Different Common Mode Voltages and Thermocouple Measurement

[mbenfer]

Hello All,

I have a compact DAQ system that I just ordered with some NI 9214 modules (not yet arrived). My application is that I am using these cards to accept type K thermocouples which are either bonded to a "hot" 120VAC line or floating. Since this module has bank isolation and not channel-to-channel isolation, I am wondering if I will have some measurement accuracy issues or potential safety issues (for the modules themselves). The link below shows the manual for this module. Since I am not too familiar with electronics/components, I am unsure as to whether or not I will have a problem.  

http://www.ni.com/pdf/manuals/375138b.pdf

The manual says it is possible to connect the reference voltage to the module common channel. Otherwise, the voltage reference for the module common is an average of the common mode voltages from each channel (i.e. some [TCs bonded to hot] with common mode voltage of 120V and some [floating TCs] with nominally 0V common mode voltage = somewhere between 0 and 120V).

Any assistance would be greatly appreciated.

Thanks,

Matt

[Paul-D]

Matt,

I want to first clarify your answers before providing a definitive answer, but I will give the information I have based on my understanding.

From my understanding, you have some TC values to be measured from lines referenced at 120V and 0V.  You want to know if you could take 120V and 0V referenced measurements on the same module, and additionally if 120V is safe.

Pages 13 and 14 of the manual mention the module features "16 thermocouple input channels and one autozero channel through a differential filter".  because all of the channels are sharing a common reference, you will not be able to use channels referenced at different levels with this module.  However, you said you were using multiple modules, so if you could group those channels going into the modules by reference level, your application can still work.

Page 36 of the manual states that the module can sustain 250Vrms of Channel to Ground continuously, which meets your 120V spec.

From my understanding of your problem, your application should work with this hardware.

Let us know if you have any further questions.

Regards,

[Paul-D]

Matt,

I also wanted to clarify something about your 120V measurements.  I assume you are tying these to wall power outlet source(s) providing 60Hz @ 120Vrms.  That being said, you will need to ensure that these 120V signals are in phase and of the same frequency.  This would be true if you had multiple TCs on a single power line, and should generally apply to multiple lines off the same power outlet circuit.  If your different lines are out of phase, then you will exceed the +/- 1.2 V common mode range specification and it may affect your measurements.  Additionally, there is some concern that if the lines are far enough out of sync or at different frequencies that they could exceed the 30V pin to pin voltage rating, which would be outside the range we officially spec as overvoltage protected from any one channel to the next.

Regards,

[mbenfer]

Thanks for the help. This is along the lines of what I was thinking. I was hoping that my arrangement would work, but alas I need to make some changes.

To clarify, all TCs are monitoring residential electrical devices wired in series (60hz, 120V), so they should be in the same phase and see the same voltage, minus a slight drop for resistance losses in wiring and the receptacles. The reason that some of the TCs are at a 0V reference is that they were electrically insulated from the conductor. This was due to the bonding method (hi temp cement) and TC placement. I do have the option to make all 0V reference TCs on the same modules and the 120V referenced TCs on the same modules, but that is not ideal for my situation without some major reconfigurations.

I want to bounce some ideas off of you to see if they would be options:

1) Would it be possible to wire each positive or each negative terminal in the TC terminal block together (i.e. Chromel wire for the positive terminal of type K TCs) in order to make every TC have the same common mode voltage? -- For my tests, it does not matter what common mode voltage the TC is (other than for hardware considerations). Also, this would have the added benefit-- if it works-- of protecting the modules in the event that a TC either becomes bonded or unbonds (i.e. changes its voltage reference) during the test. The tests are going to run for months, so this is a slight possibility.

2) Or would it be possible to wire the conductor that I am monitoring (60hz, 120V) to the common terminal in the terminal block?  ---- I am thinking that this will not work. 

[Paul-D · Accepted solution]

Matt,

First, you mentioned that having different measurements on different modules would be an issue, and although I don't think this is going to be an issue for you, given my explanation below, I just wanted to share with you that through the creation of a task, and the channel expansion capabilities of our cDAQ modules, that you could create tasks with measurement from different modules to be grouped together.

However, I don't think this is going to be necessary in your case, and you should able to group your connections however you want.

From what I am reading, with some measurements at 120V and some floating off these insulators, it looks like you should actually be able to take both these types of measurements on the same module.  I was unclear what you meant by floating in your first post.  

Per page 9 of the 9214 Manual "A connection to COM is not necessary for most applications. Specifically, leave COM unconnected for configurations in which all thermocouples are floating, all thermocouples are referenced to the same common-mode voltage", you will be able to connect these channels on the same module, as long as you do not connect to the COM.  This is because the floating channels will be pulled up to the shared reference.

Because there is concern we may still be exceeding the common mode voltage if all of your 120V sources don't match, you could programmatically check if there are common mode errors or not, our DAQmx Read Property Node offers the "Common Mode Range Error Channel Exist" or the "Common Mode Error Channels" properties, which allow you to check if there are errors, then read out the channels with the error.

Regards,

[mbenfer]

Thanks again. I think what worried me was the following from the 9124 manual:

"However, for configurations in which two or more thermocouples are referenced to different common-mode voltages, connecting COM according to the following guidelines improves common-mode rejection performance. Connect COM to chassis ground if the common-mode to chassis voltages are smaller than ±1.2 V. Otherwise, connect COM to a valid common-mode voltage reference that is within ±1.2 V of the common-mode voltages of all the thermocouple inputs." (page 9)

and 

"A valid common-mode voltage reference is a voltage that is within ±1.2 V of the common-mode voltage of the thermocouple. If COM is left floating, the internal common-mode voltage of the input circuitry is the average of all the inputs." (page 12)

So, I had thought that if, for instance, I have 8 TCs with 0V common mode voltage and 8 TCs with 120V common mode voltage, the average (internal reference) common mode voltage would be 60V. That would mean that the internal reference is more than ±1.2V different than the TC's common mode voltage (0V or 120V). But, from what you are saying, that ±1.2 V common mode voltage difference specification only applies when you attach an external common mode voltage reference. 

[Paul-D]

No problem.  Sounds like we have it all straight now.  Let us know if you have any further questions.

[mbenfer]

Paul,

I just received my NI 9214 cards and had one more question before I started wiring them up.

The overvoltage protection is +/- 30V between any two inputs. What qualifies as an input? Is this overvoltage between 

a) the + and - terminals for a single thermocouple 

b) any + or - terminal (irrespective of the thermocouple)

c) the differential voltages for each thermocouple.

Thanks,

Matt

[Paul-D]

Hi Matt,

a) Thermocouples output in the mV range, and as such our module is designed to measure in this range.  Exceeding +/-30V on this will damage your module.

b) Yes

c) Yes, this was our main concern before.  If channel 0 has a 30V difference with channel 2, it will damage the module.  We mentioned however that if the reference for your non-120V sources are floating, they will be pulled up to 120V reference, and this will be ok.

Simply put, in addition to the considerations of being careful of what you are hooking up to your module, you also need to be careful during wiring not to mis-wire, which could cause hazardous conditions.  You should probably wire up your module without the 120V source active to be sure that this issue does not arise.  Beyond simply damaging your module's ability to make readings, exceeding these voltage ranges by a significant amount can present a risk to you and others.  I recommend checking out Pages 2-5 of  the NI 9214 with NI TB-9214 Operating Instructions and Specifications so that you are aware of these hazards.