Strange problem with analog output on PCI 6251 and BNC 2110

[hmalissa]

I'm trying to control a third-party current source using the analog output connectors on my PCI 6251 and BNC 2110.

The current source needs a 0..1V input signal. I have tested it using a battery, potentiometer and voltmeter, and by manually adjusting the voltage the current supply works - the output current scales with control voltage according to specifications and is reasonably stable.

The DAQ card works too - when I connect a voltmeter to the AO0 or AO1, the measured voltage corresponds to the setpoint with high accuracy.

But when I connect the third-party current source to AO0 or AO1 of the DAQ card, the measured output voltage drops and fluctuates. This is true for both AO channels.

I'm wondering what's wrong here. I suspect it might be a grounding issue - the current source analog control is a differential input with both terminals floating. I tried to flip the FS/GS switches on the BNC 2110, but that doesn't make any difference.

Has anyone experienced a similar behavior? Does anyone have suggestions?

[johnsold]

What is the third party device model? Or, post the specifications for the external input.

Your suspicion seems to be a likely candidate.  I have seen devices with external voltage control inputs which had to be floating. By tying either input to a ground, the control signal did not work correctly.

Another possibility: Most AO lines have very limited current sourcing capability. Does the device input require more current than the AO can provide?

Lynn

[hmalissa]

Thank you very much for the reply.

The third-party instrument I want to control through its analog input is a Kepco power supply, model ATE-100-10M.

I do not have detailed specifications, but I know that it needs a 0..1V input voltage. The analog control input works, I can control the output current reliably using a battery (and potentiometer to set the voltage). The analog input draws almost no current, it's below the detection limit of my voltmeter. That means the input is high-impedance, and this is what I would expect. The input is not grounded inside the Kepco.

It is very strange that the output voltage of the PCI 6251 drops off dramatically as soon as the Kepco is switched on (it is behaving normally when it is connected to the Kepco with the Kepco being turned off).

[Henrik_Volkers]

---

@hmalissa wrote:
[...]

But when I connect the third-party current source to AO0 or AO1 of the DAQ card, the measured output voltage drops and fluctuates. This is true for both AO channels.

I'm wondering what's wrong here.

[...]

Does anyone have suggestions?

---

Did you connect a current source output to a voltage input (without additional ballast/shunt?)??  

[hmalissa]

Thanks a lot for your reply.

No, I did not. What I meant is that the third-party current source has an analog control input (0..1V). This analog control input was connected to the AO0 or AO1 of the DAQ card. I apologize for the confusion.

[Henrik_Volkers]

test:

DMM to AO0 and AO_GND and AO_GND to the Kepco neg input with a 10k resistor. How does the output behave?

Can you measure (with a DMM) any voltage across the 10k resistor (AC or DC) ?

Some DAQ AO have spikes in the output. a small RC output filter can help

[johnsold]

---

"..the output voltage of the PCI 6251 drops off dramatically as soon as the Kepco is switched on.."

---

This is almost certainly an indication that the input to the Kepco is not floating but has a definite voltage/current/impedance relationship to ground.

Try measuring with your DMM the voltage between AO GND on the BNC2110 and the Kepco input terminals with no connection between the Kepco and the DAQ devices.  Also try it with the 10k resistor as Henrik recommended.

It has been some time since I used a Kepco supply but I seem to recall that those supplies had some funny control input circuits.  There is a note on page 2-5 of the operator's manual which discusses grounding and mentions a resistor/capacitor network from the DC power circuit to the chassis.  See also the sections on DC grounding on page 3-1 and 3-2.

Lynn

[hmalissa]

Thank you very much for your help.

I connected a 10kOhm resistor between the DAQ AO ground and the Kepco ground. It turns out that the voltage drop across the resistor is not always the same, it depends on the state of the Kepco. In the worst case it is 950mV for a DAQ AO voltage of 1V, which means that the input impedance of the Kepco can be as low as 500Ohm. The control input circuitry of the Kepco seems to be unusual.

I assume that the DAQ AO cannot supply enough current. I will set up a buffer between the DAQ AO and the Kepco.

[johnsold]

Consider using an isolation amplifier. The difficulty is determining the amount of isolation required, as the Kepco documentation is obscure.

Lynn

[hmalissa]

The unity gain buffer amplifier approach did not work. The result is basically the same.

I even tried using two unity gain buffer amplifiers, one for the positive and one for the negative/ground input of the Kepco, and that does not work either. I was hoping to isolate the grounds and have the Kepco control input completely floating.

I know it works with a battery (which is totally floating). Will the isolation amplifier decouple both grounds? I'm not familiar with isolation amplifiers, where can I find an application note that describes the circuit that is suitable?

Thanks a lot.