Multifunction DAQ

cancel
Showing results for 
Search instead for 
Did you mean: 

PMT voltage measurement

Hi All,

 

I have a PMT sensor that gives me output in terms of volts. The voltage output is obtained over a BNC cable. The range of the voltage value is in milivolts (.4 to 3 mv approx).  The signal as it is a PMT signal can be varying.

 

When I measure the voltage using a multimeter, I find that that I am able to measure it, however when I connect the same to my DAQ box (in this case USB 6259) in a differential mode (when one end is referenced to overall ground) I seem to get random values which might suggest either

 

-          The measurement is floated

-          Or Labview analog to digital conversion time is so small (ADC integration time) that it picks up random noise and shows that reading.

 

I am quite stuck at this part and would really appreciate to get any inputs from you.

0 Kudos
Message 1 of 6
(4,752 Views)

 Usually the PMT output is the voltage across a resistor carrying the PMT cathode current and one end of the resistor is grounded.  Is the PMT circuit ground connected to the DAQ ground?

 

The signal variations may include slowly varying values which corrrespond to what you see on the multimeter, but include rapid fluctuations due to quantum effects as photons and electrons interact.  The multimeter performs some low pass filtering or averaging which removes most of the hig frequency fluctuations.

 

After you verify that the connections are correct, I suggest that you collect many samples and average them.  For example if you sample at 1000 samples per second and average 1000 samples, you will get one updated reading per second with much of the rapid fluctuation averaged out.

 

Lynn

0 Kudos
Message 2 of 6
(4,749 Views)

Hi,

 

Thank you for your reply. Yes, I have a common ground for system, which means the ground of the PMT and the shielding of the BNC all are connected to the same ground.

 

Even I think that the high frequency fulctuations are responsible, as a multimeter would in most cases filterout high frequency changes. I shall try to implement the average function, in the meanwhile could I ask if addidng a low pass filter after the data acquisition in Labview would be useful or  a low pass filter in the hardware side would help?

 

Regards

Avinash

0 Kudos
Message 3 of 6
(4,741 Views)

Avinash,

 

You should always have a hardware anti-aliasing filter.  Particularly in this situation becasue the PMT is capable of producing noise and inpulse type signals with very wide bandwidth.  If you are not familiar with the Nyquist criterion, it requires that all energy above 1/2 the sampling frequency be filtered out before the ADC. Any energy not filtered becomes "aliased" and cannot be distinguished from the real signals.  For wideband noise this has the effect of moving noise from high frequencies down into the lower frequencies and combining it with the existing low frequency noise.

 

If you are sampling at 1000 samples per second, then the anti-aliasing filter needs to remove all energy above 500 Hz.  Because real filters do not have abrupt band edges, the filter should be designed with a lower cutoff than 500 Hz.  Without knowing more about your data rates and how fast (or slow) your final data usage is, I cannot recommend a specific filter.

 

The averaging process is a type of low pass filter so there is probably no need for any additional software filters.  If you have specific inteference such as power line frequency signals, then a notch or bandstop filter for those frequencies might be useful.  Fluorescent lights are very bad about generating power line frequency (and many harmonics) optical energy.

 

Lynn

0 Kudos
Message 4 of 6
(4,738 Views)

Hi Lynn,

 

Thank youo for your valuable inputs.

 

Yes, it is right that I should include a anti aliasing filter for my circuit, which I tried to do yesterday.

 

However, there are a few doubts I have before which I would have to give you a quick summary of my system.

 

I have light of varying intensity shining on my PMT. The intensity of my light depends on an input voltage which I sweep from 0 - 20 V.

 

So INPUT SCANNING VOLTAGE -> CHANGE IN LIGHT INTENSITY -> CHANGE IN PMT OUTPUT VOLTAGE.

 

So this case -

 

>> Does the frequency of my measurement depend on the rate at which my voltage changes?

 

In another application, I am trying to measure the intensity over an 2D area in a point by point fashion (rastering method)

 

>> Would my measurement frequency be the rate at which I move from point to point?

 

In both the cases, do I implement Anti Aliasing filter to filter out all frequencies above these frequency.

 

Also I am in dark about the type of ADC the DAQ board uses (SAR, Dual slope integrating type or Sigma delta).

 

Regards

Avinash

 

 

0 Kudos
Message 5 of 6
(4,716 Views)

Avinash,

 

Without digging into the manuals I cannot be sure, but it is likely that the USB-6259 uses the SAR type ADC.  Dual slope integrating is typically used for voltmeters and digitial panel meters. Sigma delta is usually used for slower converters with very high resolution (24-bit).

 

Yes, the sampling frequency is related to both of those factors which change the light intensity at the PMT.  If the fastest change in your PMT output voltage due to input scanning voltage changes is X volts/second, then that corresponds to a maximum frequency component at X/2*pi (from the derivative of sin(2*pi*f*t).  To accurately reproduce that signal after sampling you would need to sample at a rate greater than X/pi.  The anit-aliasing filter (hardware) would also need to filter out any energy at frequencies higher than X/2*pi.  You can do the same kind of calculation based on the change in PMT voltage due to raster scanning.  If you signal can change from fully dark to maximum light in one raster step, then that would define the highest frequency.

 

In practice with real filters and real signals, the anti-aliasing filter cutoff and the sampling frequency may need to be significantly different from the limit values calculated above.

 

Lynn

0 Kudos
Message 6 of 6
(4,711 Views)