• Please review our updated Terms and Rules here

IBM EduQuest 30/Thirty - RTC won't keep time

kishy

Veteran Member
Joined
Aug 22, 2009
Messages
1,065
Location
Windsor, ON Canada
Between a friend and I, we have recently acquired several (somewhere between 3 and 8 once we reconcile what's damaged) EduQuest Thirty all-in-one computers. All of them except one are working (as in, POST, video out, boot, and accept input) but have various chassis damage.

There is one curious problem I've found that is common to all of them, and I'm wondering if anyone here has any useful commentary.

These units use a Dallas DS1287 RTC module. In order to get them to boot from the HDD (or even remember that it has one at all), settings must be adjusted in the (Phoenix) setup menu. Those settings are lost on the reboot after saving the settings, so a working RTC module is completely necessary to boot the unit from HDD.

Initially, I swapped in a DS12887 new-old-stock module I had, and found that the unit would not keep time. On power off, the calendar and clock would return to their default values, or close to them. I suspected maybe I had found one of the unusual cases where a 12887 cannot replace a 1287, so I dug through my things for a modded one, and put that in, and found the same thing.

Then, since I wasn't sure if my modded one might have also been a 12887, I modded one of the EduQuest original 1287s...and found the exact same thing.

So now this scenario is proven across many EduQuest motherboards, with several RTC modules of both 1287 and 12887 varieties, with multiple batteries. It's starting to look like this is intended functionality. The settings are being kept, just the clock and date are not.

All of these units had token ring cards installed, so some type of NTP implementation is not really out of reach. It's just odd that the RTC doesn't seem to do its most core function in these. I would prefer to have them internally managing their date and time, but I'm not sure how they were used in their previous lives without it.

Any thoughts, please?
 
Last edited:
It sounds as if the SRAM is keeping its content, which is to be expected, but could it be that the oscillator isn't running to make the clock "tick"? Does the clock advance between power cycles?

Otherwise, it would seem to be mostly a matter of programming.
 
It sounds as if the SRAM is keeping its content, which is to be expected, but could it be that the oscillator isn't running to make the clock "tick"? Does the clock advance between power cycles?

The clock works normally while power is applied, including through a reboot without removing power, however, if the power is turned off the clock zeroes and the date goes back to something very early (probably the first date it allows).
 
I since my last reply started playing with the one that's in my living room to see if I could make any sense of it.

First thing I did was pop in a modded and battery equipped, originally-pulled-from-an-EduQuest, DS1287, which was used in another EduQuest a few days ago.

That module has been keeping time outside of the computer I plucked it from, and the EduQuest in front of me now is keeping time and will not clear it after a bunch of reboots and power cycles.

Absolutely at a loss. This module has been seen doing it, in other EduQuests that aren't this exact one, and another module has been seen doing it in this one. Maybe it's just temperamental getting them to first begin storing the date once they have a new battery attached?

The other computers are a 45 minute drive away, so I can't explore any more than this at this time.
 
Are you certain about the cold re-boot? Do you have a program that directly reads the registers of the RTC? There should be a few programs on SIMTEL20 to manipulate and examine the RTC registers. It's always possible that a warm boot doesn't bother to read the RTC registers.

Here's what's got me puzzled: You should be able to remove the chip after power-down and it should keep ticking right along--to zero the clock (if the internal battery is good), you'd need to actually write the clock registers. The operation of the clock itself does not depend on being plugged into the motherboard.
 
Are you certain about the cold re-boot? Do you have a program that directly reads the registers of the RTC? There should be a few programs on SIMTEL20 to manipulate and examine the RTC registers. It's always possible that a warm boot doesn't bother to read the RTC registers.

Here's what's got me puzzled: You should be able to remove the chip after power-down and it should keep ticking right along--to zero the clock (if the internal battery is good), you'd need to actually write the clock registers. The operation of the clock itself does not depend on being plugged into the motherboard.

Very certain, that it was behaving as I described. I was observing the loss of date/time in the setup menu or in Win95.

Agree that the RTC module continues keeping time when removed, because it is self-contained, and that has been my experience with them in the past (+ keeping settings). That's what seemed so odd about this. But, now I have one RTC that seems to be doing this correctly in this particular EduQuest. I don't know why it works now and didn't a couple days ago. I guess this resolved itself.
 
I know this is old but here is my experience and what I think is happening. The issue seems to be with setting the time in the BIOS setup itself. There is some kind of software bug that prevents it from working. However if I set the date and time through DOS (or Windows) it works correctly.
 
Back
Top