Fake events from high speed digitizer?

[LCHEN5154]

Dear all:

 We use NI5154 1GHz digitizer to capture pulses from a detector. We observed many very narrow pulses with width less than 10ns. In general it would be a surprise to see such fast pulses from our detector. And also the associated capacitance would seem to determine a time constant around 100 ns for all detector-originated signal. Someone argued that such narrow pulses might be fake events reported by digitizer. It's really hard for me to image how that could happen? By the way, I use analog edge trigger method in the pulse capture process. We really need to clarify this issue. Any idea will be much appreciated. 

A sample figure  of such snarrow pulses was attached.

-------------------------------------------------

P.S.

A men from UK sent us what he though on our narrow pulses. Following is a copy of his text:

"My first thought is that it may be due to the digitiser itself, giving

rogue events as the pulse fall time hovers the discriminator level ?

I have seen similar effects, even at 20 MSPS with 'shaped pulses'. It is

almost as if the FPGA discriminator is 'perfect', and sometimes reports

another pulse or two immediately after a 'real' pulse."

[Henrik_Volkers]

My first idea would be to shortcut/open the input and see if the artifacts are still there and than approach the sensor. 

Shot noise in the first amplifier stage?  In fact everey capacitor can be seen as an ion chamber 😉

Next step: Move that ('old') big fast scope with >500MHz BW in the lab and have a look ?

[LCHEN5154]

Many thanks for your thoughts, Henrik.

If shoutcut the input I will see nothing. I am sure such fast pulses are not randomly generated by digitizer.

The question is can digitizer report fake events short after a genuine event?

Only one preamplifier ORTEC VT120A was used to amplify the signal. We observed such pulses even without the amplifier.

Thanks.

Lixin 

[Henrik_Volkers · Accepted solution]

---

@LCHEN5154 wrote:

Many thanks for your thoughts, Henrik.

If shoutcut the input I will see nothing. I am sure such fast pulses are not randomly generated by digitizer.

The question is can digitizer report fake events short after a genuine event?

Only one preamplifier ORTEC VT120A was used to amplify the signal. We observed such pulses even without the amplifier.

Thanks.

Lixin 

---

The Spec of the ORTEC say Output impedance is less than 1Ohm , since we talk about rise times less than 1ns: How long is the cable between preamp and digitizer? I would care about cable reflections ... somehow I have the feeling that the output should also match the cable impededance....

Did you shortcut at the input of the digitizer? How about  seeing the cable as an ion chamber or as an antenna ?  Can you try a 50Ohm termination or a short instead of the preamp? 

By looking at your signal: I see always two spikes within 5ns , how long is your cable? 5ns equals about 1.2m 50Ohm cable ....    

[LCHEN5154]

Thanks. Henrik.

After many tests, with function generator and also tail pulse generator, I didn't see evidence that such fast pulses are fake events by digitizer and also they are not from cable reflections. They can be traced back to the ORTEC preamp output. VT120A is very fast and with a large gain (200) thus it can pick up some tiny spirks.

Lixin