DAQ photodiode voltage too noisy (> 20mV)

[Bauch@Berkeley]

I am using a USB-6009 in differential mode 14Bit resolution as shown in the code and frontpanel attachment. I am acquiring two channels,one is the photodiode voltage and the other one is a ramp reference signal which is also used to trigger the DAQ using the digital input. 

The confusing results are shown in the attachment  trigger+diodevoltage_zoom.jpg: There is the ramp / trigger signal in blue and the photo voltage in green.

Whilezooming in onto the green signal it is easy to see that the noise of >20mV is way bigger than the theoretically possible 0.3 mV resoluton referring to the 14 bits.We checked the same signal with a scope and got much better signal to noise ratio. We also tried a second USB-6009 and received the same bad results and another photo diode had almost the same noise. 

Sounds like a trivial problem hundreds of times asked here, but still cannot figure out why the SNR is so bad.

Thanks for any help,

Erik 

Message Edited by Bauch@Berkeley on 09-09-2009 07:10 PM

[Henrik_Volkers]

What I would try:

-A ferrite core (Choke)  with 10+x windings of both wires (Ain AGND) from the diode to the input channel.The Analog INPUT of the ADCs are known to be noise sources.

  Nice lit: http://ts.nist.gov/MeasurementServices/Calibrations/upload/72C-2.pdf

-Build a transimpedance amplifier to get a low impedance signal (I just played around with samples (upper right corner) from linears LTC 6240)

 (takes 30min to build a dead bug bread board amp plus ?days to get the material 😉 )