I am designing a nuclear physics experiment in which I would like to be able to digitize pulse shapes coming out of photomultiplier tubes and use the shapes of those pulses to make decisions on the fly whether or not to write the pulse to disk. In this way I would be able to focus on events of interest without bogging my data acquisition system with background.
What (little) I know about FPGAs seems to indicate that it may be possible to use one in conjunction with a fast sampling ADC to accomplish this. The ADC digitizes the pulse while the FPGA analyzes the pulse in real time looking for certain shape characteristics (double peaking, peak height, return to baseline, deviation from a characteristic shape, etc.). Should it pass the shape condition, the ADC would continue digitizing the event to completion and the pulse shape would be written to disk. If the shape did not pass the set condition the event would end and nothing would be written. I do not know, yet, what conditions will be the most effective for identifying my events of interest, so the FPGA logic will need to be reconfigurable.
I have found a product offered by TI, ADC10D1500RB, which appears as though it may work for my application. However, TI was unable to give me any information about the Virtex-4 FPGA on this board. Would this reference board have the programming flexibiliy to do the kind of pulse shape analysis I have described? Or, is its functionality limited due to the preconstructed TI board that it is on?
If there are any better ideas beyond this TI board I would be very open to hearing them as well.