NI cDAQ 9188 related query

[Deepen81]

I am trying to acquire a 4-20mA signal from a water pressure transmitter installed in one of the water pipeline, from which the water is to be gradually drained out at the rate of constant 0.02 kg/cm^2 per hour. I am using cDAQ 9188 chassis and NI9208  20mA AI module for acquiring the pressure transmitter readings.

The issue is that, while a constant pressure reduction of 0.02kg/cm^2 per hour is to be captured in terms of 4-20mA output from the transmitter, the mA signal itself at any one point of time is continuously flickering (rippling) by more than 0.02kg/cm^2. So it makes overall measurement useless when that much variation was to be actually captured in a time-span of one hour.

So is there any method to conditioning that 4-20mA signal input received by NI9208 card, so that we do not get so much flickering and the 4-20mA signal captured remains fairly constant ?

Would be great to know at the earliest

[Deepen81]

sorry as I got the pressure unit wrong in my earlier post. kindly read them as Kg/m^2 instead of kg/cm^2

[santo_13 · Accepted solution]

You can implement a moving average of the data in the software layer, this way, any noisy readings are averaged out.

What is the sensitivity of your pressure sensor? kg/cm^2 -> mA conversion factor

Edit: you can also configure the DAQ to sample in high resolution mode using the method described in this article - https://knowledge.ni.com/KnowledgeArticleDetails?id=kA00Z000000P8jtSAC&l=en-US

[Deepen81]

Sir,

Thank you so much for your prompt response and the advice. Will try them and see. Might get back also, in case of any further doubts.

Pressure sensor sensitivity:

0 - 250 Bar = 4 - 20mA output @ accuracy = 0.5% FS max

Best regards,

[Deepen81]

Sir,

It was a great advice last time for some data acquisition problem that was shared then.

I have now another small labVIEW programming issue if you can help me out.

Problem Statement:

I am continuously acquiring real time 4-20mA data at the rate of 1 sample per second, set in the DAQ assistant. I would however like to have just 1 sample saved/written every 10 minutes in the "Write to Measurement File" excel sheet. I tried doing this using a case structure and a elapsed time block to control the structure. I have simply out write to measurement file inside the true statement part and wired the signal to it. The case is controlled using an elapsed time block set at elapsed time setting of 10 minutes. The false case is kept blank. The same however doesn't seem to work.

So can you kindly help me out how I can do this. Would be indeed great to hear from you.

Best regards,

[Cavarval]

If the previous answer responded your original question, you should mark is as solution. 🙂

I think this is worth a new thread, probably showing code snippets of what you are doing (maybe even the whole code) so it easier for someone in the community to spot if there are any issues there.

What you are doing sounds correct. You should be able to pull that trick so I suspect there's some issue with the code preventing it to work as expected. What does doesn't seem to work mean? Is the file empty? Does it have more data than expected?

[Deepen81]

Dear Cavaral,

Thank you for your prompt response.

Well, with regard to the previous solution advised, it actually didn't solve my issue but it did give us some additional Info. 

Our issue actually is w.r.t the very sensitively fluctuating 4-20 mA data being acquired by the NI 9208 & 9219, owing to which our intended purpose is not met.

We are actually draining a long water tunnel whose expected water pressure difference between the draining start date to end date, in a period of about 1 month is 10 kg/cm. This roughly calculated to a  pressure change of just 0.35kg/cm2 per day and 0.0143 kg/cm2 per hour.

We therefore intend to precisely measure & record the per hour change of just 0.0143kg/cm2 during the process using NI 9208 & 9219, utilizing 4-20mA Analog input feedback coming from the associated pressure transmitter in the tunnel. 

Our current labview logic (VI) and the DAQ set-up however is observed to be very sensitive with the readings fluctuating in every sample recorded. The instantaneous fluctuation most of the time are even greater than the pressure change (0.0143 kg/cm2), that is otherwise intended to be actually recorded/measured in a time span of 1 hour. So we see that unless the acquired 4-20mA data can be smoothened out (if possible through the logic) and made fairly constant upto 3rd significant digit after the decimal, the recording would not be that meaningful for our intended purpose to monitor the constant hourly pressure reduction and for a period of 1 month.

So, you may kindly suggest us if there are some solution to achieve this by way of some data conditioning option in labview.

The later query pertaining to data recording only at some selected time period in the "write to measurement file" somewhat could get solved on our own. We have used the elapsed time block and the case structure in this case. But still at some settings of longer elapsed time (say 1800 sec or 30 min) we are still encountering errors and the program getting terminated, which we are probably doubting due to loss of the acquired signal that is transferred through ethernet.

Nevertheless the VI file we have and a sample data recording TDMS file is enclosed for your reference, if you could suggest something for improvement. Would be great to hear.

Best regards

[Cavarval]

Please excuse the premature assessment.

I still think the moving average suggested before could work. Attached is a pretty simple way to implement a running average. It should work for you given you are acquiring on a point by point basis. I saved it for LabVIEW 2012 because that's the version of the one you sent. You will need the Full version at least in order to use it because it is based on the Mean PtByPt.vi.

The running average should take care of the noise impact, but there is also the fact that a higher performance device with a lower noise floor may also give you better results. Either way, test the approach VI based and let me know if it helps you.

Also, the VI returns double samples so you will loose some of the data embedded in the DAQ Assistant's dynamic data, but if you are not really using it, it will be just fine.