Hi all,


do we normally find popcorn noise in parasitic VNPN transistors? I was running some measurements on a circuit which used such transistors, and found what I could only describe as popcorn noise.


thanks,
Aaron

Hi,


thanks for the reply. I read up a little about shot noise, but discarded it as it is supposedly white, right? I have included a plot of the output noise PSD. As you can see there are what appears to be tones at a few frequencies. The only noise I know which could exhibit this is burst noise.

Either that or the noise is coming from my measurement setup, but my setup appears to be quite clean of noise (I checked). I'm powering the DUT using AA batteries connected to linear regulators with lots of shunt caps. The only equipment used are a signal generator and a spectrum analyzer.


thanks,
Aaron

Hi pancho_hideboo,


the circuit is actually an LNA and mixer, so the signal generator provides the LO. I haven't checked the spectrum of ground, but i've checked some other nodes (LNA output node and a biasing node from a constant-gm biasing circuit) and no spikes are observed. I suppose that rules out noise on the ground node right? I've also tried turning off the LO and changing the biasing of the LNA. In neither case was the 3.5 MHz noise signal affected.

Unfortunately I can't run any further measurements on this circuit (in the near future) unless I can come up with a good explanation for this noise. Its on-wafer testing at my University, and there's a pretty big queue to use the probe station.

I'm not so sure that's its burst noise any more since the same noise component appeared on a separate chip. I'm not sure if it was the same wafer.

I didn't quite understand question (1).


thanks,
Aaron

Quote:


Correct

Quote:


If the 3.5 MHz signal were on the ground line then I expect it would affect all nodes, not just the output node. Is that correct?

Quote:


Correct.


cheers,
Aaron

I just realized that it might be common-mode oscillation which has been incompletely canceled by the differential operation. This would agree with the fact that my op-amp appears to have pretty poor IIP3. Its strange though since I normally overcompensate the CMFB loop since I only use it to set the output common-mode DC value...

Hi Ken,


thanks for the reply. pancho_hideboo also pointed out that the ground line may not be clean, but as I mentioned, the 3.5 MHz noise only appeared at the output. I was only able to check two other nodes, but neither showed any evidence of a 3.5 MHz peak in the noise.


cheers,
Aaron

Hi pancho_hideboo,


unfortunately, I cannot re-assemble my test setup at the moment. I did try changing the DUT supply and I did'nt notice any changes in the frequency of the 3.5 MHz noise, but I didn't record any values so i can't say for sure. I used a supply-independent biasing which means that the gain of my amplifiers wouldn't have been affected much by changes in the supply voltage.


thanks for the help,
Aaron

Hi sos,


do you think the circuit might be picking up interference? The 3.5 MHz spike doesn't seem to move.

Anyway the probe station chamber should shield the wafer quite well from interference...i think.


cheers,
Aaron