Without handling your entire system I'm not really competent
to pronounce opinions on how it should work, and I haven't
used the MDA320 or it's code...
My back'o'envelope calculation indicates that your 10/sec sample
rate should be easily do-able without heroic measures. The HFS
program is for rates way above what you have stated, so it's
probably not much use to you.
What I would do at this point is back out to the simplest level,
maybe just sampling and tossing the values, to see where the
hangups are.
MS
raghavendra a wrote:
Hi,
I made use of the LEDs and also sent back packets at significant points
to get the time taken for the process and measured
(Timer or sampling time set as 100ms)
75-78 secs for getting 250 samples
3-5 secs for writing this data into the FLASH
I am using MICA2 motes with MDA320 boards. External sensors are
connected to it.
I am making use of the XMDA320 code for this.
Is it possible to only sample the relevant ports? and Would this help in
reducing the time taken for program execution?
I also came across the Highfrequencysampling program. Would this code
help in solving this problem?
Regards,
Ragh
On 10/17/07, *Michael Schippling* <[EMAIL PROTECTED]
<mailto:[EMAIL PROTECTED]>> wrote:
It's possible that the FLASH store takes a long time,
I've never used it so I don't know. You might try flashing
LEDs at significant points in the process and measuring
the actual time with a scope.
I'm not clear why you start a new timer with 500 interval.
Assuming your messages are not throttled you should be able
to start sending a new one from the sendDone() of the old.
As I said, if you aggregate multiple samples into a single packet
(your slippery use of sample and packet is/was confusing) you should
be able to keep up with a 10/sec rate indefinitely.
MS
raghavendra a wrote:
> Thanks for replying.
> Yes I am using call Timer.start(TIMER_REPEAT, 100); for data
acquisition.
>
> I sample at 100ms until there are 1000 samples stored in the Flash
> memory(I write an array of 250 samples four times). So this would
take
> around 100*1000*10^-3 which is equal to 100 secs or approx. 2
minutes.
>
> After 1000 samples are written I stop the timer and call
> Timer.start(TIMER_REPEAT, 500); and start reading out each of the
> samples from the Flash and send each of them as a radio packet.(I
have
> No constraints on the Reading speed)
>
> So I should start getting these messages after around 2mins, but
i get
> them after exactly 300-301 secs. Which boils down to 3 samples being
> sampled instead of 10.(1000 samples collected in 300 secs means
around 3
> packets per second).
> I am using Single hop.
>
>
> Sincerely,
> Ragh
>
> On 10/16/07, *Michael Schippling* < [EMAIL PROTECTED]
<mailto:[EMAIL PROTECTED]>
> <mailto:[EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]>>> wrote:
>
> I assume you have something like this driving your data
acquisition:
> call Timer.start (TIMER_REPEAT, 100);
> so you (should) get about 10 sample cycles per second.
>
> Are you then sending one sample per message, or aggregating
them in
> a manner like Oscilloscope? When you say "only 3 packets/sec
is being
> sampled" it sounds like you are not aggregating, and if not,
how do
> you know you're only getting 3 samples/sec?
>
> And according to my, now old, measurements a mica2 system
should be
> able to send about 25 messages per second. Are you using a
multi-hop
> layout? You (should) be able to put up to 14 2-byte samples
into the
> 29 byte message payload and at (lets be conservative, with no
hops)
> 20 messages/sec, get 280 samples/sec.
>
> MS
>
> raghavendra a wrote:
> > Hi,
> >
> > I am using MICA2 motes with MDA320 sensor boards. In my
application I
> > sample analog data at the rate of 10 samples/sec, store
the values in
> > the Flash, and then read them sequentially. I do the READ and
> send the
> > data back to the base at a slower rate since only
3packets/sec can be
> > sent at the most. But I see that only 3packets/sec is being
> sampled. How
> > can I increase the sampling rate with MICA2 motes?
> >
> >
> >
> > Thanks,
> > --
> > Ragh
> >
> >
> >
>
------------------------------------------------------------------------
> >
> > _______________________________________________
> > Tinyos-help mailing list
> > [email protected]
<mailto:[email protected]>
> <mailto:[email protected]
<mailto:[email protected]>>
> >
>
https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help
>
>
>
>
> --
> Ragha
--
Ragha
_______________________________________________
Tinyos-help mailing list
[email protected]
https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help
