JTM45 cascading gain "one-wire" mod

[Tom Phillips]

Hi Mike,

It took me a few moments to get re-familiar with this 10 year old discussion. I never used the mod myself. The OP (Larry's) question just caught my interest at the time so I gave him some ideas. The mod mimics Marshall's method to provide a high gain option by cascading the two front end stages of their earlier design. It worked for them but the layout was certainly sub optimal because, in the high gain mode, the pre-amp signal is looped back to the input section of the chassis which is a setup for instability. In addition, the component values are a compromise to allow the amp to sound OK in both the normal and high (cascaded) configuration. Since an amp built for personal use doesn't need to skimp on parts just to save a few bucks on a commercial product, a good approach would be to make each pre-amp channel totally independent. Then each channel can be optimized for overall gain and tone. The channels can be selected by using separate input jacks or the amp can be configures with a channel switching arrangement.

I'm finishing taxes too so I need to get back to that task. If no-one else chimes in perhaps I can answer your specific questions later.

Cheers,
Tom

[mikepukmel]

Hi Mike,

It took me a few moments to get re-familiar with this 10 year old discussion. I never used the mod myself. The OP (Larry's) question just caught my interest at the time so I gave him some ideas. The mod mimics Marshall's method to provide a high gain option by cascading the two front end stages of their earlier design. It worked for them but the layout was certainly sub optimal because, in the high gain mode, the pre-amp signal is looped back to the input section of the chassis which is a setup for instability. In addition, the component values are a compromise to allow the amp to sound OK in both the normal and high (cascaded) configuration. Since an amp built for personal use doesn't need to skimp on marts to save a few bucks on parts, a good approach would be to make each pre-amp channel totally independent. Then each channel can be optimized for overall gain and tone.

I'm finishing taxes too so I need to get back to that task. If no-one else chimes in perhaps I can answer your specific questions later.

Cheers,
Tom

Thanks Tom, good luck on your taxes. "Talk" to you later!

[nickb]

While Tom is having fun, here's a suggestion. I prefer to keep the both sets of jacks for versatility and have the bright channel on the high gain side. I also like to add a post PI master volume. Leaving the normal jacks in circuit tames the excessive gain. If you want it back just insert a plug with no lead into one of the normal jacks. You can thus get three different levels of gain. The MV pot is a dual gang log type.



(Fixed 'typo' on dwg)

[Chuck H]

The above drawing has a few things to be considered.

First, you've routed from the mixing resistor, then through the grid stop AND notice that the other grid stop at the (now second) stage has a path straight to ground. Effectively forming a voltage divider of 338k/68k. That's going to chew up a lot of gain. If gain is the goal? And I assume it is since you're cascading channels

Second, the shared cathode resistor is less than ideal for cascade stages. Since you're fully bypassed you could probably get away with it without strange things happening. I'd separate the cathodes. In fact...

Third, you would be coupling a 22n fed stage to a 22n fed stage at relative impedances that both have ample LF passing through the circuit. This PROBABLY isn't going to sound good. One of those stages should probably be turned into a "bright" channel, ala 1959. Complete with the different cathode circuit values. Then THAT stage would be fed to the "normal" channel grid. Perhaps even through another circuit to reduce gain a little. IMHE stacking two full gain channels is too much clipping at the third stage to sound "right" and instead sounds overly compressed and blatty.

The last thing about the drawing (that I noticed) is that it's possible to interfere with normal operation in cascade mode if anything is plugged into the inputs for the channel acting as stage two. That's not a big deal for a personal amp, but years from now when that amp has been in a closet for awhile and you've built five others since, it could be a question mark. Trust me. Better to break connection at the second stage grid so the jacks are removed from the circumstances.

I used four three different Marshall models (four heard total) during the time I was gigging and I stacked the channels on all of them. What you think sounds good could be different than what I like. So this is just presented for consideration.

Basically you would make your breaks right from the volume control and the grid and supplement your own circuit in between. When I get time I'll draw something up.

[nickb]

The above drawing has a few things to be considered.

First, you've routed from the mixing resistor, then through the grid stop AND notice that the other grid stop at the (now second) stage has a path straight to ground. Effectively forming a voltage divider of 338k/68k. That's going to chew up a lot of gain. If gain is the goal? And I assume it is since you're cascading channels

Second, the shared cathode resistor is less than ideal for cascade stages. Since you're fully bypassed you could probably get away with it without strange things happening. I'd separate the cathodes. In fact...

Third, you would be coupling a 22n fed stage to a 22n fed stage at relative impedances that both have ample LF passing through the circuit. This PROBABLY isn't going to sound good. One of those stages should probably be turned into a "bright" channel, ala 1959. Complete with the different cathode circuit values. Then THAT stage would be fed to the "normal" channel grid. Perhaps even through another circuit to reduce gain a little. IMHE stacking two full gain channels is too much clipping at the third stage to sound "right" and instead sounds overly compressed and blatty.

The last thing about the drawing (that I noticed) is that it's possible to interfere with normal operation in cascade mode if anything is plugged into the inputs for the channel acting as stage two. That's not a big deal for a personal amp, but years from now when that amp has been in a closet for awhile and you've built five others since, it could be a question mark. Trust me. Better to break connection at the second stage grid so the jacks are removed from the circumstances.

I used four three different Marshall models (four heard total) during the time I was gigging and I stacked the channels on all of them. What you think sounds good could be different than what I like. So this is just presented for consideration.

Basically you would make your breaks right from the volume control and the grid and supplement your own circuit in between. When I get time I'll draw something up.

I think it's worth keeping in mind that the subject was "one wire mod" I'm trying to keep to that.

With the 330uF on the cathodes a shared resistor is hardly a problem.

The divider is intentional. You get two advantages
(1) it tames the gain in a optional way as I explained above. Otherwise it's too easy to get it to oscillate
(2) You still get to use the second set of jacks. The first stage retains the HF pass caps and that helps the tone although you do loose the effect as you wind the gain up.

I've done this a few times now and the result is universally liked. I used to like less low end in that first stage too and it certainly a matter of personal taste, but now I leave the coupling cap alone.

[Chuck H]

I was actually talking about Mikes drawing. But I suppose your "corrections" could still apply. Your JTM45 circuit is different from the one Mike posted, using 470k mix resistors. And the way yours is wired there would actually be a 34k load following the 470k series resistance. So that's about 10dB boost. Along with a PPIMV that should make for plenty of gain. Then there's the option of plugging an open patch cable into the unused jacks to lift one or both of the 68k load resistors, depending. So your circuit actually offers four gain levels But the addition of the PPIMV does sort of take it out of the "one wire" criteria. But I think any reasonable mod would.

[dstrat]

While Tom is having fun, here's a suggestion. I prefer to keep the both sets of jacks for versatility and have the bright channel on the high gain side. I also like to add a post PI master volume. Leaving the normal jacks in circuit tames the excessive gain. If you want it back just insert a plug with no lead into one of the normal jacks. You can thus get three different levels of gain. The MV pot is a dual gang log type.

[ATTACH=CONFIG]48475[/ATTACH]

Sorry but as drawn it looks like v1b will be a oscillator that's not connected to anywhere?

Edit: ( it should be connected to v1a. )

[nickb]

Sorry but as drawn it looks like v1b will be a oscillator that's not connected to anywhere?

Edit: ( it should be connected to v1a. )

Uh oh! Thx - I did the the graphical equivalent of a typo, I guess. I've updated it.

[mikepukmel]

Thanks everyone, reading and re-reading. I don't completely understand all of the discussion, working on it. I also sent the image to one of my friends from high school, who does electronics, very busy these days as well, and his brief comment was a question if the shared cathode cap and resistor would be good for a cascaded mod. I only understand this in a descriptive way so far, he didn't elaborate. You guys mentioned that above. Does it mean that the values would need to be different on the second stage because the input would be a much bigger signal when in cascaded mode? The cathode resistor biases the tube, right? Would you need a cooler bias if the input wasn't the small mv range guitar signal? How could I calculate what the output signal (peak to peak volts) of V1a is?

What if instead of one switch, I put a few jumper blocks in, if more components needed to be different? Also, I have a bunch of free turrets, since I moved the 68k grid stops to turret lugs near the tube pins for V1, so it would be easy to add a second cap and resistor for V1b's cathode. if needed.

(Not so much worried about one wire, as getting it to sound "good", but I can't quantify that. You know, Good. Not bad.).

[Chuck H]

How about some particulars.?. Are you using a master volume on this amp now? Higher gain tones at lower levels are impossible without one. So what is your goal? Do you wish to have a very loud, higher gain amp or a controllable higher gain amp? Is your amp currently the circuit you posted with the 270k mix resistors and the 500p bright channel cap or is it like Nicks circuit with the 470k mix resistors and the 220p bright channel cap? How do you use the amp now? What channel are you normally plugged into? Would you prefer this mod on a switch or do you want to toggle input jacks? Any other circuit particulars or preferences we should know about before taking time to consider design options?

[mikepukmel]

How about some particulars.?. Are you using a master volume on this amp now? Higher gain tones at lower levels are impossible without one. So what is your goal? Do you wish to have a very loud, higher gain amp or a controllable higher gain amp? Is your amp currently the circuit you posted with the 270k mix resistors and the 500p bright channel cap or is it like Nicks circuit with the 470k mix resistors and the 220p bright channel cap? How do you use the amp now? What channel are you normally plugged into? Would you prefer this mod on a switch or do you want to toggle input jacks? Any other circuit particulars or preferences we should know about before taking time to consider design options?

Its a new build, I have most parts wired to the board, but the board is not bolted in and wired into amp. (sockets, xformers etc). By accident I got a couple of recommendations re cascading, and started looking into posts on MEF. This is the schematic the kit is built on (mostly, except a few parts):
http://raw-sewage.net/images/jtm45-readable.jpg

Its possible now to drill and put a couple of more turrets onto the board to help make the wiring neater, for a master vol mod, cascading mod, and adding cap and resistor for the second half of V1 cathode, since the board is not in yet.

This is the state its in, with the board 'dry mounted' to see that everything fits, so can still do mods:


There's no master volume, but I dug up a post someplace to add a master volume right after the tone stack, like the 2204 has (1meg, leads to the cap on the phase inverter input). Somebody said that adding a MV after the tone stack would make the tone stack 'less effective'. The kit doesn't have the master vol pot or hole drilled, so thinking about how to do that without toasting the front panel.

I have enough room on the board, since I will (or would like) to put the grid stops on the turret lugs.

****
I just looked up a few other Marshall schematics, looked at the JCM800 2204, and they add a separate cathode resistor 10k (about 3x the 2700R on V1a) on the other half of V1. But there's no capacitor on the other cathode.

I have room on the board to put another resistor and cap for the other cathode, if need be.

***
So: would like to have more flexible tone, e.g. overdrive sounds, at lower volumes, if possible. The current amp kit has 270K mixing resistors, a 500p bright cap, no mv, 820R and 330 uf on the V1 cathode. Also, the filter caps are 32 - 32 - 32 - 16, rather than the schematic 32 - 32 - 16 - 16. (i.e. 32uf phase inverter cap).

Since the amp isn't built yet, If the mod was on a switch, then I could give a listen both ways and see how both sound.

If the amp turns out to be way too loud to get 'cool' tones, then I will build a very small amp room in the basement, with double 5/8" sheet rock and a mic. When we move out, I'll just put a toilet and sink in there and make it a soundproof bathroom

[dstrat]

IMO modding is fine but it will no longer be a jtm45, I would recommend building the jtm45 and play thru it for a bit, It will be loud but thats the beauty of the amp to begin with ( Jim Marshall's take on the 1959 bassman ).
My Reasoning is you will have a working amp to start modding if you decided to.

Good Luck!

[eschertron]

To sit on the 'other shoulder': since this is a clean slate build you should make the room for the changes, split the cathode resistors as suggested, and wire up the mods with jumpers or switches (keeping in mind bad lead dress can make for bad-sounding amps). Don't wire yourself into a corner and make the cascaded stage option hard to do later. Make it part of the design goal.

[Chuck H]

I'm actually with dstrat on this (and a little with eschertron). It's not hard to add the channel stack after the build, but you should leave a place on the board to separate the cathodes if that comes up. There is simply no place on the front panel for a master volume if you're using a standard face plate. The last time I did a master on a jtm45 I stuck it in the indicator lamp hole, moved the indicator lamp over and omitted the standby switch. I had no problems electronically with this arrangement using the normal post tone stack master, but it's ugly and the face plate indications are incorrect. I use to tell people that the "indicator" indicates how loud it is A PPIMV is electronically more sensitive and many people experience oscillations after installing them. I know for certain it can be done without this problem, but I don't know if the indicator lamp hole location would work for this since I haven't done that. As far as a switch for a stack mod, don't worry about it. You can do it on a push/pull pot located on the volume control for the channel that is the "first" gain stage. No extra holes. Some guys don't like push/pull pots but I have no problem with them. YMMV.

Two other considerations...

1) Power scaling. Controls located on the back. Keeps the clean look and you might be surprised at just how much gain a cranked up JTM45 has WITHOUT stacking the channels. Certainly Angus never had any trouble getting a tone from the stock circuit, unless...

2) I saw an interview with AC/DC's equipment manager and it seems Angus has been using a channel stack+master volume mod on his amps for a long time now. Who knows how long.?. The whole time perhaps? So I would definitely look into whether that circuit is known because we KNOW that mod sounds good.

[mikepukmel]

I'm actually with dstrat on this (and a little with eschertron). It's not hard to add the channel stack after the build, but you should leave a place on the board to separate the cathodes if that comes up. There is simply no place on the front panel for a master volume if you're using a standard face plate. The last time I did a master on a jtm45 I stuck it in the indicator lamp hole, moved the indicator lamp over and omitted the standby switch. I had no problems electronically with this arrangement using the normal post tone stack master, but it's ugly and the face plate indications are incorrect. I use to tell people that the "indicator" indicates how loud it is A PPIMV is electronically more sensitive and many people experience oscillations after installing them. I know for certain it can be done without this problem, but I don't know if the indicator lamp hole location would work for this since I haven't done that. As far as a switch for a stack mod, don't worry about it. You can do it on a push/pull pot located on the volume control for the channel that is the "first" gain stage. No extra holes. Some guys don't like push/pull pots but I have no problem with them. YMMV.

Two other considerations...

1) Power scaling. Controls located on the back. Keeps the clean look and you might be surprised at just how much gain a cranked up JTM45 has WITHOUT stacking the channels. Certainly Angus never had any trouble getting a tone from the stock circuit, unless...

2) I saw an interview with AC/DC's equipment manager and it seems Angus has been using a channel stack+master volume mod on his amps for a long time now. Who knows how long.?. The whole time perhaps? So I would definitely look into whether that circuit is known because we KNOW that mod sounds good.

How cool is that, and anyone who likes AC/DC would argue that the Angus sound was incredible.

**
I don't mind push/pull pots at all, nice elegant way to add mods to an amp without drilling and trying to figure out where to put the switches.

OK settled, thanks everyone for the great discussion (going to download and save this thread).

Ordered a few packs of turrets, Have plenty of room on the board where the 4 x 68k grid stops were previously put on the board, will put a few turrets there to make splitting the cathode on V1 easier later on. Thinking about how the MV wire would be routed, and best place to put a PPIMV (just figured out that meant post phase inverter master volume) pot on the back panel. Will build it stock, get my ear plugs out, try it for a month, and see what mods to go for after the neighbors get wind of the new amp in the neighborhood. Both mods should be easy after the board is setup for it.