What do you see in this wave shapes?

elnaldo

Lunatic Member
Hello. Working in this sound interface (8 channels, analog input and analog outputs, plus firewire), and I'm having this "noise" in all the outputs and all the channels, even headphones.

https://www.presonus.com/products/FireStudio-Project

Traced with the scope and the noise is present at the very 1st OPamp after the input.

I was thinking in a power supply problem, it's a switching mode PS, but the output seems OK (at least I don't see a noise following the output pattern). Even at each IC, V+ and V- look flat, with minor noise probably from the SMPS.

One diode on the power supply output looks replaced (rectifiers before the voltage output, not at the primary side)

Another detail. If I shut down the unit and power ON again, the noise could be gone. Sometimes it's noisy, sometimes not. I've ran out of Freeze spray trying to identify a guilty part with no success.

What do you see in this 1KHz waves?

Thanks!

pic_808_12.gif pic_808_11.gif pic_808_10.gif pic_808_9.gif pic_808_8.gif
 
I might look for a cold-solder joint at the replaced diode. Touch up those solder points with your iron and remeasure. The sometimes there, sometimes not leads me to that.
 
Do you think a diode could cause this?

At very low signal level is not present. At moderate levels it appears, as said all over the circuit. Every opamp output, each channel the same.

Running the unit at 110 V (we have 220 V here) seems to tame the effect. Actually it's almost 100% stable at 110V, very low fail rate. At 220V instead, fails more.

The replacement diode is my first suspect. It's not the same part as the original. It will be replaced with a high speed or schottky diode.
 
D

The replacement diode is my first suspect. It's not the same part as the original. It will be replaced with a high speed or schottky diode.

I would get the correct spec part in there first and get a new baseline rather than put in more diodes that are not design original.
 
If it's all over everything, can you measure it with no i/p signal? What is the repeat rate, etc?

It could be EMI pickup from WiFi, mobile phone, etc nearby.

Remove all unused signal leads, in case they are acting as antennas.

Look for a dry joint that might be acting as a rectifying pickup for EMI.
 
I would get the correct spec part in there first and get a new baseline

Or at least ensure the replacement is suitably equivalent (Vf, Vr, If, Trr, etc). Sometimes replacements make the design better (e.g. replacing parallel discrete schottkys with a single, higher power device).
 
I can't pick it without a signal, perhaps it's a very small noise being amplified with the signal. I don't see a fixed freq, it rides along the 1KHz sinewave. I'll take better picts and more exaustive testing if resoldering doesn't fix the issue.

The replacement diode is a 1N4936 Fast Switching Plastic Rectifier
FEATURES
• Fast switching for high efficiency
• Low forward voltage drop
• Low leakage current
• High forward surge capability

TYPICAL APPLICATIONS
For use in fast switching re
ctification of power supply,
inverters, converters and freewheeling diodes for consumer
and telecommunication


But it's in the 48V "phantom" supply only.

I'm, recapping and retouching the soldering on the whole PS board. I'll post back the results.
 
Update: reccaped the PS, no change, BUT, discovered a part of the circuit where the signal is clean (one OPamp), so I have to keep tracing what's happening between the 2 OPamps where the noise starts (no schematic, and a double sided SMT PCB)
 
Made some progress with the tracing:

The signal is OK up to the A/D input. Then, the OPamps that show the noise, are all directly fed by the D/A output. So the problem should be the D/A chip (Cirrus Logic 4385) if the digital output is OK (I don't have the setup to test it now). If the Digital output is noisy, it must be the A/D chip.

The unit has traces of a previous work, resoldering and so, and I see some rosin remaining in the solder. I'm deep cleaning the board, in case that rosin is causing noises at sensitive points.

Is replacing the D/A possible? Do I need to program it in some way, or is it just a parts swap?
 
Is replacing the D/A possible? Do I need to program it in some way, or is it just a parts swap?

Depends how good you are at replacing LQFP surface mount devices...

I'd use our hot air gun to remove it (flood with liquid flux first), then clean up, and re-solder under our x64 stereo microscope rework station.

But first, I'd look at the data timing at the input, and the power supplies and reference voltage for it (which is the easiest way for noise to get in). Also look at the clock, but you'll need more specialist instrumentation to see if the jitter is very bad.

Your interferer could get into the DAC in both analogue and digital domains. Any obvious steps in the waveform would indicate a stuck bit in a parallel interface, but, with a serial interface, that's not going to happen.

Check out any special notes in the data sheet, especially regarding decoupling capacitors or other notes of caution.

https://www.mouser.com/ds/2/76/CS4385_F2-1141519.pdf

e.g.

"Quiescent Voltage (Output) - Filter connection for internal quiescent voltage. VQ must be capacitively coupled to analog ground, as shown in the Typical Connection Diagram."

Rosin flux is usually benign, and should have no effect on the circuit. Liquid flux residue (e.g. from previous rework) is a different matter, and should be cleaned off with flux-off and an air gun (or blowing hard though tightly-pursed lips...).

The only time I've had trouble with flux residue is under SMD xtals, from a poorly-cleaned vapour-phase reflow production process.
 
If it's all over everything, can you measure it with no i/p signal? What is the repeat rate, etc?

It could be EMI pickup from WiFi, mobile phone, etc nearby.

Remove all unused signal leads, in case they are acting as antennas.

Look for a dry joint that might be acting as a rectifying pickup for EMI.

YES!

Shut off the 1 kHz signal and trigger your scope on the glitch. Find out it's period (or frequency) and you'll have a clue as to its source...

If it's 60 or 120 Hz, it might be arising due to some inattentiveness in managing your grounds. (no idea what your overall measurement setup looks like...)
 
That looks like noise being picked up and amplified through. Is the cover on when it happens? Try a line filter? Plug it in by itself, there may be some thing on the same line causing the interference. The SMPS usually operates at 200-400kHz, the noise isn't that high in freq, so I would rule that out.
 
Without signal, the noise it's not present, the output is silent. When you feed a signal, it sounds a bit noisy, and the oscilloscope shows this "noise" riding on the sinewave. If you play an instrument, it sounds "distorted" on louder passages.

I'm pretty good soldering LQFP devices, so that's not a problem. But I replaced a USB chip in my DAC, and due to lack of some programming in one IC, I've lost some features (the device works with the default options).

I'll start with the simpler things first, perhaps the last OPamp before the output, where the noise appears. Perhaps it's just a noisy OPamp, what I've seen before a couple of times.
 
Just a wild thought.

Is your source signal clean?

You keep saying no noise without signal.
I think that rules out power supply issues.
Sort of.
 
Just a wild thought.

Is your source signal clean?

You keep saying no noise without signal.
I think that rules out power supply issues.
Sort of.
Yes, no signal, no noise. Source signal is OK, actually I tried the unit at the owner's place, with his instruments, and heard that "distortion". When plugged to my scope with a continuous sine-wave I see this. I'll try to replace the OPamp first.
 
Update: made some progress here: Replaced the OPamp fed by the DA converter, no chaneg in the noise.

BUT, with a freeze spray, I think I've located a source of noise at the AD converter. Freezing some capacitors beside the IC, the noise increases a lot. Freezing the IC itself lowers the noise, but the capacitors seems to have more effect. Since I'll not replace the AD chip, I'll do a last try replacing the capacitors.

Can I use regular caps to replace those tiny SMT devices, or could larger caps introduce more noise?

I'll use the values detailed in the datasheet. The capacitors in question are 3 small ceramics connected to pins 6 and 7: 1uF, 0.1uF, 0.1uF.

CS_5368.png
 
Back
Top Bottom