PXI-5114 Calibration

[Test_Engxx]

I have modified my code to try and perform the calibration as discussed in the last post. 

To reiterate, I am seeing an error when I go from channel 0 to channel 1 if performing them in sequence.

(Also if starting with channel 1, I get the same error going to the VCXO cal.)

My code changes were:

Close the cal session after channel 0 (Store the cals), and then reopen the calsession and proceed to channel 1.

Perform the same steps between channel 1 and the VCXO cal.

Not wanting to overwrite any cals just yet, I ran through all of that code, (using ABORT instead of STORE)

I did not receive any error messages.

Since there was some question as to wether the module would save the calibrations if everything was not

performed during the SAME session I tried another experiment.  

I ran just the first part of the calibration, (channel 0), and then closed the session with the STORE command.

I received the following error:

-1074118604

Attempt to store calibration constants without completing all the necessary external calibration steps.Consult the calibration procedure. Verify that all necessary steps are performed before closing the external calibration session.

At this point I appear to be stuck. 

If I try and do all of the steps in sequence I get the error listed in previous postings.

If I try and work around the issue by closing and reopening the session, I can't save the calibrations.

1)  Do you know why I am getting the error when trying to do the steps in sequence?

2)  Do you know of any other things I should try?

3)  To reiterate, I am using version 2.9 of the driver. (linux)

[JoshY]

Just to clarify, when you were seeing the first error between running a self calibrate from the Start handle rather than the Initialize handle, were you running through the full calibration procedure? The errors you are seeing now make it seem like you aren't able to even get to step 13 in the procedure. If this is the case, then how were you seeing this error?

Do you have access to Calibration Executive? If so, is it possible to run this procedure in Cal Exec to check these procedures? Or, is it possible to do this in LabVIEW? Even a trial version (downloaded from online) could work. We want to confirm that you are seeing these errors when following the calibration document exactly step by step.

I'm trying to verify this on my end, but I do not have a Fluke Calibrator readily available. Therefore, I was able to verify that feeding a Cal Start directly to an Self Calibrate will cause an error. However, I did not run through the entire calibration procedure, and this may cause different behavior. I do not want to run through the calibration procedure without attaining proper equipment to avoid damaging the card.

[Test_Engxx]

I am no longer concerned with the self calibration.  That question has been answered.

(I will close the cal session and reopen a session with the init command to perform self calibration.)

I was able to get to that step in the procedure by skipping the previous steps.  I was

simply trying to get an estimate of how long the self calibration would take at the time.

I am trying to perform this calibration from within an embedded test system that is running linux.  I

do not have access to calibration executive or labview. 

The problem I am experiencing now is when trying to perform the full calibration.  The error occurs

at the first step in the channel 1 calibration that requires a stimulus. (+18 Volts)

The error I see is:  External calibration constants are invalid. Perform an external calibration. Contact National Instruments if you need additional information. (-1074118603)

If you refer to message #8 in this thread, I have detailed the issue.

Essentially, the calibration of channel 0 goes smoothly with no errors reported from the driver.  In step 9 I move the

stimulus from channel 0 to channel 1.  I go to step 2, which is the cal of channel 1 with the input shorted.  This

does not report an error.  I continue to step 5 where the stimulus from the table is supplied. (+18 Volts) 

It is at that point where I see the error.  I do not believe that the error is caused by my code.  The only differences 

between the channel 0 cal at this step and the channel 1 cal is that the stimulus is now applied to channel 1

and the channel parameter in the command sent to the module changes from 0 to 1.  I can start the calibration

in my code at channel 1 instead of channel 0 and I do not see any errors during the channel 1 calibration.  It

appears that the module is not being correctly configured when going from the channel 0 cal to the channel 1

cal.  The same issue appears to happen between the channel 1 cal and the VCXO cal.  (The VCXO calibration

is performed on channel 0, which would require the module to switch from channel 1 to channel 0.)

The error reported in my last post was an attempt to 'work around' this issue by saving the channel 0 calibration

and then starting at channel 1, etc.  This was not successful since the driver will not allow you to save the

calibration unless you have performed all of the calibration steps.

I really need to get to the bottom of this issue. 

[JoshY]

I'm working with different departments to get the necessary equipment to try and duplicate this issue in LabVIEW. It may be an issue with the driver or just with the Linux driver. I've not heard of this issue before and can't seem to find prior documentation, but I'll let you know if I find something.

[Test_Engxx]

I have managed to get time on a test station in production that is running Windows, and has a PXI chassis connected to it.

I downloaded the 2.9 version of the driver and installed it on the pc.

I can see the 5114 in MAX, and can even open it in MAX and see trace data displayed in a panel.

(This is a different scope module than the one I have been working with.)

(The scope soft front panel did not seem to work for some reason, but I don't think that is an issue I am worried about.)

I have ported my linux test code into LabWindows, and have been able to verify that I get the same error reported at step

#5 of the calibration procedure on channel #1.  (18 Volts DC applied to channel 1, and the command sent to the module

has the the channel "0" replaced with a "1".)

(The error: External calibration constants are invalid. Perform an external calibration.Contact National Instruments if you need additional information. (-1074118603) )

I have, however, also seen that particular step pass on a VERY infrequent basis. (~5 out of 50 attempts)

I am wondering if there is any more information available from the driver that would be able to pinpoint a possible cause? 

It is strange that it is always at the same step in the procedure. 

Is there something that the driver is doing at that step to qualify the input, temperature, anything? 

I am using a Fluke 5720A instead of the Fluke 9500B, but  that should be more than adequate.

I have swapped cables etc., but that does not appear to have any affect.

I don't understand why I would get this error if there was something wrong with the stimulus I am

applying.  The return status for all of the previous steps is zero which should indicate success.

I am supplying the same stimulus as the channel 0 +18v step which passes.

I had noted previously that I thougt it might not be switching to channel 1, but I don't think that

is the case.  If I leave the stimulus disconnected for this step the error returned from the driver

indicates a calibration failure, with a different error reported. 

[JoshY]

When you say the error is at step 5, does this mean that the short in step 2 and the 10V in step 4 pass? Also, when you see the step pass, is anything different happening? It is strange that it would pass for some steps using Channel 1 but fail on the 18V input.

Regarding your post about the channel not switching properly, I agree that this no longer seems like the case. If you can pass the short and the 10V input, then Channel 1 is communicating properly.

I am meeting with the Calibration group to see if I can reproduce this issue. As of now, I believe the best solution would be to use Calibration Executive. However, I will get back to you with my findings.

[Test_Engxx]

Step 2 has you apply the short to the input, and step 4 then issues the cal command on the 10V range with 0V stimulus.  This returns a status of 0, and appears to take approximately the

same amount of time as this step did on channel 0.

By saying the failure is at step 5, I really mean that it was in the loop (Repeat steps 5 through 7 for the values in table 11).  The actual error returned from the driver is at

step 7, for the first value in table 11 which is +18V.

[JoshY]

I just ran the calibration of a 5114 using Calibration Executive, and I did not receive any errors through the process. Calibration Executive should be performing the same steps as detailed in the manual, so I do not understand why you are seeing this error. However, I will dig into the code more to make sure. Again, at this point I believe the best solution is to use Calibration Executive.

[JoshY]

I've checked the Calibration Executive code, and it does follow the calibration manual step for step in the Adjustment section.

[Test_Engxx]

Update:

1) I updated the NI-Scope driver to 3.1 in both my windows and linux environments.

2) I am able to get completely through the calibration with no errors.

3) I am using a Fluke 5720A calibrator instead of the 9500 calibrator called out by the procedure.

I have noticed that while performing the calibration of the 40mV range the scope shows ~4mV of noise on the

signal.  I mentioned this to metrology, and it appears to be a common occurence with 55xx series of calibrators.

They have a noise spec, but it is limited to < 10kHz.  I am not sure how much of an affect this noise will have

on the calibration of the 5114. (When running the SFP in windows and turning on the average voltage measurement

it appears to be within +/- .2 uV at 18mV even with the noise that is present.)

Has anyone evaluated the calibration when performed with a 5720A?

4) I also have a 5520A with the SC600 option available.  This provides a much cleaner signal, but the specifications

for the accuracy are not quite as good as the 9500B.   

Has anyone evaluated the calibration when performed with a 5520A SC600 option?