SSL style parametric eq for use with guitar - too noisy?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

birdman

Member
Joined
Jul 18, 2017
Messages
5
Hi all,

I'm playing around with the idea of a parametric eq pedal primarily used with a guitar rig live, so it would typically see high source impedances and be connected to high gain devices either before or after, like a distortion pedal or high gain amp.

I've built one based on the SSL 82E02 schematic (which is a Wein bridge eq I believe?..) and the eq itself is functioning well. When I have it plugged into a clean amplifier it sounds great with either a guitar or bass connected. A couple problems appear though when followed by a distortion pedal or plugged into a high gain amp.

The first issue is just overall noise level. It's quite a bit noisier than I was expecting throughout pretty much the entire frequency spectrum. I've mostly ruled out the power supply being the main contributor. I've run it on various voltages (+/- 9V, 12V, 15V) using linear regulators and charge pumps, and also my bench supply which is pretty clean. The noise doesn't change much. It appears as if the circuit itself just generates a fair bit of noise. Again, really only a noticeable issue when adding heaps of gain after it.

The second issue is that the frequency and Q pots make quite a bit of scratchy noise when adjusted. It's not really a problem when adjusting the eq while audio is running through it, and again not noticeable through a clean amplifier, but with a high gain amp it's very noticeable when adjusting the controls and not actively playing.

So I guess my questions for anyone familiar with the SSL eq's are, is it to be expected that it wouldn't work very well in my application? I've always assumed they were fairly low noise but I guess acceptable noise level depends on the application (console use vs with a high gain guitar amp). Are there other topologies that are lower noise (SVF maybe?) that might be better suited for guitar?

Anything I can do to improve the scratchy noise when adjusting the frequency and q pots, or is this just inherit to the design? I believe this is caused by changing DC operating levels but I feel like the design probably also depends on that happening..

Appreciate any advice!
 
If you're using standard carbon pots for frequency and bandwidth, try conductive plastic - they can handle some DC. The W492 project requires them, for example (though that's a SVF).

I'm not familiar with that particular SSL EQ, but if it's designed for standard line levels then unless you've modified the circuit your noise issue *might* be impedance-related. Are you stepping down or buffering the instrument output before it hits the EQ? Is the EQ output free of DC?

Fun project!
 
Is the guitar loaded with a high impedance? If not try a buffer before the EQ and see if it changes anything. That design and NE5534s in general are better suited for lower impedance.

You should also check the noise floor of the EQ. You can hook it up to a DAW and then see how the noise level changes with EQ engaged.
 
The noise might be due to the fact that console EQs are designed to work with "line level" signals....let's say approx. 1 Volt RMS. Guitar signals typically will be much lower.

This may be worth testing. Add a gain stage (with a high input impedance) before the EQ to boost the signal level to "line level" and then pad the level at the output of the EQ so that the gain through the entire system is unity with the EQ set flat.

Bri
 
scratchy pots suggest you have DC across them... I don't know much about SSL parametric but I have designed multiple PEQs over the decades.

JR
Hi John, thanks for replying. I've learned a ton from reading so many of your knowledgeable posts on here.

DC across the pots was my thinking too, and of course my first step should be to measure and see if I can spot the shifting levels. I think I've just assumed it was unavoidable as I've since found a couple posts referencing this scratchiness as inherent to the Wein bridge topology. Actually, one of them was from you in a thread from 2006! Wein Bridge (fourth post down)

I've attached the schematic I'm working from here which as far as I can tell is the real deal SSL schematic. I'm using the 4 bands exactly as shown here with some very small exceptions (50k dual rev log freq pots instead of 47k) but essentially identical. Does it seem like some pot noise in the frequency and q pots might be an inherent trait of this circuit?
 

Attachments

  • ssl_82E02.pdf
    156.4 KB · Views: 63
I'm not familiar with that particular SSL EQ, but if it's designed for standard line levels then unless you've modified the circuit your noise issue *might* be impedance-related. Are you stepping down or buffering the instrument output before it hits the EQ? Is the EQ output free of DC?

Fun project!

Is the guitar loaded with a high impedance? If not try a buffer before the EQ and see if it changes anything. That design and NE5534s in general are better suited for lower impedance.

Hey guys, thanks for your replies. I am buffering the input before it hits the filters with a standard non-inverting op amp buffer. But the noise I'm referring to just the actual noise generated from the circuit itself as I'm measuring with the input grounded.

You should also check the noise floor of the EQ. You can hook it up to a DAW and then see how the noise level changes with EQ engaged.

The noise floor of my soundcard is measuring at about -105 dBFS and connecting the eq to it measures about -101 dbFS. I know that seems like maybe not much noise but when you add a bunch of gain afterwards it's really quite noticeable. I've compared to a couple other eq pedals I have that only increase the noise floor to about -104 dBFS when engaged, and in just listening tests plugged into an amp the difference is pretty substantial. My circuit isn't crazy noisy or anything like that, just noisier than would work for me.
 
The noise might be due to the fact that console EQs are designed to work with "line level" signals....let's say approx. 1 Volt RMS. Guitar signals typically will be much lower.

This may be worth testing. Add a gain stage (with a high input impedance) before the EQ to boost the signal level to "line level" and then pad the level at the output of the EQ so that the gain through the entire system is unity with the EQ set flat.

Bri
Thanks Brian, I hadn't considered that. I'll definitely try that out! I'm not sure it would make much difference to the noise I'm referring to though as I have the input grounded so really just listening to the noise generated by the circuit itself.
 
Thanks Brian, I hadn't considered that. I'll definitely try that out! I'm not sure it would make much difference to the noise I'm referring to though as I have the input grounded so really just listening to the noise generated by the circuit itself.
If you attenuate the output by the same amount you increase the input, the intrinsic noise of teh EQ will be attenuated as much.
Once DC has been eliminated from the potential list of suspects, "scratchy noise" in a Wien EQ is due to the fact that the impedance seen by the opamp varies, so it modulates the noise resulting from "input noise current".
You may find that replacing the 5532 (bipolar) with a OPA2134 (JFET)) reduces the scratchy noise, at the detriment of the idle noise.
 
Hi John, thanks for replying. I've learned a ton from reading so many of your knowledgeable posts on here.

DC across the pots was my thinking too, and of course my first step should be to measure and see if I can spot the shifting levels. I think I've just assumed it was unavoidable as I've since found a couple posts referencing this scratchiness as inherent to the Wein bridge topology. Actually, one of them was from you in a thread from 2006! Wein Bridge (fourth post down)
The scratchiness I encountered was due to wiper bounce, long story and unlikely to be your specific problem.
I've attached the schematic I'm working from here which as far as I can tell is the real deal SSL schematic. I'm using the 4 bands exactly as shown here with some very small exceptions (50k dual rev log freq pots instead of 47k) but essentially identical. Does it seem like some pot noise in the frequency and q pots might be an inherent trait of this circuit?
I don't know, are SSL PEQs scratchy?

Your original post refers to a pedal application. Does that infer a modest low voltage PS? Many pedals run from single 9V.

rather than reading a SSL schematic, how about a schematic of your project.

Abbey's suspicion that bipolar opamp (like 553x) input bias current could introduce DC to a high impedance circuit. Bifet opamps may nominally have higher noise voltage but lower bias and noise currents.

JR
 
The Frequency pots are included in the DC 'bias' path of the op amp so altering those WILL cause some noise while being adjusted.
The EQ is intended to be run at desk internal 'system' level and as such would normally have next to no gain applied afterwards by which time the EQ noise will be among the noise due to the summing stages. This is another case of 'borrowing' someones design and using it in the 'wrong context', like so many that grab modules from old mixing desks and fail to appreciate the originally designed operating levels. Changing op amps MIGHT improve the 'scratchiness' but then if the DC offsets of different op amps are worse than 5534 you may not get far with that either. The op amps will need to be able to 'drive' quite low 'load' impedances (some are quoted as 600 Ohm capable, as a 5534).
 
As I wrote earlier, DC offest currents are not the main cause of scratchiness. It is the noise current modulation that varying impedance generates.
That's interesting, noise current is by definition AC so varying pot resistance will modulate the amplitude of this AC noise. l would not expect this to sound scratchy unless the pot resistance changes are large discontinuous steps.

I encountered scratchy noise from an EQ pot with inadequate wiper design (long story). DC (voltage or current) imposed on a pot will cause the familiar scratchy sound as changing R causing changing V.

Caveat Lector... I am just speculating, others posting here have more hands on familiarity with the actual circuit in question.

JR
 
That's interesting, noise current is by definition AC so varying pot resistance will modulate the amplitude of this AC noise. l would not expect this to sound scratchy unless the pot resistance changes are large discontinuous steps.
I first noticed that when Soundcrafte chose to replace the TL072 with a 5532 in their post-fader amp.
Many customers complained about it. I did a few trials, reverted to TLO and submitted the result to customers. They unanimously said it cured the problem. The fact that the idle noise had slightly increased was generally not noticed. It resulted in an ECO.
I wished I had the OPA2134 at the time.
 
If you attenuate the output by the same amount you increase the input, the intrinsic noise of teh EQ will be attenuated as much.
Ah yes, right...duhhh! Thank you for pointing that out. I guess this will come at the expense of headroom but I'll give this a shot.

Once DC has been eliminated from the potential list of suspects, "scratchy noise" in a Wien EQ is due to the fact that the impedance seen by the opamp varies, so it modulates the noise resulting from "input noise current".
You may find that replacing the 5532 (bipolar) with a OPA2134 (JFET)) reduces the scratchy noise, at the detriment of the idle noise.
This is a huge help, thank you. Opamp noise is an area where I definitely need to improve my knowledge. Looking at the datasheets it's quite a massive jump in current noise specs when comparing the 5532 to any of the FET input opamps. I'm going to order some and do some experimenting. I'll report back what I find!

This is another case of 'borrowing' someones design and using it in the 'wrong context', like so many that grab modules from old mixing desks and fail to appreciate the originally designed operating levels.
Oh absolutely. Sure is fun to experiment though! Best way to learn.
 
Ah yes, right...duhhh! Thank you for pointing that out. I guess this will come at the expense of headroom but I'll give this a shot.


This is a huge help, thank you. Opamp noise is an area where I definitely need to improve my knowledge. Looking at the datasheets it's quite a massive jump in current noise specs when comparing the 5532 to any of the FET input opamps. I'm going to order some and do some experimenting. I'll report back what I find!


Oh absolutely. Sure is fun to experiment though! Best way to learn.
When evaluating op amp ein (noise) it is useful to consider the impedance the op amp input is seeing. Bipolar op amps generally trade lower EIN noise voltage for higher EIN noise current. BiFET op amps are the opposite, higher EIN Noise voltage with much lower EIN noise current. What this means in practice is that BiFET op amps can be quieter working in high impedance circuits, bipolar can be quieter with low impedance. The line dividing which is better where involves math.

JR
 
Hi all,

I'm playing around with the idea of a parametric eq pedal primarily used with a guitar rig live, so it would typically see high source impedances and be connected to high gain devices either before or after, like a distortion pedal or high gain amp.

I've built one based on the SSL 82E02 schematic (which is a Wein bridge eq I believe?..) and the eq itself is functioning well. When I have it plugged into a clean amplifier it sounds great with either a guitar or bass connected. A couple problems appear though when followed by a distortion pedal or plugged into a high gain amp.

The first issue is just overall noise level. It's quite a bit noisier than I was expecting throughout pretty much the entire frequency spectrum. I've mostly ruled out the power supply being the main contributor. I've run it on various voltages (+/- 9V, 12V, 15V) using linear regulators and charge pumps, and also my bench supply which is pretty clean. The noise doesn't change much. It appears as if the circuit itself just generates a fair bit of noise. Again, really only a noticeable issue when adding heaps of gain after it.

The second issue is that the frequency and Q pots make quite a bit of scratchy noise when adjusted. It's not really a problem when adjusting the eq while audio is running through it, and again not noticeable through a clean amplifier, but with a high gain amp it's very noticeable when adjusting the controls and not actively playing.

So I guess my questions for anyone familiar with the SSL eq's are, is it to be expected that it wouldn't work very well in my application? I've always assumed they were fairly low noise but I guess acceptable noise level depends on the application (console use vs with a high gain guitar amp). Are there other topologies that are lower noise (SVF maybe?) that might be better suited for guitar?

Anything I can do to improve the scratchy noise when adjusting the frequency and q pots, or is this just inherit to the design? I believe this is caused by changing DC operating levels but I feel like the design probably also depends on that happening..

Appreciate any advice!
You're right, if there's DC on a pot, it always crackles! Try to insert the EQ in the part with the highest possible signal level!
 
Taking any 'chunk' of circuitry from anywhere without considering HOW and WHY it was used originally is always a mistake. DC (bias curent on pots which are 'obviously moved) will cause scratching and indeed the changing impedances will alter noise levels too. Op amp manufacturers used to claim 'low noise' as a 'headline without saying current or voltage noise or specifying tthe conditions until you read and understood the graphs and applied them to the circuit YOU want. Of course for an EQ where you want wide sweeps of frequency you need high resistance pots BUT this will simultaneously give you variable noise figures which, depending on the rest of the EQ will result in variable noise outputs (possibly filtered or enhanced by other sections.
So, for a guitar EQ where the guitar (or suitable flat gain preamp to reduce the effect of a complicated (multi amplifier) EQ section is used) can then run into voltage headroom issues. After all an SSL Channel EQ dealt with signals that were pre formatted (restricted) practically all the time and the noise of one channel of EQ section neatly disappears in the overall mix noise. You have to decide where your compromises are going to be and design proper strategies to achieve what you want.
 
Back
Top