voltage signal from loadcell in usb 6008

[touheed]

hi, 

i am using usb 6008 to acquire the voltage signal from usb6008.[ loadcell is comp&expansion type with 0-200kg capacity and outputs 20mv at rated load ] so i used the amplifying circuit using texas INA128 instrument amplifier with gain of 100.

problem is iam able to read the voltage signals (in range 2to-2) but the the signal is smooth only at the times when load is varied and at fixed load it oscillates with small values along the actual reading.

Do i need to reduce the resolution of my DAQ? if yes please help me how to do it. if not please suggest me a way to get rid  of this problem 

[johnsold]

Reducing the resolution will only give the appearance of smoothing the signal and you might end up losing important information.

The first thing to do is to determine whether the oscillation you are seeing is due to a real, but small variation in the load or some kind of electrical interference being picked up by the wiring or your amplifier, or something else.

Please post your VI with some data showing both smooth and oscillating signals. Also tell us the sampling rate you are using.  What produces the load? How is it varied? What will you be doing with the data after you acquire it?

Lynn

[touheed]

thanks for ur reply. i have atteached the vi along with the graph by varying load. 
thing is i just want the output to be correct upto 2 decimal place (output limited to kg variation) but its showing me variation even for small gram changes. 

[johnsold]

You do not need the shift register and Insert Into Array for the graph. Just uncheck the "Clear data on each call" button in the popup for Build XY Graph. It will do the accumulation of dat for you.

It is really hard to tell much about noise from an image.  Please run your VI until you have data similar to that shown in the image. Stop the VI. From the Edit menu select Make Current Values Default. Save the VI (possibly Save As... under a different name). Then post that VI.  The actual data will be saved in the graphs and charts and we can examine it.

However, it appears that the noise is always there. It is harder to see during the rapid transistion regions but from the noisy signals I have looked at over the years I think it is persistent.

From the image it appears that you are using the single ended +/-10 V range with 11 bit resolution making 1 LSB = 9.8 mV.  If you switch to differential and the +/-2 V range you would have 12 bit resolution with 1 LSB = 0.98 mV.  The system noise is specified as 5 mV rms on the 10 V range and 0.5 mV on the 1 V range. Although that specification is at the maximum sampling rate and it appears that you are sampling much slower, the noise is comparable to the specification.

When you say 2 decimal places do you mean 2 significant digits? You also mention kg variation. 1 kg on a 200 kg compression/tension load cell is 1 part in 400 which is finer than 2 significant digits. Nine-bit resolution = 1 part in 256, so 1 part in 400 could be obtained by using a moving average with 4 elements from your 11-bit data. That will not remove all the noise but will reduce it somewhat.

Lynn

[touheed]

thanks for it. i have removed the shift register and yeah i mean 2 significant digits. coz when i calibrated the loadcell it gave me 0.01v change for every kg. so i want to make my measurement with this degree of precision not more than that i mean if its [ 0.0132v, 0.0474v, 0.3753 ] than i just want the daq to measure it as [ 0.01, 0.05, 0.38 resp] as i dont require the or no more need to measure the changes below that.

as said i have stopped the vi from while loop stop button and made the values as default and have attached the saved vi.

thanks in advance for ur precious reply

[touheed]

daq and amplification circuit

[johnsold]

I gave you some defective advice. I rarely use charts so I was unaware of this. Apparently Make Current Value Default does not save the data.  The only data which came with your VI is the data in the XY Graph.

When I look at that data, it appears that something really strange is happeing in your data acquisition.  I separated the X array and the Y array for the XY graph.  I would expect that the X array would be approximately a straight line corresponding to increasing time. This is what is in the data:

The straight segments near the baseline consist of the same three values repeated many times.

These are the first few X values. 

I have no explanation for how this is happening but it is consistent with the XY graph.  It may have something to do with how the DAQ Assistant is configured.  I do not have the DAQ VIs so I cannot see how it is set up.

Lynn

[Henrik_Volkers]

I recovered both channels and the 2k sample rate 🙂

Detail:

What you see is the resolution of your DAQ card togetther with the inherent noise. 🙂

One word to your bridge amplifier: You need to measure (or need to know quite well that it's stable) the bridge voltage (10V I assume)

Your INA would like to have some decoupling capacitors CLOSE (!) to the supply pins. Additional 10 Ohms 1µF||100nF decoupling is so common, that it isn't shown in the documentations 😉

The power spectrum show a clear peak at 50Hz, so there is room for your amplifier improvements.  (Well the 50Hz could already be in the bridge exitation 😉 )

What you can try: (Read with 20k and)* apply a low pass filter.

*BTW what is the highest frequency you want to measure?