0-10v analog VFD control is non-linear

I'm running into an issue with my 0-10v control on my Emerson Commander SK 2HP VFD. The VFD puts out 10v on one terminal and 0v on the other and accepts the voltage you want on the third terminal. The problem is, around 5v it starts to behave non-linearly.

Example: my spindle max rpm is 10000. I ask for 500rpm and it sets the voltage to .50v. I ask for 5000rpm the voltage is 5.00v. I ask for 6000rpm and the voltage is 5.65v. It continues to sag from there. I think the max I achieved was 8 volts.

I have my VFD programmed so the max HZ it will spin the motor at is 167HZ, or 5000rpm (I have a 1:2 ratio from motor to spindle).

I would like to know if the non-linearity is common, or something odd that I might have missed. The manual indicates possibly a 10k resistor on +10v but I tried that and the voltage dropped to 1.76v instead of the 10v it was supposed to be. I'm using a Mesa 7i76 for the 0-10v output.

I have worked around the issue by doubling the VFD's max HZ so the working range is now 0-5v, I essentially don't use anything above 5v. That works but seems wrong, there has to be a better way.

Is it wired like this?

No, It is wired like this to the Mesa card

T1 == 0v == TB4.1
T2 == variable == TB4.2
T3 == 10v == TB4.3

B7 is hooked to my motor thermistors and configured as such on the VFD.

I only thought of the 10k resistor because they mentioned it for analog 2 output and I'm clueless about basic electronics.

Pretty sure that's wrong because you are driving the
VFD's 4-20 mA current input (T3) with the 7I76 output,
not the VFDs voltage input (T4)

This is close to a short so overloads either the 7I76's analog output
or the 10V power source from the VFD (not sure who will cry uncle first)

Well, analog output 1 can be changed from current mode to voltage mode and I have done that.

I could try using output 2 I suppose. Maybe output 1 is damaged. I did buy the VFD used.

possibly, but since their examples only use A2 as voltage input and
A1 as a current input I think that's your best bet (and B7 open)

I tested using A2 as the input and the result is the same. Linear up to 5v, non-linear past that and it maxes out at 8v.

What is the resistance of the poteniometer between the TB4.1-4.3 pins on the Mesa 7i76? The manual very clearly states you need at least 4kohm and they recommend 10khm.

I am supplying the 0v and 10v from the VFD and I really suck at electronics. If the 7i76 is low where would I insert a 10kohm resistor? I'm going to call Emerson tomorrow and see if they have any ideas.

It may be that their 10V supply cannot power the 7I76s isolated
analog output in which case you may need an external 10V supply

If that's the case how firm is the 10v requirement. Can I feed it 12v and it will be okay? I have plenty of 12v current because that's what I run all the field io off.

I checked and the 10v output is limited to 5ma. Seems kind of low, would the 7i76 opto need more current?

The 7I76 isolated output uses about 10 mA max

12V should be OK (15V is maximum)
you may want to change the scaling in hal so the output is 10V max