No output from PXI-6733 in buffered mode

[Roger]

Hi,

I seem to be encountering an odd issue where the analog outputs from a PXI-6733 do not respond to pseudoclock pulses in buffered mode and was wondering if anyone else has encountered this issue.

In my setup, the pseudoclock is connected to PFI0 of the PXI-6733.

• Manual mode: output voltages change as expected (using the blacs tab)
• Buffered mode:
  • Output stays at 0 V when using constant()
  • Using ramp() results in an error during transition_to_manual

Some items I've checked:

• Verified that the HDF5 shot file does have values for PXI-6733 analog output
• The PXI-6733 does register outputs from the pseudoclock using the NI MAX test panel

As a sanity check, I used an the analog outputs from a PXIe-6361. Both constant() and ramp() produced the expected voltage outputs.

Attachments:

• connection_table.py - snippet of the connection table
• testing.py - snippet of the labscript code using the PXI-6733 that hasn't been working (the  ramp() parts were commented out as was working on just testing constant()).

I'll work on compiling better notes on this issue tomorrow when I'm back in the lab.

Thanks,

Roger Ding

[dihm....@gmail.com]

Roger,

You have already done the checks I would normally suggest. This is perplexing. I have two small requests in an attempt to help:

1. Can you post the full traceback for the error you get in transition_to_manual with the ramp() command?

2. Have you confirmed that the pulseblaster slow clockline output during the shot lines up with what you expect? (ie tee it into a scope, make sure its timings line up with what runviewer says they should be)

-David

[Roger]

Hi David,

At the moment, the problem is likely a software issue with the particular hardware I'm using (PXIe-1085 chassis with the PXI-6733; more details below). I'm waiting to hear back from NI technical support but I expect it'll be slow responses over the holidays.

Longer summary:

I had been troubleshooting the PXI-6733 issue in Python and labscript for a while without much progress. On a whim, I decided to try using NI MAX's Test Panel to see if the hardware behaved as expected, which is where discovered that the PXIe module (PXIe-6361 and PXIe-6738) analog outputs work as expected but the PXI-6733 analog output does not. In particular, the PXI-6733 DC analog output worked but using "Voltage Sinewave Generation" raises Error -200016 (see picture below):

After discovering this, I used used "Voltage Sinewave Generation" in the NI MAX Test Panel to try a PXI-6733 in different combinations of chassis to see if the problem is specific to the hardware I'm using.

Main system:

• Computer-side interface: PCIe-8383
• Chassis: PXIe-1085 (firmware 20.0.0f0)
  • Slot #1: PXIe-8381 (PXI interface)
  • Slot #2: PXIe-6361
  • Slot #3: PXIe-6361
  • Slot #4: PXI-6733
    • Raises Error -200016 when using "Voltage Sinewave Generation"
  • Slot #5: PXI-6733
  • Slot #17: PXI-6704
  • Slot #18: PXI-6738

Test system #1:

• Computer-side interface: PCIe-8383
• Chassis: PXIe-1082 (firmware 1.0.0f0)
  • Slot #1: PXIe-8381 (PXI interface)
  • Slot #4 (first hybrid slot): PXI-6733
    • "Voltage Sinewave Generation" output works and does not raise an error

Test system #2 where the PXI-6733 raises Error -200016 when using "Voltage Sinewave Generation":

• Computer-side interface: PCIe-8383
• Chassis: PXIe-1085 (a complete different chassis with firmware 1.0.0f0)
  • Slot #1: PXIe-8381 (PXI interface)
  • Slot #4: PXI-6733
    • Raises Error -200016 when using "Voltage Sinewave Generation"

Notes:

• The issue seems to be related to using the PXI-6733 in a PXIe-1085 chassis.
• I just noted if "Voltage Sinewave Generation" raised Error -200016. I didn't check to see if the outputs had the correct amplitude and frequency as specified in the NI MAX Test Panel.
• I didn't get around to testing labscript with the PXIe-1082 system to see if the PXI-6733 worked as expected (although I suspect this problem underlies the labscript issues).

I submitted an NI support request after trying to "repair"/reinstall various NI software using the NI Package Manager without success. An NI representative responded with (truncated response below):

Thank you for providing a detailed description of your system setup. Based on the information, the hardware appears to be compatible—PXI‑6733 is a hybrid‑ready PXI module, and the PXIe‑1085 chassis supports hybrid slots for PXI peripheral modules.

However, the reported error (‑200016 during sine‑wave generation in NI MAX) suggests that while the device is recognized, it cannot properly initiate output in this configuration. This typically points to a resource initialization or clock configuration issue at the software layer rather than a hardware incompatibility.

To proceed effectively, we need to confirm that your driver and platform software versions are fully compatible with the PXIe‑1085 firmware. Please generate and share the NI MAX Report. This will provide details about your software environment and operating system, helping us narrow down the root cause and offer a targeted solution.

I responded by sharing my main system's NI MAX Report and am waiting to hear back.

Roger Ding

[Roger]

A small update as I'm continuing to troubleshoot with NI Technical Support:

On my experiment computer (a Dell 5820 workstation), NI Technical Support recommended installing the MXI-Express BIOS Compatibility (2025 Q4) and changing the BIOS DIP switch on the PCIe-8383. Doing this seems to get the PXI-6733 working in an NI MAX Test Panel but ends up causing PXIe modules to throw errors (e.g., trying to use the Self-Calibrate function in NI MAX results in the Status Code -50152).

Getting a bit frustrated, I borrowed a colleagues HP Z2 G9 and setup a test system using a spare PXIe-1085 chassis with a PXIe-8381 (MXI interface), a PXI-6733, and a PXIe-6361. I also installed the PCIe-8383 (MXI interface) from the Dell 5820 into the HP Z2 G9 (and reverted the BIOS DIP switch since the HP computer did not have the MXI-Express BIOS Compatibility software installed). Both the PXI and PXIe modules seemed to work (at least in NI MAX).

As a result, I think the issue may be a weird incompatibility between the Dell 5820 workstation and the PXIe-1085 chassis.

[dihm....@gmail.com]

Roger,

Well that is a wild story. When NI stuff goes wrong, it really does go wrong in strange ways. Glad you have at least gotten somewhere.

You have probably tried these things, but figured I'd add them since my lab uses Dell 5820 workstations almost exclusively. They can be very quirky devices, especially with expansion chassis. Things that have improved our success in the past:

1) Make sure your bios is up to date. They are still fixing things to this day. Dell Command Update is a decent utility for managing that stuff.

2) Make sure you are using the correct PCIe slot. All slots use a full size connector, but not all of them are electrically connected (ie a x16 physical slot with only x4 electrical connections). The MXI card probably needs at least a x8, but may partially work on a x4.

3) If you are using ECC RAM, be sure to follow the manual's valid layouts. Also, the cold boot times (especially after modifying RAM) are absolutely glacial. You just have to wait them out, unfortunately.

4) Make sure you disable relevant power saving features on the PCI bus. I forget what they call them, but there are a couple options that need to be disabled when using an expansion chassis for reliable operation.

-David

[Roger]

Hi David,

It seems a bit odd that the Dell 5820 (with a Xeon processor) has issues with the PXIe-1085 (24 GB/s) chassis but seemed to work with the PXIe-1082. At the moment, I'm waiting to hear back from NI Technical Support if they have other solutions, but I'm also waiting to get an HP to replace the Dell that will (hopefully) avoid this issue altogether. I added some additional details below if that's helpful for others.

1. The BIOS was updated to the most recent version using the Dell Command Update utility.
2. I originally had the PCIe-8383 in Slot 1 (PCIe x8), but NI Technical Support recommended moving it to Slot 4 (PCIe x16) and increasing the PCIe bus limit in the BIOS (Advanced Configurations > PCIe Bus Allocation > Option 3).
   • PCIe Bus Allocation in Dell Precision 5820 | 7820 | 7920 Tower
3. I'll need to double-check but I believe am running ECC RAM. I did notice inconsistent boot times during my testing, but I don't recall which configurations resulted in noticeably slower boot times since the 5820 generally boots slowly (and I often find myself checking if DVD drive opens to make sure the computer is on during a cold boot).
4. NI Technical Support also recommended disabling Active State Power Management in either the BIOS or through Windows.
   • Error -200279 when Using Thunderbolt Hardware Configuration
   • May need to go into the BIOS and enable the Block Sleep option (Power Management > Block Sleep > Enable) to get the Windows power options to show up as they do in the KB.

Roger Ding