Noisy circuit, can a PCB trace or solder joint cause it? stumpped...

[pucho812]

trying to wrap my head around this, is it possible for a motherboard to cause noise issues? it has no components on it, it is  just a bunch of ribbon cables, headers and soldered connections on it.

I ask as I was trouble shooting a neotek the other day that has noise on the right side of the stereo buss.  it sounds like a fried/failing IC.  Swapped IC's on he master card from left side of the stereo buss to right side and the sound stayed on the right side. Replaced the master cards with new ones, and the sound was the same, noise on the right side. 
I get noise at the stereo bus insert send, and if I patch into the return, the noise goes away.  so at this point I know noise is pre-insert.

I ran signal into a channel and fed it to stereo buss. If I disconnected the ribbon carrying all the signals to the mother board which then splits it off to the master cards and monitor cards, I still get the noise.
If I disconnect the output of the master cards  I loose the noise.

PSU voltages are with in spec,  the PSU has low ripple.

not sure where else to check really, and if it turns out to be a trace or some other thing on the master card, it will be a huge PIA to replace. 

 

[ruffrecords]

Just to be clear, you swapped master cards but the fault remained on the right side with either card?

In which case the master cards are not at fault (unless in the extremely unlikely even that both cards have the same fault).

Check the noisy output with a 'scope and look for any high frequency oscillation - these often sound like noise.

I don't suppose you can plug the master card in another slot??

Cheers

Ian

 

[JohnRoberts]

The noise is punishment for spelling bus wrong. 

A (partially) shorted bus (to ground or a low z) could increase noise gain and express as higher noise...

JR

 

[pucho812]

sorry john. I know misspelled bus. I  corrected it.

Ian yes, his happens with both master cards.  So I suspect something else in the desk.  I have attached an audio file of the noise  as I took a video over the weekend. The noise changes as I filmed settings on the master. I started off  in front of the right speaker.  then I moved the master to show settings and how the fader  effects volume as well as the cr master. The cr master is  obvious as the noise is in the stereo buss.
Here is the noise in question
https://soundcloud.com/joevezzetti/noiseneotek/s-6Eqpm

now as mentioned the master section and monitor section ribbon to a motherboard. the channels ribbon to that mother board as well.  that mother board also has a lot of wiring to connectors on it.  Some of he ribbon connections are soldered  and glued to place and then a connector is on the other end.  Or on some cases there is actually a had connector on that mother board

 

[ruffrecords]

John makes a good point. A solder splash or partial bridge could do this. Visual inspection of the ribbon connectors and PCB connectors would be a good idea.

Cheers

Ian

 

[pucho812]

John makes a good point. A solder splash or partial bridge could do this. Visual inspection of the ribbon connectors and PCB connectors would be a good idea.

Cheers

Ian

While that is true, why would it take over a year and a half to develop? waiting out the warranty?

 

[ruffrecords]

While that is true, why would it take over a year and a half to develop? waiting out the warranty?

It's mechanical and therefore not subject to the normal laws of physics  ;D

Cheers

Ian

 

[pucho812]

  good point. 

According to the owner, the noise increases and decreases randomly.  So some days it is worse then others.  That had me thinking  the board or a joint flexes from heat.  But it doesn't appear to get that warm under he hood.  But then again I do not know how hot i tdoes get, when I look at the desk, the board has been off all morning.

 

[Walrus]

Has a foreign liquid (coffee, beer, wine etc) got under a pcb ribbon connector? 

I've certainly had something similar in the past where sweet coffee had been spilled down the desk, and although we thought we had cleaned it off OK, a good while later some that had got under a capacitor started to become conductive as the sugar content dried and "shorted" a psu rail to the signal path with all sorts of strange noises
Removed the cap. cleaned the board, and new cap. fixed it. (Took a lot off head scratching to find though!)

 

[pucho812]

Nope, no liquids ever in the desk and I cleaned all the connectors when possible

 

[Monte McGuire]

Stranded cable, ribbon or otherwise, that is soldered into PCB pads can fail due to flexure. If all of the strands fail, they could still conduct by mechanical contact, but over time, the strands will get oxidized, making the connection tenuous. I had a Neotek Series II that had some of these type of connections to the control room module, and they did fail from flexure. The problem is that there's no simple way to provide strain relief with such a connection. So, you just have to re-make these connections if they fail and then try not to ever move the module, which would happen if you pull it for repair or odd testing.

 

[pucho812]

Stranded cable, ribbon or otherwise, that is soldered into PCB pads can fail due to flexure. If all of the strands fail, they could still conduct by mechanical contact, but over time, the strands will get oxidized, making the connection tenuous. I had a Neotek Series II that had some of these type of connections to the control room module, and they did fail from flexure. The problem is that there's no simple way to provide strain relief with such a connection. So, you just have to re-make these connections if they fail and then try not to ever move the module, which would happen if you pull it for repair or odd testing.

A- HA,!!!! yes, that makes sense. Would flowing the joints help or do you just replace the ribbons?

 

[buildafriend]

Have you pressed on the headers cables and boards while listening?

Does the console use a rear I/o panel? If so can you bypass it?

Visual inspection is critical. Sometimes having a friend look with you or listen while you move things is all you need..

 

[buildafriend]

Sometimes you can swap ribbon cables from L to R too

 

[pucho812]

I didn't swap left and right ribbons due to length.  Pushing on any ribbon and or connector did not yield any changes in the noise as heard.

 

[ruffrecords]

Can I ask how do you know this noise does not originate in one of the channel modules?. If it is not the master modules (because you have swapped them) and it is not the ribbon, then it must be on the bus from one of the channels.

Cheers

Ian

 

[Brian Roth]

Can I ask how do you know this noise does not originate in one of the channel modules?. If it is not the master modules (because you have swapped them) and it is not the ribbon, then it must be on the bus from one of the channels.

Excellent point!  Pucho never mentioned what series Neotek this desk is, but glancing at the Elan schemo, I see post-pan buffer opamps hanging into the stereo buses.

Bri

 

[pucho812]

Excellent point!  Pucho never mentioned what series Neotek this desk is, but glancing at the Elan schemo, I see post-pan buffer opamps hanging into the stereo buses.

Bri

Yes brian it is an elan. Yes that do but there is a ribbon cable which carries all the signal of those channels to a mother board, which then feeds it off to the master card. I know the noise is not on a channel as I disconnected the ribbon carrying all  those signals to the motherboard.
About the only thing I didn't do was disconnect each channel one at a time , remove them all from the desk and then put them back in one at a time.

 

[Brian Roth]

Looking at the Elan "master" module schemo, I see two stereo "direct input" circuits....used for echo returns IIRC.  Each has a mute switch that hard disconnects the return channels from the stereo mix buses.

Bri

 

[Monte McGuire]

A- HA,!!!! yes, that makes sense. Would flowing the joints help or do you just replace the ribbons?

Soldering to stranded cable is problematic since the solder will wick back up the cable to a certain point, destroy the annealing, and make a new stress concentrator at the spot where the solder stopped wicking. Further flexure will start the problem again, but at a new spot. The 'right' answer is to somehow connectorize the situation and use IDC (crimp) connections to attach the ribbon or stranded cable to the IDC connectors. Most IDC connectors have some form of stress relief built in as well, so you can generally avoid the whole problem.

If re-engineering a connector system is too complex, then replace the cable, somehow fasten it to the PCB that it's going to  be soldered to, and then make the solder connection. In this way, the mechanical mounting will be stable before you've soldered the cable, so once it's soldered, it can't flex at the solder joint, and can't fail so easily. Might be difficult, but it's an annoying design consequence - shoulda been connectorized in the first place. I love Neoteks, but this is one of the larger PITAs about their overall design.