| ??? 03/10/09 22:48 Read: times |
#163330 - There's nothing confusing about the procedure Responding to: ???'s previous message |
Erik Malund said:
Presumably, it matters not in what order/sequence you light them, so long as the entire 1280 LED's are scanned, whether on or off, within 8 ms. Scanning individual LED's means 6.25 microseconds per LED.
this is where you are confoosed not ONE, but a row of LEDs are lit at a time, giving, in the OPs scenario, 1ms. I'm not confused about how you want to do this. It'll require a current-limited source at each anode, though, else your driver goes poof if an LED shorts. What confuses me is why you'd recommend such a costly part when a couple of ubiquitous parts such as a '2803 and a '595 will do essentially the same thing. You're wanting to drive the cathodes with your 80 bit-wide controlled-current registered SIPO (10 concatenated 8-bit registered SIPO registers) that apparently sinks current when turning the LED ON. Are you using a bipolar transistor to power the anodes, or a mosfet? I'd guess you drive resistors to current-limit the flow into the anodes, and your cathode-driving SIPO register sinks the current for each column of 8 LED's, as each row of anodes is, in turn, energized via some semiconductor. What does this cost? What does the 10 8-bit Allegro parts cost when you buy then <25 units? Presumably the cost of the 8 bipolars or mosfets and the 80 current-limiting resistors on the 80 anodes is negligible by comparison. Remember, the O/P has to drive two such arrays of 8x80. I found the '595 for 65 cents and the ULN2803 for 39 cents US via Octopart.com. I found the A6833 at $6.53 each and only Newark claimed to have them. I believe a lot of money and exasperation can be saved by using more common parts available around the world, e.g. 'HC595 or 'HC4094 buffered with ULN2803 or equivalent to drive the columns at their cathodes. The rows would be driven exactly as you'd drive them when using the Allegro parts, and you'd simply put a resistor to the common anode supply (from whatever switches the row-drive on and off) to limit the current. Since the '2803's are just darlingtons with a common emitter connection, and since each one would drive one cathode in each column, it could easily be connected to GND through a resistor that would then, being an NPN darlington used one at a time, control the current based on the base voltage. The datasheet describes the inner circuitry, a couple of resistors, and I'd guess that combination would be considerably cheaper than the 10 Allegro parts, and easier to obtain, no matter where one buying them. the sequence is
enable row 8 anode drivers and CONCURRENTLY clock row 1 into the column drivers NOTE1 disable row 8 anode drivers latch row 1 into the column drivers enable row 1 anode drivers and CONCURRENTLY clock row 2 into the column drivers disable line 1 anode drivers latch the column drivers enable line 2 anode drivers and CONCURRENTLY clock row 3 into the column drivers ... without using a double register serial to parallel you will be lost regardless Erik NOTE1: this, of course, requires the (...) cycle to be complete The procedure is clear enough. What's unclear is why one would want to spend so much money on this function. RE |
