Really Low Gain switch - SOLVED

[Tyrannocaster]

A few days ago I posted asking about how to do a switchable Hi/Low input setup with just one jack and that's what I was going to do as the last addition to this FrankenPro (which is really just a 5E3 at the moment since it's using 6V6s, albeit at pretty high voltages) but now that I can use 12AX7s (the crappy Fender branded 6L6s misled me and I don't have any other 6L6s on hand right now) I find myself constantly switching between a 12AX7 in the first preamp tube and a 12AU7. I can't get good cleans with a 12AX7 and of course the 12AX7 doesn't have much gain for distortion. Just turning the amp down isn't the way to go here because of the interactive volume controls, which change the amp's tone with the volume. Is there a way to switch the gain such a massive amount and just keep a 12AX7 there? Can I get enough drop by an alternate plate resistor and bypass resistor/cap combination?

EDIT: Scroll down to read the final working solution.

[sluckey]

You could use the hi/lo switch circuit I suggested but make the bottom 68K smaller, like 47K, or 33K, or 27K, or 22K, etc. Smaller = lower signal.

Or you could put the cathode bypass cap(s) on a switch.

Or you could use geezer's bypass cap attenuator. Very effective and easy to do since you already have a separate resistor/cap under each 6L6. Look at page 3 of this pdf...

http://home.comcast.net/~seluckey/amps/misc/Amp_Scrapbook.pdf

[tubenit]

Look at the full/bright/fat switch here in "switches, bells & whistles"

http://www.el34world.com/Forum/index.php?topic=5621.0

With respect, Tubenit

[Tyrannocaster]

Thanks for all the options, Tubenit! Correct me if I’m wrong: the Hi/Low switch would drop the gain before the signal even reaches the tube, so the S/N ratio would suffer, while actually changing the gain in the tube should create less of a hit there - or is that just gasbaggery?

Geezer’s bypass attenuator would reduce the gain on the power tubes, which is too late in the chain. I need to tame the gain in the preamp.

The full/bright/fat switch looks like the best option here; can it drop the gain as much as I’m asking about - ie, the difference between a 12AX7 and a 12AU7? If so, are those the values you’d try? I didn't think you could get that much change with just the bypass cap value, plus doesn't the higher value bypass cap emphasize bass more? Ideally, I'd like to keep the amp's current bright jangly tone.

[John]

If you have way too much gain, try lifting one leg of whatever bypass cap you have now and see how that sounds. Then use a couple test leads to clip in a .1, 1u, 2.2, 4.7, etc. I'm running a 6sl7 in V1 and the bypass cap is only .1 which is much less than I thought I would need. Naturally, yours will vary with the tubes you're using. HTH!

Oh, the bypass cap value might also affect the treble response.

[Tyrannocaster]

The bypass cap is already down to 1 uF. I don't think reducing it is going to do that much at this point. And you mentioned a crucial point, the change in tonal response.

I like the way the amp sounds now and after experimenting with a bunch of different preamp and power tubes this morning I think it will be great for its intended purpose, which is playing early Beatles tunes live in a friend's band. I have it to the point where I can live with the distorted sounds, while the clean (well, mostly clean - it's a tweed Fender, LOL) sounds are good, indeed. But it sounds a lot better to me than a stock tweed Fender; the low end isn't farty at all at any volume and it isn't as spongy either, although it still has that tweedy quality. That's the most important thing. It'll never be a Plexi style amp.

I think it's probably done. 

[thelonious]

Sorry I'm super late to this party - you've probably discussed a lot of this in your other threads, too, so take this FWIW, which is random late ideas that might spark future experimentation for anyone reading this thread.

I think it's probably done.

Yep, if you like it as much as it sounds like you do now, you're probably right.

But... since you've got quasi-parallel triodes at the input (or at least the possibility for them), you could try experimenting with a switch that would leave everything as-is for clean, but send some of the input signal to the grid of the unused paralleled stage when you want extra gain (thus making up the gain difference between the 12AU7 and the 12AX7 with the flip of a switch).

The opposite idea is to go with the 12AX7 for the distorted sound, but use the switch to switch in and out local negative feedback from the cathode of the functioning triode stage to the grid of the unused stage in order to reduce the gain when the switch is flipped. That NFB might actually get a "cleaner" clean, too, I guess, but I've never tried it, so I can't speak from experience.

Either way, you would get to keep one of the sounds you like now (either the clean one, with the 12AU7, or the distorted one, with the 12AX7), but have more flexibility available with the flip of a switch... but yeah, you probably can't keep both as they currently are, as you've already pointed out.

[thelonious]

By the way, the reason I'm suggesting these methods is because they both work with the quasi-parallel/interactive controls thing you've got going and don't want to lose. I don't think you'd have to add much signal to the unused grid either way (either in-phase or out-of-phase) to get a little more or less gain, so theoretically it's just slightly pushing or pulling what you've already got going on. Maybe?

[Tyrannocaster]

Sorry I'm super late to this party - you've probably discussed a lot of this in your other threads, too, so take this FWIW, which is random late ideas that might spark future experimentation for anyone reading this thread.

I think it's probably done.

Yep, if you like it as much as it sounds like you do now, you're probably right.

But... since you've got quasi-parallel triodes at the input (or at least the possibility for them), you could try experimenting with a switch that would leave everything as-is for clean, but send some of the input signal to the grid of the unused paralleled stage when you want extra gain (thus making up the gain difference between the 12AU7 and the 12AX7 with the flip of a switch).

Well, I don't think there is ever any unused paralleled stage; that sucker is at work all the time, mostly via triode to triode feedback which is volume dependent and which acts rather like a "Q" control, or a notch/peak filter. It is truly weird and I can see why some people hate these things and change the tone controil wiring. The so-called unused channel has an effect even when set to zero, depending on where the used one is set. In fact, with some preamp tubes I got oscillations at the very low end of the "unused" channels sweep so I actually put a 250 ohm resistor between it and ground; problem solved and it had the nice bonus of taming some of the really strong interactive effects and making it more easily manageable. My point is that I see this as a feature and not a bug so I wouldn't want to mess with it.

The opposite idea is to go with the 12AX7 for the distorted sound, but use the switch to switch in and out local negative feedback from the cathode of the functioning triode stage to the grid of the unused stage in order to reduce the gain when the switch is flipped. That NFB might actually get a "cleaner" clean, too, I guess, but I've never tried it, so I can't speak from experience.

That is really interesting and I hadn't thought of it. In fact, I think that is a truly nifty idea and I have to think about it while the amp is still apart!   Thanks for mentioning that! Who knows how strange that might get, though; perhaps it would be better to confine the feedback loop to just one stage...I have no idea. Wow, there is always something.

I found a pair of 5881s that I didn't know I had and put them in. Interesting that the plate voltage (383) is a full 15 volts lower on those than on the 6L6s (398) or the JJ 6V6s, which were the highest at 401. Those JJs sound really good, though - I've read that they are not that much like a "real" 6V6 and since they seem to be happy with that voltage (and dissipate 16.5 watts, which is 2.5 over the design limit for 6V6) now I have to decide which ones to use. The Fender branded 6L6s are just horrible sounding, suitable only for putting in an amp when you fire it up for the first time and fear redplating.   The Sovtek 5881s sound good and are a little louder than the JJs although they don't dissipate quite as much wattage as the Fender tubes but I suspect they might last longer. I think I'll try to find some more 6L6s to try in this.

I wish I felt safe using my pair of EL37s in this amp but I don't know what the filament capability of the PT is; I would guess it's fine, since I ran the EL37s in another amp that also had an organ PT (that one was from a small Baldwin, but this one is from an old Hammond AO-29 and it is pretty big) but I don't know how you tell other than by checking the external temperature. Judging by the tubes in the AO-29 I would guess it would be marginal at best: a 5U4 (doesn't count), two 6V6s, a 6BA6, two 12AU7s, a 12AX7, two 6AU6s, and two 6C4s. That adds up to 2.85 amps filament current and the EL37s make for 2.8 just by themselves. Adding in the three preamp tubes would take it to 3.75 amps; who knows how much safety margin Hammond left? I'd rather be safe than sorry in this case.

Big thanks to everybody on all my threads about this amp! You guys have been really helpful and I appreciate it.

[Willabe]

Well, I don't think there is ever any unused paralleled stage; that sucker is at work all the time, mostly via triode to triode feedback which is volume dependent and which acts rather like a "Q" control, or a notch/peak filter. The so-called unused channel has an effect even when set to zero, depending on where the used one is set. In fact, with some preamp tubes I got oscillations at the very low end of the "unused" channels sweep so I actually put a 250 ohm resistor between it and ground; problem solved and it had the nice bonus of taming some of the really strong interactive effects and making it more easily manageable.

I don't think that's right. I go along with thelonious on this.
    
In a 5E3/5E5 the volume controls are interactive. But I don't think it's because of changing the loading on the unused triode by rotating that channels vol. pot. That triode is unused as long as you don't plug into that channel.

The signal from the channel you plug into on it's way to the next triode grid has a choice of either going to that grid or going back through the other (1M) vol. pot. As you turn up the other (unused channel) vol. pot (unused channel) there's less resistance and more signal can go back/be bleed off through the coupling cap, then back through the plate load R (100K) then to ground through the B+ rail's filter cap. The coupling caps value is setting the "Q" notch.

If you were to disconnect the plate wire from the unused triode and leave everything else in place it would still work the same, ie, interactive controls.

As far as getting osc. with some pre amp tubes and not with others, I think it's either bad lead dress, bad grounding, a bad tube or a little of any/all of them?

The only reason I left both channels in on my 5E3 was so I could bridge the 2 channels if I want and didn't have another use for it.

Let's see what others have to say, maybe I'm wrong? Wont be the 1'st time.    

                        Brad      

    

[HotBluePlates]

The full/bright/fat switch looks like the best option here; can it drop the gain as much as I’m asking about - ie, the difference between a 12AX7 and a 12AU7?

No, not exactly.

Likely gain of a 12AU7 stage with a 100k plate load will be in the range of 10-14. Same stage with a 12AX7 will have a gain around 50-60.

When you don't have a bypass cap, typical gain stages have about half the gain of the bypassed case. So the unbypassed setting of the bright/full/fat switch can give you a gain ~25-30, which is still double what you'll get from a fully bypassed 12AU7.

The bypass cap is already down to 1 uF. I don't think reducing it is going to do that much at this point.

Reducing that cap's value will make that stage brighter. Right now, the -3dB point for 1uF and 1.5kΩ is around 100Hz (estimate a bit low due to an approximation). Essentially everything above that frequency is getting full gain. So you really ought to taste-test with the cathode bypass disconnected to see what you think.

... switch in and out local negative feedback from the cathode of the functioning triode stage to the grid of the unused stage in order to reduce the gain when the switch is flipped. ...

That might work. A much simpler method would be to use the switch to switch in/out a resistor from the plate of the used triode to its grid.

For example, assume you have 1M from triode grid to ground (if you have a grid stopper or use Sluckey's Hi/Lo switching setup, it will be a bit higher , but close enough). If you add a 10M resistor from triode plate to grid, along with a blocking cap of 0.001uF (or 1000pF), triode gain will drop to 10. You can leave the cathode bypass in place for its tone-shaping effect, as 10M and 0.001uF has a -3dB point of 15Hz, which won't interfere with the bypass cap roll-off point.

This will allow you to use 12AX7s and switch to 12AU7 gain. Well, that's the theory... try it and see if that trips your trigger.

[DummyLoad]

what HBP describes is basically a plate follower. you can vary the mu of a plate follower by varying the values of the input resistor and feedback resistor.

with the 1Meg pots, you're somewhat limited in the range of the gain, however, it may produce acceptable results for you.

example; 12AX7 gain stage B+ 250V, Ra is 100, Rk is 1.5K - pretty much fenderish values. EXCEPT we have a grid stopper, like the first stage that many of our amps have, only this stage can be ANY gain stage, i picked a 47K grid stopper b/c it won't kill off much of the HF. see attached schematic. there's DC on the grid, so there'll be DC on the wiper of the pot and it may be scratchy. for clarification, the schematic represents what a first stage might look like.

tube CAD shows the following results with 250K load with the above conditions and an adequate bypass cap.

with Rfeedback:
500K - gain is ~15
550K - gain is ~17
600K - gain is ~19
700K - gain is ~21
800K - gain is ~24
900K - gain is ~27
1Meg - gain is ~30
1.1Meg - gain is ~34
1.2Meg - gain is ~38
1.3Meg - gain is ~42
1.5Meg - gain is ~46

add a disconnect switch (switch on a pot) in series or switch in larger (=>1.8Meg) resistor and the FB loop is disconnected or Rfb is high enough and NFB has little effect and gain shoots to over 50.
with the Ck and Cc values shown in the attached schematic, f -3dB low is 25Hz, however, at maximum gain, f -3dB hi is ~55K and that might be the achilles heel of the plan. in other words the higher the value of the Rfeedback, the lower f -3dB hi roll-off becomes [worse HF response]. there is a trade off: if Rin is a LOWER value HF roll-off improves, however, minimum gain is higher. 

note: if you use a switch as indicated in the schematic, it's probably going to pop without a shunt across the switch.

i haven't fiddled with this circuit personally, but i know it works - RicharD modified a bass-pre we designed some time ago by adding a 1M pot to the feedback loop of the plate follower and he's quite happy with it. this was done so he could vary the gain of signal feed to the mixing console independently of the stage amp gain level.

-or- you can shave off gain by lowering the load resistance - kind of like a cross-line MV, 33KR under a 250K pot wired as a Variable resistor will vary gain with the conventional 12AX7 with 100K/1.5K gain stage from ~10 to ~25 if Rk is not bypassed; if Rk is bypassed, gain is ~28 to ~54. switch Ck in or out for greater variance.

--DL

[Tyrannocaster]

Very cool, HBP and DL; and they both look doable - thanks. That's exactly what I was looking for. Plus, because of my own stupidity I hit a snag with the amp and will have it out opened up for considerably longer than I expected to.

Lucky that the guy's birthday isn't until the end of August! Dumb Tyrannocaster installed an OT that won't quite clear the speaker frame! There's no room in the cab to move the speaker (it's an old, trashed Acoustic 109 solid state amp, very compact) and at this point I'm just not going to take the amp apart to move the OT - I don't think there is room to move it back anyway; it's pretty close to the 6L6s as it is. So I either have to make an entire new cab (head or combo) or cut down the existing cab into a head. The cab is too small to sound really good - they made it to be as small as possible and it's truly compact. Because there was no OT (solid state) they could get away with that and I flat out forgot to check the clearance before I put the unknown brand OT in; it's pretty beefy and if it were smaller I wouldn't have this problem but the amp might not be as loud either.

So if there are Darwin awards for amp builders I'd be eligible for one if the mistake were dangerous but luckily it's not. Just time consuming and irritating. But it does give me time to play with the circuitry and these ideas look fairly easy to try out.

Thanks!

[Tyrannocaster]

HBP, I tried your 10 meg/.001 idea but it didn't seem to have any effect. Since I don't have any 10 meg resistors here I used three 3.3 megs in series, and I can't imagine that would affect things. So I also tried with 6.6 megs and then finally 3.3 megs but while there is a difference with 3.3 megs it is so subtle as to be not worth doing - mainly I hear a slight decrease in hiss from the amp but nothing noticeable in terms of a volume drop from the guitar. So it is having some effect with 3.3 megs, but not the desired one. I suppose the cap value should change with the resistor but I don't know what the new value should be.

I also tried Dummy Load's 470k/.047 idea; this works better - there is a volume drop, though still not that much. There is less high end but that may be because of the volume drop and human hearing, however I can't turn the amp up to compensate for it without changing the amp's tone because of the interactive nature of the two channels. But the main problem with this solution is that when it is in the circuit the guitar's volume pot becomes very scratchy.

I wonder if the feedback between the channels cancels out some of the effect of this tinkering.

[thelonious]

... the guitar's volume pot becomes very scratchy...

Weird. Is there DC present at the jack?

[Tyrannocaster]

I didn't check while I had the temporary alterations tacked in and now I have the amp put back together because I'm going to show it off today and won't be able to work on it until at least tomorrow. All I can say is that without them, there's no scratchy guitar pot issue.

[HotBluePlates]

The cap in series with the resistor from the plate should block d.c. and prevent a scratchy guitar volume pot.

DummyLoad's drawing and application of the idea works better than what I suggested. Note that it requires the series 47k grid stopper; that's an integral part of the feedback mechanism which I didn't include/explain.

I also tried Dummy Load's 470k/.047 idea; this works better - there is a volume drop, though still not that much.

Volume drop isn't so much the goal, but drop in gain is. Did gain drop, similar to using a 12AU7? If yes, but not enough, you can reduce the feedback resistance to something like 220-270kΩ. If you do this, it might be a good idea to up the series cap to 0.1uF (half resistance and double capacitance = same bass roll-off point).

I wonder if the feedback between the channels cancels out some of the effect of this tinkering.

You're trying a great many changes from a typical amp. It is a good idea when doing this to breadboard first; but since you're already at the build stage, it helps if you don't have a tight deadline, so that you can experiment and tweak things. A major change in one part of the amp will always present a chance of upsetting the balance you have, possible requiring a corresponding change elsewhere in the amp.

But I'd argue that the interactive volume control effect you've noticed is not feedback, per se. Rather, the backwards wiring of the volume controls allows the signal of one channel to bleed backwards through the other volume pot when it is set above ~halfway. The used-channel's signal then sees a cap on the other side of that cap and an impedance to ground. Which is describing something akin to a guitar volume control, even if the resulting tonal effect is different.

Most new-intermediate amp tinkerers have a hard time looking at a plate load resistor as an impedance to ground, but for a.c., that's exactly what it is. The plate load connects to a filter cap, which looks like a short-circuit to ground for the audio signal; that makes the filter cap "same-as" ground, and explains why the volume control interact in a way that alters the tone.

[Tyrannocaster]

I would have thought the cap would block the DC, too. Perhaps the cap is bad? Just because it is new doesn't guarantee anything, I guess. I'll make sure I use a different one next time.

I did use the grid stopper, although mine is a 33k; is the 47k value Dummy Load referred to critical?

Honestly, I did not note a gain drop, at least not in terms of how early the amp breaks up, but the volume drop was not nearly what I get when switching to a 12AU7. The drop was noticeable, but only that; let's say 6 dB, which is something you can hear but isn't earth shaking to the ear even though it represents a 75% reduction, or 2x. Of course, I made up that figure, but you get the idea. As you noted, the difference between a 12AX7 and a 12AU7 should be more like 50 or so, depending on several variables, but at least 50.

I will try it again with a 270k and a .1 cap just to see, but I am somewhat skeptical. I think there may be too many variables here for an easy fix.

Thanks again for your generous help.

[Tyrannocaster]

Well now, things are getting interesting. I used a 240k resistor and a .1 cap - different brand of cap, too. (.1 orange drop) No scratchy pot now AND there is a good volume/gain drop when throwing the switch. The switch doesn't seem to be very noisy, either - no pops, although you can hear it. I can't really crank the amp yet because it's too early in the morning but this looks like a keeper to me. It's only tacked in right now but after I get to play with it at gig volume if I'm still happy with it I'll solder the circuit in permanently and call it done.

So once again, thanks to HBP and DummyLoad for the help with this circuit. The amp has turned out really nicely.

We demoed it last night with a bunch of other amps and it held its own ground: it's basically a light distortion amp made for cleanish sounds but it can break up when necessary, although it can't even come close to the higher gain 18 watt Marshall clones I've built in that respect. But it doesn't have to and to be fair, they can't touch this thing for its tweed clean vibe either. And speaking of vibe, the tremolo came in for special praise, despite its lack of a footswitch, LOL. The tremolo has enough depth even at the lowest speed (a common problem with trems in my experience, cf the new Fender Excelsior) and enough speed at either end of the dial - fast is fast and slow is slow. Outside of the footswitch the only thing you might ask of it is more complexity but that would require a totally different circuit and I wanted a preamp tube wiggler that only used one tube. And the amp takes a volume boost pedal really well, so you can get louder, more complex dirt easily if you want. It's a good compromise all around.

Incidentally, both this amp and my 18 watters (I have three with different preamp circuits) are considerably louder than the gorgeous Custom Shop Deluxe Reverb reissue (the one with the maroon vinyl and wheat colored grill cloth) that we had there. Lots louder, even when running through the same Eminence Legend GB. But that Deluxe is a beautiful piece of eye candy, I have to say it. It's apples and oranges, of course, but I'm pleased with how well my ugly little amp does against that competition.

EDIT: Tested at full volume and the switch is a keeper. Now I am able to use a 12AX7 and still get clean tones from it so no need for the compromise tubes I was using last night and no need to swap tubes, either. I suppose you could add this switch as a footswitch, too, because the effect is very much like hitting a booster pedal, although I don't know if you would get any oscillation or other problems because of the long length of cable run. But I'm going to leave it as a switch on the amp, just to reduce the number of options; options can get in the way, I think. BTW, when I wired this up this morning, I put the circuit on the wrong side of the grid stopper, ie, right to the grid. Before I caught the error I tried it out and it works great there.

[thelonious]

Ok, now you've gone and made me want to build one of these! Sometime when you get the chance, could you possibly post a quick sound clip demonstrating the effect of the switch? It would be helpful to have as part of the thread, for posterity's sake...

And can I just say once again that I love the collaboration that takes place around here?

[Tyrannocaster]

Yeah, I hate it when I read all this stuff but don't get a chance to hear what it actually sounds like. As it happens, I recorded the difference the switch made this morning when I was testing it before I played it full up through a speaker. I wanted to make sure I wasn't kidding myself, I guess. This is through the amp's line out (the one with the tone control, LOL - but it was set on full here, obviously) straight into the board, which had some delay and EQ running from earlier. It's not music, but it does show off two things - the gain switch, and the amp's nice tremolo. In this clip the amp is using a 12AX7 and the amp's volume (channel 2 - the "on" one. I don't remember what the so-called "unused" one was set at) was turned up to about Noon IIRC - enough to really start to break up with the normal gain. And there is more gain left if you keep going. The difference with the switched lower gain should be readily apparent. I must emphasize that the amp sounds very different through a speaker!

Clip: https://www.box.com/s/d40fcfc828573ff72697

[HotBluePlates]

... Tested at full volume and the switch is a keeper. Now I am able to use a 12AX7 and still get clean tones from it so no need for the compromise tubes I was using last night and no need to swap tubes, either. I suppose you could add this switch as a footswitch, too, because the effect is very much like hitting a booster pedal, although I don't know if you would get any oscillation or other problems because of the long length of cable run. ...

I would strongly advise against running wires out to a footswitch to take the place of the mechanical switch. It's begging for oscillation and treble loss, because the long wire run is effectively an antenna for the first preamp tube grid.

If you want to make that footswitchable, we have to go a different route for switching: LDRs or JFETs.

You'd want to use one of these two to take the place of the mechanical switch, the use the footswitch to switch a control voltage to either of these, which turns them from their high-resistance state to a low-resistance state.

The LDR/JFET and control voltage should reside inside the amp, with the LDR/JFET right at the circuit area for this tube. The control voltage can be placed elsewhere in the amp (maybe near the power supply), with footswitch wires doing something like switching a ground or controlling a transistor which allows the control voltage to the LDR/JFET switch element or not. That ensures the footswitch and cabling has no impact on the sound or operation of the switch and tube stage.

It is possible to have both a footswitch and panel switch, each of which can control the gain setting, but the circuit could get a little complicated. That is probably best devised/planned at the pre-build stage, when you can leave space for all the related parts. Regardless of what you think of the tone of Mesa's amps, you will eventually note that on many models most of the complexity is more about switching functions rather the actual tone-shaping parts of the amp.

I did use the grid stopper, although mine is a 33k; is the 47k value Dummy Load referred to critical?

Yes, sorta.

When I suggested this, I wasn't thinking the term "anode follower" that DL mentioned. What I had in mind was feedback around an opamp, which is a way you can look at a tube stage. The grid is the inverting input (-) of our opamp; the cathode would be the non-inverting input (+) of the opamp. The plate is the output.

Opamp feedback theory tells you if you have a resistor in series (R1) between the signal source and inverting input, and another resistor between the output and inverting input (R2), then the gain of the opamp is R2/R1. There is a limit in that the gain set by feedback cannot exceed the gain of the stage without feedback (if you use a 12AU7, you ain't gonna get it to have a gain of 120). You also have to be judicious in your selection of resistances (I wasn't the first time around with my erroneous post).

A real tube stage has some extra complications, which is probably why DL's numbers don't exactly equal R2/R1. For example, the formula predicts a gain of 10.6 with a feedback resistor of 500k (R2) and a grid stopper of 47k (R1); 500k/47k = 10.638. DL's chart shows TubeCAD comes up with a gain ~15. If you reduce the value of the grid stopper, you'll increase the gain. Lesson there is use a pot to trim to the value that works like you like, then sub the pot with a fixed resistor.

See diagram of an opamp feedback circuit below. Pretend the d.c. voltages shown are really a.c. voltages. As drawn, 47k grid stopper and 470k feedback resistor result in a gain of 10; 2vac input results in -20vac output (gain of 10, inverted).

[DummyLoad]

as HBP says, use an LDR  or use a small signal relay with a DC coil to simulate switch.

happy to throw 2cents worth now and then; much more significant is that you now have added knowledge, something useful and you hacked out the solution to your problem.

happy rosin smokin'

--DL