Increasing Interchannel Delay on DAQExpress / USB 6002

[BertrandBru]

Hello,

I purchased a NI USB 6002 a couple a days ago as I need to record the analogue signal coming from two S-Beam load cell sensors and one torque sensor. All sensors will deliver signal between +/- 10V.  Although I don't have the sensors yet, I was doing some testing using an accelerometer, so I could play with the software, in this case DAQExpress.

Although I understand that Labview offers for more option, I decided to go for the DAQExpress software (free) as my requirements are very simple.  I just want to record the analogue signal coming from different sources for 10 seconds at a sampling rate of 500Hz. 

I have been able to create a VI that allow me to do exactly that.  I attached a screen shot of the VI to show you how it works.  You will also note that the code is doing a couple of other things like switching one digital output high after 2 seconds, and low after 7 seconds.

All the above works as I need, except that there is a large crosstalk recorded between the different analogue channels.  Below is an example. A dual accelerometer was connected to ai 1 an to ai 2.  When streaming the voltage individually (both connected on the NI USB 6002, but only reading one channel), the voltage is 1.9704V for ai1 and 1.17148 for ai4.  When reading both channels at the same time, the readings are 1.4444V for ai1, and 1.17106V for ai4.  Using a voltmeter, the voltage stayed the same as when read individually (1.97V for ai1 and 1.17 for ai4).

I looked into the community forum and found out it is not uncommon when using a multiplexed system, which I believe the NI USB 6002 is. (https://knowledge.ni.com/KnowledgeArticleDetails?id=kA03q000000YHy6CAG&l=en-GB). Long story short, the capacitor has enough time to discharge when the multiplexer switches from one channel to the next one and show remnants from the previously scanned channel when the measurement is taken by the ADC.

It seems that one solution is to increase the interchannel delay. This should give enough time to the capacitor to discharge.  The procedure is explained to increase the interchannel delay is explained for Labview (see link):

https://knowledge.ni.com/KnowledgeArticleDetails?id=kA00Z0000019NFQSA2&l=en-GB

However, this option does not seem to be available for DAQExpress (see picture below). There is no drop menu: More>>AI Convert>>Rate as specify in the link.

Is it a software limitation, a hardware limitation or is the procedure different with DAQExpress?

Some help would be appreciated.

Thank you

Bertrand

[BertrandBru]

with the VI attached.

[Frozen]

This knowledge article explains some other options that you  might explore as well.

https://knowledge.ni.com/KnowledgeArticleDetails?id=kA03q000000YHy6CAG&l=en-CA

[BertrandBru]

Thank you for your reply.  I really appreciate it. Although I only mentioned one of the options, I read the other option but I thought that changing the interchannel time was the easiest to implement.  

From what I understand, there are about 3 options, the other are more recommandations (which I am taking on board):

• Reduce sampling rate and increase interchannel delay
• Implement a voltage follower or buffer circuit to decrease source impedance to less than 1 kΩ.
• Read the desired channel twice by implementing a "dummy channel".

As mentioned in my email, I am using DAQExpress, I have not been able to find how to create a dummy channel. I would appreciate it is some could lead me towards the right direction.

I still have not found a way to the increase interchannel delay using DAQExpress. Do you know if it is possible using DAQExpress and the NI USB 6002?

Unfortunately, I am not an electronic engineer, and implementing a voltage follower is above my skills.

However, I tried to understand what I was doing.  I found out from the datasheet of the acceleraometer that it has an output impedance of 32kΩ, which is well above the recommendations. I then tried to use a different set of transducers with lower source impedance. I selected 2 diffrerent battery pack, one with 3 x 1.5V battery, and one with 6 x 1.5V battery.  I cacluated their output impdeance, and there are well under the recommended 1kΩ:

• pack 1: Open-circuit voltage: 4.35V, Load resistance: 100Ω, Loaded circuit voltage: 4.14V --> impedance: 5Ω
• pack 2: Open-circuit voltage: 7.91V, Load resistance: 100Ω, Loaded circuit voltage: 7.35V --> impedance: 7.6Ω

I then connected the 4.35V to channel a0 and the 7.91V to a1.  The same crosstalk effect is observed. Do you have idea why?

Last thought, does Ni sell DA system that offers similar spec than the NI USB 6002 (8 AI +/-10V, and a few digital output) but with simultaneous sampling instead of multiplexed.

Thank you in advance.

[Frozen · Accepted solution]

Yes, NI sells all sorts of DAQ devices. I think pretty much all of them have the +-10V range. However, the USB6002 is one of the cheapest ones. They go up in price substantially. I don't know which one does simultaneous sample off hand, and that is something that google (or an NI rep) can help you with.

I would recommend setting up a dummy channel. This is actually pretty simple to do. Looks like you figured out how to had the "real" channels so adding some extra ones would be the same way.

Simply wire up AI2 and AI3 to the same sensor and AI4 and AI5 to the same sensor. (you can use different analog in channels if you want)
Then you can measure all four  channels in a "row" AI2, AI3,AI4, and AI5 (in that order)

Then you ignore (throw away) AI2 and AI4. (these channels pre-charge the ADC with the correct voltage so the subsequent channel will have a correct measurement)

Hope this helps

[BertrandBru · Accepted solution]

Thanks for explaining how dummy channel works. It all make sense. I will try on Monday and fingers crossed, this trick will be enough!

Thanks a lot.

---

@Frozen wrote:

Yes, NI sells all sorts of DAQ devices. I think pretty much all of them have the +-10V range. However, the USB6002 is one of the cheapest ones. They go up in price substantially. I don't know which one does simultaneous sample off hand, and that is something that google (or an NI rep) can help you with.

I would recommend setting up a dummy channel. This is actually pretty simple to do. Looks like you figured out how to had the "real" channels so adding some extra ones would be the same way.

Simply wire up AI2 and AI3 to the same sensor and AI4 and AI5 to the same sensor. (you can use different analog in channels if you want)
Then you can measure all four  channels in a "row" AI2, AI3,AI4, and AI5 (in that order)

Then you ignore (throw away) AI2 and AI4. (these channels pre-charge the ADC with the correct voltage so the subsequent channel will have a correct measurement)

Hope this helps

---