[moved] Ground spikes in scope observation

[Muralidharan]

HI ,

I have an electronic product with metal parts and pcbs. Metal parts are connected to ground while parts of the pcb are also conectd to ground. I observe using oscilloscope probe the voltage level at pcb ground w.r.t.metal ground. I get a small patch on the scope which I can imagine as coming out of return currents /noise etc and it is ok.

When another device in the vicinity connected to the same ac line is switched on, I find a large spike appearing on the pcb ground w.r.t metal ground(observation points same as first described..
Can anyone describe exactly why this large spike is observed when another device is swithced on?

Thanks,

Murali

 

[D.A.(Tony)Stewart]

Induction?
Or ground shift by conduction?
And Induction probe error.

use low impedances, shielding and filtering to cure this.

 

[Muralidharan]

electromagnetic induction due to currents in ac line getting switched in ? May be..how do you figure it out what it is due to? What is ground shift?

 

[amit.kumar11]

Hi,

Tour board is powered on by Ac supply or DC supply?


Amit

 

[Muralidharan]

ac then converted to dc through smps which powers the pcb under question, Amit

 

[D.A.(Tony)Stewart]

RLC impedance ratios are critical to understand for load/source Attenuation of Stray E or H fields and conduction current attenuation. Every part has conduction resistance. Even groundwires have ESR + ESL.

Impedance is a function of risetime which is ~1/3 f for reactive components or stray coupling of E (v) or H (i) fields.

Choose 100mV current shunt to test for current and conduction noise, e.g. ESR ripple in caps


Next time put detailed schematic and layout photo in question, so we dont have to guess or use crystal ball.

 

[Muralidharan]

RLC impedance ratios are critical to understand for load/source Attenuation of Stray E or H fields and conduction current attenuation. Every part has conduction resistance. Even groundwires have ESR + ESL.

Impedance is a function of risetime which is ~1/3 f for reactive components or stray coupling of E (v) or H (i) fields.

Choose 100mV current shunt to test for current and conduction noise, e.g. ESR ripple in caps


Next time put detailed schematic and layout photo in question, so we dont have to guess or use crystal ball.

Thanks for the general hints..Do you propose to break the ground path and include 100mv shunt and capture the waveform across the shunt? Using oscilloscope probe, I am afraid will induce voltage across the lengthy scope probe itself and it may not capture the event across the shunt exactly. because the shunt resistance is too small ,compared to the probe that may be in parallel with it, do you expect that this will not happen?..It is a product with multiple pcbs and a metal frame for ground..Schematics Layout discussion are not practical. I have already put 6mh Inductor in ground line to prevent emi exchange between my product and the ac mains. In terms of prevention, is there anything other than a series inductor in ground path, overall enclosure shielding for preventing E field interference etc that can be done. I am not able to e field interference (V) but chances of M field interference seem more likely. This observation is present even if the product is connected through an isolation transformer..which I think reduces the effect of conductive coupling ..in all probability it looks like more of magnetic interference caused by surge currents of smps used in switching device connected to the ac mains . HOw exactly to identify which type of interference is predominant and what filtering can be done beyond what i have suggested above