Strain Gage - value draft

[Žiga]

Hello Balazs!

First of all thanks for clearing some things out. Now I understand.

"I recommended to connect the excitation pins of the strain gage directly to the AUX excitation pins of the 9237." => I tried what you suggested.

Firstly I connected EX+ from strain gage directly to 9237 AUX EX+ and tested. Then I also connected EX- from strain gage to 9237 AUX EX-. The results were still the same.

I used female pins (attached pic) instead of  9942 connector (still waiting for shipment). When I get the 9942 I'll try this solution again.

I'll try using an external power supply (I'll try with Agilent E3634A).

"Where do you connect the EX-?" => Combinations I tried:
    
    1.connected strain gage EX- to 9237 AUX EX-
    2.strain gage EX- to 9237 AUX EX- and to 9949 shield
    3.strain gage EX- to 9949 EX- (7th pin)
    4.strain gage EX- to 9949 EX- and 9949 shield to 9949 EX- (7th pin)

RS+ in RS- match up with EX+ and EX-.

I use the original RJ50 from NI. 194612B-02, 2 meters 1046 12506, made in

China.

I did not try to measure with any other RJ50 cable.
I use MAX 5.0 and NI DAQmx 9.3.

Regards,

Žiga

[BalazsNagy]

Hi!

"Combinations I tried: 

    
    1.connected strain gage EX- to 9237 AUX EX-
    2.strain gage EX- to 9237 AUX EX- and to 9949 shield
    3.strain gage EX- to 9949 EX- (7th pin)
    4.strain gage EX- to 9949 EX- and 9949 shield to 9949 EX- (7th pin)" => In my opinion excitation and shield sholud not be connected.

I think the connection on this page, with some modification sholud be suitable for you:http://digital.ni.com/public.nsf/allkb/ED161F9867C820998625796D00769DBC 

Modification: The EX+ and EX- signals of the strain gage should be connected to the AUX EX outputs of the 9237 module. 

"I use the original RJ50 from NI. 194612B-02, 2 meters 1046 12506, made in China." => I suggest you to try the measurement with an other RJ50 cable.

A read about some similar strain measurement problems, with the same configuration, that you have. In a few of them the main reason of the measurement error was the RJ50 cable, with part no. 194612B-02.

Because this an other revision is avaliable with part no.194612C-02. Maybe both versions are beeing shipped.

"I'll try using an external power supply (I'll try with Agilent E3634A)." => Please let me know what did you experience.

I think your softwares are up to date.

Best regards,

[GPeter]

Hi Žiga,

Have you tried the suggested configuration?

Please give a short summary, how it is working for you! If your problem is solved please indicate it by closing the topic.

Thanks,

Peter

[Žiga]

Hello!

Sorry for a long wait!
I got 9942 connectors from NI and I tried the previous solutions again with this connector and external excitation (used PS Agilent) -> Drift occurs..

"http://digital.ni.com/public.nsf/allkb/ED161F9867C820998625796D00769DBC " -> didn't try this one yet.

I have tried with both cables, 194612B-02 and 194612C-02. Sadly no difference. I'm still waiting for another RJ50 cable (not from NI), sadly it takes too long to get it..


Best regards,
Žiga Pristov

[BalazsNagy]

Hi!

Do you have the opportunity to try the measurement with an other NI 9237 module?

Did your module worked before? Maybe it's broken somehow, and should be sent for repair.

Please let me know if you could try the above suggested connections.

I'm sorry that the problem is still present.

Best regards, 

[Žiga]

Hello!

Sadly we do not own any other NI 9237 module.I can't order a new module on a hunch that the new one will eliminate the problem.

I have two NI 9237 and two carriers, it works the same.

Currently I am working on a full bridge. I hope I can get all four gages on the axis.

Well nothing is "ideal" otherwise we would all be out of work.

Regards,

Žiga Pristov

[BalazsNagy]

Hi!

The full bridge configuration should elimintae the drift, in case the reason of the drift is the varying cable impedance when the RJ-50 cable is disturbed.

In full bridge mode, both sides of the bridge see the same voltage change and the remote sense capabilities of the modules means it is not an issue.

If drift will be present in full bridge configuration too, the reason has to be something else, and we have to escalate your problem.

Please let me know how the measurement works in full bridge mode.

Best regards,

[Žiga]

Hello!

I understand what you mean and I hope I get expected results out.

Currently I am still waiting for RJ50 cable to arrive and my equipment to be modified.

Don't close the topic yet. Thanks.

Regards,

Žiga Pristov

[Arlon]

Hello,

Sorry if I'm speaking out of turn here, but I have several years of experience with these modules and I think I might be able to help out here.

First of all, if you are trying to measure bending in a beam, it is always a good idea to use a half or full bridge to compensate for any axial stress which may be present (gauges on adjacent arms of the bridge on top and bottom of the beam). This also helps to compensate for temperature effects.

Now, if you decide to go with a quarter or half bridge, I highly recommend completing the bridge with external resistors (preferrably precision ones with low temperature variation) and soldering all connections and always keep the leads within the bridge as short as possible. Any change in resistance or any interference inside the bridge can have a very large effect on your measurement (I've seen this just by waving my hand over a long coil of leads). If at all possible, avoid connectors, screw terminals, etc. within the bridge. If you do have any connections, make sure the leads do not move relative to the connection!

We started out with the RJ50 cables with "A" or "B" in the part number and found out the hard way that you can get significant drift when the cable was moved. The story we heard was the conductors in the cable were too small for the connector meaning if you bent the cable at the connector, you would get a change in resistance. Having a full bridge does not make it work because the change in resistance may only occur on one conductor. After we got the "C" cables, and thoroughly tested them to make sure they work (they did for us), we decided to toss them altogether and just use regular network cables with the RJ45 (8P8C) connectors so that we didn't have to stuff around with adapters and difficult to find hardware. NI does not recommend using these connectors due to the possibility of damaging the outer pins used for shunt calibration. Our experience after plugging and unplugging them hundreds of times is that those two pins may not spring back quite as far as the others...

One other thing if you haven't done it already: I would double check the scaling factor and make sure the drift is significant relative to what you are trying to measure. I still get tricked by the autoscale on a graph every now and then.

Hope this helps,

Arlon.

[Žiga]

Hello!

Okey so I changed RJ50 cable for another RJ50 cable (not shielded and it looked poorly made so I'm looking for another better one, got any ideas where to get a good RJ50 cable?). Unfortunately the drift still occurs. The connectors plugged harder in card so the moveable space was reduced. Somehow resistance still changes.

"http://digital.ni.com/public.nsf/allkb/ED161F9867C820998625796D00769DBC" -> tried this solution, still didn't get rid of drift.

Thanks for your information Arlon, I will look into it.

I got nothing new on full bridge configuration.

Regards,

Žiga Pristov