Experiment: Ridding of pulse generator noise

[physguy09]

My research involves using a detector that I am trying to test with a pre-experimental setup. To test the functionality of the detector, I am pulsing a light source near the detector using a Picosecond Pulse Labs pulse generator. Here is where my troubles/question comes in: the pulser is generating too much noise, such that the noise generated is greater than the signal I am sending. After swapping out the pulser, I have noticed that when I drop to the nanosecond scale on any pulse generator, the oscilloscope picks up noise as soon as I have the pulse generator output near my detector (whether the pulser is plugged into the light source or not).

My set-up is as described in the diagram below

[PLAIN]https://webspace.utexas.edu/xythoswfs/webview/_xy-25006560_docstore1

NOTE: The LED and the detector are in the same box.

Now I have enclosed the detector in a dark chassis box hoping that it would act as a Faraday cage, to no avail. Does anyone have any suggestions as to how I may eliminate/reduce the noise? I need the pulse to be in nanoseconds in length, so extending the pulse duration is unfortunately not an option I have.

Many thanks

 

[Hepth]

Are the oscilloscope and pulse generator on the same power supply/outlet/strip? Common ground? Etc.

What kind of power supply are you using?

Try to isolate them if possible.

 

[physguy09]

I had not thought of that. I can put them on different power supplies, but the ground will still be common. I am using a high voltage power supply for the detector (Keithly 2400, but the noise is seen on the scope whether I have the detector on or off.
Thank you

 

[likephysics]

What is the amplitude of the signal?
All instruments have some inherent noise.
When you measure pulse gen output, how are you connecting the scope probes?
Directly to the pulse gen output or is it on some breadboard/pcb?

 

[physguy09]

What is the amplitude of the signal?
All instruments have some inherent noise.
When you measure pulse gen output, how are you connecting the scope probes?
Directly to the pulse gen output or is it on some breadboard/pcb?

The amplitude of the signal is 10 V (I am using an attenuator to get this amplitude from the pulse generator, otherwise output would be ~50V), with a width of 10 ns (the amplitude of the noise when I place the connector in contact with the chassis box is 2 mVpeak-peak with a width of ~10 ns, when I actually connect the generator to the chassis box, the noise is 120 mVpk-pk with a width of 160 ns).

To measure the pulse generator output I am connecting the pulse generator to the scope via a bnc cable. To measure the noise, I am connecting the chassis box to the scope via a bnc cable, that itself is connected to the detector via probes that are shielded with aluminum tape. The pulse generator in turn is connected via a bnc cable to the LED in the chassis box.

Thank you for the help

 

[likephysics]

The 2mV 10nS noise you mentioned, is it riding on the pulse?
2mV sounds negligible.
Is there any difference when you check the pulse gen output vs. output at attenuator?

The 120mV, 160nS noise is really strange.
I guess you are using a linear LAB power supply.
Is the chassis box grounded?
Any decoupling capacitors at the power supply.

I reread your original post-
Noise is seen on the scope when you go to nanosecond scale on the pulse gen even when the pulse gen is not connected to the LED.
But when you connect scope directly to pulse gen, you don't see any noise?
Can you run the scope on Batteries, just to isolate it from power line noise.

 

[physguy09]

The 2mV 10ns noise is just picked up by the scope when the pulse generator is on, result of the EMF induced by the generator I believe.

As for the power supply itself (the power supply I am using, not the power supplied by the outlets), the noise is present regardless of whether the power is on or off (it only powers up the detector). The chassis box is grounded only by the bnc connectors that are grounded to the power supply and pulse generators themselves, I know not a better way to do so, and any suggestion would be welcome.

What I meant by the scope picking up noise from the generator is the 2mV 10ns noise mentioned above. When I connect the scope to the generator there is some noise at the peak of the pulse, and the pulse doesn't exactly die out after 10 nanoseconds, rather there is a ~1.5V tail that drags on for about 15ns yielding a total pulse width of ~25ns. For the purposes of the LED, since it is off if running <1.8V, I did not consider it a big issue. I can see however how this might contribute to some of the undesired noise. Ideally, I am trying to work with a pulse width of ~10-15ns.

Many, many thanks for the help