power-line noise trouble, either diff/NRSE/RSE cann't avoid

[wonwe]

We encountered a trouble during our DAQ
 
physical connection:

A photodiode (worked as a photocell) was shielded in a copper sheath, then connected to AI0 channel of BNC-2120(Q9 coxial jack) through a 75-3 coxial cable( RG-59/U ), we use simplest way to realise an light intensity measurement using our PXI-6221 card.

When isolated Photodiode from light, which only dark-current of photodiode accounted for, we got waveform like follows:

 setting "Terminal Configuration" as "differential", we got:

http://forums.ni.com/attachments/ni/250/19905/2/diff.PNG
 setting "Terminal Configuration" as "RSE",we got:

http://forums.ni.com/attachments/ni/250/19905/3/RSE.PNG
 setting "Terminal Configuration" as "NRSE", we got:

http://forums.ni.com/attachments/ni/250/19905/1/NRSE.PNG
as we can see, we could found an power-line frequency components (50Hz ,China) in our readings, I think our photodiode is just a floating source, so either of three measurement systems is ok, Why there is an power-line noise in our test? how to avoid it? and why "NRSE" is so different from the other two? I mean it did existed an obvious time-dependent voltage drift in the reading(higher reading-out as time passed), which measurement system shall we take in our coming measurement?

I got awfully puzzled , and need your help as soon. Our DAQ Assistant configuration is listed as following picture

http://forums.ni.com/attachments/ni/250/19905/4/config.PNG, we use--

chassis:  PXI-1042Q
Controller: PXI-8186 @ slot0
Card: PXI-5124 at slot1
      PXi-4461 @ slot2
      PXI-6221 @slot3( we use this card as our A/D in this case)
     BNC-2120 (as BNC Connector block of 6221)

      LV PDS 7.1
      NIDAQ7.4.0

帖子被wonwe在03-04-2006 02:06 AM时编辑过了

帖子被wonwe在03-04-2006 02:07 AM时编辑过了

[AnalogKid2DigitalMan]

Have you tried connecting the copper sheath to earth gound? Does it fully enclose the photodiode (except for the window of course)?

[wonwe]

The copper sheath is fully enclose the photodiode as you said, but I haven't tried it to the earth ground yet. Does that work?

whether the power-line noise should goes to the monitor or not?

'Cause our cable and BNC-2120 are just  highly close to our LCD monitor ?

Somehow, why existed an obvious time-dependent voltage drift in NRSE measurement?  THis puzzled me as well.....