> -----Original Message-----
>
> the silence is starting to freak me out a little!
>
> Somebody say something!
I've been working on something, too, but I didn't want to post about it until I
was fairly certain I had the kinks worked out.
I put a supercap in my T200 (1.5F, same as I was using in the M100) to see
whether or not the issues Kurt talked about having with his T200 could be
overcome. I'm happy to say that overall it is a success, but with a couple of
caveats to be aware of. The T200's power on/off circuitry is sensitive to the
memory backup battery (or cap) voltage, and from what I've read this is already
a known issue where a T200 with a bad NiCd battery won't turn on and just
displays horizontal lines (the lines on mine pulsated and were accompanied by a
pulsating 'ticking' sound from the piezo speaker). It sounded like sometimes a
T200 in this state could be coaxed into working by leaving it with batteries or
an external supply connected for some time before trying to turn it on, and
that's what led me to trying the supercap.
Since the voltage rises linearly as a capacitor charges, and since the NiCd
battery was just being charged off the 6V supply through a resistor, the 1.5F
capacitor takes quite some time to build up enough voltage to be an effective
memory backup - much longer than the voltage would ramp up on a (properly
functioning) NiCd battery. When I first put the cap in my M100 it turned on
just fine (because the M100 doesn't care what the memory backup voltage is),
but I was dismayed to find that the clock and RAM reset when I turned it off
for a few seconds. Checking the cap voltage I found it was only about 0.3V, so
I decided to leave the machine alone for a while. After I think 20 minutes or
so it had hit 1.6V and I tried it again - this time it would hold RTC and RAM.
After a total of nearly an hour it was up over 3V and it eventually settled
around 4V.
Kurt's prior experiments with his T102 showed that he could leave the batteries
out for about 24 hours and not lose RTC or RAM. I tried mine 7 hours overnight
and it was fine - I haven't tried testing it for longer than that.
With the T200, I thought the key to success would be to leave the machine off
until the capacitor charges up to a voltage close to what one could expect from
the NiCd battery. It turns out I was right, but I did get ahead of myself a
couple of times due to impatience (which allowed me to collect some useful
data, oh well).
When I first put the batteries in my T200 after putting in the cap, it switched
itself on. I pushed the power button right away, thinking I should turn it off
and let the cap charge, but it went into the twitching/ticking two-lines
'seizure' if you can call it that. I turned the memory backup switch off for
10 seconds and back on and the machine switched back on again. I went into
BASIC and gave the 'POWER CONT' command so it would stay on while the cap
charged.
I waited about 20-25 minutes until the cap voltage was over 1.5V and tried to
turn the machine off, and was immediately punished for my impatience. Not only
was it now stuck twitching, but I couldn't recover by turning the memory backup
switch off. The machine was still open, so my solution was to discharge the
cap (using a flashlight bulb with two wires soldered to it, which was laying on
my bench and is handy for all sorts of things) and start all over again.
This time I gave it 45 minutes and the cap voltage was I think around 3V, and
power on/off worked fine. Cap voltage eventually settled around 4.0V, which is
what the NiCd was sitting at when I removed it.
I did try taking batteries out of the machine for 7 hours and it lost the RTC
and RAM. It also wouldn't turn on right away when I put the batteries back in,
which initially worried me, but after leaving it alone for a good 45 minutes or
more I tried again and it turned on just fine. (I had already reassembled the
machine so I couldn't follow the cap voltage to know for sure when it would be
ready.)
I just realized now, in writing this, that I had meant to do further tests of
incrementally increasing the amount of time without batteries to see how long
it does last as an RTC/RAM backup. I seem to recall reading somewhere that the
original Tandy 24k memory expansion modules are a bit power-hungry compared to
the RAM used in the 100/102 so that might explain the fact that it wouldn't
last overnight for me. I'll check the real runtime and report back, but
really, as long as it lasts long enough for a battery change I'll be happy, and
I'm quite certain it will last more than long enough for that. :)
So, the caveats in summary:
1. If you let the cap run dead, you'll need to give the machine around 45
minutes (or more, to be safe) with batteries or an external adaptor connected
before it will turn on. If the machine turns itself on when you put the
batteries or external adaptor on it, you'll need to user POWER CONT from BASIC
so it doesn't turn itself off while the cap is charging.
2. If you try to turn it on when the cap is at a low-but-not-quite-dead
voltage, you might not be able to turn it on at all without further draining
the cap, which should be doable by leaving the memory backup switch on and the
batteries out for some lengthy period of time (at least overnight, maybe more),
or open the machine up and drain the cap yourself with a suitable resistor.
(Note that some T200s with a bad NiCd have this same problem turning on.)
> I'm afraid I only have 2 M100 things to say:
> 1) I got about 4 solid hours of m100 gaming time in last week. TSWEEP
> is hard, ROG is also hard :-), the .CO tetris clone is evil fast, and
> the keys for Starblaze100 make no sense to my hands :-(.
I actually bought a Starblaze 100 cassette off ebay a month or two ago, and my
son and I have been having a lot of fun with it. It really opened his eyes as
to what is possible wrt performance and efficiency when one writes a program in
assembly rather than BASIC. :)
Which Tetris were you playing (and where did you get it)?
jim
