Analog help: Splitting guitar signal and eliminating cross-talk

[Scott D]

Hey everyone,
First ever post, so be kind. I am hoping somebody with analog experience would be able to help. I have taken some advice from an EE I know, but have exhausted his analog knowledge.

I am not an EE, but know enough to be dangerous.
In simple terms, I have an integration problem. I have designed and built an analog circuit that will be housed inside a guitar and requires the AC source input from the guitar, whilst maintaining its standard output. So, I want to split a guitar signal from inside the cavity in more or less a "Y".

In essence, the circuit is for all intents and purposes, an amplifier in its own right, which does frequency filtration. As expected, the output power drops when split. I also understand there are inherent issues with respect to both amps may attempt to drive each other etc. (although the theory/math hurts my brain).
The issue I have seems to be one of cross-talk between the circuit board and the guitar output to the amp. There is a noticeable fuzz/distortion which is often talked about in this scenario of splitting signal/driving two amps at once.
Relevant portion of circuit: " alt="" class="bbc_img" />

I ran a "Y" split through 10k resistors on the active/hot line to circuit/guitar output. When scoping the circuit board vs the guitar output, I can clearly see one signal seems to be out of phase with the other. Which does make sense, as I am hearing a classic "fuzz" distortion sound through to the guitar amp.
" alt="" class="bbc_img" />

The leading theory so far, had been that the passive pickups (and its field) are picking up the circuit board and introducing noise this way. I have tested the circuit boards at least ~1m away, with the signal being introduced by alligator clips directly to the POT, needless to say, it doesn't seem to make any difference.

To anticipate some questions:

• In terms of my circuit, it works as expected independently, 9v battery power source
• 9v power source and circuit input source etc, ground's are all common to guitars ground
• Circuit does have DC noise filtering and is not in and of itself 'noisy'
• Guitar uses passive pickups, not active pickups - 500k pots.
• guitar is fully shielded, noiseless in its own right, HSH pickup configuration (with awesome custom switching  )
• Audible distortion/humming etc is removed if grounding the circuit's signal input, when switched on
• No distortion or interference when circuit is turned off
• no distortion, or loss of signal etc if circuit input source is substituted, eg from piezoelectric or signal generator

What I would like help with, if possible - is splitting and isolating the guitar signal to prevent cross talk.
Any solution needs to be able to fit in an already over-crowded guitar cavity, So a big chunky transformer wont work.
It would also need to ensure the same frequency response, ideally from 50h to ~1.6khz
it would also, preferably involve passive components and not "another preamp" or filter, or otherwise turn the pickups to active. - That being said, I would still like to know!

In other terms, how can one SPLIT and ISOLATE an AC signal?
is there a level of impedance matching or system that must be in place for normal operation of the amp etc, and if so how do you do it? - I don't have an LCR or way to measure impedance.

I have done some reading on opto-couplers / transformers etc but not sure how best to proceed. My alternative would be to use an isolated and separate input. that is a Piezoelectric sensor/element but having just tested that today, it does not have the frequency response/power required to drive my circuit - although I only tested one.

I would appreciate any help with solving this one.
Scott

[Scott D]

Sorry, pics were meant to be clickable to larger view.

It's not an effects circuit, and must be inside the guitar for this application.

[capt bullshot]

Your circuit has a quite low input impedance (1k / R20). Try adding a unity gain / high input impedance buffer - just another Op-Amp with input signal on "+" and "-" tied out its output node.

Is the 9V battery supplying this circuit only? So consider moving the reference ground to your AGND (virtual ground) node, you can get rid of a lot of AC coupling capacitors then.

[Audioguru again]

Your filters pass only a very narrow band of low frequencies, like truck horns.
The 0.1uF input capacitor feeding the extremely low value of the 1k input resistor cuts frequencies below 1600Hz.
Here is a simulation of your Multiple-feedback-Bandpass filters:

[Scott D]

Hey Capt Bullshot,

I did forget in my original post - thanks for reminding me. I did try a unity buffer as you suggested (LM358 - which may not be suitable). That being said, it was on a breadboard, but it didnt have any effect. I built the circuit originally with consideration that I may have to have this additional buffer in place and it would resolve all issues. 
9v is indeed for the circuit only.

Correct me if I am wrong, but if I was to make my ground reference to 50% of V+, then my AC can only swing within 25% of V+. I would get clipping?!
eg 9v input, ground reference at 4.5v, then the AC swing can be between 4.5v and 9v.

R20 @ 1k, and R16 @ 1M gives me a nice hefty gain which upon testing was ideal. What is the relationship with the 1k resistor and impedance? ie reducing the value/gain of the op-amp increase or decrease impedance? I do not know how to measure or calculate it. I had thought that the op-amp / circuit would be high impedance, or at least within similar to a guitar amp

[Scott D]

Your filters pass only a very narrow band of low frequencies, like truck horns.
The 0.1uF input capacitor feeding the extremely low value of the 1k input resistor cuts frequencies below 1600Hz.
Here is a simulation of your Multiple-feedback-Bandpass filters:

Thats correct. I needed this to be a high Q multiple feedback bandpass filter.
Cool simulation graph 

[Audioguru again]

You said your guitar has a "passive" pickup with 500k pots. Is the pickup a magnetic or is it a piezo?
The extremely low input resistance of 1k (R20) of your circuit will kill the output level of the pickup which might be why your voltage gain is so high at 1000 times. A guitar amp input resistance is 1M or more.

Your 1st opamp is an inverting type with an input resistance of 1k ohms (R20). If it is changed to be a non-inverting type that has a 2.2M resistor to ground and a capacitor to ground in series with R20 then its input resistance will be 2.2M and the input capacitor can be 0.022uF producing an output flat down to 16Hz. Then this opamp can be the high input resistance buffer needing a fairly low gain.

I forgot to tell you before that the resistors to ground at the outputs of the filters are way way too low which will kill the output levels and level of low frequencies. The datasheet of your opamps shows a minimum load of 2k ohms but you have 69.8 ohms, 169 ohms and 1k ohms killing the output levels and killing the level of frequencies below 488Hz, 201Hz and 34Hz for the filters.
If the loads are 10k ohms then the output level of the opamps will be almost 0V to almost +9V. Then the max output will be about -4.4v to +4.4V if the 9V battery is new.

Why do you have diodes in series with the outputs? They rectify the signal and create extreme distortion.

You said you are "splitting" the outputs. A splitter connects two amplifier inputs to one output of your circuit. The amplifier input resistance is probably 1M or more which will have no effect on the levels.
You said you want to "isolate" something. What?

Again, why do your filters pass only three frequencies like truck horns?

[Scott D]

You said your guitar has a "passive" pickup with 500k pots. Is the pickup a magnetic or is it a piezo?
....

Why do you have diodes in series with the outputs? They rectify the signal and create extreme distortion.

You said you are "splitting" the outputs. A splitter connects two amplifier inputs to one output of your circuit. The amplifier input resistance is probably 1M or more which will have no effect on the levels.
You said you want to "isolate" something. What?

Again, why do your filters pass only three frequencies like truck horns?

The pickups are standard magnetic passive pickups. HSH configuration, Vol 500k, tone 500k
The diodes you are seeing on the end of the schematic picture are the first part of rectifiers. The final output is actually DC.

I have two amps I have tested this with, a large 80w Marshall valve, and a small 20w generic portable. Each have the same problem with cross-talk/hum/noise etc (depending on gain/levels and if I use a resistor in series or not).

Isolation I want, perhaps incorrect nomenclature is:

                     Guitar signal out:
                    /                       \
             Circuit                    Guitar amp

I want both to have the same input signal from the pickups, but not be able to 'drive each other' (again nomenclature?)

Truck horns (love it), the input gain of the first amp is 1000, This is for having a desired level of response from the player, such as grating the pick along the strings, tapping, soft strumming etc. As you know each frequency has a different amplitude etc. So compared to the simulation posted earlier, its actually closer to ~3-5Vpp on that output, which then is reduced again at the next stage etc.
In any case, the high Q is needed as I have a specific requirement for a narrow filter around those frequencies - its DC not audio.
The 69.8, 169, 1k resistors are more or less 2nd order filtering - which again is by design.
 
I dont want to get bogged down too much about the rest of the circuit/filtering and why etc, but if it does relate to the problem im trying to solve - then I am not understanding how?
It sounds like I need to re-re-re-read the datasheet, are you saying that I am running the op-amp out of spec? And if so is that going to cause "serious problems" or is it just, poor practice?

So, here is a question based on what you are saying about the impedance:

Assuming that I need an input resistance of 1M or more, then if I were to swap the values from 1k / 1M to 1M / 10M respectively, - would that then satisfy the impedance requirement? That is that both amp's on each side would be at a level that is expected? (or is there a gain/impedance relationship that exists?!)
OR leaving the values as is, and putting a 1M resistor in series with the input, would that also do the trick?

[Scott D]

Ahh, hang on 1M / 10M drops the grain to 10. Don't want that. But still,
Is the requirement to get the impedance I need - contingent on the value of R20 being 1Mohm or higher?

[Scott D]

Further to my previous questions earlier;

With reference to http://whirlwindusa.com/support/tech-articles/high-and-low-impedance-signals/
I am trying to answer this question myself - generally I see that the rule of thumb are

1. Guitars output impedance is typically 7-15k or higher for passive guitar pickups. I have read upwards around 50k from other references though.
2. You always want the amp input impedance to be higher, typically 10-100x higher is the best fit.

If I assume that I have the typical guitar pickup ~7-15k, at 7000/2 = 3.5K
so linking that to what was said earlier about my circuit, R20 being too low, I have an input impedance far too low even at 3.5k.
However, if I had even a 10k resistor in R20, that would 2.8x higher which would in theory improve the situation. 100k again, and 1MB even better.

Am I on the right track here?

[capt bullshot]

Yes, increasing R20 is the right track (kind of). To keep the ridiculously high gain (1000) you'd have to increase the feedback resistor to keep their ratio the same. So if you set R20 to 1M, the feedback resistor would be 1G (that won't work in practice).
The better way is to re-wire the first OpAmp into a non-inverting configuration, that way you can keep the 1k / 1M to set the gain and have high impedance (set by a single resistor to GND). Audioguru said the same above.
If your input is high impedance, there's no need for signal splitting / isolation, low signal output impedance vs. high input impedance allows for parallel connection of the inputs without circuits "talking back" to the source.

[Scott D]

Yes, increasing R20 is the right track (kind of). To keep the ridiculously high gain (1000) you'd have to increase the feedback resistor to keep their ratio the same. So if you set R20 to 1M, the feedback resistor would be 1G (that won't work in practice).
The better way is to re-wire the first OpAmp into a non-inverting configuration, that way you can keep the 1k / 1M to set the gain and have high impedance (set by a single resistor to GND). Audioguru said the same above.
If your input is high impedance, there's no need for signal splitting / isolation, low signal output impedance vs. high input impedance allows for parallel connection of the inputs without circuits "talking back" to the source.

unfortunately, the PCBs are already fab'd. Yeah a 1Gohm isnt going to work. Using 0402 SMD too. So I cant change the design without a full re-design and re-fab, but I can easily change the values.
On hand I have 1k, 10k, 15k, 100k, 1M, 10M in 0402, So I am thinking 10k/10M as a test, and then 15k/10M thereafter and hope that solves it.
I can use a potentiometer in series with the input and measure the resistance to work out the min required value.

I may be able to get away with a lesser gain for my purposes, but its a matter of testing, as I could perhaps also increase the gain at the next stages if needed. Would be a bit of testing involved though.

I will be able to test this hopefully tomorrow; is there anything else I should consider/look for if I continue to have issues with this? - Or will it be a re-design hereafter? 

[Audioguru again]

Don't you agree that the input resistance of an electric guitar amplifier is 1M or higher? Isn't that why the guitar's pots are as high as 500K? Then that is why you must re-design the 1st opamp to have an input resistance of 1M which is 1000 times higher than you have now. Your 1k ohms value of R20 is shorting away the output level of the guitar pickup which is why the gain of your 1st opamp needs to be as high as 1M/1k= 1000 times.

The best way to increase the input resistance 1000 times is to change the 1st opamp from its existing inverting type to a non-inverting type:
A 22nF input coupling capacitor feeds into the (+) input pin 3 of the opamp that has a 2.2M resistor to ground (its input resistance) and a voltage divider of two feedback resistors from the output to pin 2 sets the gain. The resistor from pin 2 needs a capacitor in series with it to ground. Then the gain can be much less but the circuit will have lots of output level if the load is increased from its existing extremely low resistor values to about 10k.

Without a re-design you can try changing R20 to 2.2M and changing the feedback resistor to 4.7M to 22M.

What will you use the rectified "truck horns" DC to do? Light up LEDs? 

[Scott D]

Don't you agree that the input resistance of an electric guitar amplifier is 1M or higher? Isn't that why the guitar's pots are as high as 500K? Then that is why you must re-design the 1st opamp to have an input resistance of 1M which is 1000 times higher than you have now. Your 1k ohms value of R20 is shorting away the output level of the guitar pickup which is why the gain of your 1st opamp needs to be as high as 1M/1k= 1000 times.

The best way to increase the input resistance 1000 times is to change the 1st opamp from its existing inverting type to a non-inverting type:
A 22nF input coupling capacitor feeds into the (+) input pin 3 of the opamp that has a 2.2M resistor to ground (its input resistance) and a voltage divider of two feedback resistors from the output to pin 2 sets the gain. The resistor from pin 2 needs a capacitor in series with it to ground. Then the gain can be much less but the circuit will have lots of output level if the load is increased from its existing extremely low resistor values to about 10k.

Without a re-design you can try changing R20 to 2.2M and changing the feedback resistor to 4.7M to 22M.

What will you use the rectified "truck horns" DC to do? Light up LEDs?

What you suggest makes perfect sense, I am also having a bit of a lightbulb moment with respect to the relationship with the output power being a product of load/vs gain (right?)
And you are right, I started with a gain of 10-100 which worked, but I found I wanted a bit 'more'.
I committed myself to the inverting design quite a while back.. so long I cant recall what the deciding factor was, perhaps it worked out to a lower part count, better or easier filtering, or it .. it worked.
I did reference a few amplifier designs etc but obviously there is a lot to consider.

so noted for Version 2.0
Thanks Audioguru

[radiolistener]

I have designed and built an analog circuit that will be housed inside a guitar

that's fail, electronic components, especially capacitors will be affected by guitar vibrations and will make distortions. Some times ago Dave published a video, where oscilloscope shows terrible pulses when you knock on his body. The same thing will happens with your circuit inside guitar.

So, if you're want to get best sound quality, you're needs to move your analog circuit far away from guitar.

[Scott D]

I have designed and built an analog circuit that will be housed inside a guitar

that's fail, electronic components, especially capacitors will be affected by guitar vibrations and will make distortions. Some times ago Dave published a video, where oscilloscope shows terrible pulses when you knock on his body. The same thing will happens with your circuit inside guitar.

So, if you're want to get best sound quality, you're needs to move your analog circuit far away from guitar.

How significant was the noise? and was it at any specific frequency? It didnt seem to be an issue or significantly measurable when I tested with just the guitar.

[Audioguru again]

Electronic components, especially capacitors will be affected by guitar vibrations and will make distortions. Some times ago Dave published a video, where oscilloscope shows terrible pulses when you knock on his body. The same thing will happens with your circuit inside guitar.

So, if you're want to get best sound quality, you're needs to move your analog circuit far away from guitar.

A cheap ceramic capacitor is microphonic and does what you say. Audio circuits should use film, not ceramic coupling capacitors for low distortion.

[Buriedcode]

I have designed and built an analog circuit that will be housed inside a guitar

that's fail, electronic components, especially capacitors will be affected by guitar vibrations and will make distortions. Some times ago Dave published a video, where oscilloscope shows terrible pulses when you knock on his body. The same thing will happens with your circuit inside guitar.

So, if you're want to get best sound quality, you're needs to move your analog circuit far away from guitar.

Well that depends, but my response to that is - not really.  There are plenty of examples of circuits inside guitars - active pickup preamps, buffers, onboard distortions (fuzz factory I believe has been built into some guitars) and I haven't heard any reports of these making distortions - except the fuzz factory where it is actually required.

I wouldn't worry at all about it, if you have a very high gain amp in there, and it picks up vibrations, then there are ways to mitigate that.  Also, "guitar" and "best sound quality" don't really go together, even the cleanest guitar amp has oodles of harmonic distortion.  Into a HIFI amp guitar sounds pretty terrible.

[Scott D]

Big believer in closing the loop - to that end, I can confirm that the problem was well identified and resolved by replacing the signal input resistor to 1Mohm.
Thank you everyone helping out, and explaining it.

 

[Audioguru again]

Big believer in closing the loop - to that end, I can confirm that the problem was well identified and resolved by replacing the signal input resistor to 1Mohm.

Then the guitar pickup is no longer loaded down and the gain of the 1st opamp is only 1 instead of 1000.
The 0.1uF input capacitor feeding the new 1M resistor value will pass earthquake frequencies down to 1.6Hz, the capacitor value can be changed to 0.01uF and still pass all audio frequencies. Use a film capacitor, not electrolytic and not ceramic.