Revisiting the Abductor debug...

Luke51411 · July 21, 2015, 09:39:39 PM

So I built Dead Astros abductor delay a while back and it was really cool... But I was getting a problem of excessive white noise. I built a second one and it managed to be even worse with some distortion on the echoes... I think I may have other issues with the second one... I'd like to at least get one of these working properly as it was promised to a friend. I think I isolated the problem to thi in and out buffer ic as I removed all the pt2399s and the noise was still quite prevalent. The 2399s do amplify the noise a bit especially in the long delay mode but they are not the source of the noise. I'm baffled by it as the signal path from in to out is very short when the pt2399s are cut out of the circuit... Where is this noise coming from? I've audio probed it and it gets noisy at pin 1 of IC5.
https://dl.dropboxusercontent.com/u/7464107/ABDUCTOR%20II%20DELAY.pdf
I suppose I need to get fresh voltages... Ugh so many pins... I'm baffled by this one.

snap · July 22, 2015, 02:45:52 AM

R4=R5=470k = inverting opamp resistor noise generator.

Luke51411 · July 22, 2015, 09:03:47 AM

Quote from: snap on July 22, 2015, 02:45:52 AM
R4=R5=470k = inverting opamp resistor noise generator.

That's what he has in the build doc and others have been built without this issue... What values should they be?

Luke51411 · July 22, 2015, 09:24:30 AM

I've looked at a few different pt2399 delay schematics and most of the others have a small cap across r5 should I do something like this? They also have different values for these two resistors... I guess I can try a few things since it seems like this is where the noise could be coming from.
This one has 470k for r4 and 1M for r5 (relative to the abductor) http://www.madbeanpedals.com/projects/ZeroPoint/ZeroPointDD.pdf
This one has both as 220K with a 10pf across r5 http://www.madbeanpedals.com/projects/ZeroPoint/ZeroPointMini.pdf

samhay · July 22, 2015, 09:49:46 AM

Luke - do some reading about inverting op-amp stages. There are 3 things to consider: gain, input impedance, and frequency response.

Luke51411 · July 22, 2015, 10:18:40 AM

Quote from: samhay on July 22, 2015, 09:49:46 AM
Luke - do some reading about inverting op-amp stages. There are 3 things to consider: gain, input impedance, and frequency response.

Good idea. I always learn more reading about something when I have a specific question to ask.

Luke51411 · July 23, 2015, 08:37:01 AM

Quote from: snap on July 22, 2015, 02:45:52 AM
R4=R5=470k = inverting opamp resistor noise generator.

Snap, could you elaborate on this?

Samhay, I've been reading some general info on opamps on the google and starting to understand a little bit. Is there a suggested source for this info? The problem I came across is there is so much info I wasn't quite sure where to start digging in. In the mean time I'll keep reading whatever I find.

Luke51411 · July 23, 2015, 09:32:28 AM

Ok I found the AMZ article on buffers http://www.muzique.com/lab/buffers.htm and that is pretty helpful in understanding what is going on. I'm still not sure how it may relate to my noise issue.

samhay · July 23, 2015, 10:16:02 AM

I started writing this a while ago:

You will get thermal noise (hiss) from a resistor. The larger the resistance, the larger the hiss. If you want more detail look for Johnson–Nyquist noise.

The gain of an inverting op-amp (input fed to the (-) input is the ratio of the feedback resistor ( (-) input to output) and input resistance. In the case of the Abductor, the gain = 1 (470k/470k).
Rob used quite large resistors here as the input impedance of an inverting op-amp is determined by the input resistor(s) and we generally want high input impedance (100's k) for guitar stompboxes - here the input impedance = 470k.
As a result, we have a conundrum - we can lower the thermal noise by decreasing the value or R4 and R5, but if we do, we lower the input impedance, which will load a guitar pickup, causing you to lose treble content.
If you will always have a buffered effect before the Abductor, then you can lower the values of R4 and R5 to lower the thermal noise. However, you will have to look at the value of the input capacitor as this forms a high pass filter with R4.

Luke51411 · July 23, 2015, 11:40:54 AM

Quote from: samhay on July 23, 2015, 10:16:02 AM
I started writing this a while ago:

You will get thermal noise (hiss) from a resistor. The larger the resistance, the larger the hiss. If you want more detail look for Johnson–Nyquist noise.

The gain of an inverting op-amp (input fed to the (-) input is the ratio of the feedback resistor ( (-) input to output) and input resistance. In the case of the Abductor, the gain = 1 (470k/470k).
Rob used quite large resistors here as the input impedance of an inverting op-amp is determined by the input resistor(s) and we generally want high input impedance (100's k) for guitar stompboxes - here the input impedance = 470k.
As a result, we have a conundrum - we can lower the thermal noise by decreasing the value or R4 and R5, but if we do, we lower the input impedance, which will load a guitar pickup, causing you to lose treble content.
If you will always have a buffered effect before the Abductor, then you can lower the values of R4 and R5 to lower the thermal noise. However, you will have to look at the value of the input capacitor as this forms a high pass filter with R4.

Thanks for the explanation, I'm starting to understand some of the principles you explained but you really tied it all together in an easily understandable way. Would 220k be a good compromise for resistor value? I'm assuming it would be advised to use metal film resistors as well.

samhay · July 23, 2015, 02:29:22 PM

Metal film resistors should be quieter than carbon film. For reference, this is one take on resistor noise: http://www.aikenamps.com/index.php/resistor-types-does-it-matter

As Rob and others don't seem to have an issue with noise in this circuit (I haven't built it), I would start by replacing R4 and R5 with 470k metal film resistors. If that doesn't reduce the noise to your satisfaction, then reduce their values until you find a reasonable compromise.

Luke51411 · July 23, 2015, 02:57:51 PM

Quote from: samhay on July 23, 2015, 02:29:22 PM
Metal film resistors should be quieter than carbon film. For reference, this is one take on resistor noise: http://www.aikenamps.com/index.php/resistor-types-does-it-matter

As Rob and others don't seem to have an issue with noise in this circuit (I haven't built it), I would start by replacing R4 and R5 with 470k metal film resistors. If that doesn't reduce the noise to your satisfaction, then reduce their values until you find a reasonable compromise.

Ok, I'll try it and see how it goes. I still think I have another issue somewhere... I suppose I should try another power supply just in case but I haven't had a problem using it with other pedals.

PRR · July 23, 2015, 04:28:36 PM

> thermal noise (hiss) from a resistor. The larger the resistance, the larger the hiss.

Over-simplification.

The thermal hiss POWER is the same for any value resistance.

A high-value resistor has larger hiss VOLTAGE, but low hiss CURRENT.

A low-value resistor has larger hiss CURRENT, but low hiss VOLTAGE.

Which matters most depends on the circuit.

Yes, a simple opamp inverter with 470K is liable to output a lot of hiss. (Which is the immediate point you were addressing.)

However a follower with a 470K input resistor may NOT cause much hiss, because the low hiss Current is often shunted by other audio impedances.
__________________

There is also "excess noise" in poor resistors. Carbon COMPosition is pretty bad (yet we used them widely for many decades). Carbon Film is very-nearly ideal. Metal Film is a little more ideal, but hardly any places in audio would show a difference. (Especially around any quantized "delay" {BBD, digi-delay} which typically have "hiss" (quantization noise) far higher than mediocre linear stages.

samhay · July 23, 2015, 06:01:39 PM

>Over-simplification.
Agreed. Thanks for clarifying.

Luke. Might be a good idea to audio probe the output of the input buffer. This will give you an idea of whether the noise is coming from here or elsewhere.

snap · July 23, 2015, 06:24:53 PM

That`s what he did (pin1 of IC5). Maybe "deadastronout", the creator of the circuit can clear things uo?

Luke51411 · July 23, 2015, 09:16:08 PM

Quote from: snap on July 23, 2015, 06:24:53 PM
That`s what he did (pin1 of IC5). Maybe "deadastronout", the creator of the circuit can clear things uo?

I chatted with DA a little bit when I first built it and I couldn't figure anything out then. I looked at the 470k resistors and they looked like tayda resistors so I swapped them out for xicon but I don't have metal film 470k at any rate it didn't fix the noise problem. Subjectively it may have gotten a little bit better. I removed the voltage regulator so as to deactivate the pt2399 and they are definitely contributing to the noise or amplifying it but there is still a nice bed of static... I can kind of hear a hum that seems like it goes slightly up and down in pitch when I turn the volume control up pretty high.

samhay · July 24, 2015, 04:20:58 AM

What else do you have in your signal chain, and what type of amp are you using?

deadastronaut · July 24, 2015, 04:57:36 AM

I cant get to my bench right now, but when i do i,ll retest it ok...

john danna , jdansti, built one recently, itll be interesting to see if he has the same issue too...

back later guys.

anotherjim · July 24, 2015, 07:00:54 AM

My two pence...
I doubt resistor noise is the problem.
I question the position of D1 - I think it should be before the 100uF input capacitor across the power supply. As it is, D1 is preventing the regulator from properly benefiting from the input capacitor as it only sees it on one half cycle (yes, it's DC here BUT the current has an AC component due to the PT2399 clocks, signal content and other fluctuations).

Temporarily short D1 out and see if that helps. Get some 100nF ceramic caps and temporarily solder them on at...
7805 In to Ground
7805 Out to Ground
Across C3 (4.5V to Ground).

The fact that one builder has a problem and others don't might be down to the power supply they have or length/quality of the cable from their power supply.

Luke51411 · July 24, 2015, 09:30:58 AM

Thanks everyone. My power supply is a Godlyke Wall wart. I've used it for a lot of other pedals including up to 8-10 in the chain without any issues. I need to get a hold of another supply just to verify. I'm running it in to a Gibson Falcon with nothing else in the chain. The regulator gets quite warm but as I mentioned, the bed of static is still there when the regulator and/or pt2399s are removed. I'm going to double check all the values more closely for everything. I haven't really done that yet as the problem seemed to be isolated to the input/mixing section and I've verified all the values in that area.

News:

Revisiting the Abductor debug...