bad signal output

[kiarash]

Hello

for my research I have made a sensor and its circuit for signal conditioning but when I 

wanna use    NI USB-6008  for data logging I found out that the signal is somehow strange.

the sensor is working good because I can see in the graph if I simulate the sensor

the graph shows that puls but the appriance of the graph is not what I expect, I mean there are too many

noise (I think) and the gerenal apprianece of the graph is poor ( I have changed

different rate and samplying points but still it s poor in shap

I have attached the graph and my vi. hope it helps

Moreover Someone told me I should use something like avraging the signals

to avaoid those too many fluctuations which I know they are not right

would you please hepl me

Thank you in advanced

Kiarash

UBC

[tbd]

Hi Kiarash,

      It's hard to tell if there's really a noise problem.  It looks like your DAQ device acquired only four unique values in the 1000 samples of your example - this isn't intrinsically bad (in fact it might be very good).  Your sensor is a big unknown here - do you know what variation to expect from it?  If you have a digital-scope and take a one-second sweep, does it yield similar results?

Before looking at your data, my first guess was that your DAQ wasn't properly grounded.  DAQ Assistant is configured for "differential" inputs, so you'll need two wires for each signal, and a missing ground will definately result in noise, but, having looked at your graph, I doubt that's a problem here!

Maybe it's simply a matter of perspective?  "Zoom-out" a bit - increase the upper-limit and decrease the lower-lmit of your graph.

Cheers!

[johnsold]

I agree with tbd that you do not have much of a noise problem. 10 mV peak to peak on a 1.4 V signal is >40 dB signal to noise ratio. What kind of sensor and signal conditioning circuit are you using? Has it been verified to work independently from the USB6009 and LabVIEW? What is the sensitivity of the sensor? How much process variation do you have?

Lynn

[kiarash]

Hello

Thanks for your information. the sensor is a hot wire anemometer for measuring air velocity

the resistance of a thin wire will change by tempraure and if the air flows over the wire

beacuse of the cooling effect the resistance will change so based on a wheatson bridge

I can measure the voltage changes (in the circuit I have jut amplified the signal)

I have connected the 2 output wire to the abaloge part of the Data Aqc. and I also grounded

the NI-USB 6008 simply by wireing to the ground of the plug (am I right?) but still I get the sam signal

I attached 2 graph one of them (1) is when I blow air over the sensor so you can see

that there is a peak but as you can see in the graph this peak in not smooth

I think it shoud be smooth (like Sine) and in the second graph it is when ther is no ait flowing

so do you think I can modify the signals (I mean the 1 graph somehow to see a smooth signal)?

thanks for your time

I really apprecite your help

I am exhausted with this part of my project

Regards

Kiarash

UBC

[johnsold]

Typically hot wire anemometers use a feedback circuit to maintain the wire at a constant temperature and generate an output which represents the change in resistance with airflow. Is that the way yours is connected? What is the air velocity range for which the device is calibrated? What is the sensitivity of the device? Is it compensated for changes in air temperature and humidity? (When you blow on it, your breath is probably warmer and more humid than the ambient conditions.)

Lynn

[kiarash]

Hello

yes I have used the feedback circuit for this home made hot wire as you mentioned. the air velocity range would be from 1 to 30 m/s but I have not calibrated that becasue I have just made that at first I wantted to know if the output signal is OK or not. and about the temperature and humidity compensation I should say no I do not have in my circuit . you are right but as I am a chemical engineeri I have no experience to make complecated circuit.

thanks for you time, I really apprecite your help

Regards

Kiarash

[RavensFan]

It looks like you are getting a baseline with spikes of about +/- 5 mV.  What is the DAQ range set at?  Is it a 10V range?  If so 5mV error out of 10 V is .05% which is pretty small.  Since there doesn't seem like there are any intermediate points, maybe it is such digitilization error of +/- 1 or 2 bits.  An NI USB-6008 is 12 bit resolution, that's 1 in 4096.  A 5 mV step out of 10V is 1 in 2000.  So your steps are only 2 bits.  If you wanted to do better, you would need a device with a higher resolution DAQ.