If all you do is put a relay or some other type of switch in parallel with the op amp, the input and output of the unpowered op amp wind up in parallel with your drum pad, which is a high impedance device, and can't tolerate much loading. The input and output impedances of an unpowered op amp are unspecified, hard to predict, and hard to simulate accurately, due to the nature of the op amp spice model.
Do you have a datasheet, schematic, or anything else that will tell me what the drum brain's input impedance is? How long is the cable between this preamp circuit and the drum brain? How about between the drum pad and the preamp?

I think I saw a similar thing, I was attempting to use a function generator to get a low voltage pulse into the brain directly, but when the generator was plugged in, even when turned off, it would trigger rapid, inconsistant drum hits. this only happened when the unit was plugged into the wall, sadly this is the only way a function generator is useful lol.

no op-amps were involved. I have a bunch of waveform captures of the drum hits if your interested.

 

What company makes the drum brain? Does it have a model number?

 

Its ION Audo, no model number or anything, its the xbox 360 drum brain.

I believe the actual maker however is either Alesis, or Medili (Medeli?)

 

One thing I'm notcing is that anything beyond about 3.8V(Max) input to the op-amp results in a clipped signal on the output (about 7.6V[Max]).

Does this matter to the brain? I honestly don't know, but what I do know is that I have seen upwards of 10V(Max) being put out from the drum pad (10V input to the op-amp)

I could be wrong but there should be a resistor/pot combination I could use that would result in max unclipped output at 100% pot value. The problem I see with this is that the drum pad is capable of higher output. Since I only know the simulation I can only assume that when the drumpad puts out 10V (one of the harder hits) it won't reach the brain and wouldn't register as a loud hit.

Part of this problem is due to not knowing the exact threshold at which the brain determines the hit is loud or less than loud, or quite frankly HOW it goes about determining the hit level.

am I overthinking it? most likely! but I just really want to know this thing, I want to produce a quality unit.

My real life test had good results, better than I could have hoped for to be honest so it's probably a lot less complicated than I'm making it

 

I created a discrete transistor version of LM324, and ran a zero power sim on a single stage. I'm sure the model isn't perfect, but its performance with zero power has to be closer to real life than the macromodel. If you only had one op amp, and the power supplies were disconnected (as opposed to being zero volt sources), which I assume would be the case, it would probably work with a simple bypass switch of some sort. I'm not so sure of the performance when there are 8 op amps with their supply pins all connected to each other.
I might have some time in the next few days to work on the full circuit sim, but you would need patience. I have a life outside this forum.
BTW, if it works, the bypass can be done (I think) with one transistor, a resistor, and a capacitor, per stage.

 

the LM324 is a quad package so there are only two IC's on the board (I get the feeling you knew this though)

do you have a schematic?

 

the LM324 is a quad package so there are only two IC's on the board (I get the feeling you knew this though)

do you have a schematic?

Here is a sim I ran on LTspice, using a simplified LM324 model with discrete transistors.
If you want to run this sim, I can attach the extra files needed to do so.

 

When I build and run that in multisim it appears to work, but when the power is turned off it still looks to be amplifying the output, controllable by the pots still.

I've built a couple of circuits and actually tested them and they work a hell of a lot better than multisim said they would (this is to be expected) so I'm not sure what to think of this in reality.

The simulated results still lead me to believe that there may be some power backfeeding from somewhere, but I can't even be sure of that.

I was going to include a picture of this with the power ON but the LED still doesn't light up to indicate such and it would still look the same anyway.
The results are the same on an amp that doesn't have this passthrough on it as well though, so it may not be related to this.
Would a 2N3904 work instead of the jFET? Those are expensive.

 

Hey Ron could you (or someone who knows their stuff) explain exactly whats happening with your bypass?

It looks to me like it would be sending the unamplified signal through to the output even when the unit is on.

with propogation delay this could result in double hits being registered or other odd errors. since this is being used mostly for a video game having that signal passthrough with the amplified signal COULD cause a high score to go kaput if it detects a second, delayed signal.

there is an in game calibration though that adjusts the games delay because audio and video equipment causes delay. I believe there is a 20ms or so delay already thast I've had to calibrate for.

 

Hey Ron could you (or someone who knows their stuff) explain exactly whats happening with your bypass?

It looks to me like it would be sending the unamplified signal through to the output even when the unit is on.

with propogation delay this could result in double hits being registered or other odd errors. since this is being used mostly for a video game having that signal passthrough with the amplified signal COULD cause a high score to go kaput if it detects a second, delayed signal.

there is an in game calibration though that adjusts the games delay because audio and video equipment causes delay. I believe there is a 20ms or so delay already thast I've had to calibrate for.

When power is on, the JFET's Vgs=-9V, so it is off. When power is off, Vgs=0V, so the JFET is on (I probably should have added 1Meg or so from Vee to GND to ensure this). The cap from the input to the gate ensures that Vgs=0 even when the input signal varies above and below GND. It also causes the delay when power is applied and removed. I didn't think this would be a problem.

BTW, I chose 2N5432 from a plethora of JFETs because it has low Rds(on), and is available in thru-hole. I just realized that Vgs(off) can be as great as -10V. You might want to choose another JFET.

In my simulation, the output is doubled when the power is on because the op amp gain is 2, but the bypass gain≈1. You can change the op amp gain.

 

Thank you, hat makes sense now.

Would the 2N7000 be an acceptable replacement? or do I need a JFET?

What would be a good value for Vgs? I don't really understand FET's so I need a little guidance here.

would I want a lower Vgs? like -8V? I'm having a little trouble trying to wrap my head around how the -ive voltage will keep it off and why -10V is not ideal?!

For op-amp gain would I want to make the gain higher or lower? the gain right now is 1.5 on the op-amp so does that mean with the bypass it's adding to, subtracting from, or doing nothing to the op-amps gain?

 

A MOSFET won't work. N-channel JFETs are suitable for this application because the are off when the gate voltage is more negative than Vgs(off), and are fully on when the gate voltage is zero. Note that source and drain are interchangeable.
Your negative power supply is -9V, so the magnitude of Vgs(off) must be less than 9V, or it will not turn off when power is on.
The bypass switch doesn't affect the op amp's gain. It is just a switch from input to output.
You can set the op amp's gain to any value you want.
Jameco has J109, which should work here. The minimum order is 20, but since you need at least 8 of them, and the unit price is 19 cents, you can probably swing it.