k36853p@saha.hut.fi (Tapani Otala) (10/18/90)
I bought my 48SX around July. The first set of batteries drained around
August. At that time I thought the batteries that came with the machine
were old/crap anyway, so I didn't think anything was wrong.
But, couple of days ago the second set of batteries drained (_naturally_
about 20 minutes before mid-term physics exam...). Now this is not
exactly what I would have expected, so I measured the batteries --
seems that one of the three batteries is drained, while the other two
are in good condition.
Any suggestions as to reasons why this happened? I could _almost_
accept the fact if all three batteries would be drained, but...
Could it be a) a single bad battery, b) a bad contact somewhere?, or
c) an extraordinary stroke of bad luck...
Couple of questions: how much does the serial I/O drain the batteries?
Do the Kermit commands in 48SX leave the serial port open? The manual
states something about closing the port when not in use (so as to
conserve batteries), but I thought this applied to the low-level I/O
commands only. The reason I'm asking this is that I've been doing
some pretty heavy serial I/O using the 48SX as a Kermit server.
Also while I'm certainly not an expert on 48SX, I should think that the
checking of pending alarms does drain the batteries somewhat even when
the machine is turned off. I _used_ to have a quite large alarm list
(around 20-30 alarms)... so if this has any bearing on the problem...
Any comments are appreciated, via e-mail or as replies to this
article. (I don't mind bying new set of batteries every other
year or so -- but I _do_ mind bying them every two months...)
-- k36853p@saha.hut.fi ! Clothes make the man. Naked people have little
-- tpo@otax.tky.hut.fi ! or no influence on society. -- Mark Twain
-- Tapani_'TpO'_Otala ! edp@jareth.enet.dec.com (Eric Postpischil (Always mount a scratch monkey.)) (10/18/90)
In article <1990Oct18.013921.28895@santra.uucp>, k36853p@saha.hut.fi (Tapani Otala) writes: >Also while I'm certainly not an expert on 48SX, I should think that the >checking of pending alarms does drain the batteries somewhat even when >the machine is turned off. I doubt it. HP posted specifications on power drain, and the 48 uses very little when it is turned off. Observe that the number of alarms you have is irrelevant; all that is necessary is to check the earliest pending alarm. I'd expect the hardware to have a register for the current time and a register for the earliest pending alarm, and some clock circuitry updates the current time and generates a power-on interrupt of some sort when the pending time is reached. -- edp
prestonb@hpcvra.CV.HP.COM (Preston Brown) (10/19/90)
>how much does the serial I/O drain the batteries? The normal run current for the 48SX is about 10mA, and using the I/O raises that to about 12mA. Please note these are just averages, your calculator may be different. >Do the Kermit commands in 48SX leave the serial port open? The manual >states something about closing the port when not in use (so as to >conserve batteries), but I thought this applied to the low-level I/O >commands only. The reason I'm asking this is that I've been doing >some pretty heavy serial I/O using the 48SX as a Kermit server. Yes, they leave it open; the current drain is about 12mA when doing I/O but drops to about 5mA while waiting. >I should think that the >checking of pending alarms does drain the batteries somewhat even when >the machine is turned off. Nope, the 48 sets the harware to wake it up when the next alarm is due and then goes to sleep. In general, the battery life for the 48 may be only a few weeks if you use it alot. It sounds like your unit is behaving normally. Preston
ds@dodger.ATT.COM (Doug Sturim,1A-138,2034425999,TRW) (10/20/90)
In article <1990Oct18.013921.28895@santra.uucp> k36853p@saha.hut.fi (Tapani Otala) writes: > >I bought my 48SX around July. The first set of batteries drained around >August. At that time I thought the batteries that came with the machine >were old/crap anyway, so I didn't think anything was wrong. > >But, couple of days ago the second set of batteries drained (_naturally_ >about 20 minutes before mid-term physics exam...). Now this is not >exactly what I would have expected, so I measured the batteries -- >seems that one of the three batteries is drained, while the other two >are in good condition. > I too am having a battery drain problem on my rom A HP48. My batteries last about one month and that's with no alarms and little usage. However all of my batteries drain to about 1.1V. I called HP on this and they said that I would be entitled to warranty service.
ds@dodger.ATT.COM (Doug Sturim,1A-138,2034425999,TRW) (10/21/90)
Now I am pissed. My 48's batteries went dead in 3 days. Because I have 128K ram card, I lost 85K of programs. The real pisser is my batteries went right when I was doing a full memory backup. SO NOW I have no archive of my memory. Now the 64 dollar question: If I send my calculator back, will I get a ROM E? Does anyone at HP know what versions of the calculator the warrent dept. is sending out now? Or would I be Better to tough the battery problem out until I am offered the new upgrade? Comments, criticisms, and flames are welcome. Doug ------------------------------------------------------ Doug Sturim TRW New London, CT att!chinet!mcdchg!clyde!ds ======================================================
akcs.michaelv@hpcvbbs.UUCP (Michael VanLoon) (10/23/90)
Although I have no way of knowing this for sure, my assumption for the alarms is this: the 48sx has a countdown timer that operates off a small oscillator when the 48sx is off (and on, as well). Whenever an alarm is set, the calculator looks to see which one is the next one to be activated. It then sets the countdown timer to do in interrupt in so many clock ticks, which will set off the alarm (and wake up the machine, if needed). Note that it only sets one timer no matter how many alarms you have set. Once the most recent alarm goes off, it resets that timer for the number of ticks until the next alarm. Consequently, the calculator would use the same amount of power no matter how many alarms you have set. This is just my theory, but it seems like the most logical way for the alarm mechanism to be implemented. Michael V ISU