How to calculate individual periods of an incoming signal

[majoris]

So let's assume you use the 6601, and you set your photons to show up roughly every 1 microsecond. Drive the counter on the 6601 (i.e. the source) with the 20 MHz clock, then do a buffered event count by taking the photon detection event positive edge on the gate of the counter. This will give you an array of counter values like

100

120

141

160

182

You can then subtract adjacent elements of the array to find the time in counts between photons. Finally you can multiply by 50 ns to get the time between photons in nanoseconds. Note that your resolution is only to the nearest 50 ns.

That sound accurate? If you want more resolution, you will have to space out your photons even more in time.

[Romasp]

thats the thing. Its not the statistics that I need. I need to know every time not just the mean and so on because the correlation program I am working on reqiures a series of time differences without any gaps otherwise it will create artifacts.

I have found out a little more information, the pulse with of the detector is 10ns however as far as I understand, this is not important. I think that a pulse will be detected in the sample even if the sample size is larger (due to the counter being perfectly capable of counting over a more macroscopic time). I think the resolution needs only to be greater than the time between photon arrivals.

However I tried to get the digital output of the counter and it did not seem to see every photon which confuses me further. It was only able to see one in about 10,000. Would the detection of 2 photons override the detection of one photon? so 2 photons would become an output of 1, and 1 and 0 photons would gove an output of 0.

thank you all for your replies

[majoris]

How did you end up setting up your acquisition? Both in the hardware and in the software (i.e. what did you use for your gate and source terminals, etc)?

Can you show the VI at all?

[Romasp]

It is only an completely basic program to check the information that I could get from it using the slower counter and there is no way that the data is correct. I also tried looking at the analog signal but that only gave me some random signal that was the same even without a source of voltage.

Here is the program:

[johnsold]

According to the specifications for a PCI 6601 the minimum input pulse width to the counters is 25 ns.  It is not surprising that your 10 ns pulses from the PMT are often missed.

You either need to add a pulse stretcher between the PMT output and the counter or get a faster device.

Lynn

[Romasp]

Thanks for pointing that out. I nearly ordered the connection box for the PCI 6601!

I have not at this stage even used the PCI 6601. I have been using the PCI 6014 which has a min pulse width of 10ns and its capable of counting. I have another counter running along side that is giving the same values for countrate which I know are correct.

[majoris]

While the pulse width might be an issue, let's assume it isn't for a minute (since you have the 6601, you obviously can test to see if your pulse width is sufficient):

There's a fundamental problem with your diagram and it's that you aren't even using your counters! 

Goto Help >> Find Examples >> Hardware Input and Output >> Counter Measurements >> Count Digital Events

You want something in this folder. Notice these use external clocks! Are you using an external clock? If not, you can potentially generate a clock using

Goto Help >> Find Examples >> Hardware Input and Output >> Generate Digital Pulses

One of those examples will help you there. Of course, if you have a 20Mhz reference already, just go ahead and connect it to the appropriate PFI port.

(If you look on the forth page, you can see the PFI channels that are by default routed to the sources and gates of each counter: http://www.ni.com/pdf/products/us/4daqsc386-393_386-387.pdf)

The source increments the counter (it should be your 20Mhz reference). The gate takes the acquisition (it should be your photon pulse). Good luck!