3
\$\begingroup\$

I am using an NE5532 as a voltage follower/buffer in an audio circuit and am having some noise issues. I’ve simplified the circuit to isolate the issue. The audio input signal comes in through an audio transformer to the +IN on the amp. The amp output is currently going directly into a commercial mixer for testing.

With the mixer gain and volume turned up and no sound signal, rather than quiet white noise/background static I’m getting a relatively loud/fast ticking/flapping sound … sounds something like shuffling cards, but more of a ticking. With an input signal, it’s loud enough to interfere with the music. There are also some blips of loud static every second or so. For testing I replaced the op amp with an LM411 FET amp, the noise was mostly gone, except for the ticking being very very faint at full volume. I have tried all different combinations of grounding, and transformer isolating outputs, nothing helps, it seems to be an op amp issue and the noise is in the signal itself.

I’ve tried scoping the output, and can’t distinguish the ticking noise and blips from noise measuring from only the original source.

Any idea what is causing this and how to fix it?

Another thing: This amp is used as a buffer prior to output with line level signals up to 2Vrms. According to the datasheet, excessive current is drawn at a differential input of 0.6V due to the input transistors. Should I be modifying my circuit with input resistors or something, or should this amp only be used with small input signals?

enter image description here

A few things: The final circuit is much more complex, this is basically what I've simplified it to just for isolating the noise issue. The final circuit is single-ended +12VDC, with opamp biased to +6V, and power supply is +15VDC switcher with heavy filtering and +12V linear regulator. Regardless of how I power the circuit, whether +15V switcher with filtering or a linear 12V bench supply, I get exactly the same noise as described. Input to mixer is balanced.

Update: Here are the additional tests I've completed trying to isolate this noise problem. All have been unsuccessful so far: 1) I was able to measure from 3-12mVAC between completely unused traces even with no power on the breadboard and discovered that this was causing some hum, so I deadbugged the circuit. Same result. 2) I've tried adding an LC filter w/ 500kHz cutoff after the opamp to eliminate RF interference, same result. 3) I've tried a different brand mixer after discovering the jack shields on the first were not chassis grounded. The original used a linear supply, this one a switchmode supply. No change either way. 4) After measuring 45VAC to ground on the new mixer chassis, I've tried variations of earth grounding the mixer chassis through the bench supply. This eliminated the massive hum I'd get without doing it, but the opamp noise is still there. 5) I've tried both unbalanced and balanced inputs, no change. 6) I've tried multiple variations of different cables, plugs and adapters, headphones and shielding the cables. No change.

The only thing I can think of left is that it's something with the opamp itself. As originally mentioned, an LM411 reduced the noise significantly, although it was still there. I'm attaching an overview of the system setup, I apologize for the poor quality. The drawing is mainly to isolate any potential ground loops, so the internals of the devices may not be entirely accurate, I only verified where GND and COM were connected. Is there a way to attach an audio file to this post? Maybe someone will recognize the noise. Not experienced with the phenomenon, but I think it may be motorboating.

Thanks so far for the troubleshooting input, any other comments are greatly appreciated.

enter image description here

\$\endgroup\$
12
  • 2
    \$\begingroup\$ Maybe post a circuit diagram? \$\endgroup\$ Jul 26, 2017 at 15:56
  • \$\begingroup\$ Most op amp configurations, including a buffer, have nearly zero differential input voltage. It isnt clear what your buffer concern is. \$\endgroup\$
    – Matt
    Jul 26, 2017 at 16:00
  • 1
    \$\begingroup\$ Post a schematic, including the full details of the power feed to this opamp. I have some suspicions, but without a schematic there is nothing to talk about. Closing as unclear until then. \$\endgroup\$ Jul 26, 2017 at 16:47
  • \$\begingroup\$ So, why aren't you showing the correct circuit? Why aren't you showing the split rail generator and naming the op-amp you are using for this? What does R5 do? \$\endgroup\$
    – Andy aka
    Jul 26, 2017 at 17:04
  • \$\begingroup\$ This is the correct circuit. Its the simplified circuit I'm using to test, which is giving the EXACT same noise as the final circuit. I initially intentionally left out the schematic because I was trying to leave out the things I've already isolated as not causing the issue. The actual input impedance seen by the transformer secondary is 47k. The NE5532 is the opamp i'm using for this. \$\endgroup\$
    – User7251
    Jul 26, 2017 at 17:06

4 Answers 4

3
\$\begingroup\$

So, I was finally able to solve the motorboat problem by placing an LRC filter with a 50kHz cutoff frequency directly at the output of the transformer. The transformer has a self-resonance frequency at around 280kHz but amplitude starts growing around 90kHz, not sure if this is the cause specifically but killing that resonance and whatever else was making its way into the opamp seems to have eliminated the noise. Thanks everyone for their input.

\$\endgroup\$
2
\$\begingroup\$

You say in a comment that no noise is exhibited if the transformer is connected directly to the mixer's balanced input.

However, I suspect that the intended role of this transformer is as a "balun": converting unbalanced to balanced.

The mixer, being balanced, is a three-terminal input. If the transformer's secondary is connected so that the centre tap goes to the mixer's GND, and the other two ends of the coil to the + and - inputs, then that is a very different topology from the single-ended scheme you have going with the op-amp.

In the configuration shown, the transformer's coil is a single ended input source. I would expect its windings to pick up EMI which gets amplified. The mixer's balanced input is not being used as a balanced input. In the balun configuration, noise from the transformer is suppressed because it is split into two coils of identical number of windings, whose noise cancels out by the opposite polarity inputs of same impedance.

One way to proceed is to ditch the transformer and implement the unbalanced-to-balanced conversion with op-amps, perhaps with a circuit like this one.

Implied in the circuit is an end-to-end ground connection.

If you want the isolation of the transformer (no end-to-end galvanic connection), you can work it into a different two-op-amp circuit. In a nutshell: extend the + and - inputs of the mixer with a pair of identical pre-amplifiers, each implemented by an op-amp. (This preserves the property of the balanced inputs being identical in impedance.) So now you effectively have a new mixer input which is still balanced but whose balanced inputs have more favorable impedance characteristics and level for your laptop's line out thanks to the two active buffering stages. In front of these stages, you engage the balun transformer in the usual way.

\$\endgroup\$
5
  • \$\begingroup\$ Kaz thanks for the response. So the idea of this device is to take in either balanced or unbalanced audio signals, hence the transformer, which also helps eliminate any common mode noise picked up by the input line, and galvanically isolates this device from whatever it is connected to on the input side. The transformer currently used is a 1:1 TY-145P from Triad Magnetics, with both primary/secondary center taps floating. As I understand it, for dealing with input signals, transformers are far superior to active circuitry/diff amps in terms of CM noise reduction. Continued ... \$\endgroup\$
    – User7251
    Jul 29, 2017 at 6:32
  • \$\begingroup\$ With regard to the outputs, you are also correct. The original circuit was designed with balanced impedances at the opamp output and at the output return to ground. There is only one opamp output, so in that sense it's not a typical 2-amp inverted output signal, but the only difference would be the signal amplitude going into the next device, assuming it uses differential inputs. And for the case of my circuit, I'd actually prefer that the amplitude is simply "passed-through" to the next device". The above circuit is simplified in the sense that I have removed the extra RC ... \$\endgroup\$
    – User7251
    Jul 29, 2017 at 6:38
  • \$\begingroup\$ network to ground at the return output in order to solve this noise problem and eliminate as many components as possible to isolate the source. As mentioned above, the sound coming out of the amp resembles "motorboating". I have tried adding larger bypass caps with no success, it seems the sound is just unrelenting and nothing I change seems to make any difference. When connecting directly from laptop to mixer through the transformer, I have tried both unbalanced and balanced inputs (theoretically this should make very little difference from the transformer), and there was just the expected... \$\endgroup\$
    – User7251
    Jul 29, 2017 at 6:49
  • \$\begingroup\$ white noise. It's really the opamp itself that seems to be getting me into trouble, at least that's what all of my experimenting leads me to believe. But I'm hoping someone with much more experience can point out some nuance I haven't considered/am missing. \$\endgroup\$
    – User7251
    Jul 29, 2017 at 6:51
  • \$\begingroup\$ I just placed the output to the mixer's unbalanced input directly at the opamp input and directly at the output, with ground in the same location. At opamp input, quiet white noise. At opamp output, motorboat. I tried scoping both before and after and cannot detect any difference between the signals. I checked with a voltmeter, I do get a 25mV higher DC at the opamp output. I'm starting to think I'm going insane. \$\endgroup\$
    – User7251
    Jul 29, 2017 at 7:41
0
\$\begingroup\$
  1. Make sure the opamp isn't oscillating.

  2. Lower the input impedance. 47k is the max I would go.

  3. Clean the power supply.

\$\endgroup\$
1
  • \$\begingroup\$ Input impedance of the final circuit into the buffer is actually just that, 47k. I added it into the test circuit as well. \$\endgroup\$
    – User7251
    Jul 26, 2017 at 17:02
0
\$\begingroup\$

In the absence of a schematic, I would speculate that you are getting noise coupled into the circuit from your power supply. Try operating it from a battery to see if the problem goes away.

Failing that, short the input of the transformer and move the circuit away from all other electronics. If that eliminates the noise, undo the changes one step at a time to discover the source of the noise.

Edit after OP posted schematic:

Your transformer secondary is 600 ohms. You should have a 600 ohm resistor across it to properly terminate it since the op amp input impedance is very high. Remove the 47k resistor. I would include a decoupling capacitor in series with the output of the op amp to ensure no DC offset is passed to the mixer. Make sure you have phantom power turned off on that mixer input.

\$\endgroup\$
13
  • \$\begingroup\$ I've tested the transformer without the opamp, just directly into the mixer, no noise issues at all. Also, I kept thinking it must be power supply as well ... but the planned supply is so heavily filtered I've gotten the ripple to below that of my bench supply, and either way I still get the noise issue. \$\endgroup\$
    – User7251
    Jul 26, 2017 at 17:03
  • \$\begingroup\$ Shorting of the transformer input is to eliminate pickup through the input, not to test the transformer. For the power supply, I am thinking more like EMC/RFI than ripple. Can you post a schematic of your circuit? \$\endgroup\$
    – Glenn W9IQ
    Jul 26, 2017 at 17:14
  • \$\begingroup\$ Glenn, I edited the original post to add the schematic. +/-15VDC is coming from a linear bench supply. Ahh, missed the shorting of the input, will try that and post back. \$\endgroup\$
    – User7251
    Jul 26, 2017 at 17:17
  • \$\begingroup\$ Shorting the transformer primary doesn't change anything, the noise is still there. \$\endgroup\$
    – User7251
    Jul 26, 2017 at 17:21
  • \$\begingroup\$ Tried powering from a 9V battery, same result \$\endgroup\$
    – User7251
    Jul 26, 2017 at 17:29

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.