Need to Measure Light Freq. using photodiode and have it analysed by FFT, details below.

[Julesy]

I a new to LV. Project is to measure light freq. using photodiode heald against a TV screen. Am required to have it analysed by FFT, I have read lots but still at dead end of which VI I can manipulate, tried Spectrum Analysers but not right. Need it to collect at rate of at least 100/sec. I am using a CB-68LPR I/O and PCI-MIO-16E-1. Photodiode is connected to an amplifier and filter before it is hooked into the connector board so signal is clean. I will be measuring freq. of ciliary beat where photodiode placed against a TV screen.

[Rainer Ehrt]

Julesy wrote:

> I a new to LV. Project is to measure light freq. using photodiode
> heald against a TV screen. Am required to have it analysed by FFT, I
> have read lots but still at dead end of which VI I can manipulate,
> tried Spectrum Analysers but not right. Need it to collect at rate of
> at least 100/sec. I am using a CB-68LPR I/O and PCI-MIO-16E-1.
> Photodiode is connected to an amplifier and filter before it is hooked
> into the connector board so signal is clean. I will be measuring
> freq. of ciliary beat where photodiode placed against a TV screen.

########

Hi Julesy,
are you shure what you want ti do?
TV screens don't show pictures, they show a point (3 points on a colour
TV) with changing intensity passing by very fast.
For instance, if you h
ave the PAL Standard, each picture is built with 625
lines, in 20 ms, interlaced. This means, 64 ?s line with the leading and
trailing overhead. The visible part of a line is aubout 52 ?s long.
If you attach your fototransistor to the screen, and its diameter may
cover a circular area of maybe 3 mm and the visual part of the line would
be 520 mm, then: And if this covers four of the screen lines, then you
will receive pulses of 0,3 ?s length if the line passes the center of the
fotodiode. Not passing the transistor in the center would give back an
even shorter pulse.
And if you get all this working (which I don't think) then you would
measure 2 groups of pulses every 10 ms with 2 pulses each (remember
interlacing).

Even if all the DAQ would work, is your fototransistor fast enough. Think
about miller capacities. Its not so simple at it seems first to solve this
measurement problem.

Please let me know what you intend to do. I was an TV- enginear many years
ago, and I don't remember ev
erything relevant anymore, but there are some
hardware- related things to be aware if if you want to measure TV-
signals.

Bye,
Rainer

[Julesy]

Rainer,
Thanks for your input. This technique has been proven to work but using which software I am unsure. I will be placing the photodiode into a probe that only has a 1mm hole at the end. It is a low-noise, quadrant silicon photodiode (RS stock no. 652-027), it is apparently high blue sensitivity and shunt resistant, suitable for low light level applications. This has been researched by another person who assures me this diode will be appropriate. Does this clarify anything further?
Thanks
Julesy

[Armando]

This is an interesting topic. There is a document on developer zone that explains about light measurements:

http://zone.ni.com/devzone/conceptd.nsf/webmain/56203F38296D9B0686256B480083098E?OpenDocument

[Labviewguru]

,

My solution to problems like this is always to put the signal up on an oscilloscope and examine it. From this you can determine your amplitude and signal type (AC/DC, voltage, etc). When you combine this with knowledge of your nyquist frequency and other data, you can determine how to setup your MIO card. Then, you simply read in the waveform, and perform your FFT. You can't do any of this until you know what your signal is however.

If you need further assistance, please email me or post here.

Good luck

[JPG]

If I understand correctly you are trying to measure frequency of light. If this is visible light the frequency should be around 4.3 x 10e14 - 7.5 x 10e14
HZ depending of color. To accurately measure this means that you have to sample twice as fast(10x10e14).
This sample rate is much too fast for PCI-MIO-16E-1.
I do not know what is ciliary beat and its frequency. Maybe it can be measure by a daq board.
If you need to measure light frequency,you could possibly use a monochromomter.

[PdB]

My brain is also not able to sample my eye at 10E+15 Hz, but it can detect color and intensity changes.
A photodiode converts the AC light signal into a DC component. Followed by a high-GB low-noise current-amplifier one can measure light-intensities by a DAQ-board you mention.
As for a past project I worked on, I know that by measuring variations in light-intensities, one can construct interesting sensors.

[chrisd]

I guess you just want to measure the screen frequency. This, according to
my experience, can be done with a photodiode. I used to work for a company
where we developed a device for the colour calibration of computer
screens. We just used a photodiode and measured the pulses (connect one to
your scope input if you want to see) and averaged for a while in order to
determine the screen frequency. Using an FFT is a rather complicated way,
I would say.

Ciao,
Chris

Julesy wrote:

> I a new to LV. Project is to measure light freq. using photodiode
> heald against a TV screen. Am required to have it analysed by FFT, I
> have read lots but still at dead end of which VI I can manipulate,
> tried Spectrum Analysers but not right. Need it to collect at rate of
> at least 100/s
ec. I am using a CB-68LPR I/O and PCI-MIO-16E-1.
> Photodiode is connected to an amplifier and filter before it is hooked
> into the connector board so signal is clean. I will be measuring
> freq. of ciliary beat where photodiode placed against a TV screen.

[JPG]

Since you are new to Labview,
taking a LabVIEW BASICS courses can get your project going much faster. There are many developers and consultants out there.

http://labviewtraining.netfirms.com/

[Christopher Mitchell]

Since your instrument converts a frequency to a DC signal you should be all right sampling at 100 Hz or so. Keep in mind that at that sampling frequency your ciliary beat cannot exceed 50 Hz. For a better represntation of your signal I would sample 5 to 10 times the highest frequency in the ciliary beat.

There are several signal analyzer examples that come with LabVIEW 6.0 that might help get you started. They are found under Help >> Examples >> I/O Interfaces >> Data Acquisition >> Signal and Spectrum Analyzer.