Multifunction DAQ
Analog out grounding question
Do you know the Field Wiring and Noise Considerations for Analog Signals? Have a look at Table 1, and let us know if that answers your questions!
The correct wiring of the NI-USB's output to your temperature controller of course on the temperature controller as well. Could you please share it's model number if the guide I linked does not answer your question?
From what I understand now, there are two cases: The signal from the temperature controller towards the USB-6343, and the signal in the opposite direction.
controller -> USB-6343
Yes, I'd also say you have a grounded signal source and have a NRSE configuration. From your posting I understand that you already set up the USB-6343 and your wiring correctly.
USB-6343 -> controller
Unfortunately, your temperature controller does neither have a differential input nor a "fix" for this in terms of a sense type input. As your temperature controller is on a different ground level, it is impossible to send a single-ended signal to it without the possibility of creating a ground loop. Basically you have the not recommended "grounded signal source, RSE" case in the earlier-referenced Table 1 in front of you.
That's another +1 from me for differential signals.
There are some possible solutions to this problem. Have you considered using
- a different controller,
- a signal level converter (isolating your usb-6343's output from the controllers input), or using
- the USB-6343's power adapter to supply the controller as well?
Regarding your other question: According to USB-6343 Screw Terminal Pinout, "All ground references—AI GND, AO GND, and D GND—are connected on the device, they are connected by small traces to reduce crosstalk between subsystems. Each ground has a slight difference in potential." So yes, it is common to use the AI GND / AO GND when the hardware devices are on the same GND plane. If not, one uses AI Sense instead of AI GND for the Inputs. Unfortunately, your temperature controller does not have a similar sense input.
Thanks guys for your feedback.
What i also can do is to power both the temp.controller and the usb-6343 from the same linear power supply as mentioned, so they have a common "-" ground (not connected to mains earth).
Then we have all ground levels on the same level ...
The manual says:
Single-ended input connections can be used when all input signals meet the following criteria.
- Input signals are high level (greater than 1 V)
- Signal cabling is short and travels through a noise-free environment or is properly shielded
- All input signals can share a common reference signal at the source
Differential connections should be used when any of the above criteria are violated.
Well, all input signals and output signal are greater then 1V, so according the manual i should use RSE or NRSE.
In the meantime i did some measurements with normal cables (not shielded ... will do that later) and got some good results with NRSE.without the bias resistors.
I will do the same measurement later with the bias resistors connected, to see if it makes some differences.
Later i will do the same measurement in RSE mode.
Above is a 16 hour temperature measurement. The blue line is the set temperature (from AO of the usb-6343) and the orange is the actual measured temperature (some pikes).
@Henrik_Volkers wrote:
The additional GND connections mentioned in my first post (spare 4th paire) would create a gnd loop, but with an old network cable (4 pairs twisted and shielded) not that critical.
The critical GND loop is the controller power gnd to controller GND to 6343 gnd to 6343 power gnd ..
That's correct, except the "not that critical" which is always something up for discussion. Eric wrote that the voltage for both devices comes from different power supplies. If they are connected to "the same" electrical outlet Eric's set-up should work. He could also have additional wiring from controller-PSU GND to NI-PSU GND to make sure a possible high-current flows "around" the devices and not through them via the pins you mentioned.
I saw a lot of cases where ground-loops between devices even connected to different building power systems in labs with big machinery around did work for years, however I would never recommend someone to do such a thing. Only Eric knows the bigger picture of his setup; if he feels safe about it (and compared the PSUs GND levels to each other), he could wire the devices. Taking into account that the temperature controller is obviously meant to be used in such a way (no differential pins provided), it should survive small currents coming from a ground loop and its measurements should be insensitive to voltage changes from a ground loop. At least I hope...
@ikaiser
... Only Eric knows the bigger picture of his setup; if he feels safe about it (and compared the PSUs GND levels to each other), he could wire the devices. Taking into account that the temperature controller is obviously meant to be used in such a way (no differential pins provided), it should survive small currents coming from a ground loop and its measurements should be insensitive to voltage changes from a ground loop. At least I hope...
Ok, first ... i typed a lot of information this morning .... but it seams my posting dissapears !!!
First of all, thanks guys for you information !
First thing i will do is to connect both Temp.Controller and the USB-6343 to the same 12V linear power supply to have a common "-", so they will have a "floating" dc output with a common ground (should be RSE or NSRE connected because in and output voltages are greater then 1V, according the manual).
The temp controller is a very sensitive device which responses in mV.
In the meantime i did a 16 hour temperature measurement in NSRE without the bias resistors connected.
The blue line is the set temperature (EXT SET T) and the orange line is the actual measured temperature (ACT T MON).
I see some little spikes, but all where connected without shielded cables.
