回复： why differnent channel can affect each other in voltage measurement using 2532

[木虫]

I use PXI 2535 to measure lots of channels (connected with electrodes) voltages using the sample vi: niswitch DMM swtich synchronous scanning.vi. Among these channels, one channel was  -0.6V and other channels were about -1.0V. But after running the vi, the -0.6V turned into about -0.9~-1.0V. It is very strange. I connected another electrode (-0.6V) to a channel (all other electrodes still connected), whether I used the soft front panel or the sample vi its voltage all turned into about -1.0 V. I didn't find this phenomenon when I used 2532 finished the above measurement. I suspect there was current flew between channels with high and low voltages.    

The red line (HI) was connected to r0, and the black line (LO) was connected to a reference electrode. The V diffence between different  channel vs the reference V were recorded.

When I measure voltages with low difference (lower than 100-200mV), each voltage can be measured correctly (using 2535).

thanks advance

小木虫

[木虫]

 I used a setup shown in Fig. 1 to measure potentials of 121 electrodes (120 zinc wire and 1 iron wire, 1mm in diameter, embedded in epoxy resin, each electrode allowed one surface exposed) in NaCl solution against a reference electrode (with a constant potential). The HI line of DMM was connected to r0 and the LO line was connected to the reference electrode.

Fig.1

These 121 electrodes were connected to the terminal of 2535 with 5 pin Ribbon cables, each of which has 25 lines.Generally, the potential of Zn is about -1.1V and -0.5-0.6V of Fe in NaCl solution. I used the niswitch DMM switch synchronous scanning.vi, sample of LV8.5 to measure these potentials. But I found that the potential of Fe was about -1V, which was similar to that of Zn. This did not occur when PXI 2532 was used. I used a portable DMM to measure the potential of Fe in the 121 electrodes and the results was -0.56V whether I run the sample vi or not. Then I suspected that the switch bring some effects on the measurement.To clarify this problem, I connected a Fe wire to c135 channel (as shown in Fig.1) and measure the influence of other 121 electrodes on the potential measurement of the c135. 1. I did not connect pin Ribbon cables to the switch and used the DMM waveform demo to express the potential of the c135 channel evolution. I opened the soft front panel of the switch and made the c135 connect to r0, which was connected to the HI line of the 4071. When the demo was run, the potential was constant and about -0.56V, as shown in Fig. 2.

Fig.2

2. I connected one pin Ribbon cable (25 zinc wire connected to switch terminal). When the demo was run, I found that the potential was constant and about -0.56V, as shown in Fig.3.

Fig.3

3. When two pin Ribbon cables (50 zinc wire connected to switch terminal) were connected to the switch, similar result was obtained, which was shown in Fig 4.

Fig.4

4. When three pin Ribbon cables were connected to the switch, the potential of c135 fluctuated, which were shown in Fig 5 a-c.

Fig.5 (a)

 

Fig.5 (b)

 

Fig.5 (c)

We conclude that the number of the Zn wires have an effect on the Fe potential measurement using PX I 2535. But why not using PXI 2532 ?We also found other interesting phenomenon: when one or two pin Ribbon cables were connected to the switch (25 or 50 wire electrodes connected to the 2535), the waveform of the potential of c135 (Fe wire) can be affected by the nearby body action, e.g. stand up and sit down, go by the test table. The waveform changes were similar to the Fig.5.  I’d like to know the reason for this phenomenon and how to prevent it? Thanks advance.

[chericks1]

Hi 小木虫,

First of all, thank you for the detailed information. 

I would like to explore the possibility that the phenomenon you describe is caused by the FET switch's fault protection mode. As shown in the PXI-2535 specifications, the module has a maximum switching voltage (channel-to-ground) specification of ±12 VDC, 8 VAC. Once this maximum voltage specification is exceeded, the FET switch's fault protection mode is triggered.  

From your description, we can be confident the signal being switched has a magnitude that is within the allowed voltage range. However, the switch could still be exhibiting an intermittent ‘fault’ behavior if the amplitude of the signal in relation to the FET module's ground reference exceeds ±12 VDC. Please help me see if this is the case. Verify a) the applied signal is not referenced to a different ground potential than chassis ground, and b) the applied signal cannot ‘float’ to a potential that is above the fault protection circuitry’s limit.  Both of these scenarios, (a) and (b), can result in triggering the FET to go into fault protection mode.

The following KB shows common ways these two scenarios can result with resistance measurements. 

NI PXI-2535 or PXI-2536 FET Switch Doesn't Close
http://digital.ni.com/public.nsf/allkb/FD603AF66DCC934386257403006C4109?OpenDocument 

Looking at the information you provided, Figure 1 looks like the system is floating in relationship to earth and chassis ground.  If this is true, then I suggest referencing the system to PXI chassis/earth ground.  My recommendations are as follows:

1. Best - The container with the NaCl solution: is it a glass or metal container?  If it is glass, then perhaps you can insert into the solution a path to chassis ground via a 10 kOhm or greater resistor.  If it is metal, then you can wire to chassis ground from the outside of the container (once again through a 10 kOhm or greater resistor).
2. 2nd Best - Create a path to chassis ground on the HI (between PXI-4071 and PXI-2535) or LO (between PXI-4071 and Reference electrode) line via a 10 kOhm or greater resistor.
   

Both of these suggestions will prevent the signal lines from carrying a common mode voltage that exceeds ±12 VDC and triggers fault protection.

Let me know if you have any questions or concerns.  Hope this helps!

Chad Erickson

Switch Product Support Engineer

NI - USA

[小木虫]

Hi,Chad Erickson,

I will have a try next afternoon.

Best regards

小木虫

[木虫 · Accepted solution]

Hi,Chad Erickson,

       Good news, the floating V disappeared when a line was connected between the ground of chassis and the solution.

       Best regards.

     小木虫

[小木虫]

Hi, Chad Erickson,

I have another question now: what is the difference between the best and the second best suggestions? I think they are equal to each other because the NaCl solution is conductive and the electrodes connected to every channel (including the reference electrode connected to Row 3) are conductive to each other. Then I think connect a path from chassis ground to the solution, Hi line or the reference electrode (LO line) will get the same results, is this right? In a electrochemical test, the ground line are requested to connect to the working electrode (working electrodes were sequentially connected to the HI line throuth 2535 in this test ). 

The last question: the highest input resistance of 4071 is greater than 10G (1*e10)Ohm. I'd like to know that whether this highest input resistance can reach 1* e12 Ohm?

Best regards

 小木虫

[chericks1]

Hi 小木虫,

The 'Best' path references a common point to ground (GND) in an optimal location. Refer to System Configuration - Best and 2nd Best.JPG attached below.

The '2nd Best' path also references the system to GND, but does so at a point closer to the FET switch matrix. Although this GND reference would work just as well in a perfect world, if the tank of solution had less than an infinite impedance to GND, the voltage potential difference between the '2nd Best' reference point and the NaCL solution could create a small leakage current between the '2nd Best' point and the NaCL solution. This leakage current would travel through the 'weak path' noted in the attached drawing. Just so you know, I don't think the small amount of leakage would greatly affect the resulting voltage measurement.

After speaking with the DMM group, I believe the input resistance is likely to remain between ten and hundreds of GOhms.  It is not expected to reach 1 TOhm.

Hope this helps!

Chad

[小木虫]

Hi, Chad Erickson,

       Thanks for your help.

       I'd like to know that whether in current measurement through DMM and 2535 (also other swtiches), the path between the solution and the ground of the chassis is necessary? I think it should be. Because the current measurement in fact is a voltage measurement.

     Best regards

[chericks1]

Hi 小木虫 ,

The fault protection feature of the PXI-2535 makes it necessary to reference the system to chassis ground.  You are correct in assuming this is also true for a current measurement.  The only other NI switch module that has fault protected FET switches is the PXI-2536.

Hope this helps!

Chad Erickson

Switch Product Support Engineer

NI - USA