Hi team,
Here is a confusing item of OPA350 current noise spectral density curve.
As OPA350 datasheet show:
So why the current curve performance is as above current noise spectral curve show? Rather than like voltage noise curve, but 1/f noise small and Bandwidth noise large? As frequency increase, current noise increase?
If all CMOS amp current noise performance like above? Or only OPA350?
Thank you, hope to hear from you.
Best regards,
Mia Ma
Well Mia,
That increasing current noise going to higher frequencies shows up in the device level cadence sims for JFET and CMOS input op amps. It is, however, impossible to measure (having tried over several months one time) as you get swamped out by the resistor noise used to gain it up and/or bandlimited by that R parasitic C.
I assume you are talking about this plot,
If I try to simulate the input current noise for the updated 2019 OPA350 model, it is not included I don't think - that increase at higher F in this closed loop sim, might be a sim artifact of the output response rolling off.
If you think about trying to measure say the 1pA level (about 250kHz) in the datasheet plot, at a closed loop gain of 1, to exceed the voltage noise of 7nV you need an R to ground on the V+ node say at 7kohm to get equality. That 7k itself adds 10nV noise at the input, so you need to increase it even more to see that noise. So you end up chasing your tail with higher sense R. Say you go to 100kohm, then the input V+ CM cap of 6.5pF will bandlimit the noise contribution of the R and Ibias noise at about 244kHz rolling off your measurement. Anyway, I could not find an app that would be impacted by a noise curve like Figure 3 - so I stopped trying to put it into the datasheets. I would guess Figure 3 in the OPA350 datasheet is cadence sim generated. And yes, I even tried transimpedance (including tee networks) on the bench to pull this increasing Ib term out in test - no luck.
The last time I tried was for the OPA655 as shown in Figure 3 here - that part was upgraded to the OPA656 where that datasheet abandoned that effort shown in fig. 13 of that datasheet.
Hi Mia,
Q: If all CMOS amp current noise performance like above? Or only OPA350?
All CMOS op amplifiers behave this way. The op amp input current noise spectral density can be expressed as follows (shot is short for Schottky noise)
OPA350 input current noise spectral density at low frequency is near negligible in CMOS amplifier. However, at the higher frequency, the input capacitance of the op amp causes an increase in the input bias current, and results in the increases of the input current noise spectral density.
Best,
Raymond
