Replacing a switching opamp with a relay?

[audiovisceral]

I have a Marshall 6100 guitar amp I love, but it features four cheap little switching opamps to handle channel switching, and I would much rather replace these with more transparent relays.

The switching opamps are M5201, also known as NJM2120D and, with a bit higher slew, NJM2121D. I've attached the basic pin schematic. From what I can tell, pin 1 receives a "switch control" voltage, which tells it whether to put through the A or B input to the one main output.

I'm not experienced with implementing relays, so I'm not sure how they could be used to replace these components, or if it would be easy.

Any help would be great.

Thanks.

 

[audiovisceral]

Here attached is the Marshall power amp schematic, with the M5201 IC's I'm trying to replace in red.

Thanks again.

 

[Samuel Groner]

Are you sure that with this power amp afterwards you'll need to worry about the transparency of the switches?

Samuel

 

[audiovisceral]

Are you sure that with this power amp afterwards you'll need to worry about the transparency of the switches?

Only if it wouldn't be too difficult to replace them. I don't like the idea of my entire guitar rig filtering through these crummy 2V/us opamps. If at all possible, I'd like to get rid of them.

 

[abbey road d enfer]

Only if it wouldn't be too difficult to replace them.

Replacing them with relays would be a major undertaking, because you would need to reconstruct all the circuit around different opamps, thus needing an additional PCB.

I don't like the idea of my entire guitar rig filtering through these crummy 2V/us opamps.

Considering the amplitude of signals there, slew-rate limitation happens at around 200kHz. Is it really a problem?

 

[audiovisceral]

Replacing them with relays would be a major undertaking, because you would need to reconstruct all the circuit around different opamps, thus needing an additional PCB.

Ya that was more or less what I was thinking. I guess no dice.

Considering the amplitude of signals there, slew-rate limitation happens at around 200kHz. Is it really a problem?

Theoretically, perhaps not. In practice? We'll see. The ones I have in there are the 2 V/us M5201/2120D older versions. I have some of the newer 4 V/us 2121D's coming from Mouser in a few days since one of my originals is fried. From other opamp experiences, I wouldn't be too surprised if there is an audible difference, even with just that small step up.

 

[PRR]

> Are you sure that with this power amp afterwards you'll need to worry about the transparency of the switches?

This ill-controlled power beast is part of the instrument's "sound board". Yes, it is far from "transparent". However it will render damaged signals differently than un-damaged signals. Slewing flaw has a distinct character different from tube nonlinearity and transformer droop. And electric guitar needs selected nonlinearities: the raw sound of a naked steel string is boring.

> cheap little switching opamps

IC205 5201 drives IC206 TL071 which has gain of 6 (or more) and 13V/uS slew.

Anything over 13/6= 2.28V/uS in IC205 will get trimmed by IC206 anyway.

If 2.0 is not enough, then 2.28 won't bring joy either.

What IS the signal level here? The schematic is incomplete. (IMHO, if a geetar amp won't fit on one page, it is too complicated.) I suspect it is unity-gain from IC206 out to V7 input. (What, six tubes before the driver? PLUS all dem opamps? WAY too complicated.) The power-stage gain is a nasty function of (too many) resistors and switches; taking the one path I grok, gain is something under 16 to the 4 ohm output. Four 5881 bottles at over 400V is a "100 Watt" amp, 28V peak output, say 2V peak at V7 grid. Assuming unity-gain to V7 in from IC206 out, we have 0.35V through the FX loop (very reasonable) and out of IC205. OTOH, assuming that CON306 CON207 conceal some loss-pad, the only reason is to add distortion, and slew is a form of distortion.

BTW: the "slew rate" of the V7 stage is OTOO 7V/uS. The peak signal is 50V, four times higher than IC206 TLO71 can possibly swing. We could suspect that IC206 only needs 2V/uS and that IC205 only needs 0.4V/uS to keep V7 working to the max.

There may be other flaws in the 5201, but it has ample signal level, both peak and slew, for its position in the gain map.

I say someone did the math.

> I'd like to get rid of them.

Saw-off everything before V7. Jam RL201. Add one 12AX7. You can give it a sexy name like, oh, "5F6a".

And get rid of ZD1. That's just stupid. Some kid put that there. G1 bias *must* at least track B+ and G2 bias with supply variation. (I think the whole bias-supply has been "improved" beyond reason or sanity.)

 

[sodderboy]

Physically it would not be too difficult to add a board.  You replace the IC's with high quality machined pin sockets, and use a long-pinned wire-wrap socket through a breadboard to connect the new relays circuit.  It really depends on how high the surrounding components are.  If a direct board cannot fit, then use the wire-wrap as wired connector.  You solder wires to the top pin sockets and wire to the bread board.  You make one board for each chip.  If it does not work for you, put the original chips back in the sockets and play the blues while you consider your next project.
Depending on your solder skills, if you remove the chips intact and install sockets, you can simply put small wire jumpers in the sockets to listen how it sounds in one mode before breadboarding.  If you do not hear a diff, pop the chips back in the sockets.

If you do like the difference, then the next consideration is if the control FET's can power the relays instead of the chips.  Stealing power from the audio rails and adding driver transistors is only indicated if it really makes a difference.
Mike

 

[audiovisceral]

Physically it would not be too difficult to add a board.  You replace the IC's with high quality machined pin sockets, and use a long-pinned wire-wrap socket through a breadboard to connect the new relays circuit.  It really depends on how high the surrounding components are.  If a direct board cannot fit, then use the wire-wrap as wired connector.  You solder wires to the top pin sockets and wire to the bread board.  You make one board for each chip.  If it does not work for you, put the original chips back in the sockets and play the blues while you consider your next project.
Depending on your solder skills, if you remove the chips intact and install sockets, you can simply put small wire jumpers in the sockets to listen how it sounds in one mode before breadboarding.  If you do not hear a diff, pop the chips back in the sockets.

If you do like the difference, then the next consideration is if the control FET's can power the relays instead of the chips.  Stealing power from the audio rails and adding driver transistors is only indicated if it really makes a difference.

That's what I was thinking at first, but it seems like it would be way too much work, which is what I was posting here to find out. This is especially the case since it's PACKED already in these amps - 3 channels and channel 2 has 3 modes.

Some pics here:
http://www.timeelect.com/mar30.htm

Naaa, I guess I'll just keep it the way it is. On the plus side, I got the NJM2121D's here, and whatever the math says about slew or distortion, the amp does sound a bit more lively with them in instead of the old factory M5201's.

I also grabbed a a few TLE2072CP's and a TLE2071ACP to replace the TL072CP's and TL071CP in there. Quick question before I pop those in: The TLE2071ACP is substitutable for the TL071CP, right? As far as I can tell the only difference between the ACP and CP is the ACP has half the input offset voltage, which is always a good thing, right?

Also, PRR, damn man you know your stuff! If you're curious, the full schematics are here, but I don't think I'm going to bother digging in any further.

http://www.drtube.com/marshall.htm#AN30

 

[Gus]

You need to think about what you might ADD with the relays.
Coil switching noise
contact bounce noise
Doing relay switching correctly is not always as simple as it first appears
If the supply you use is at the edge you might cause clicks via power supply coupling to other stages.

The signal from a guitar is limited.  
Maybe you like the amp sound because of the switching chips.
Maybe the "better" opamps will hurt the sound, remember a guitar amp is about controlled distortion and EQ not clean, as PRR posted "the raw sound of a naked steel string is boring."

I have looked at a number of Marshall schematics and it appears to me they have a good designer(s).  Your improvement might make it worse but the only way to know is to try it.