Light pen

How is it okay, Kirill? One day *25 Feb 01 10:17:02* I saw Kirill Frolov writing to Sergey Svetlov and here... SS>> layout (with your “tips”)... And simpler, this is: Change SS>> frames 50 times per second, in one frame the beam passes across the screen 15625 SS>> pulses per second. This is what you need to catch with a photocell! KF>> This is in the Y coordinate. What about X? KF> SS> Yes, what does the Y coordinate have to do with it? At a certain quantity SS> pulses across the screen, location is determined SS> photocell! KF> KF> Here is the algorithm for determining the X and Y coordinates, draw here with diamonds ^^^ - ? KF> block diagrams or whatever you want. Or don't write nonsense. KF> SS> And for speck (screen usage ~70%), count pulses and SS> even less. Look carefully above! KF> KF> And how many of them should be counted? Let's assume that KF> there are at least 100,000 dots on the screen. So that they fit into the frame KF> information must be transmitted at a speed of 100000*(frames per KF> second)=5 million points per second. Actually more. Our TV the standard is 625 lines. From smart books I deduced that in 1 second. 2 full frames are obtained, that is, 25Hz per frame. (easy to check, dividing the line frequency per frame and multiplying by 2) Now a little theory: The beam scans the screen from left to right and top to bottom in 1 frame (approximately!). A table of impulses is being compiled (it may not be called that way - I’m not a programmer ), upon reaching a certain impulse, the photocell determines coordinates from the table. Considering that the definition frequency is in the audio range area (!), I can assume that this is not so expensive for the processor. KF> The position of the pen is defined as the time from the beginning of the frame to the flash of the beam KF> under the pen. Only one ray _sequentially passes through all points KF> times - that means to get accuracy to one point you need to be able to KF> count time and monitor the state of the photodiode KF> faster than the beam passes one point on the screen. It turns out 200 KF> nanoseconds (actually even less) per point! Spectrum is resting here ^^^ - Where did such data come from - justify it. KF> however, like any Pentium-III. It must be done or with iron logic ^^^ - you line frequency, at 85 Hz (at least 800x600), do you know him? KF> or on a specialized microprocessor. KF> SS> PS: well, I don’t know how to explain anymore... KF> KF> Because you don’t know how it works. Read smart books... * With the best... Sergey. *

Hello, Sergey! Once, Tue Feb 27 2001 12:17, Sergey Svetlov wrote to Dmitry Lomov: KF>>> And how many of them should be counted? Let's proceed from the fact that KF>>> that there are at least 100,000 dots on the screen. So that they KF>>> fit into the frame, information must be transmitted at speed KF>>> 100000*(frames per second)=5 million dots per second. Actually KF>>> more. DL>> SS>> Our TV. the standard is 625 lines. From smart books I deduced that for SS>> 1 sec. 2 full frames are obtained, that is, 25Hz per frame. DL>> DL>> what? SS> Maybe you can explain how the TV works... no, I won't be here. When I teach children in a circle, the process is superficial familiarization with the raster and the principles of its formation on digital chips takes about 4 hours, during which I wave my arms wildly and write a bunch of boards, along the way asking a lot of leading questions, so that they themselves everyone guessed ;) in one second 25 full frames are obtained. In Spectrum, lowercase alignment no pulses are used, so the frame rate is about 50 hertz. SS>> upon reaching a certain impulse, SS>> photocell, the coordinates are determined from the table. SS>> Considering that the definition frequency is in the audio region SS>> (!), DL>> DL>> detection frequency - 50 hertz, detection accuracy - +-140 ns ;) DL>> "spectrum" definition - about 7 MHz. SS> Yes, I'm not talking about the full video signal! This is not about SS> catching pixels, and the method determining the coordinates (of something) SS> on the features of the TV. How do you differentiate between the sensor being in the center of the screen and being alone? a centimeter to the right? by its sound frequency? no, just a sensor impulse will be ~1 µs later. what is the sound frequency here? ;) SS>> I can assume that this is not so expensive for the processor. DL>> DL>> for a processor this is unrealistic. SS> Does it have difficulty processing audio frequencies? see above. DL>> does this postulate also need to be substantiated? SS> Yes, desirable. there is nothing to erase quotation. I don’t remember what needs to be justified ;) All the best. Dmitry.