Drift Problems using NI 9237 as Half Bridge

[striegler_gerd]

When using the NI 9237 Amplifier to measure half bridges my signal drifts in not acceptable ways (up to 300 µstrain).

Another problem is that the signal changes with up to a couple of hundreds µstrain when you touch or move the RJ50 cables next to the module.

I could not find any logic in the drifting problem. Some sensors drift more then others. Others are drifting one day but do not the drift the other day.

I suppose the problem is the connection of the bridges to the amplifier with the RJ50 cables.

Does anybody has the same problems? Can you suggest a solution to this?

Is there a other way to connect strains to the 9237 without RJ50 cables?

[PeterAdelhardt]

Hi Gerd,

in this article you will find various tips to improve your strain gage signal quality. However, since you are using half bridge configuration, you should not see any temperature related drift effects. 

Are your connections properly shielded? 

On this page you find the various RJ50 Adapters you could use with the 9237.

There is another version of the 9237 using a 37pin DSUB connector, see also the manual.

Regards, 

Peter

[striegler_gerd]

Hi Peter,

the drift issue should not be temperature related cause we use half bridge completion right next to the active strain.

We are using the original RJ50 cables provided by NI and in another application we use the RJ50 to screw adapter (196809-01) and have the same problems.

Last week i tested the cabling with different connectors ( orignal RJ50 NI, Connector of Bürklin crimped with and without shield and the new 37 DSUB Module)

So far i get the best results using the DSUB connector. There is no disturbance when touching (moving) the cables next to the module and also no abnormal drift.

The test between the different RJ50 cables had different results. The original cable is very sensitive to movement of the cable next to the module. The self-crimped cables differ if you connect the shielding or not. I tested one cable without shielding and a ribbon cable which works fine. No effect when moving or shaking the cable. This cable is sensitive to noise though. When crimping a round cable with shield to the connector i got results similar to the orinal NI cable.

So far my results show me that i can use the DSUB connector for upcoming applications. But i found no solution for the problems with the RJ50 connection in the running applications.

Regards,

Gerd S.

[Matthew000001]

Hi Gerd,

I can confirm that I have had exactly the same issue regarding the 300 - 400 microstrain change from touching the cables, except I am using a 1/4 bridge and the NI 1/4 bridge completion accessory.

In my application, I was using the NI-9237 to collect data from several strain gages. So, I would plug in the first 4 cables to the 9237 and aquire data, and then un-plug these and plug in a new set of cables from an additional 4 strain gage channels.

(Note: Before I started testing, I had saved all of the bridge balancing and gain calibration constants to a file, and then appropriately set them prior to aquiring each set of 4 channels)

So, this was an application where I needed to keep touching the cables. My strain gages indicated strains on the order of 300 microstrain to about 2500 microstrain, so the effect from the cables was significant for me.

I currently don't have a work around, but I am going to have to come up with something. My initial thought is that this may not have been the best connector to have been chosen for the module. A good screw terminal such as the one on the completion accessory (and consequently on the new NI strain module) may have been a better choice by NI.

[jbuckmaster]

I can confirm drift problems with the above configuration as well...  Some more data points:

* We are using 'external excitation' for our 9237's and I can confirm that ALL channels drift when I heat up the power supplies with a shrink wrap gun.  The channels do NOT drift when I heat up the cRIO.  Still working on a solution to the power supply temp drift...

* Separate from the power supply issue, we observe ghost drift on Channel 2 of each of the 9237's we have plugged into the cRIO backplane.  Don't know why it only shows up on a single channel -- one of our engineers thinks it's a high frequency noise ailising problem - our host program runs at only 10 Hz while the 9237's are screaming at 2000Hz.

Any suggestions?

[rb_mi]

When any of the channels are balanced, the 0 reading drifts significantly. We have the 350 ohm quarter bridge strain gage is attached to the 9337 module via the 350 ohm bridge completion module and a RJ50 cable. When the completion module or cable is moved the signal becomes even more unstable. We have tested all 4 channels with 2 different completion modules - cable combinations, and 2 different strain gages. The strange part is that when a Lebow 1000lb load cell is attached to the NI9337 module and there was minimal zero drift once balanced.

To verify the condition of the gages I used an old Vishay conditioning amp and found there was no significant drift.

The module and software worked in the past with out issue which leads me to believe the module has an issue.