|
In lugnet.robotics, Claude Baumann wrote:
> 2. I think that the RCX firmware should survey
> the battery status and more and more disable
> current-intensive functions... We could reduce
> the consumption to a few mA, if my memory doesn't
> leave me.
That sounds wonderful, although well beyond what I can do (which is why I
love reading your projects & posts).
> 3. As to the sleep mode, only the non-maskable
> interrupts can awake the RCX... we'll play with
> the watchdog-timer solution first...
Which sounds good, although ultimately I worry about what happens when the
power drops really low, and the watchdog timer keeps powering the system up to
check the power, draining it down faster than a power-controlled system would.
But I completely agree with applying the SW hack before the HW hack in this
case.
> 4. However, his battery-charging doesn't convince me.
> With our rover, we just can use 2 LEGO solar panels.
I's disagree, for a number of reasons. The LEGO solar cells are not great
efficiency, and even for their area are rather heavy, which means powering the
motors to move them will "cost" more. Notice it's not that a LEGO rover powered
by two stock LEGO solar cells isn't possible - it's just that using the lower
efficiency cells, coupled with the greater weight, implies even more time
recharging (sitting still) than moving. It comes down not to power management,
but power to weight... and both the LEGO solar cells (high weight, with low
efficiency) and batteries (high weight and *very* poor efficiency during
charging... notice how hot batteries get while charging... plus poor response
time to short transient power pulses from the panels) and poor in this regard.
Shuffling between voltage and current doesn't change the fact that what you
get out of a solar cell is power (current times voltage)... and worse,
converting & conditioning that power induces loosing some as well.
--
Brian Davis
|
|
|
../snip>
> Shuffling between voltage and current doesn't change the fact that what you
> get out of a solar cell is power (current times voltage)... and worse,
> converting & conditioning that power induces loosing some as well.
Please consult http://www.convict.lu/htm/rob/mars_IV.htm. I propose to do the
discussions in this forum.
|
|
|
Question to H8/3292 specialists (Kekoa Proudfoot, Michael Anderson, Markus Noga,
Ralph Hempel, John Hansen, Dick Swan ... these are the ones I know, ... are
there others... probably?)
I tried to plunge the RCX into sleep mode (not the normal RCX software stand-by
mode), which, according to the datasheet, halts the CPU, but keeps on running
the clock. The hardware manual explains that the H8/3292 can be exited from this
state through an internal or external interrupt.
But, I was not able, after several tests of doing this. I first cleared the 7th
bit of SYSCR, then configured an interrupt, added a dummy interrupt service
routine, finally called "SLEEP". The current dissipation corresponds to the one
noted in the datasheet, but none of the interrupts -neither external, nor
internal, woke-up the RCX. Has anyone got an idea, what I might be doing wrong
here? Am I missing a configuration sequence? Is there a special timing that
needs to be respected?
This is quite intriguing, because it should work according to the datasheet.
http://www.convict.lu/htm/rob/mars_IV.htm
|
|
|
In lugnet.robotics, Claude Baumann wrote:
> Question to H8/3292 specialists (Kekoa Proudfoot, Michael Anderson, Markus Noga,
> Ralph Hempel, John Hansen, Dick Swan ... these are the ones I know, ... are
> there others... probably?)
>
> I tried to plunge the RCX into sleep mode (not the normal RCX software stand-by
> mode), which, according to the datasheet, halts the CPU, but keeps on running
> the clock. The hardware manual explains that the H8/3292 can be exited from this
> state through an internal or external interrupt.
>
> But, I was not able, after several tests of doing this. I first cleared the 7th
> bit of SYSCR, then configured an interrupt, added a dummy interrupt service
> routine, finally called "SLEEP". The current dissipation corresponds to the one
> noted in the datasheet, but none of the interrupts -neither external, nor
> internal, woke-up the RCX. Has anyone got an idea, what I might be doing wrong
> here? Am I missing a configuration sequence? Is there a special timing that
> needs to be respected?
>
> This is quite intriguing, because it should work according to the datasheet.
>
> http://www.convict.lu/htm/rob/mars_IV.htm
Perhaps it's not the best habitude to answer one's own post, but the good news
is that we were able to fix the problem:
The issue was that we were using the Assembler code
bclr #7,@0xFFC4
in order to select sleep mode instead of software stand-by mode. We already had
troubles with direct bit setting or clearing with other microcontrollers.
Sometimes the hardware requires that certain registers are first read and then
set. This also is the case with the SYSCR register of the H8/3292. So, we used
the following code instead:
mov.b @0xffc4:8,r6L
and #0x7F,r6L
mov.b r6L,@0xffc4:8
==> And everything works now. We already tested the NMI-interrupt that is
generated through the watchdog timer overflow and IRQ0 interrupt that is
produced through the RUN-button pressed.
|
|
|
In lugnet.robotics, Claude Baumann wrote:
|
../snip>
|
Shuffling between voltage and current doesnt change the fact that what
you get out of a solar cell is power (current times voltage)... and worse,
converting & conditioning that power induces loosing some as well.
|
Please consult http://www.convict.lu/htm/rob/marsIV.htm. I propose to do the
discussions in this forum.
|
Hi Claude,
If you need to model your rover, I just LDrawed the solar panel and the
supercap...
Philo
|
|
|
In lugnet.robotics, Philippe Hurbain wrote:
|
Hi Claude,
If you need to model your rover, I just LDrawed the solar panel and the
supercap...
Philo
|
Hi Philo, this is amazing. Thanks a lot.
|
|
|
I've been away a long while.
How are RCXs currenlty programmed? I used a C type
program before and I liked it a lot, is that still
used?
Also, is there a good central site for RCX bot builds?
Thanks!
Rich
____________________________________________________________________________________
Be a better friend, newshound, and
know-it-all with Yahoo! Mobile. Try it now. http://mobile.yahoo.com/;_ylt=Ahu06i62sR8HDtDypao8Wcj9tAcJ
|
|
|