Greetings, all.
I am designing a particular piece of training/test equipment that was originally specified to handle a signal at 10 Mhz, divided by 10 so we are sampling only 1 in 10 data to display (1 MHz). Display is an X-Y type, with magnitude on the y-axis and the x-value is simply clocking along the x-axis (and loops back. The operator makes keypad entries defined by where the data are indicated (LED lights up in appropriate quadrant). Responses are stored by microcontroller, and reports of percentage correct, etc. are produced.
And, 2 weeks ago, the specs change. We are still only sampling at 1 MHz (or below), but the input data now may be as fast as 100 MHz. (I am still waiting for the call when the client says "Here, I am sending another $10,000 and extending your deadline 3 weeks so you can really have time to vet the prototype thoroughly before we see it." (Which we do anyway.))
Sorry, back to the problem at hand. The simplest way that I can think to do this without resorting to an ECL divider is to find a TTL counter that can handle an incoming data stream at 100MHz, divide by 10, and output a 10 MHz signal, and feed that into the existing circuitry. There may be other ways to do it (I have considered multiplexing x 10 and only send 1 output to the existing circuitry. But finding a multiplexor with the relevant specs has not been easy, either.
If I go the counter/divider route, this encapsules what I am looking for:
decade counter/divider that is:
-synchronous
-can handle Freq(in,max)=100MHz
-clocks on positive edge
-a reasonable way to get the value of the data stream out every 10 counts (hi-speed A/D?, 9x-ring counter started with a 1 with an OR to the data value?(Frankly, I do not think we could find a walking ring counterthat could be pushed that fast))
and finally, a tight, jitter-free 100MHz clock to time the chip with. At 10 MHz the was simply a buffered TCXO. Do such devices exist at 100MHz, or do I get out my PLL book?
Any and all suggestions are appreciated. I have been through TI's counter list (about 2/3's of it) and found 2 counters that will run at the necessary speed, but lacked any of the other points. And each "look" requires opening the datasheet and hunting out the details, because in thespec matrix, not even the freq(max) is listed. So that is a time consumer for sure!
Regards to all,
Tom Bellows
