simultaneous Fluke 8845 and M series card measurement

[dkuhinek]

Hi, I need some help and would be very grateful for any idea.
I need to make simultaneous measurement with Fluke 8845 and M series card (6221 or 6281) of equal DC voltage. Fluke is connected using serial port (baud rate 115200) and I am using drivers found on NI web page. Examples inside driver work fine for DC voltage but results do not have time base. So I decided to output trigger signals from M card to Fluke.
I made a vi that generates rectangular voltage with 1Hz frequency on DAC0 of DAQ card and measures voltage on one DAQ card input and 8845. Signal from DAC0 is connected to PFI7 (AI sample rate) and to 8845 trigger input, and is used for timing/triggering. As vi runs data is collected but when I open “lvm” file it is obvious that 8845 time is lagging. For 800 second measurement this difference is about 1,7 seconds. Is there some better way of data acquisition with 8845, that will not produce this problem? Theoretically triggering should do the job.

Another application I need should take simultaneous measurement of AC voltage with Fluke and DC voltage with card with sample rate of 1S/s. If I change DC to AC voltage configuration, the shortest period between samples is 4 seconds. When using Fluke View Forms software, the function of logging with delay of 1s works perfectly. So faster acquisition is possible but something is not well optimized in my vi. I have tried few different implementations but for now cannot find solution.

Thanks again for any help or idea.

Dalibor

P.S. A vi is in attachment.

[Dennis_Knutson · Accepted solution]

Since the Fluke is not returning any time information, what are you basing this 'lag' on? There will be nothing in the lvm file that is valid for the dynamic data that you are creating for the Fluke.

Also double check that you are configuring the Fluke correctly for an external trigger and that it's not just triggering on the initiate function. I also don't see where you are configuring the DAQmx task to acquire on a trigger. Have you looked at the shipping examples?

Message Edited by Dennis Knutson on 03-05-2009 08:09 AM

[dkuhinek]

The lag is evident in recorded data in lvm file which makes me doubt if the VI works fine. The time for fluke is created inside VI. Also when changing measured value it seems like fluke is making one measurement sooner than M card.

Also I have tried to make another VI so that M card is in charge. I have routed M card sample rate out and connected PFI7 to fluke trigger input.

When this VI is run data are recorded without previously mentioned lag but it seems that fluke is taking one measurement sooner. Delay function didnt compensate for that. The vi is attached.

I was hoping that when Fluke receives trigger, it will make measurement and send it to labview through serial port. I dont need its time base

but have to be sure that the measurements received are simultaneous. It would be nice to see an example of measurement with external triggering, in a loop, with recording. Since there are no detailed explanations of fluke driver functions, its hard to optimize data acquisition.

Maybe my task cannot be done by serial communication?

Thanks

Dalibor

[Dennis_Knutson]

I still don't understand how you are calculating the lag. In the VI you posted, there is no time for the Fluke acquisiton being passed to the Write to Measurement File. The only thing from the Fluke that you are passing is the measurement value. That is the single orange wire.

[dkuhinek]

Loop is repeating every second (defined by mx read sample rate) and I am assuming that in the process of data merging time is added to Fluke sample value. This seems true, as can be seen in xls file from one set of measurement where X value of second measurement increases from zero up and delta t is equal to one. The same measurement is shown in attached picture. Both graphs have time that is increasing.

Just as I said, this will be irrelevant if I can be sure that the acquisition is started simultaneously. Then I can save only one time column.

From further research I can say that Fluke is making measurements with sample clock, but cards conversion clock is delayed by 3 ticks (which is probably reason for different values in the picture). I can lower this to 2. If it is less than two, I don't get any measurements on my graph.

I think I will have to use conversion clock as trigger for Fluke but already had some difficulties since this clock is active low opposing to sample clock which is active high. Can the clock be inverted before routing out?

If you see flaw in my logic please let me know. Thanks.

Dalibor

[Dennis_Knutson]

Is is absolutely not true that any time information is being added to the Fluke data. It is impossible to add time data to the Fluke data.  I don't know how you can think that. You are passing a single DBL to the Write to Measurement file. This is the measured value only. A time does not just magically jump to the Write to Measurement file. It has to be through a wire that you pass to it. Any time information associated with the Fluke Measurment is some sort of default being generated inside the Write to Measurement file and is completely bogus.