OK, my backup plan didn't work either (I was going to post to facebook and then post link, as suggested (and tried), with IMG tags/brackets, in the above posts).
So scratch that. My direct question, in words, is this. I am trying to design a power supply for a ridiculous number of chips (complex chips; image sensors/microprocesor). (1) Does one use Kirchhoff's laws to make a parallel power supply circuit and (2) can one treat the dynamic behavior of a functioning IC as that of a variable resistor, i.e. variable load?
I am told image sensors are like little "light-processing-microprocessors". I think that is an accurate sort of statement. CMOS tech. Probably need a "GND-island" underneath them? Mine are VGA, old chips, low tech. They have analog and digital supplies, and GNDs to match (for decoupling purposes; separate analog and digital; but i have read that all GNDs should connect at one (narrow) point, even if they are spearate analog and digital).
The circuit I envision includes many FIFO chips and FIFO-USB bridge chips. Both these chips require their own XTAL, and so, with an array of (say, right eye only) 6 CMOS imagers, 6 FIFOs, and 6 FIFO-USB bridges, one clock circuit can run all 6 CMOS imagers (an XTAL, fed to a PLL and/or fanout buffer chip, to synchronize all image chips), and the imagers feed out a pixel clock with the valid data on the 12 bit bus.
The six FIFO chips will "catch" the data, according to the pixel clock and the 12 data lines from the microproc/image sensors. The XTAL for the FIFO will govern ITS OWN internal operations, and (probably?) won't interact with the image sensors, or the FIFO-USB bridge (?). The power supply for the FIFO chip is somewhat complex, too (see below).
Finally, as far as clock circuits are concerned, the FT600 (FTDI Inc.) chip wants its own XTAL too, and this will govern its own internal operations and not interfere with the other two chips. Also, the USB will "talk" to the USB devices and the USB-host, according to usual protocol. That is the beauty of USB; it is universal!
So, clocks "taken care of", here is the real question, about power supplies. I am assuming that Kirchhoff's Law are a paramount, here? I mean, If I use, say, the LT3080 (1A; 3.3V output; but also, perhaps, can be designed to put out the 2.5 V that the FIFOs need?), and use them in parallel with each other, I can get the 3A that I ?might? need (worst case amperage scenario?). Now, I can use "one trace", set to 3.3V, "capable of up to 3A", and split it into six (branched, like a tree), and then there are "six traces" , each "set to 3.3V, by the linear regulator", and each "capable of up to 0.5 A"? Right? I suppose, maybe I should study Thevenin again? I dunno?
Furthermore, my "notion" of how ICs work in a circuit, is as "active element", or really, as "aggregate active elements", since there are about a zillion transistors on it, each one potenitally functioning as an active element. (again I am a biologist, and inexpert at these things!). Put another way, I also view them as a variable load; because they are capable of drawing variable current, I assume they are capable of "showing" variable resistance to the rest of the circuit; and crucial to my question, showing variable resistance to the power supply? Is this correct?
So, with a well designed power supply, and each of the 18 ICs dynamically changing the load they present to the circuit (in this case, a "six branch" trace, according to kirchhoff's law, 3.3V, as set by me on the 3QTY LT3080s--in parallel for heat spreading and increased amperage output--each of the six traces capable of 0.5A). I forget the combined amperage i calculated for each of the six "camera head-boards", but one power supply board can probably supply all six camera-boards. Another option would be to power with the USB 5.0 V power supply; but that seemed more complicated to me. I am a beginner and need to do things in the standard fashion.
So, finally, if this ("six-branch") is adequate, conceptually, I just need a regular AC/DC transformer, full bridge rectifier circuit, with some smoothing capacitors (to feed the LT3080s circuit). Well, not exactly. I still will need a 2.5 V supply for the FIFO memory. (This whole question would have been much easier with diagrams. I need to fix my laptop; got a 2000-0142 HDD error message; Dell says no warranty, buy a new HDD. it works, but its slower than a booting commodore 64! But that is an entirely different complaint!)
So to recap, I just need to know, am I using cKirhhoff's laws appropriately? Dropping 500 mA across each camera-board, at 3.3V, indicates that the EACH camera board should show a minimum of 6.6 Ohms to the power supply. (these numbers aren't correct... I will try to post EXACT numbers from the spec sheets, below this post.). AT LEAST this much is true; IF ALL 6 camera-boards showed less than 6.6 Ohms, simultaneously, THEN, the power supply circuit would be drawing more than 3A and something would burn or break or stop functioning.
OK, below will be the relevant numbers from the spec sheets, but "I am new" and don't know if I have to include ALL of the "input/output pin amperages"
|