Analog Monochrome Video Acquisition With Daq 6025e

[Mic_Scale]

Hello,

My monochrome video signal is seen here (bottom of picture):

http://forums.ni.com/attachments/ni/200/7468/1/H_Vid.gif - Analog (0 - 0.2 V).

I'm trying to acquire this signal, using the 6025e board. The H-Sync signal, seen on the picture (on upper part of the picture) is used as a retriggable trigger for starting the acquisition for each line, then getting a number of samples (=pixels). This process is continued until a  sufficient number of lines is acquired.

The video signal is connected to one of the AI, AI7, and the H-Sync to the PFI6.

I've used an example for this application in Labview, sent to me by one of the NI Application Engineers, MarcoPolo5, given here:

http://forums.ni.com/attachments/ni/200/7499/1/DAQ_Image2.vi

I've also built my own example, using the "stop acquisition" function (which I'm attaching).

On both of these examples, the image I'm getting is strange. I get perfectly horizontal grayscale stripes, no matter what the rate of my sampling clock (but I've noticed that the higher this rate gets, the wider these stripes get), no matter how many samples (=pixels) for each line I get, the pixels on each line are the same grayscale color. The stripes are of different grayscale color, and width.

Very similar to what is seen here, only the stripes are horizontal in my case:

http://icons.wagmedia.com/files/desktops/stripes/gray_stripes_1280.jpg

Looking at my video signal, it's obvious that it is rather similar on each line (between the H-shyness), meaning the lines should be similar to each other, and not sharp stripes. Secondly, on each line pixels should be varying in their color, since the video signal is not a horizontal DC line between two H-Syncs.

Could anyone please help me understand what is going on ?

P.S

Attaching my example.

[Tom_W_[DE]]

Hi Mic_Scale-

The first example you posted will be necessary if you want to collect a finite number of samples on each sync signal pulse and also retrigger the operation.  I would first suggest that you look at the data being captured by the DAQ card on a graph in LabVIEW.  Does it match what you are seeing with your oscilloscope?

Thanks-

[Mic_Scale]

Thank you very much for your answer,

You were right, I looked at the captured signal trough a graph and then with MAX. I've been seeing a strange thing. The 6025e board showed a different signal then was seen on the scope. It was as though the board was trying to make the input signal into a DC form.

I took an oscilloscope and generated a sinus signal with rate of 1 Hz. Then I've set the "Test Panels" in MAX for the AI5 to finite samples of 5000 with the sample rate of 1000 Hz. I saw a pretty good picture of sinus wave. Then I started to increase the rate of the signal, and the image was getting so much worse, at mere 10 Hz, it was not possible to see the sinus, it looked like a picture of a simple hair brush, or white noise.

The problem was occurring with other AI's too.

We use an Analog Devices 5B01 backplane for analog signal isolators to the board:

http://www.analog.com/en/prod/0%2C2877%2C5B01%2C00.html

with the 5B31 isolated V input module:

http://www.analog.com/en/prod/0%2C2877%2C5B31%2C00.html

Could the problem be connected to them ?

[AnalogKid2DigitalMan]

The 5B31 module has a 4Hz bandwidth with a pretty steep rolloff. Thus you cannot expect to acquire any signal with a frequency content >4Hz with any appreciable amplitude, I think they also make a 10KHz version of the module. Or if safe, just eliminate the isolation if possible.

[Mic_Scale]

Hello, thank you very much for answering.

If I would like to use Isolation, will these 10K BW modules not interfere my analog signal ?

Given that my H-Sync is within the range of 60-500 Hz ? And my board is the 6025e Daq, with the limited rate of 200K per seconds.

Thanks again.

[e-Cash-Money]

Hello Mic_Scale,

If your signal is going to be 500 Hz (at a maximum), then a 10KHz bandwidth should be more than enough.  Sampling at 200KS/s would be 400x greater than the signal frequency, so I think the DAQ board will be ok as well.

[Mic_Scale]

Hello E.Lee,

Well, my H-Sync is 500 Hz at the most, but on each line the analog video signal is not limited to this band width, isn't it ?

I mean, suppose the picture is a half black and half white pattern, then this is a step function on the scope, and is has all the spectrum of frequencies, no ?

Or do you say that 10K is enough to represent this kind of image ?

Thank you for your help.

[Mic_Scale]

Hi, just wanted to emphasize that I'm asking about the analog video signal isolation with the 10K BW, and not about the digital isolation of the H-Sync pulse.

Thanks.

[BruceAmmons]

If you have a filter with a 10 kHz cutoff, you can probably use signals up to about 8 kHz.  After that, the signal amplitude will start to be reduced.  The exact place where this starts varies with each filter.  This is the same type of effect you observed with your 4 Hz filter.

Your pixel clock is the critical frequency to look at.  If the pixel rate is higher that 8 or 9 kHz, you will lose some details with the 10 kHz filter.

If you use the 10 kHz filter, you will want to acquire your data at a minimum of 2x10 kHz.  25 kHz would probably capture any data that is useful.  50 kHz would be overkill but the signal might look better visually.  Anything above 50 kHz is a waste of data.

Bruce

[Mic_Scale]

Thank you for your advice, Bruce.

Since we currently are in the configuring part of our system, meaning there are 3 factors we're working with:

The size of the area to be scanned by the camera, the pixel number in a line acquired to the image and the pixel clock rate, I cannot tell exactly what the pixel clock rate will be. One thing I'm sure of is that the rate is planned to be high above 10Khz, it could easily reach 150Khz and above (up to 200 Khz).

Is there another suitable solution for isolating the AI channels, without hurting the bandwidth  ?

You see, just before the analog signal reaches the 6025e board, there's a control box which controls contrast and brightness of the signal. A slight rotation of contrast could easily boost the voltage to very high values in a very short period. I'm afraid that a not careful hand could damage the system.

Thanks.