Use PCIe-6351 as a counter

[haominw]

Hello there,

I'm using PCIe-6351 with a BNC-2110 block to count TTL pulses from a single-photon counting device. The input signal has 10 ns pulse duration and 2.2 V amplitude. I'm using a LabView VI (see attached pics) to count pulses at every 10 ms. I also use shift registers to save and plot the data in the monitor (waveform chart). The idea is to get photon counts per milliseconds. 

However, my signal is contaminated with periodical spikes (see attached pics). I'm wondering whether these spikes are from my LabView code, or from the hardware limitation of PCIe-6351 that being not capable of counting TTL pulses with 10 ns duration and 2.2 V amplitude? 

Any suggestions would be appreciated!

[GerdW]

Hi haominw,

---

@haominw wrote:

counting TTL pulses with 10 ns duration and 2.2 V amplitude

---

These pulse specs are rather near the definition minimums as described here...

[haominw]

Hi Gerd,

Thanks for the link. It says the pulse width requirements are "

• Devices with NI-STC are 10ns.
• Devices with NI-TIO are 5ns."

I'm wondering which one is used for PCIe-6351?

In addition, I browsed some literature and found someone was using PCI-6602 to count signals from the same photon detector. Does PCI-6602 have better performance than PCIe-6351 in terms of counting 0-2.2 V and 10 ns pulses?

Thank you for your time,

Haomin

[Kevin_Price]

The specs I found online for the 6351 don't answer those questions definitively.   NI's desktop MIO boards are commonly TTL-compatible, so the 2.0 minimum voltage for high (that GerdW already linked to) suggest you are right on the verge of compatibility.

As to pulse width, I have only a vague memory from a thread over a decade ago that leads me to *suspect* (but not know) that the NI-TIO boards are the outlier with a 5 nanosec min pulse width.  If STC specs 10 nanosec, I would speculate that the STC-2 on the M-series as well as STC-3 on your X-series are probably similar.  So again, as GerdW already said, you're approaching the limits.

Having said that, the bigger issue is you're using software timing for your acquisition, and it'll be quantized to the nearest 1 msec out of 10 nominal.  Note that the 10% quantization corresponds pretty well to the 10-12% spike size. 

Search the forums for more threads on photon counting and you'll find ways to do this much better with regular DAQmx functions and hardware timing instead of the limited DAQ Assistant and software timing.

-Kevin P

[haominw]

Hi Kevin,

Thank you so much for the detailed reply. I've changed my codes to DAQmx functions and now using an external clock (1 MHz TTL) to better read photon counts without using shift registers (use finite sampling mode) and without spikes. 

A follow-up question, I'm wondering whether are the counter specs of PCI-6602 better than the PCIe-6351? Because the PCI-6602 is designed to be a specialized counter/timer device?  When comparing their specs, I didn't see much difference for the required pulse width (the lower threshold for PCI-6602 is 10 ns). There is only a slight difference between sample clocks, which is 100 MHz for PCIe-6351 and 80 MHz for PCI-6602. I guess if I was using the 1 MHz external sampling clock this mismatch makes little difference. 

A summary of my question would be, where can I find more counter specs for PCIe-6351? I do see the counter specs (e.g., TTL pulse width requirements, amplitude threshold, etc) in the manual of PCI-6602 but for PCIe-6351 these specs are missing. If you have any clues on that please let me know, thank you very much!

Haomin

[Kevin_Price · Accepted solution]

One really important advantage of X-series counters (like your PCIe-6351) is the much larger on-board FIFO.  It makes a HUGE difference for the maximum sustained sample rate for counter tasks.   I'm only going by memory, but I seem to recall benchmarks for the 6602 topping out somewhere in the low 100's of kHz whereas an X-series board could sustain > 10 MHz.

The 6351 will be a better choice for most situations, unless you need more than 4 counters or need to interface with a very high frequency external signal.  The 6351 is only spec'ed for up to 25 MHz external clocks while the 6602 is spec'ed for 80 MHz.

-Kevin P

[haominw]

Hi Kevin, 

Thank you, you are so cool! I asked a similar question to NI service at Newark and did not get any satisfied answers. We'll stick with PCIe-6351 for the photon counting. 

Haomin