Signal Conditioning

cancel
Showing results for 
Search instead for 
Did you mean: 

NI 9205 differential voltage measurement

I am using a NI 9205 in differential mode measuring voltage, plugged into a EtherCat chassis and then to a PXI chassis.

 

If I have say 5V on channel 1, and nothing connected to channels 2 through 16, channel 1 will read 5V, but so will channels 2 through 16 without a Voltage present.

 

All of the wiring up to the module checks out, no shorts anywhere.

 

The main problem would arise if say I has a sensor connected to channel 2, and for some reason had a bad connection to that sensor, I may never know it, as I would be recording the

voltage that is on channel 1, on channel 2.

 

Is this the normal operation for the NI 9205?

0 Kudos
Message 1 of 8
(9,416 Views)

This is expected.  The NI 9205 module is a multiplexed module.  This means there is only one ADC and a multiplexer connects each channel to the ADC in turn.

 

As a result, after taking a measurement a certain amount of charge is built up on the multiplexer from the channel that was being measured.  Normally, when you switch to the next channel this charge dissipates into the next channel during the settling time.  However, in the case you mention, the effective impedence of the next channel is infinite and as a result the charge can't dissipate during the settling time.  The end result is that you see the same signal on the second channel that you did on the first.  This effect is refered to as Ghosting.  You can learn more here: How Do I Eliminate Ghosting From My Measurements?

 

This is unfortunately a standard trait of multiplexed analog input modules.  Are you input signals floating or grounded?  If they are grounded, it's likely that there will still be a sufficient discharge path that you will at least see a signifigantly reduced signal in the events of one of the terminals having a bad connection.  If they are floating and you use the recommended bias resistors to COM, there will also be a discharge path for the built up charge and you should still be able to detect if there is an issue with a channel.

 

Hope this helps,

Seth B.
Principal Test Engineer | National Instruments
Certified LabVIEW Architect
Certified TestStand Architect
0 Kudos
Message 2 of 8
(9,407 Views)

How large of a resistor do you suggest?

0 Kudos
Message 3 of 8
(9,386 Views)

For floating sources, we recommend between 10k and 100k in the Field Wiring and Noise Considerations for Analog Signals article.  The resistor should be between both the positive lead to ground and the negative lead to ground.  The specific value depends on your source. The larger the resistor the greater the input offset error (due to input bias current), but the smaller the resistor the more the source is loaded down (may cause some gain error).  Both should be quite small, but you may want to try a few values to see what gives you the best result.

 

Regards,

Seth B.
Principal Test Engineer | National Instruments
Certified LabVIEW Architect
Certified TestStand Architect
0 Kudos
Message 4 of 8
(9,382 Views)

In your 1st post, you recommend floating resistors to com.  In your second, it is to ground.  Can you please clarify?  Also, do we allow the ADC common to float or tie that to ground as well?

0 Kudos
Message 5 of 8
(9,329 Views)

My apologies, I misspoke in my previous post.  The bias resistors should be to COM on this module.

 

The COM terminal (essentially the common reference for the isolated front end) shouldn't be left floating.  The module is only specced for 60VDC isolation from earth ground (250V for the Spring Terminal version).  This can be seen on pages 27 and 28 of the operating specifications.  COM should be tied to something in order to ensure that the front end of the device does not float more than 60VDC away from earth ground.  If you don't need to measure signals more than 10VDC from earth ground, the most effective method is to tie the COM terminal to the cRIO chassis ground.

 

Additionally, no signals to this module should exceed 30V from any other signal on the front end (channel to channel maximum voltage, page 27 of the specifications).

Seth B.
Principal Test Engineer | National Instruments
Certified LabVIEW Architect
Certified TestStand Architect
0 Kudos
Message 6 of 8
(9,312 Views)

I want to know how to define  floating signal and grounded signal. if the low side of the signal tied to the zero of the circut but not the actual groud, is it can be called ground signal? can the com be tied to the zero directly? by the way the zero of the circut is defined by the source,from the Switching power supply.

0 Kudos
Message 7 of 8
(9,011 Views)

I would recommend taking a look at our Field Wiring Guide for an explanation on the difference between a grounded and floating signal.  You may also find the article about Ground Loops and Returns informative which is the sixth section on our page on Analog Fundamentals.

 

David A

0 Kudos
Message 8 of 8
(8,991 Views)