How to correctlly set up the reference position of NI 5122?

[cirrus]

Hi,

I need your help. I am try to program the NI 5122 to acqusite the backscattering signal of laser puse.  I use external analog trigger. I want to set the pretrigger 100, total record length 2500.  and I wire the (100.0/2500.0)*100.0 to the board, I couldn't get what I want, instead, I got around 30 points. When I tried 200 points pretrigger, and wire (200.0/2500.0)*100.0, I got around 130 points. I have no idea why?

Could you guys give me some suggestion?

Thank you in advance,

Cirrus

[Matt_E.]

Hello Cirrus,

I would like to verify your setup quickly just to make sure I am correctly thinking about what you are doing.  If your setup differs from my description please let me know. 

You have a signal A which you are feeding into the external trigger BNC on your 5122 and a Signal B which you are feeding into either channel 0 or channel 1.  You are trying to do an analog edge trigger on signal A and acquire 100 pre-trigger samples of Signal B with a total record length of 2500 samples at 100 MS/s.

I am assuming that since you are using the external trigger for signal A, Signal B does not have an event you can trigger on directly but contains the information you want to digitize.  I am also assuming that since you have said you are only getting 30 pre trigger samples when you fetch your record that you must have a guess about where the trigger event should have occured in your fetched waveform of signal B.

If this is true then your problem could be related to cable delays.  Basically since the paths taken by signal A and B are completely separate any extra delay added by a longer cable in one of the paths would translate to an offset in the trigger time.  There is a simple check you can do to see if your board and software are working correctly.

Put A BNC splitter on signal A and run it to both the External trigger and Channel 0 inputs.  Try to keep the cable lengths of each path the same.  Next run your VI and check to see if your waveform shows the trigger edge occurring at the specified point in your record.  If it does then your board and software are fine and there is some extra delay in your system that you will need to calibrate out.  If the edge does not occur in the correct part of the record with this setup then disconnect all signals and run self calibration from MAX.

If this does not work repeate the same setup using the configured acquisition VI from the NI scope examples just to rule out a software problem. 

If the setup I described above does not correctly describe your situation please let me know.

Regards,

-Matt

Message Edited by Matt E. on 01-10-2008 09:54 AM

[cirrus]

Hi Matt,

I am very thankful that you replied me so quickly. I test my borad by using your method wich means split a single into two, one goes to channel0, another goes to external trigger. It turn out that this time, pretriger setting is correct. You are correct. Because of cable delay, we miss some points. e.g, 25M, 2500 sample poinrs, 100 points of  pretrigger, We lost 70 pionts, the delay time is  70/25Mhaz = 0.28micro sec.

By the way, my set up is like this:

I have two signal channels from the PMTs, the parallel channl  to ch1, the perpendicular channel to ch0. I have additional signal which is signal from energy detector  for detecting laser  pulse out. This additional  signal is used as  external  analog edge trigger. 

I tried to short the cable of the trigger signal, it do not change much.  Is there any other factor to disturb my pretrigger point?

Thank you very much!

Cirrus

[Matt_E.]

Since the test setup produced the trigger event in the correct position of the record we know your board is working correctly.  Because the delay in the system is not from the board you will need to manually calibrate the delay out of your system.  Because you said that you are getting too few pre trigger samples this means that your signals from the PMTs are being delayed relative to the signal from your energy detector.  Some of this delay could be due to the cable and it is also possible some could be from another physical delay (such as in the time it would take a sound wave to travel from a speaker to a microphone across a room).  The cable delay is pretty easy to characterize and compensate out.  Place a BNC plitter at the output of your power detector and connect your cable you were previously using for the power detector to the external trigger.  Connect the other output of the splitter to one of the cables you were using for the PMT output to channel 0.  Run an acquisition and look at where the measured edge on channel 0 appeared relative to your specified reference possition.  The error in number of samples * sample clock period is the delay of your cable.

For example if you specified a reference position of 50% in your record of 2500 samples then you would expect to see the edge at sample #1250.  If you actually saw it occur at sample # 1000 and were sampling at 100MS/s then you would need to compensate for a delay of (250 * 1e-8 s) = 250e-8 seconds.  You should be able to program this value into the Trigger Delay input of the niScope Configure Trigger API VI

If this still does not correct your reference position problem then there is a delay in your system from somewhere other than the sable and you will need to find a way to characterize the delay and add it to the value calculated for cable delay.  Also if for some reason the delays to channel 0 and channel 1 are not the same you will have to change your acquisition to acquire more pre/post trigger samples than you actually need and write your own code to re-align the waveforms correctly.

Hope this helps!

Regards,

-Matt