of an offset reissue, there is no cap, plus the lead dress is not exactly the same.) AFA squeal suppression, there are a couple of ideas I see in a Lee Jackson amp schematic (looks basically the same as a Marshall) like using shielded wire for the V1a *plate* lead, the cathode wire for the stage immediately before the cathode follower/tone stack, and the tone stack cathode wire (the amp amp also looks to use the more orthodox shielded wire on the grids). Anyway, need to organize my squeal :D and try out some more ideas.  
4/24/2003 2:42 PM
Dai Hirokawa
more experimentation...
-shielded cable for bright ch. input to grid for V1 with shield connected to the cathode  
-same thing but shield connected to plate  
-moving negative lead of 8uF filter cap which was connected under the PCB from B+ side of V2b (cathode follower stage) to ground side of the cathode R on V2a to the immediate ground side of the 100k for the cathode follower  
-coupling the green wires on V3 together with a piece of heatshrink (I thought about this, and actually, it'll never be anywhere near the capacitance of the probe <470pF>, so nowhere near as effective in phase-canceling highs--but I suppose it won't hurt to try to kill the upper frequency garbage)  
-using a shielded wire for V2b pin 8 (cathode follower output--yellow wire)  
The weird stuff at the very upper ranges of the TREBLE and PRESENCE control rotations is still there. I guess the 1+1 stage being DC coupled makes it amplify a wide range (so it works well to cut some of the high freq. response there?)?  
Also, for kicks:  
-an 82nF (measures 86.3nF on my *cheap* cap meter so probably not exact but should be thereabouts) instead of 100nF for the V3 grounded grid side  
Didn't notice anything particularly different.  
-tried a Toshiba 12AX7 I bought recently (cheap--I don't particularly like them in Marshalls) and it seemed to make the high freq. problem a lot worse (didn't sound good either--bad tube?)  
Sound-wise, I think I'm on the right track, but can still be better.  
(Stuff I forgot to mention before: )  
-tube shields didn't work to kill the squeal  
-470pF bypass on br. ch. 470k mix R  
-1500pF bypass on br. ch. pot (1Mlin). Big seems good (i.e. the mid boost/high mid excitation a bigger cap helps to provide seems good with this setup). No particular reason for 1500pF, just checking out a recently acquired old mica cap.  
-There is an additional 50nF film paralleled with the first 50uF filter cap (directly off of the rectifiers)  
4/25/2003 6:05 PM

Dai: have you checked the value of resistors for plate and cathode..and those 470k? Also may have to pull the board and check for loose parts..i know its a pain,but the only way to make sure lol.. also make sure the input jacks are good and switching contacts clean.. or use a shorting plug. Sounds like you have one identical to one i just had took alot of time to finally cure the amp..but it can be done.. even some of the coupling caps can cause problems..those little green ones?? Keep us posted on what happens.  
4/26/2003 2:58 AM
Dai Hirokawa

Well I was up last night playing with it again, and that 220pF (actually tried a 320pF this time) kills the oscillation. Listening to the sound for a while and it seemed kind of fuzzy and weak, which I didn't like (worn out preamp tubes? Have to figure that one out...). Not sure where that was coming from or whether it was because my ears were tired, etc. So so far, those two measures seem to kill the oscillation at the upper reaches of the TREBLE and PRESENCE controls (the shielded wire on the one run and the cap). So I guess it's the additional gain from the big bypass cap I added across V2a cathode plus the DC coupling and consequent wide range amplification that causes the oscillation. I think I need to figure out which part is feeding back against which part. I'm interesting in knowing if I can kill it another way and whether if that will sound different than the cap (or maybe it sounds the same?--maybe the grid stopper for the cathode follower Matt mentioned above will work). (I also moved the 100k plate directly on the V2 socket like the original amps and not on the PCB.)  
As a side note, I measured the .1uF that I took off and replaced with the 82nF (measured higher at 86.3nF). Turns out the .1uF measured 90nF, so really, I wasn't changing the value much at all.  
As far as the caps, the 22nFs out of V1 are these mustard cap look-a-likes off some old PC board (no leakage). The tone stack 22nFs are the original white square metal polyesters (treble cap is a 320pF ceramic--experimenting with a slightly off value cap), the 22nF into the PI is a new polyester Orange Drop, the 22nFs out of the PI are cheap looking yellow ECI (metal polyester also, I think), and the 82nF is also an Orange Drop. Getting under the board of these MVs is actually easier than the earlier ones (though it is still kind of a pain, ha ha). I've added wires coming out of some of the PCB holes so I can top solder some of the parts without having to get under the board. Haven't completely figured out the best way to do this but it seems if I put a kink in the piece of solid core wire at the hole by bending 90 degrees then completely opposite (180), I can get a kink that keeps that piece of wire there and stops it from moving. Then above the board, the two pieces bend out in opposite directions somewhat, and I tack solder the part leads to these wires from the top.  
Okay, just listened to it again, and I figured out that I had somehow knocked the 3-position speaker simulation switch (on my Palmer spk. emulator) from "MELLOW" to "NORMAL" (which is brighter), which was causing the fuzziness (whew...). BUT, I'm getting the squealing again (LOL). Ever more squeal kill to be done...  
4/26/2003 3:00 PM

Do you have anything between your guitar and the amp besides the cord? On a previous home brew I was getting a squeal at high volume settings and/or treble and presence settings. Turns out it was my pedal board and it happened with another amp too. Just a thought.  

  Page 1 of 4 Next> Last Page>>