Bad measured data at high frequency

[TarPista]

Hi all!

I'm using PCI-6251 and receiving a strange problem. I'm creating a task for reading analog input on two channels. The first one is for measuring current the second one is for voltage.Here are the function calls in order:

DAQmxCreateTask(..);

DAQmxCreateAICurrentChan(...);

DAQmxCreateAIVoltageChan(...);

DAQmxCfgSampClkTiming(...,SampleRate,...);

DAQmxStartTask(..);

(...)

DAQmxReadAnalogF64(...);

The problem is the following. If the SampleRate is lower than (approximately) 12800 the measure data on the Current channel is right (the measuerd current should be about 5 Ampere and it is). But as I raise the SampleRate frequency the measured current is shifting "upway" as much as I'veraised the requency.

I attach 2 jpg files showing the data, the first one is the well measured data (at 10240Hz), the other one (at double frequency, 20480Hz) is the wrong. Any idea for the reason?

By the way I've tried to measure voltage also on the first channel, and calculating the current manualy: got same results. The other bug I've realized, that if I change the order of the channelcreating instructions (call DAQmxCreateAIVoltageChan first, after it DAQmxCreateAICurrentChan) I also get wrong datas on the current channel.

Are these problems caused by a bug of the card's driver? I dont know which is the newest driver, as I've found out my one only uses the C:\WINDOWS\System32\drivers\nipalks.sys which version is 1.9.0f0.

Thanks for your help!

[K_Dinnes · Accepted solution]

The newest DAQmx driver is the verison 8.3

But I don't think that this is caused by the driver.

The behaviour you explained above can occur, when you acquire multiple channels on a multiplexed DAQ device and when the output impedance of the signal source you connect to the AI channels is very high.

In this case there is no chance for the input of the AD converter to unload fast enough.

If my assumption is true you should get the right values if you only acquire one channel.

The KB below provides some work arounds for this issue.

https://knowledge.ni.com/KnowledgeArticleDetails?id=kA00Z000000P84RSAS&l=en-US

Thanks

Klaus Dinnes

NI Applications Engineer

[TarPista]

Thank you for your answer! Your assumption was right, we've built an amplifier only by now, and it seems to work fine!

thx again,

Zsolt Selmeczi