Finding noise…

[pucho812]

This is a weird one.
unit passes signal, everything is testing out to spec except I am measuring a noise floor that is 45dB louder then spec. Any tips for finding noise in a unit that is performing as it should?

 

[rock soderstrom]

This is a weird one.
unit passes signal, everything is testing out to spec except I am measuring a noise floor that is 45dB louder then spec. Any tips for finding noise in a unit that is performing as it should?

My crystal ball asks what kind of device it is?

 

[abbey road d enfer]

unit passes signal, everything is testing out to spec except I am measuring a noise floor that is 45dB louder then spec.

What are you measuring noise with? If the BW of measurement is too large, noise increases.

Any tips for finding noise in a unit that is performing as it should?

Just the usual noise tracing, stage after stage... Short the last stage input, then short the stage before, and so on.
With such a discrepancy, I would be temped to accuse the first stage, however it can be a gross defect in teh PSU.
What is the noise spectrum like?

 

[ruffrecords]

Try altering level controls if it has any. Does the noise change. If it does then the source is likely upstream of the control otherwise it is downstream. Divide and conquier.

Cheers

Ian

 

[pucho812]

Noise is measured with an ap 515x. The circuit is by all accounts working as it should. I should see -90dBu noise floor, I see -42dBu. Signal tracing with a a sine wave shows a clear sine wave from top to bottom on a scope. But clearly there is audible noise and measurements are off. I would have suspected filter caps but they all measure low esr and within the capacitance value on the cap, with a cap tester in circuit.
luckily I have a unit to compare it with.

 

[radardoug]

So you wont tell us what the device is. You haven't confirmed that your test setup is working perfectly. You say you are doing cap measurements in circuit. This is normally not possible.
We are not mind readers. We cannot know what you are doing unless you give us as much information as possible.

 

[pucho812]

So you wont tell us what the device is. You haven't confirmed that your test setup is working perfectly. You say you are doing cap measurements in circuit. This is normally not possible.
We are not mind readers. We cannot know what you are doing unless you give us as much information as possible.

Can measure caps in circuit with my cap tester.

 

[radardoug]

Hmm, Google search for cay tester brings up nothing. Model number? Brand?

 

[pucho812]

Hmm, Google search for cay tester brings up nothing. Model number? Brand?

Peak electronics

 

[Brian Roth]

So the noise should be very audible when it is connected into a monitoring system.Hiss? Hum? Whistles?

Bri

 

[radardoug]

Getting information out of you is like pulling teeth! What model number?
And, again, what is the peice of equipment that you are having trouble with? Brand and model number please.

 

[abbey road d enfer]

Can measure caps in circuit with my cap tester.

An ESR tester won't tell you how leaky the cap is when submitted to its operating voltage.
Measuring ripple in operation is the conclusive test.
The type of noise (hum, hiss, grouch, bark, squawk) would determine which direction to investigate.

 

[tony hunt]

Another question, kind of been asked before:
Was the the noise audible before measuring began?

 

[pucho812]

Another question, kind of been asked before:
Was the the noise audible before measuring began?

Yes. With no input signal and the input grounded as is normal when not Connected, I hear noise and measure noise.

 

[JohnRoberts]

IIRC Peak was a name associated with developing Peavey Mediamatrix , but multiple consumer products have used the name.

The obvious suspects are 1) bad IC, 2) bad active component. 3) solder bridge causing op amp to operate at excessive closed loop gain.

Can you probe backwards from the output and see where the noise first increases. Is product stereo? are both channels noisy? If yes look for common noise source like PS.

Did noise start coincident with some event? It will probably be last thing you check.

JR

 

[pucho812]

IIRC Peak was a name associated with developing Peavey Mediamatrix , but multiple consumer products have used the name.

The obvious suspects are 1) bad IC, 2) bad active component. 3) solder bridge causing op amp to operate at excessive closed loop gain.

Can you probe backwards from the output and see where the noise first increases. Is product stereo? are both channels noisy? If yes look for common noise source like PS.

Did noise start coincident with some event? It will probably be last thing you check.

JR

it's a mono unit. no solder bridges to be found. 2 is possible, going over all the transistors as we speak. 3. only ic's in there are for dc servos and appear to be working.
The noise started before the unit arrived. unit was picked up use by a friend. Upon first use noticed the buzz, noises, in the audio and when he reached out to me, I got sent a photo of the insides and noticed a blown capacitor I replaced the blown cap and made sure the surrounding components in the circuit were functioning as I am working without a schematic. I also confirmed the correct PSU output as I have notes I have made before. From tests with the AP machine, I can see the noise the moment connect the output. I can also see what looks like a blend between a triangle wave and a sine wave on a scope at the output with out a test signal on the input. connected. So clearly it's there, it is not effected by any volume control.

 

[abbey road d enfer]

I can also see what looks like a blend between a triangle wave and a sine wave on a scope at the output with out a test signal on the input. connected.

What frequency?

 

[JohnRoberts]

A rogue DC servo could be dumping noise into the audio path...

JR

 

[radardoug]

Still no brand/model numbers! Why will you not give us this information? Is the output active balanced? Failure of the one side driver would put dc on the output on one leg, but still pass audio. Or JRs suggestion.

 

[pucho812]

break through. Found a faulty dc servo ic. Now the unit is back up to expected noise floor specs from the factory data. No more funny wave at the output, no more nothing, just a unit working as it should. Persistence pays off. As for make and model, unfortunately leaving that to guesses this time unless people want to dm me on it as not long ago I got a strongly worded letter from some company about discussing their tech. While I normally don’t cave to that, since a friend works there don’t want him in trouble