Fluctuating high-pitch noise in op-amp circuit

I've seen a very nice filtered power strip in the local ads, sold for very cheap; I'm gonna buy it and give it a try - not to "solve" anything but because this whole issue made me read a lot about noise and I'm eager to see if such a device actually brings some benefits.
Your power cord is an antenna. Power strip is too far away to kill RF. Your cell phone is RF. Most appliances have a hi-freq microprocessor in them, emitting a little RF. FCC compliance means they only emit a little RF, not a whole lot.
The IEC filter you talked about is a partial solution. I was looking at one in a computer monitor I'm re-e-capping. Has 1.2 microHenry on each of hot & neutral in, plus a .047 cap across. That is sleazy compared to the 11 turns on a toroid I was recommending at your equipment entrance. That is more like 10-14 microhenry. Look at any flat screenTV on the curb for the garbage pickup
- two such toroids are by the power cord entrance. (which is one place to get them) Also a blue MOS supressor to keep lightning & motor shut off spikes down to ~150 v, also a yellow type X2 cap across the line. That is serious entry projection against RF. Also keeps switcher supply in the TV from emitting RF out the power cord.
This device power transformer with iron E core can kill RF, but it is 3" from the outside wall. Too far, the traces to it emit RF. Also the power transformer doesn't have a steel jacket, like my dynaco and peavey equipment. Those power transformers are right by the power fuse, too. So with a 68 pf cap across the RCA entrance, and a 11 turn coil at the speaker jack, I don't get dogs barking the song "Dixie" when the yellow pickup truck drives by as I used to. CB radio sleazebags.
 
Last edited:
Your power cord is an antenna. Power strip is too far away to kill RF. Your cell phone is RF. Most appliances have a hi-freq microprocessor in them, emitting a little RF. FCC compliance means they only emit a little RF, not a whole lot.

Look at most digital data cords and DC supply cords... that lump is a Ferrite Choke intended to block RF. It needs to be close to the casings to work.

You can also add them to standard AC power cords safely since there is no electrical connection. They just clip on.

They do work and they are Plenty Cheap on Amazon
 
A proper RCA cable has it's shield attached to ground at both ends. And, to be honest... that could well be your problem.
...........................................
Exactly!
Plastic insulating rings around RCA chassis connectors are an invitation for noise and interference.
Jim Brown (past Audio Engineering Society co-chair of the EMI/RFI committee) often writes about this problem. And those bypass capacitors are more of a band-aid than a solution. The camel's nose is already in the tent.

Others that have also written on this problem include Neil Muncy(RIP), Keith Armstrong and Henry Ott.
 
Look at most digital data cords and DC supply cords... that lump is a Ferrite Choke intended to block RF. It needs to be close to the casings to work.

You can also add them to standard AC power cords safely since there is no electrical connection. They just clip on.

They do work and they are Plenty Cheap on Amazon

I have plenty of those ferrite clamps at home. I have installed them on my DAC IEC cord, this problematic x-over as well (can't say it helped much), and on my SPDIF coax cable.

I tried on the IC cable that goes from preamp to crossover and I can't say I heard a world of difference, I took them out because apparently it kills the dynamics of the sound (?), which I didn't really hear (but I was focused on noise).
 
I have plenty of those ferrite clamps at home. I have installed them on my DAC IEC cord, this problematic x-over as well (can't say it helped much), and on my SPDIF coax cable.

I tried on the IC cable that goes from preamp to crossover and I can't say I heard a world of difference, I took them out because apparently it kills the dynamics of the sound (?), which I didn't really hear (but I was focused on noise).

Ferrite beads snapped onto a shielded cable won't have any effect on sound. They generally don't have much effect below a couple of megahertz, up in radio land. What they do is block RF on the cables from getting into your devices.

Given that it appears you were having an RFI ingress problem, I'd suggest you leave them on.
 
Ferrite beads snapped onto a shielded cable won't have any effect on sound. They generally don't have much effect below a couple of megahertz, up in radio land. What they do is block RF on the cables from getting into your devices.

Given that it appears you were having an RFI ingress problem, I'd suggest you leave them on.

I will put them back and give it another try.

My whistling noise is still there; it is now much fainter since I put those small caps on the input jacks (before it would drive me out of the room, amplified by the horn tweeters it was just unbearable) but to my regret it is still there, sometimes. More often in the evening. I would really like to know what external source is creating this and blow it up.. haha
 
I will put them back and give it another try.

My whistling noise is still there; it is now much fainter since I put those small caps on the input jacks (before it would drive me out of the room, amplified by the horn tweeters it was just unbearable) but to my regret it is still there, sometimes. More often in the evening. I would really like to know what external source is creating this and blow it up.. haha

Most likely it's something with a transmitter in it... a phone, a router, a laptop or tablet ... but I've had problems from some pretty weird sources.

I had a buzzing problem in one of my amplifiers a year ago. Went high and low trying to improve grounding, used up half a bag of ferrite beads, it drove me crazy for about two weeks. Finally I just went around and one thing at a time unplugged everything in my apartment ... turns out it was the clock in a new stove causing the problem. A couple of caps on the AC line going into the clock module fixed it right up.

The big problem is that most companies now simply ignore RFI issues or don't even know about them. Since the death of CB and Ham radio as popular hobbies, there doesn't seem much reason. Well, except that with the new proliferation of short range radio devices --WiFi, Cordless, Cell, BlueTooth, etc.-- these problems will be coming back to haunt us.

Finding the cause is not easy and often requires very specialized equipment.

Input bypassing will help. Chip bypassing both on power pins and on input pins will help. but once you've got a device that is susceptible it's always going to be susceptible. The best you can hope for is to reduce it to a bearable level.
 
Most likely it's something with a transmitter in it... a phone, a router, a laptop or tablet ... but I've had problems from some pretty weird sources.

I had a buzzing problem in one of my amplifiers a year ago. Went high and low trying to improve grounding, used up half a bag of ferrite beads, it drove me crazy for about two weeks. Finally I just went around and one thing at a time unplugged everything in my apartment ... turns out it was the clock in a new stove causing the problem. A couple of caps on the AC line going into the clock module fixed it right up.

The big problem is that most companies now simply ignore RFI issues or don't even know about them. Since the death of CB and Ham radio as popular hobbies, there doesn't seem much reason. Well, except that with the new proliferation of short range radio devices --WiFi, Cordless, Cell, BlueTooth, etc.-- these problems will be coming back to haunt us.

Finding the cause is not easy and often requires very specialized equipment.

Input bypassing will help. Chip bypassing both on power pins and on input pins will help. but once you've got a device that is susceptible it's always going to be susceptible. The best you can hope for is to reduce it to a bearable level.

yesterday I tried to use some interconnect cable (VanDenHul) that has the shield connected only at one side, and as you predicted the noise was like 100x stronger; I might try to re-terminate it with the shield attached at both ends, in the meantime I'm using my cheap shielded cable (I don't wanna go into a pro/anti-cable debate, but the cheap one doesn't sound half as good)

You mentioned putting small caps from input to ground on each chip: could you please specify which type and value capacitor I should use? I suppose I can just put them under the PCB like I did with pSU bypass caps?

Thank you!
 
yesterday I tried to use some interconnect cable (VanDenHul) that has the shield connected only at one side, and as you predicted the noise was like 100x stronger; I might try to re-terminate it with the shield attached at both ends, in the meantime I'm using my cheap shielded cable (I don't wanna go into a pro/anti-cable debate, but the cheap one doesn't sound half as good)

You mentioned putting small caps from input to ground on each chip: could you please specify which type and value capacitor I should use? I suppose I can just put them under the PCB like I did with pSU bypass caps?

Thank you!

You would have to first know where the RF ingress is happening... Then add very small (33pf?) caps from the chip inputs to ground. If you don't know the ingress point you're going to end up using a ton of them, most of which do nothing.

I too do not wish to get into the cables debate. I lose friends every time I do... but I can tell you this much... expensive cables are not better.
 
You would have to first know where the RF ingress is happening... Then add very small (33pf?) caps from the chip inputs to ground. If you don't know the ingress point you're going to end up using a ton of them, most of which do nothing.

I too do not wish to get into the cables debate. I lose friends every time I do... but I can tell you this much... expensive cables are not better.

we won't get into the cable debate ;)

It seems to me the RF ingress comes trough line input (since using a differently shielded INPUT cable changes perceived noise); so I suppose adding those caps on the input buffer op-amp will help. No need to do the following chips I suppose?
 
we won't get into the cable debate ;)

It seems to me the RF ingress comes trough line input (since using a differently shielded INPUT cable changes perceived noise); so I suppose adding those caps on the input buffer op-amp will help. No need to do the following chips I suppose?

First I would try the easy fix ... RCA cables made from good quality microphone cable with proper soldered connections at both ends. (i.e. the standard cables)

Also run the thing with the covers on to ensure maximum sheilding. I think we already established the case is grounded but it can't help to chek it again.

Bypassing at the rca inputs probably has the input chips covered. but it is possible the RF is entering any of the chips. I would "touch test" the rest of them until I knew where the problem was by simply pressing the cap leads against the input pins of the chips...
 
You should be using coaxial cable for RCA interconnects.

IF you mean RF coax such as RG59 or RG56 ... no.

Most RF cable is actually very poorly shielded, 70% is typical.

Even dirt cheap microphone cable will be 80% or better shielding. Plus it will be softer, more easily soldered and the odds are it will stand up better.

So far the best RCA cables I've found have been the ordinary Amazon Basics cables that sell for like $10.00 each.
 
The coverage percentage is much less important than whether its braided or lapped screen - small holes don't leak RF for wavelengths significantly larger than the hole, they cannot radiate. Long thin gaps in a shield can act directly as an antenna, which is why lapped screen is never used for RF.


At audio frequencies noise is essentially capacitively coupled to the screen which has far too low an impedance to ground to produce a problematic voltage from the interference. The inner conductor thus doesn't see the interfering signal.


RF interference can get into a coax cable at a low level, but usually the connectors are the points of ingress of significant levels of interference - bad or patchy screen connection on an RCA socket is very common as the connector design is so poor (some expensive connectors manage to provide decent contact pressure on the shield, but standard ones are hopeless.


Also for RF you want the shield of the socket to bond to the chassis seamlessly. For avoiding ground loops in audio signals you don't necessarily want this. This allows easy ingress to an enclosure for RF in many bits of equipment. The ideal would be an RF feed-through capacitor arrangement for the socket's shield connection, keeping RF currents on the outside of the box.
 
The ideal would be an RF feed-through capacitor arrangement for the socket's shield connection, keeping RF currents on the outside of the box.

You can get very close to this with clip on ferrite beads on the cables themselves.

Our friend might find that multiple solutions are needed, depending how bad the RF design of the circuit is.

Of course the most probable solution would be to get it away from the source of the RF... either by moving the wireless router or moving the crossover away from the router.
 
Last edited:
Of course the most probable solution would be to get it away from the source of the RF... either by moving the wireless router or moving the crossover away from the router.
The OP lives 1 floor below an office with multiple computers running he said 24 hours.
The problem occured when he changed from 5532 ICs to OPA2132, going from 9 v/usec slew rate to 20. He has found it too difficult to find the feedback resistors between pins 6-7 and 2-1 to install parallel 33 or 39 pf disk capacitors. These capacitors reduce the RF gain of the fast op amps. I explained in post 61 how to identify pins 6-7 and 2-1 from the top.
I suggested he probe with a meter power off to find zero resistance between those pins and the ends of a nearby resistor. He had a problem with breaking the case of the disk capacitor when bending the leads. This requires a needle nose plier to hold the leads near the case when bending.
 
Last edited:
The OP lives 1 floor below an office with multiple computers running he said 24 hours.
The problem occured when he changed from 5532 ICs to OPA2132, going from 9 v/usec slew rate to 20. He has found it too difficult to find the feedback resistors between pins 6-7 and 2-1 to install parallel 33 or 39 pf disk capacitors. These capacitors reduce the RF gain of the fast op amps. I explained in post 61 how to identify pins 6-7 and 2-1 from the top.
I suggested he probe with a meter power off to find zero resistance between those pins and the ends of a nearby resistor. He had a problem with breaking the case of the disk capacitor when bending the leads. This requires a needle nose plier to hold the leads near the case when bending.

In that case the solution seems painfully simple ... Rather than putting in faster chips then limiting them with capacitors, just put the 5532 ics back in and call it a day.
 
Last edited:
The OP lives 1 floor below an office with multiple computers running he said 24 hours.
The problem occured when he changed from 5532 ICs to OPA2132, going from 9 v/usec slew rate to 20. He has found it too difficult to find the feedback resistors between pins 6-7 and 2-1 to install parallel 33 or 39 pf disk capacitors. These capacitors reduce the RF gain of the fast op amps. I explained in post 61 how to identify pins 6-7 and 2-1 from the top.
I suggested he probe with a meter power off to find zero resistance between those pins and the ends of a nearby resistor. He had a problem with breaking the case of the disk capacitor when bending the leads. This requires a needle nose plier to hold the leads near the case when bending.

Back from two weeks non stop work... let me correct you on one point: the fluctuating nature made me believe at first I didn't hear the noise when the original NE5532 were in there, so I put them back. but then I heard the noise again, loud and clear, so I put back the OPA2132 which, to my ears, sound better.
It seems what really influences the situation is the type of cable I'm using at the input.
 
I have same issue, Sound isn't coming out of my left channel. Used to be able to fix it by wiggling the plug going into the receiver from my turntable. I have a few extra RCA cables I can cut up for a spare male plug, can anyone guide me through fixing it or have a link to a good guide? Not the most experienced with electrical repairs.