Spite Build - Marshall JCM 800 2204

DIY Studio Stuff
User avatar
Greg_L
Posts: 20668
Joined: Wed Jan 04, 2017 1:07 pm
Location: Where the knuckle meets the poophole

Re: Spite Build - Marshall JCM 800 2204

Post by Greg_L »

Picking away at this thing....out of spite!

My buddies out here in the meat world are pretty interested in this build. They like the other stuff I've been doing, but this is a 50w JCM 800. This amp is one of the godfathers of modern rock and roll tones. And they know the spite story so they think it's funny I'm doing this. I guess it is funny a little, but I'm building this thing to be bulletproof. Everything I've picked is overspec'd and extreme top quality. The goal is for this to be the amp I grab for gigs. Solid, dependable, and sound great....and built by me. It will get abused.

So anyway, just been chipping away at it over the last few days, being meticulous, super careful. I've had to readjust my mindset. I'm back in the Marshall way of doing things. It's hard to be as neat and orderly with the way the Marshall circuit is designed and flows. Those damn giant multisection filter caps....I might, in future, break this thing down into individual filter caps like the Fenders have, but for now I'm gonna work with it the way it's supposed to be.

So not much to see, just working layers and doing the rear panel stuff first. That seems to have worked out for me in the past and I like doing it this way. Heater wiring, grounding locations, power tube socket wiring, power supply, much of that stuff is now done.

Image

Image

Image

Front panel stuff is next.
Rebel Yell
User avatar
rayc
Posts: 8488
Joined: Wed Jan 04, 2017 6:31 pm
Location: South of Bundaberg North of Brisbane

Re: Spite Build - Marshall JCM 800 2204

Post by rayc »

Bloody hell that's NEAT. I know it makes good engineering sense to do it that way but you do go the extra mile to do things right. It's a wonder to behold. Uncle Doug's latest is pretty cool for neatness too.
Cheers
rayc
User avatar
Greg_L
Posts: 20668
Joined: Wed Jan 04, 2017 1:07 pm
Location: Where the knuckle meets the poophole

Re: Spite Build - Marshall JCM 800 2204

Post by Greg_L »

rayc wrote: Mon May 23, 2022 3:36 am Bloody hell that's NEAT. I know it makes good engineering sense to do it that way but you do go the extra mile to do things right. It's a wonder to behold. Uncle Doug's latest is pretty cool for neatness too.
Uncle Doug's new vid is good. He goes into nice detail about the pentode/triode thing.

As far as neatness goes, it's got benefits beyond just looking good. First off, if you're neat you're careful and that usually means better work. I'm not soldering the wrong thing to the wrong spot or having errant blobs of solder touch things they shouldn't. But as neat as I try to be I still get fuckups. There's just so much stuff to do.

On this thing, neatness will hopefully reduce the inherent flaws in Marshall's design of this amp. If you look inside of a real vintage 2203/2204, there are wires flying everywhere. Marshall did not skimp on wire. They used as much as they could and had it going all over the place. They also had terrible grounding. And they also often pick up random foreign language radio stations. And many of them came from the factory with little quick-fixes installed to cure the oscillation that can happen with a high gain amp designed and wired poorly. So while they are some of the greatest rock and roll amps of all time, they have some major flaws that can be mostly fixed. I aim to do that to the best of my ability. Some of it can't be avoided, like some of the wire routing. You just have to make it work the best you can.

Most importantly, the grounding. The grounding can't be made "perfect" on this amp while also using those giant can caps, but it can be improved. That alone should pay huge dividends on the happy and healthy functioning of this amp. We'll see.
Rebel Yell
User avatar
Lt. Bob
Posts: 6577
Joined: Wed Jan 04, 2017 3:02 pm

Re: Spite Build - Marshall JCM 800 2204

Post by Lt. Bob »

your usual impressive work Mr. Drumble
User avatar
liv_rong
Posts: 1953
Joined: Thu Jan 19, 2017 7:32 pm

Re: Spite Build - Marshall JCM 800 2204

Post by liv_rong »

I have a question, and I'll first say I dont know anything haha, so its likely a stupid question. But regarding wire length, those black and orange wires, there is no "extra", does one have to worry about future repairs or mods? I'm just thinking with having an old house, before it was re-wired, there were a couple times where there was just barely enough to do what I had to do because so many people before me had like cut or shortened them for whatever reason. With solder you just heat it up and the it's good? Like what if you need to cut it and strip a little etc?

And I am not critiquing, looks like great work! I am asking out of curiosity.
User avatar
Greg_L
Posts: 20668
Joined: Wed Jan 04, 2017 1:07 pm
Location: Where the knuckle meets the poophole

Re: Spite Build - Marshall JCM 800 2204

Post by Greg_L »

liv_rong wrote: Mon May 23, 2022 3:07 pm I have a question, and I'll first say I dont know anything haha, so its likely a stupid question. But regarding wire length, those black and orange wires, there is no "extra", does one have to worry about future repairs or mods? I'm just thinking with having an old house, before it was re-wired, there were a couple times where there was just barely enough to do what I had to do because so many people before me had like cut or shortened them for whatever reason. With solder you just heat it up and the it's good? Like what if you need to cut it and strip a little etc?

And I am not critiquing, looks like great work! I am asking out of curiosity.
So the black/orange twisted wires....those are the filament wires. They are what make tubes glow. There is 6.3 volts AC cruising through those wires. They are coming straight out of the power transformer and going to the very first power tube socket. I used the heavy 18 gauge wire that came with the transformer for the first two power tube sockets as they have the highest amperage demands. Then I step it down to more manageable 20 then 22 gauge red/black for the rest of the string. It's all connected and bounces from tube socket to tube socket.

As far as mods/repairs, you're not wrong, there is no extra slack or wiggle room with those wires. And we don't want any. Those wires can be some of the worst offenders at causing unwanted noise and hum in an amp so we want them short, neat, tightly twisted, and tucked the fuck out of the way. The upside is there is very little you'd ever have to do with them during the lifespan of an amp. They have no connection with actually passing guitar signal so they are unaffected by mods nor do they typically get in the way of mods. And as far as repairs go the only time you'd ever need to mess with them is if you replace a tube socket and in that case you'd just put them back right back where they were on a new socket. De-solder the wires out of the damaged socket, install the new socket, and re-solder the wires right back in. So if done right, the filament wires are pretty much set-and-forget forever.

And in the unfortunate circumstance that you just have to repair or replace a section of filament wire, you'd just cut some out and splice some back in. As long as you use the right gauge of wire and do clean splices it's no problem. Electrically it's one long string of wire, but in reality it's little sections between each tube socket. You could relatively easily just replace a section between sockets if you had to.
Rebel Yell
User avatar
Greg_L
Posts: 20668
Joined: Wed Jan 04, 2017 1:07 pm
Location: Where the knuckle meets the poophole

Re: Spite Build - Marshall JCM 800 2204

Post by Greg_L »

Semi related...

The JMP-era headbox I ordered months ago finally arrived.
Image

IMO this is the best looking era of Marshall heads.

So the amp nestled in that headbox is a 1979 JMP 2204. Same amp I'm building now in this thread but my version is based on the JCM 800 era higher voltages. This 1979 JMP 2204 chassis has been in the wrong headbox ever since I got it and it's always bugged me. Now it's in the right kind of headbox.

The new build 2204 will go in the old headbox. I needed to order a headbox anyway for the new build, so I got a proper one for the old 2204 instead.
Rebel Yell
User avatar
liv_rong
Posts: 1953
Joined: Thu Jan 19, 2017 7:32 pm

Re: Spite Build - Marshall JCM 800 2204

Post by liv_rong »

Thanks for the explanation @Greg_L
User avatar
Greg_L
Posts: 20668
Joined: Wed Jan 04, 2017 1:07 pm
Location: Where the knuckle meets the poophole

Re: Spite Build - Marshall JCM 800 2204

Post by Greg_L »

Preemptive....this is gonna get stupid

Image

Let's talk some tech and how I'm gonna try to improve the hum, buzz, and noise of a classic Marshall circuit. This might work great, it might make no difference, but it's worth a try on these otherwise inherently noisy amps.

This is how Marshall usually grounds their amps - just pics from the net.
Image
They run a little plain wire soldered to the back of every pot and just stick things to it. Not good.

And the filter caps....
Image
See the little black wires? That's the ground wire and they just make them as short as possible straight to chassis no matter where it is.

None of this is good. Sure it has worked for decades but it always could have been way better. The problem with the typical Marshall setup is it creates ground loops and grounding "eddy currents". There are just too many random and different paths for circuit grounding to happen. Having circuit grounds on the bus wire is okay in theory, but it's connected at every pot and every pot is connected to chassis. That means multiple paths to ground. And the filter caps...they're being grounded out of sequence and just for convenience. All of this adds up to a grounding mess, and Marshall's higher gain amps are prone to hum and buzz and oscillation partly because the grounding is so poor. At the very least get that bus wire off the pots and have it flow to one chassis point. That's a huge improvement on it's own.

Now, for mine, I'm going one step further. I talked about filter caps and "nodes" in the Assman thread and we have the same thing with Marshalls, except they're a little different. Where Fender often uses a single capacitor for each power node, Marshall uses multi-section can caps. In Marshall's case it's usually two capacitors in one can. With Fenders you can ground each filter capacitor along with the circuit grounds it's working for. This is the ultimate primo best way to ground an amp. You can't do that with Marshalls because the filter caps are two-in-one and each cap only has one single ground tab. So after some research here's what I'm gonna try.

This is my schematic divided into filter cap sections.
Image

The red circle is the first filter cap. It's the first to see B+ voltage and it's the first stage of filtering for the whole amp. I'm using both halves of the filter cap as one so it's grounding is not a big deal. It can ground to chassis right where it is. The red squares are parts of the circuit that coincide with that first filter stage and they will all ground right at the cap. That will be, the filament center tap, the high voltage secondary center tap, the main power plug ground, and the bias supply. These will ground with that filter cap. Technically I think the output tubes cathodes should ground there as well, but I'm just gonna keep them grounded right at the tube socket for other reasons.
Image

Moving on, on the schematic at the top is the filter choke, and it sort of acts like a boundary between the main filter node (red side) and the next filtering node (blue side). As we travel down the B+ rail there is another multisection cap that filters the B+ for the screen grids in the power tubes and the phase inverter. The blue circle in the schematic and the cap circled in blue in my pic is the same capacitor. So everything that coincides with that screen/phase inverter capacitor will be grounded together. That would be the phase inverter cathodes which ground through the presence circuit, the master volume pot, and the entire tone stack. But I'm not grounding this stuff way off in the corner where the capacitor actually is. I'm grounding it all to the chassis "midway" through the amp as it physically exists in the amp. And it's gonna be right under the presence pot.
Image

So now we're getting to the preamp - the green section of the schematic. The dividing line between the blue phase inverter section and the green preamp section basically are the tone stack capacitors. They decouple the two sections. Coupling caps are found all over an amp but the tone caps are what divide these two filtering stages. So on the schematic the preamp filter cap is circled in green and all related grounds are green. This filter cap provides the super important filtering for these very sensitive and finicky gain stages that make this amp what it is. So I'm gonna group all of the preamp grounds together. That includes all of the cathodes from both preamp tubes, the preamp volume pot, and the input jack grounds. I'm not running all of this shit to a bus wire soldered to the pots like Marshall does. I'm grouping them all down to one single chassis ground tab near the input jacks faaaaaaarrrrrrrr away from the other high current noisy stuff.
Image

Basically, this is racial segregation if electrons were a race. I'm keeping things that are alike with their own kind and hopefully there is no mingling.

And no bus wire!
Image



I've you've made it this far, I applaud you. :coolstorybro: :illdrinktothat:

If it works, I applaud myself. :wank:
Rebel Yell
User avatar
JD01
Posts: 15855
Joined: Wed Jan 04, 2017 1:11 pm
Location: Wales, UK

Re: Spite Build - Marshall JCM 800 2204

Post by JD01 »

I got partway through the first long paragraph then gave up and just enjoyed the pretty pictures. Amazing work.
User avatar
Greg_L
Posts: 20668
Joined: Wed Jan 04, 2017 1:07 pm
Location: Where the knuckle meets the poophole

Re: Spite Build - Marshall JCM 800 2204

Post by Greg_L »

JD01 wrote: Wed May 25, 2022 3:12 am I got partway through the first long paragraph then gave up and just enjoyed the pretty pictures. Amazing work.
Lol I don't really expect anyone to read it. It's a lot. If it was flipped I doubt I'd read it. To be honest though breaking down the schematic and typing it out is like talking to myself and it really helps me think it through better. I've had a few aha! moments typing this kind of stuff out.
Rebel Yell
User avatar
Lt. Bob
Posts: 6577
Joined: Wed Jan 04, 2017 3:02 pm

Re: Spite Build - Marshall JCM 800 2204

Post by Lt. Bob »

lol .... I read every word and so, a question.

No bus wire and it looks nice and clean but in your other builds, like mine, don't you essentially use a bus wire when you run that coat hanger wire across the pots?.
User avatar
Greg_L
Posts: 20668
Joined: Wed Jan 04, 2017 1:07 pm
Location: Where the knuckle meets the poophole

Re: Spite Build - Marshall JCM 800 2204

Post by Greg_L »

Lt. Bob wrote: Wed May 25, 2022 12:17 pm lol .... I read every word and so, a question.

No bus wire and it looks nice and clean but in your other builds, like mine, don't you essentially use a bus wire when you run that coat hanger wire across the pots?.
Yes but that coathanger bus wire is floating. It collects grounds and funnels them to one single chassis connection near the input jacks. It is not connected to the pot casings which would create ground loops. The floating bus collects circuit grounds in sequence and only allows them to go in one direction. I use it on everything. I could do that same thing on this Marshall build but I'm taking it one extreme step further to see what happens. I'm eliminating the floating bus ground and grouping grounds with their respective circuit and filter cap.
Rebel Yell
User avatar
WhiskeyJack
Site Admin
Posts: 11413
Joined: Sun Jan 12, 2014 11:48 pm
Location: Canada
Contact:

Re: Spite Build - Marshall JCM 800 2204

Post by WhiskeyJack »

JFC. The more amps you build the more i get lost in what it is you are doing. I am with JD here. I really like the pictures.

And FYI, depsite my not knowing what it is exactly you are doing i can definitely see the care and attention you put into the work you do. that much is obvious. I have seen the inside of enough electronics to know that half the time people just don't give a shit.

There is a lot of shit giving going on in your builds and that is cool.
:happytrees:
User avatar
Greg_L
Posts: 20668
Joined: Wed Jan 04, 2017 1:07 pm
Location: Where the knuckle meets the poophole

Re: Spite Build - Marshall JCM 800 2204

Post by Greg_L »

WhiskeyJack wrote: Wed May 25, 2022 1:43 pm JFC. The more amps you build the more i get lost in what it is you are doing. I am with JD here. I really like the pictures.

And FYI, depsite my not knowing what it is exactly you are doing i can definitely see the care and attention you put into the work you do. that much is obvious. I have seen the inside of enough electronics to know that half the time people just don't give a shit.

There is a lot of shit giving going on in your builds and that is cool.
Thanks. I'm just trying to make sure they work right from the get go. I don't wanna do all this work and then spend another giant blob of time trying to figure out why it doesn't. An extra 30 minutes of care and attention now makes more sense than about ten hours of frustration later.

The grounding thing came to me as I got a better understanding of how these things work. I've been fixated on grounding ever since I jumped into this shit. As I learned more and dug deeper I found other examples of people doing the same thing. I'm not reinventing the wheel here, I take no credit for any of this, but I realized that I might be on to something and so far my experiments have worked well.

This is a good amp for this grounding experiment because they are so bad! I know how they work and sound in stock form, and the grounding is not good. They're noisy and very prone to oscillations and electrical pollution. Maybe this way will be better. Maybe it will suck and then I'll just arrange it closer to it's stock form. Everything is on the table.

Another option that I will consider is "lifting" the filament center tap. I'm not gonna get into all that right now, but it basically involves creating an elevated DC voltage reference for the heater voltage to ride on top of. Think of it like a high occupancy vehicle express lane overpass for the noisy filament AC voltage. The elevated DC voltage is like a highway overpass that the heater AC voltage cruises along above all the other stuff. This gets the heater voltage away from the "0" voltage ground reference that the rest of the circuit has to use. The elevated DC highway will be about 50 volts, and that will be zero as far as the heater wiring is concerned. But I'm not going there unless I think I need to.
Rebel Yell
User avatar
Lt. Bob
Posts: 6577
Joined: Wed Jan 04, 2017 3:02 pm

Re: Spite Build - Marshall JCM 800 2204

Post by Lt. Bob »

Greg_L wrote: Wed May 25, 2022 2:30 pm

Another option that I will consider is "lifting" the filament center tap. I'm not gonna get into all that right now, but it basically involves creating an elevated DC voltage reference for the heater voltage to ride on top of. Think of it like a high occupancy vehicle express lane overpass for the noisy filament AC voltage. The elevated DC voltage is like a highway overpass that the heater AC voltage cruises along above all the other stuff. This gets the heater voltage away from the "0" voltage ground reference that the rest of the circuit has to use. The elevated DC highway will be about 50 volts, and that will be zero as far as the heater wiring is concerned. But I'm not going there unless I think I need to.
so sorta like a bias current in a stereo amp that the signal rides on top of
User avatar
JD01
Posts: 15855
Joined: Wed Jan 04, 2017 1:11 pm
Location: Wales, UK

Re: Spite Build - Marshall JCM 800 2204

Post by JD01 »

Greg_L wrote: Wed May 25, 2022 9:38 am
JD01 wrote: Wed May 25, 2022 3:12 am I got partway through the first long paragraph then gave up and just enjoyed the pretty pictures. Amazing work.
Lol I don't really expect anyone to read it. It's a lot. If it was flipped I doubt I'd read it. To be honest though breaking down the schematic and typing it out is like talking to myself and it really helps me think it through better. I've had a few aha! moments typing this kind of stuff out.
Yeah, I totally get what you mean. I run a lot of training courses in geotechnics for my staff, sometimes I'm explaining stuff that I've been doing for years but taking the time to turn in into a presentation really crystallises my own understanding.
User avatar
Greg_L
Posts: 20668
Joined: Wed Jan 04, 2017 1:07 pm
Location: Where the knuckle meets the poophole

Re: Spite Build - Marshall JCM 800 2204

Post by Greg_L »

Lt. Bob wrote: Wed May 25, 2022 4:29 pm so sorta like a bias current in a stereo amp that the signal rides on top of
Yeah, I suppose.

Here is an expertly crafted visual aid. The main thing to understand is that AC and DC can live in the same wire.

Let's pretend the red wave is the 6.3v AC filament voltage riding on a 0v reference. With the transformer's filament winding center tap grounded, that center tap provides a 0v reference.
heater 6.3v - 0.jpg
So what can be done to reduce heater hum getting into stuff, sometimes an elevated voltage reference can be created and that's where the heater voltage gets it's reference point. It's very common in really high gain amps. What you do is take a feed somewhere off of the high voltage B+ supply and create a voltage divider network which is really just two resistors. Without getting into all the math, you choose two resistors that will yield the voltage you want - say 50 volts DC. At this 50v point you created is where the filament center tap will be attached. Now instead of riding on 0 volts, the heater voltage will be riding "elevated" on a 50v DC reference point. It's still 6.3 v AC for the filaments, but it's not referenced to 0v ground anymore. It's riding up above everything on a 50v reference.
heater 6.3v - 50.jpg
This is a very loose and sloppy generalization, but it's the basic point.
You do not have the required permissions to view the files attached to this post.
Rebel Yell
User avatar
Greg_L
Posts: 20668
Joined: Wed Jan 04, 2017 1:07 pm
Location: Where the knuckle meets the poophole

Re: Spite Build - Marshall JCM 800 2204

Post by Greg_L »

JD01 wrote: Wed May 25, 2022 5:00 pm
Yeah, I totally get what you mean. I run a lot of training courses in geotechnics for my staff, sometimes I'm explaining stuff that I've been doing for years but taking the time to turn in into a presentation really crystallises my own understanding.
Yes, that. I don't have anyone in the meat world to bounce this stuff off of, so I'm flying solo on all of it. Trying to explain it, even if no one cares or understands it, actually helps solidify my own understanding.
Rebel Yell
User avatar
Greg_L
Posts: 20668
Joined: Wed Jan 04, 2017 1:07 pm
Location: Where the knuckle meets the poophole

Re: Spite Build - Marshall JCM 800 2204

Post by Greg_L »

So getting away from hypotheticals and what might be, I need to get back to what is and what has to be.

Designing another board. Again, like the Assman, I'm using a generic board and installing my own turrets and layout. I'm not having to get too creative though. I used my JMP 2204 and JCM 800 2203 as reference and drew it up myself. Those are PCB amps but the "flow" is pretty intuitive. They have a very logical signal path even for PCB amps. They're easy to decipher. This is a fairly standard layout with a few tweaks for convenience and efficiency.

Image

Image

Turrets installed and this is how it will sit in the chassis.
Image
Rebel Yell
Post Reply