This is a brief update re: our stop meeting at CIT March 26th.
I have benchmarked the readout time and it looks like my guess of
.3 MByte/sec was conservative. To remind you, the "fastest" readout
rate I have gotten is 3 MByte/sec. I modified that code to do some
reasonably realistic things like count rollovers to acquire a 1000 us
window of data; this has only slowed it down to about 2 Mbyte/sec.
That is with my fast computer; the uVaX+VIC might be 50% slower. Also,
it might be neccesary to add to the code I have- it works but is
minimal in what it checks. Anyway, I am not so worried about rates
<< .3 Mb/sec now.
By the way, I lost a factor of two in one of the numbers I told you.
It is 328 us between rollovers (2^16 x 5 ns). This is different from
the 160 us minimum that the memory will hold (with zero suppress off).
So in a readout algorithm where I count rollovers, the granularity
is actually 328 us; in fact, it is more like 656 us to avoid fencepost
error. I think I will try to revise my readout to work with an actual
calculated time and see how much slower that makes things. It may be
a good tradeoff against data size.
That's all for now. If anyone has thoughts about some of our estimates,
I'd like to hear them. I look forward to seeing rzl's scheme for the
stop manager card.
Ciao,
Ed