Baseline Shifting from Analog inputs on the T7

6 posts / 0 new
Last post
fieryfire
fieryfire's picture
Baseline Shifting from Analog inputs on the T7

Over the course of the tests with the Labjack T7 with the architecture of strain sensor->wheatstone bridge->amplifier->filter->Multiplexer->analog input of T7. After having acquired data from over 100 captures we observe that there is a shift in the baseline of the signal.

1)For Ex: When our strain sensors are in unexcited state, the baseline reads Voltage 'x', if the sensors are excited by a value of 'y', then we expected the value to peak to 'x+y' and upon release of the excitation, the value from the sensors come back to 'x' , but the odd behaviour we see in our acquisition system is that the values once excited to 'x+y' does not come back to x but instead to an 'x+z' where z<y.

 

We have not observed this phenomenon before and from our team discussions it does not appear to be a mechanical phenomenon from the sensors but more from a remanent effect in the analog chain.

Below you will find a live capture of a signal at a baseline of -1330.8uE, but after the excitation, it shifted to a baseline of -1307.2uE, a shift in the baseline of 23.6uE.

 

 

2)

Another strange pattern that we observed is the oscillation pattern of the baseline shift. From the picture below, you will find that after each excitation of a sensor, the baseline shifts a alue of +z and -z even though the are excited from the same direction. Once again, this does not appear to be a mechanical property or behaviour of the sensors embedded to the mat.
 

 

Could you kindly advise us, if you have observed such a pattern in the past or if not, what could be the possible causes of such an oscillation

 

Thanks,

Vivin

LabJack Support
LabJack Support's picture
I don't see an attachment,

I don't see an attachment, but anyway for it to have meaning to us we need to see raw volts not scaled values.

What are you using for software and do you see the same problem with LJLogM?

What are you using for excitation?  Are you using an AIN channel to measure the excitation in real time to confirm it does not change when you see the problem?

Do you see the problem if you only have 1 or 2 channels connected, and nothing else except power/comm?

 

 

fieryfire
fieryfire's picture
Hello, Labjack, sorry for the

Hello, Labjack, sorry for the delay. I just noticed that i was not able to upload the images. I have attached them now once again.
The LJLogM and LJStreamM has the same issue. We are using the python code on your website to extract the data to our backend software. 

Our sensors are strain gauge sensors, so we use mechanical load for excitation.. And indeed we are using the AIN channel in real time.

We see the same problem even with 1 channel connected
 

LabJack Support
LabJack Support's picture
You shift attachment shows a

You shift attachment shows a change of 23 uE.  What is this in volts?

What are you using for excitation?  Use another analog input to acquire your excitation voltage at the same time so you can see if it has shifted also.

For one channel experiencing shift remove your signal and instead connect a wire from GND.  Does this channel still experience shift?

 

fieryfire
fieryfire's picture
23uE corresponds to 0.0575V.

23uE corresponds to 0.0575V. Infact the scale from uE to Voltage is (/400).

We use mechanical application for the excitation. In fact we use 76 channels via the multplexer ADG408 to be able to simultaneosuly capture from 76 channels. And we see the same for ALL channels.

If we connect the channel to ground it seems to go to 0V. and there isnt a shift. It seems to be occuring only with the strain sensors

LabJack Support
LabJack Support's picture
So 23uE is 58 mV?  That seems

So 23uE is 58 mV?  That seems like a lot.  Often the entire span of a bridge circuit is only 10 mV or so.

I'm assuming you have bridge circuits and you use an excitation voltage for those bridge circuits?  That is the excitation I am talking about.  You should acquire that with another channel so you can see if it is changing.

Try connecting 1 bridge that is just sitting on your desk and not attached to the test specimen such that this bridge is not subjected to any strain.  Then do your baseline shift experiment again.  Do the readings from this bridge change at all or stay totally flat?