Ground connections - PXI Express

[krivan]

Hi,

I have designed a board for pxie-6358 cards but I guess I got the ground connections wrong...

On the PCB board I connected all AI X- and AI GND connections together and these connection points connect further to analog ground. The DGND terminals are also connected together and connected further to digital ground. The board receives its vdd [3.3V] and ground from an Agilent E3631A power supply, the analog and digital ground points of the board are connected to the power supply, creating basically a common ground point. And this common ground point is connected to the green terminal in the middle of the power supply that is the ground of the electrical network.

To supply an analog input signal [square waveform] I use a function generator that is grounded to the electrical network.

As far as I know the PXIe-6358 does only differential measurements so I use the attached vi for analog data acquisition. The problem I experience is that if I fire up the board and start the VI, I see a square wave but it is constantly sinking until -5 V and reaching that the singal becames constant DC at -5V.

However if put a wire into AI0+ and another into AI0- and I connect it up with the function generator running the same VI I get a very clear and correct square wave signal between zero and 2V [I randomly set 2V as amplitude].

Can somebody please point out where I did something wrong.

If I redesigned the board, could I resolve all the ground issues if I took the +5V terminal of the PXIe-6358 and would power the circuit from there creating the necessary 3.3V by resistors? The circuit consumes not more than 40 mA. I would connect then all AI X-, AI GND, AOGND, DGND terminals together because they are supposedly all joined up inside the PXIe-1073...

Does it sounds to be a correct idea?

Thanks for the suggestions,

Krivan

[alpha1]

Hi Krivan,

In differential mode, the AI X- channels are not directly connected to AI GND. In differential mode, channels are allowed to float from the GND level on the board. However, this should not cause your measurement to sink to -5 V, but rather just be offset by a small amount. Do you have a full wiring diagram that we can check?

[tbob]

As Imtiaz said, do not connect AI- to ground.  The function generator output should also not be connected to any ground.  Connect generator output hi to AI+ and generator output lo to AI-.  Also, I have never connected anything to the green terminal of the power supply.  This is earth ground, the third prong (the round one) on a standard AC plug.  You do not have to ground your circuit to earth ground.  If you have other instruments elsewhere connected to earth ground, with the instrument inputs or outputs connected to the circuit, you could introduce common mode voltages because of the resistance between the two earth grounds.  Unless you absolutely need earth ground, don't use it.  The power supply negative terminal takes care of the return path for the positive terminal.

Actually, the word "ground" is a misnomer.  It should be called "return".  The power supply puts out voltage and current on the positive side and that current returns to the negative terminal (well, techincally the electrons leave from the negative and travel to the positive, but lets not go there).  It does not have to be grounded.  All of your AI, analog, digital "grounds" are actually return paths for the signals.  They should all tie togethert at some common point, most likely the negative terminal of the power supply.  Maybe "common" would be a better word than "return".  I know everyone uses the word "ground", so do I.  But I know it isn't really an earth ground.

[krivan]

Hi,

thank you for the helpful replies so far!

Fair eonugh, I won't use earth ground however I am a bit confused now regarding the connection of AI X-...

As I wrote when I tried to measure the square wave directly I did exactly what tbob said: "Connect generator output hi to AI+ and generator output lo to AI-.". It worked properly, no problem.

Let's consider the example when I would like to measure voltage signals on the board. In this case no function generator is present, only one power supply delivering GND, 3.3V, 5V to the PCB board []. 

On the board I need to measure the positive voltage signals with respect to a reference point, a ground [or "return" if you like]. Doing differential measurements I need to connect the signals to AI X+ and to make the reference point AI X- should be then connected to ground. As far as I know the PXIe-6358 calculates the differential signals from the relation of AI X+ and AI X-, so as a consequence AI X- need to be connected to ground.

By the way on the redesigned board AI X-, AI GND, AOGND, DGND terminals are tied together and go to the common ground point. Imitaz said that "In differential mode, the AI X- channels are not directly connected to AI GND". However as I see the situation AI X- should be connected to the common ground, it can of course be that AI GND shouldn't be connected to the common ground point but if it's true then why?

[tbob]

You are attempting to make a single-ended measurement.  You cannot use a differential input to do this.  Single-ended:  only one line supplies the signal with respect to ground.  Differential: two lines supply the signal and both are floating above ground.  You cannot use the PXIe-6358 to measure a single ended signal.  It has differential inputs only.  You will have to buy a different DAQ card, one that has single ended analog inputs.  You could convert the single ended signal to a differential one with special op-amps, but your best bet would be to buy a regular DAQ card with analog inputs that are not differential.  The PXIe-6341 has 16 analog inputs (single-ended) plus some other functions.

[krivan]

Hi tbob,

I still don't see why my solution would be wrong. The PXIe-6358 does have only differential inputs, true. However as differential it calculates the difference between the two terminals no matter what are the signals on them. Even if I connect the signal of interest to AI X+ and the ground to AI X- it should work.

As you also suggested and as I wrote it in the previous posts I connected the positive terminal of the function generator to AI 0+, the negative to AI 0- and I saw the expected clear square wave signal in LabVIEW. The negative connection of the function generator is in [internal] connection with the ground of the power system.

The X series manual also says that it is possible to measure "Ground referenced Signal Sources" in differential mode, page 4-46 - Table 4-6 right side.

According to the table if the board gets its ground from the power supply and AI GND is not connected to the common ground point but AI X-, DGND, etc. are connected to that, the measurement as I see should work. Or am I wrong?

[tbob]

---

@krivan wrote:

Hi tbob,

According to the table if the board gets its ground from the power supply and AI GND is not connected to the common ground point but AI X-, DGND, etc. are connected to that, the measurement as I see should work. Or am I wrong?

---

In your original post, didn't you say that you connected all grounds together, including the poser supply negative?  The statement above states that the power supply negative is not connected to the signal grounds. 

It could possibly work, but I think you are opening a big can of worms.  Differential signals are usually produced by differential op-amps.  Normally, the op amp takes in a postive power supply and a negative power supply, and there is no ground connection.  Everything is floating above (or below) ground).  Differential means the difference between two non grounded points.  By introducing a common ground, you are causing disturbances to the op amp circuitry.  You are introducing a return path where there should be none.  If an op amp uses +/-15v for a supply, the + side supplys the voltage and the - side is the return, for the op-amp circuitry only.  Now you are grounding an input or output, providing a different return path to some other power supply or signal generator, or something.

Bottom line is that maybe you could get it to work, but that is not the way it is intended to work, and you will run into problems.  Use differential equipment for differential signals and use single ended equipment for single ended signals.  Don't try to adapt one to the other to save the cost of buying a new PXI card.  You will spend more in time and effort in the long run.

[krivan]

Hi tbob,

I formulated misunderstandable. Everything is connected together to one common ground point at the moment, yes. However when the board comes back to the fab I will correct the mistake disconnecting AI GND from the common ground point. I see now what you exactly meant with the differential measurements. Unfortunately the cards wasn't my choice and cannot change them so I must somehow convince my PXIe things to do what I want...

[tbob]

You need to make your differential input floating above ground.  Try this circuit: