Reduce white noise in output signals? (synth op-amp circuit)

Pulling out your hair? Don't know what to do or where to go? Ask in here.
Forum rules
READ: VSE Board-Wide Rules and Guidelines

If your Help request has been solved, please edit your first post in order to select the Image Topic Icon to let others know your topic has been solved.
Post Reply
db0451
Junior Member
Junior Member
Posts: 147
Joined: Thu May 08, 2014 12:06 am

Reduce white noise in output signals? (synth op-amp circuit)

Post by db0451 » Sun Dec 28, 2014 2:03 pm

Not going to go over the background (again), but I have at least one synth where the post-DAC circuitry generates an absolute tonne of white noise/hiss, which is worsened because it then mixes a bunch of the multiplexed and noise-addled signals together. The fully mixed output is bearable, just about, but it makes me wish for proper SNR.

I replaced a representative sample of old op-amps that are apparently known to be noisy and bad (not going to quote their names and possibly contribute to that idea because:), but there was literally almost zero audible difference, with the noise seeming just as bad as before.

Thus, I must assume something else in the circuit is generating the hiss. But what? I have done a little bit of background reading, but specific things like this are hard to find good info on, or the 1 page of 300 per textbook.

Now, I found explanations that higher resistances promote thermal/Johnson noise and might be best replaced.This link is good. However, I’m really looking for practical comments on this theory. I see in the circuits here a LOT of 10 kOhm resistors for weighting/summing/dividing, as both discrete resistors and arrays. Assuming I determine that the op-amps (or fitting replacements) support the new impedance, is replacing as many possible 10 kOhm resistors with lower-value ones likely to reduce the noise?

Or is there a much more plausible solution that I have not yet stumbled into?

I suppose a main concern of mine is that perhaps the DAC itself is inherently a bit hissy (I mean constantly, including silence, not quantisation noise) and that will always be an issue when multiple outputs are summed. But I’m not sure that’s how noise-mixing works, anyway. ;)

All ideas will be appreciated. Thanks!

User avatar
meatballfulton
Moderator
Moderator
Posts: 6099
Joined: Wed Apr 13, 2005 9:29 pm
Gear: Logic Pro X

Re: Reduce white noise in output signals? (synth op-amp circ

Post by meatballfulton » Sun Dec 28, 2014 5:00 pm

Got an oscilloscope?

With the synth running idle, check the DAC output and see how noisy it is. The compare it after the op amps.

If multiple sources are being summed, check each source individually for noise levels.

Changing the resistor values is not a good idea unless you really understand the circuit. It is true that higher impedances are more prone to noise but that's more of an issue with circuits like inputs to preamps.
I listened to Hatfield and the North at Rainbow. They were very wonderful and they made my heart a prisoner.

User avatar
Stab Frenzy
Moderator
Moderator
Posts: 9723
Joined: Tue Jun 06, 2006 5:41 pm
Gear: Eurorack, RYTM, Ultranova, many FX
Location: monster island*
Contact:

Re: Reduce white noise in output signals? (synth op-amp circ

Post by Stab Frenzy » Mon Dec 29, 2014 2:41 am

Could be the power supply as well, that's what I'd check first. Noise on the power rails is gonna come through everywhere.

db0451
Junior Member
Junior Member
Posts: 147
Joined: Thu May 08, 2014 12:06 am

Re: Reduce white noise in output signals? (synth op-amp circ

Post by db0451 » Mon Dec 29, 2014 2:48 am

Thanks both.
meatballfulton wrote:Got an oscilloscope?
Sadly not. Much needed, though.
If multiple sources are being summed, check each source individually for noise levels.
I can do this through the separate pre-sum outputs. They are not ideal noise-wise but obviously much less so. The problem then is that I have to sum them, whether through the mixed outputs or a separate mixer, to get any useful polyphony!
Changing the resistor values is not a good idea unless you really understand the circuit. It is true that higher impedances are more prone to noise but that's more of an issue with circuits like inputs to preamps.
Yeah, I figured there would be a risk of throwing off the rest of the output circuits. One of my forays into op-amp replacement led to a meandering high-pitched oscillation, though I think that might be due to a busted solder pad that has not been bridged properly, rather than mismatched components. Anyway, the original types of op-amp will be going back once I finally get them from HK. Since the PCB is currently a bit messed-up anyway, I might try replacing the 10 kOhm resistors on the affected outputs. It probably won’t get worse ;)
Stab Frenzy wrote:Could be the power supply as well, that's what I'd check first. Noise on the power rails is gonna come through everywhere.
Yeah, but would that not centre at 50 Hz or some harmonic thereof? This is pure white noise. I never found a rationale for the PSU introducing white noise, but I might simply have missed that explanation. Anyway, I replaced all the electrolytics with new (and physically much smaller!) ones, as well as some elsewhere on the PCB (including around the DAC), and it made precisely no difference. Useful exercise, though.

User avatar
Stab Frenzy
Moderator
Moderator
Posts: 9723
Joined: Tue Jun 06, 2006 5:41 pm
Gear: Eurorack, RYTM, Ultranova, many FX
Location: monster island*
Contact:

Re: Reduce white noise in output signals? (synth op-amp circ

Post by Stab Frenzy » Mon Dec 29, 2014 7:54 am

If the power supply filtering caps aren't doing their job then the result will be white noise, not an earth hum.

db0451
Junior Member
Junior Member
Posts: 147
Joined: Thu May 08, 2014 12:06 am

Re: Reduce white noise in output signals? (synth op-amp circ

Post by db0451 » Mon Dec 29, 2014 2:16 pm

OK, thanks, but I assume they are fine as they have all been replaced with modern equivalents. Also, I previously had a second unit that had the same noise levels (within about 0.4 dB), so that probably indicates a more general problem elsewhere?

User avatar
madtheory
Supporting Member!
Supporting Member!
Posts: 5399
Joined: Mon May 01, 2006 12:45 pm
Real name: Tomas Mulcahy
Gear: Wurlitzer Opus 1536, Model F, Morovdis Arpeggiator, Maplin My First EQ, Jeff Wayne Thunderchild rack, Thermostat, Buck Owens' Moog.
Location: Cork, Ireland
Contact:

Re: Reduce white noise in output signals? (synth op-amp circ

Post by madtheory » Mon Dec 29, 2014 3:08 pm

Stab Frenzy wrote:If the power supply filtering caps aren't doing their job then the result will be white noise, not an earth hum.
I don't think this is correct? How do they contribute to white noise? These are designed to reduce AC ripple, which would manifest as a hum in the audio outs. Now this hum can be anything from a loud buzz to a very quiet hum, depending on level and harmonics. So "hum" is a pretty vague term!

White noise is Johnson noise, which is always caused by resistors. The trick is to have the smallest possible number of resistors at each gain stage. But it's pretty much impossible to change those design decisions without ending up building a new circuit. If a synth designer couldn't get that right then it's unlikely to be much good as a synth because they probably mucked up other stuff too. The only way to retrospectively improve noise performance is to replace op amps and transistors with modern equivalents that are genuinely equivalent. Even then you'll probably still have to replace capacitors to comply with the modern design. That's why you had the oscillations.

Douglas Self "Small signal amplifier design" will explain all of this stuff really well.

Anyway, what synths are you working on? I've never found it to be a problem in practice (i.e. a mix) with the DX-7, D-50, CZ-101 etc. all of which I've heard people complain about over the years. The Casio VZ-1 has a ganky noise gate circuit because the DA is rubbish by modern standards. Ya, they hiss when you listen closely in isolation, or at high gain (higher than you'd set it for normal use) so I think maybe the problem is not as bad as you might feel it is? Hiss is easily masked in a mix.

Post Reply