Are these PSU spikes enough to cause trouble?

[Chris Wilson]

I am battling with a highly complex HP spectrum analyser that shows noise spikes and jitter from time to time on it display. Thanks to the generous help of others I am slowly trying to track the cause down. I have measured the power supply lines after the regulator and on the -15V rail I see the spikes shown at http://www.gatesgarth.com/SA/scope-shots.zip   The last two photos are of the scope looking at the +15V test pin. Whenever i see noise on the SA screen I also see it on the -15V test pin. I never see spikes on + or -5V, or +20V or +15V supplies. The PSU in the SA is a linear, the machine an HP8568B.

My test methodology is probably poor, with a fairly long ground lead, but the fact only -15V shows these spikes is surely significant? The probe was set to X1. The scope an old Tek 7854


Thanks for looking.

[slateraptor]

It sounds and looks and smells like noise coupling, but I suspect only you can tell us if |?V/V| = 0.27% is trouble or not.

[T3sl4co1l]

Can you focus the scope trace? 

Also zoom on the spikes, they look ~sort of~ consistent, not that there should be any reason for such a thing to be produced randomly from a linear supply regardless.  A DSO (or single sweep + long exposure in dimly lit room) would be helpful here.

Possible culprits: electrical breakdown (capacitor? diode/transistor avalanche?), coupling (AC line noise? nearby transients?), frayed wires (intermittent shorting or breakdown), crummy resistors..?!

Aside from aging capacitors (electrolytics are always prime suspect), almost anything else is a stretch.

Tim