Question on the AD8138 input referred noise of the Vocm input noise specification:
The datasheet lists this as 17nV/rtHz, which is more than 3 times that of what the datasheet claims in the features. Does this noise always apply and scale with BW and gain? I am using this device to convert a single ended signal to differential, the Vocm pin will be held at 1.2V. Graphs shown in Figure 37 and 38 show the voltage and current noise (RTI), the broadband voltage noise isn't anywhere close to 17nV. If the 17nV/rtHz always applies, these graphs are very misleading.
As an example, say we have a case where the gain is unity, and the signal bandwidth is 100MHz. The noise bandwidth is ~ 50MHz. The broadband voltage noise at the output (just considering voltage RTI noise, neglect current noise and thermal resistor noise)
RTO Vn = 5.7nV/rtHz (50MHz)^.5 = 40.3uVRTO Vocm = 17nV/rtHz (50MHz)^.5 = 120.2uV
Total RTO = [(40.3uV)^2 + (120uV)^2]^.5 = 127.8uV
Please answer ASAP, thanks for your time.
Instrument Electronic Design Branch
I was searching through EZone on a similar topic and was stunned to find that this thread has been neglected for all this time. Please accept my sincere apologies. Since your question is a good one and I think others would have a similar question, I'm going to answer it as best I can.
First, just for reminder, Vocm in the input voltage that sets the output common mode voltage of the differential amplifier outputs. The common-mode feedback loop is closed internally and therefore would not have the dependence of gain and bandwidth that I think you are asking. The internal Vocm circuit itself is slower than the differential signal path as you can see by comparing figures 21 and 36. That all being said, Vocm noise will appear as common-mode noise on the differential outputs assuming the external differential feedback paths are balanced. The attached calculations might be helpful to illustrate.
The section on "Estimating The Output Voltage Noise" on pages 17-18 of the datasheet go into this in a bit more detail.
You may also find the ADI DiffAmpCalc helpful as it provides a summary of noise contributions for a specific configuration and includes the component of noise from Vocm.
Additionally, it's important to add bypass capacitance on the Vocm pin to help reduce any unwanted signals from coupling into the this pin.