Two stages sharing grid load resistor?

[wizard333]

I see a lot of channel switching designs where the input goes right to the first stage of both channels in parallel and the shunting of one channel occurs somewhere down the line.

Is there any issue with the input stage of two channels sharing the same single grid load resistor? Probably the standard 1M? Some designs seem to use separate grid load resistors even though they are in parallel; I can't figure a good reason for that.

Some other designs run into one channel (usually the clean channel) input stage no matter what, and have a 1M grid load at the input, but relay switch in the gain channel in parallel when the clean is shunted down the line for the gain channel to be active. So you're still running into 1st stage of the clean channel in parallel when the gain channel is selected. Yet there is a separate grid load resistor for the gain channel, even though at that point it seems to me that is not necessary since the input for both stages can use the clean channel's grid load. Seems redundant, or am I missing something? A good example of this is the Mesa DC series.

If you think about the old 4 input Marshall plexi trick of using a patch chord between input 2 of channel 1 and input 1 of channel 2, you're running into both channel's input stage in parallel, but you also now have 2 1M grid load resistors in parallel, so wouldn't both stages be seeing a 500K grid load now? (ignoring the 68k input resistors for simplicity).

[Merlinb]

You are correct that if the grids are essentially connected together then you could use only one grid-leak resistor. (And yes, when you jumper a Marshall's inputs you effectively get 500k, more or less).
It is probably more of a layout issue, as the two valves are not always close to each other and you don't want long distances between the grid and the grid-leak, or you may want to keep separate ground schemes for each channel, so the grounds only meet at the input jack.

[wizard333]

Its not really a separate ground scheme if they are in parallel is it? Even given layout issues; we're probably not talking more than a few inches here.

So, in the case of a circuit where the 1st stages are in parallel at all times, if you for some reason wanted 2 grid resistors you'd want 2x2M to get the standard 1M. Carvin for some reason uses a single 2.2M in amps like that, I guess they just wanted a really high grid resistor.....

In the case where one channel is always connected to the input jack and the other channel gets relayed in to sometimes have the 1st stage in parallel, is there a reason channel 2 needs a grid resistor when its not active? When its switched in it will 'see' the other channel's resistor; when its not switched in is there a reason to care? For some reason Mesa uses a 470k grid resistor for channel 2 in that scenario, and I'm not sure why since 1) it seems un-necessary, 2) that gives a really low grid load at the input when that channel is active (1M and 470k in parallel), which will tend to dull the sound of passive pickups due to loading.

[Merlinb]

Its not really a separate ground scheme if they are in parallel is it? Even given layout issues; we're probably not talking more than a few inches here.

It depends on the design of the amp. Two separate channels are sometimes supplied from separate B+ chains, so deserve their own ground schemes.

In the case where one channel is always connected to the input jack and the other channel gets relayed in to sometimes have the 1st stage in parallel, is there a reason channel 2 needs a grid resistor when its not active? When its switched in it will 'see' the other channel's resistor; when its not switched in is there a reason to care?

It would be very bad practice to leave the grid of the unused valve floating. It would be liable to develop grid leakage problems, or run unecessarily hot (a couple of milliamps instead of ~0.5mA say), reducing its lifespan. Bad things generally start to happen when electrodes are left floating

[wizard333]

That makes sense.

That also makes it more problematic trying to get a std value on the input grids of both channels. If you set the channel who's grid is always there to 1M; you'll get something less when the other channel is activated. If you set both at 2M, the channel that is always on will have the whole 2M there when its the only one on, which could make it sound brittle.

The clever scheme there would be to use two relays, one that switches the 2nd channel's default grid resistor out when the channel goes active.

Hmmm.

[Merlinb]

That makes sense.

That also makes it more problematic trying to get a std value on the input grids of both channels. If you set the channel who's grid is always there to 1M; you'll get something less when the other channel is activated. If you set both at 2M, the channel that is always on will have the whole 2M there when its the only one on, which could make it sound brittle.

The clever scheme there would be to use two relays, one that switches the 2nd channel's default grid resistor out when the channel goes active.

Hmmm.

I think you may be overanalyzing it. There's not much difference between a 1Meg input impedance and 500k, usually.

Alternatively, you could have a single input jack with a 2Meg to ground. This would go to the relay arm. The two relay contacts would each go to a grid, each with its own 2Meg grid-leak resistor. That way you will always have 1Meg input impedance.

[wizard333]

I'd disagree on the 500k vs 1M thing. There may not be much difference on paper but there is a noticeable difference at the ear drum when using passive pickups. A difference in feel to if you are sensitive to what happens during your pick attack. Just like there is a difference between a 250k and 500k volume pot on the guitar, which goes into parallel with that grid resistor when plugged in.

That idea with the 2M at the input and having a 2M on both channels separated by a relay is an interesting one though. I like it!

[Merlinb]

I'd disagree on the 500k vs 1M thing. There may not be much difference on paper but there is a noticeable difference at the ear drum when using passive pickups.

I suppose you might hear the diff with nothing else changed, but if you're changing channels anyway, you're gonna get a whole different sound whether the input impedance stays the same or not!

[wizard333]

Different sound yes; but loading the pickups that way creates a certain change in "feel" no matter what sound you have going at the speaker. Things will be more dull, and you can "dial the high end back up" later in the amp, but its not the same thing, and it doesn't create the same feel at the strings.

[brucec59]

...For some reason Mesa uses a 470k grid resistor for channel 2 in that scenario, and I'm not sure why since 1) it seems un-necessary, 2) that gives a really low grid load at the input when that channel is active (1M and 470k in parallel), which will tend to dull the sound of passive pickups due to loading.

I would submit that this is intentional. Mesa tends to scatter bleeder caps throughout their amps to reduce over-brightness and increase stability, so to them dulling the signal at the input is a good thing.

And Wizard I agree with you about "feel." Very important factor in an amp.

[wizard333]

Yes they do, and its probably intentional, though not well executed. They tend to end up with amps that sound like the tone control on the guitar is almost all the way down, when its all the way up.

Marshall has done the same thing, but generally with better results. They did it to great effect in the 800 series channel switching amps. Other companies have done it to varying effectiveness; some dont bother so you end up with extremely bright and shrill amps as the gain goes up. I've done it myself and its a hard thing to get right, especially since the high end cut required varies by the amount of gain you're going to be dialing in. I tend to use a high end cut that works for low-medium gain levels, then have a push/pull to option in more high end cut for higher gain levels. You have to use your ears and play the amp as you change things, there is no formula for getting it "rignt on paper". There are varying ways to do it, the absolute WORST way to do it is direct caps to ground off your main signal path, in terms of killing the feel of the pick attack. I like a combination of caps across the plate resistors and caps to ground burried in the tone stack but not right off the main signal path.

Of course, if you dont care how the amp feels, its easy to just cut high end. Lots of designers are not serious players and they end up making that mistake. Then you end up with amps that just dont "feel" right, that don't respond well to subtleties in picking or especially to fast picking. Your average slow blues player probably will never care; more technical players will notice it immediately.

The problem I have with pretty much all Mesas is that they dull up the tone so much that you are stuck with a dilemma. Either you turn up the high end to the point where chording sounds good but single notes sound thin, or you set the high end so single notes arent thin, but then chording is a wall of mud. There tends to be only ONE semi-happy medium present in those amps, and its in the John Petrucci/James Hetfield land of tone, and if thats not your tone, fahgettaboutit. Not very versatile at all. People also end up turning the gain way up because the amps sound dull otherwise and the extra gain brings back some high end harmonics. Also very un-optimal in my opinion as cranking the gain kills dynamics and doesn't cut a mix very well.

When I get some time I'm going to dig into some Mesas and see if I can fix that issue.

[brucec59]

It is possible to get lots of gain without resorting to bleeding off highs. My amps have no caps across the plate resistors or shunting to ground.

[wizard333]

You can get lots of gain yeah, but how does it sound? You can get all the gain in the world and end up with a very bright amp, I see it all the time. My personal tastes go more to the big fat mids sort of tone, with plenty of mid/low mid/and thump, with just enough high end to bit and respond to pick attack but not enough to make me run for cover.

[brucec59]

Right. I mean it is possible to have lots of gain, no bleeder caps and sound killer, just as you describe. But it involves more than tweaking component values in the preamp - it is the whole approach to the amp.

[wizard333]

Ok I'll bite! Clips?