• Please review our updated Terms and Rules here

Tandy 2500SX/20 BIOS error: Time of day clock stopped

K

keenerb

Veteran Member
Joined
Mar 11, 2016
Messages
861
Location
Georgia, USA
Pulled my old 2500SX out of retirement to tinker with for a bit. Bios battery was dead, so I replaced, and I'm getting the above error during bios boot, and have to hit F1.

Nothing seems to help; swapped batteries and manually grounded RCLR pin, I've reset time 20x.

Any idea what's up? Maybe I have a bad DS1285Q?

It keeps time perfectly while powered on. VBAT is correct. It maintains CMOS memory while power is disconnected.
 
Last edited:
With some computers, two actions are required: Replace battery, then go into SETUP (CMOS SETUP) to change/set the time there.
 
Yeah I've done that a ton. I think something's gone wrong. I've swapped out the battery a half-dozen times in the past years.
 
Last edited:
keenerb said:
It keeps time perfectly while powered on. VBAT is correct. It maintains CMOS memory while power is disconnected.
Click to expand...
That is new information. I was still showing as the last person to have edited the thread. Perhaps you should have created a new post rather than editing the original after I made my earlier post.

The RTC component of the DS1285 is 'ticked' via one of two methods:
* A suitable crystal tied to the X1 and X2 pins (pins 2 and 3), together with exciter circuitry in the DS1285; or
* A 32.768 kHz clock signal applied to the X1 pin (pin 2).

The 'clock signal' method does not appear in the data sheet that I see of the DS1285, but the data sheet indicates that the DS1285 is a "Drop-in replacement for IBM AT computer clock/calendar", and the IBM AT uses the clock signal method.

Which of the two methods is your DS1285 using ?

If the 'clock signal' method:

As I stated, the IBM AT uses the clock signal method. Per [here], there is an oscillator circuit providing the subject clock signal. A 'clock not ticking/advancing when AT powered off' symptom is known to happen in the IBM AT when the battery voltage drops below the point where, although still suitable for the AT's RTC chip, is not enough voltage for the oscillator circuit to work.

Volume 1 of the 'PC Engineers Reference Book' makes reference to the situation with, "The minimum standby voltage for the 146818 is 2.7v, but your settings can remain even down to around 2.2v. Usually, the clock will stop first, as the oscillator needs a higher voltage to operate."

Are you in a position to verify that the oscillator circuit is still oscillating when the computer is powered off ?
 
modem7 said:
That is new information. I was still showing as the last person to have edited the thread. Perhaps you should have created a new post rather than editing the original after I made my earlier post.

The RTC component of the DS1285 is 'ticked' via one of two methods:
* A suitable crystal tied to the X1 and X2 pins (pins 2 and 3), together with exciter circuitry in the DS1285; or
* A 32.768 kHz clock signal applied to the X1 pin (pin 2).

The 'clock signal' method does not appear in the data sheet that I see of the DS1285, but the data sheet indicates that the DS1285 is a "Drop-in replacement for IBM AT computer clock/calendar", and the IBM AT uses the clock signal method.

Which of the two methods is your DS1285 using ?

If the 'clock signal' method:

As I stated, the IBM AT uses the clock signal method. Per [here], there is an oscillator circuit providing the subject clock signal. A 'clock not ticking/advancing when AT powered off' symptom is known to happen in the IBM AT when the battery voltage drops below the point where, although still suitable for the AT's RTC chip, is not enough voltage for the oscillator circuit to work.

Volume 1 of the 'PC Engineers Reference Book' makes reference to the situation with, "The minimum standby voltage for the 146818 is 2.7v, but your settings can remain even down to around 2.2v. Usually, the clock will stop first, as the oscillator needs a higher voltage to operate."

Are you in a position to verify that the oscillator circuit is still oscillating when the computer is powered off ?
Click to expand...

You have read my mind. I had just gotten to the datasheet where it's talking about oscillators. I have a meter that should be able to read that crystal for certain.
 
osc.PNG

It definitely has a dedicated oscillator tied to pins 2 and 3. I'll see what I can detect on those pads!
 
My VC99 can't detect a frequency across the two pins of the oscillator or between either pin and ground.
 
keenerb said:
My VC99 can't detect a frequency across the two pins of the oscillator or between either pin and ground.
Click to expand...
Different meters have different abilities in what they can measure. The alternating waveform (which probably has a DC component) may not be the kind of signal that a VC99 can determine the frequency of.

keenerb said:
I believe the leg closest to the chip (X1) is supplying 3v to the crystal?
Click to expand...
Crystals (as distinct from oscillator units) do not work that way; oscillations do not result from simply applying a DC voltage to them. The crystal is part of an oscillator circuit, with the remainder of the circuit being in the DS1285.

BTW. Re crystal replacement. Note that apart from frequency, crystals have other electrical properties. Chip data sheets usually provide the required crystal properties.
 
keenerb said:
It keeps time perfectly while powered on.
Click to expand...
Presumably, you are using DOS' TIME command to determine that.

The question is, which of the following is the case:
* In an AT-class computer, each time that the TIME command is used, DOS reads the time from the RTC chip.
* In an AT-class computer, at power-on time, DOS sets the DOS clock to the RTC clock. Then, each time that the TIME command is used, DOS displays the DOS clock.
 
modem7 said:
Presumably, you are using DOS' TIME command to determine that.

The question is, which of the following is the case:
* In an AT-class computer, each time that the TIME command is used, DOS reads the time from the RTC chip.
* In an AT-class computer, at power-on time, DOS sets the DOS clock to the RTC clock. Then, each time that the TIME command is used, DOS displays the DOS clock.
Click to expand...

Using dos time command and MTCP's sntp command. It'd DR-DOS 7.02 if that clarifies anything.
 
modem7 said:
The question is, which of the following is the case:
* In an AT-class computer, each time that the TIME command is used, DOS reads the time from the RTC chip.
* In an AT-class computer, at power-on time, DOS sets the DOS clock to the RTC clock. Then, each time that the TIME command is used, DOS displays the DOS clock.
Click to expand...
I did an experiment. I fired up one of my IBM 5170's into DOS (IBM DOS 3.10). As expected, periodic use of the TIME command showed the current time. Also as expected, the MC146818 RTC's clock output pin (CKOUT), pin 21, outputted a 32.768 KHz signal. (Pin 21 on the DS1285 RTC is used for a different purpose.)

I then disabled the 32.768 KHz oscillator signal going into the MC146818 RTC. As expected, the MC146818 RTC's clock output pin no longer showed a signal. Periodic use of the TIME command continued to show the current time.

Therefore, in an AT-class machine, DOS' TIME command reads the DOS clock (clock 'ticking' triggered by an interrupt), not the RTC's clock.

[ When I then warm booted the 5170, I saw the POST display a '163-Time & Date Not Set' error. One of the known triggers for that error in a 5170 is the MC146818 RTC's clock not advancing. I then enabled the 32.768 KHz oscillator signal going into the MC146818. The error did not appear when I then warm booted the 5170. ]

Putting everything together, it appears that the clock in your DS1285 RTC is not advancing at all, computer powered on (DS1285 RTC powered by motherboard) or powered off (DS1285 RTC powered by battery). When the computer is turned on, the POST does a couple of reads of the DS1285 RTC's clock, sees that the clock is not advancing, then displays 'Time of day clock stopped'. So, crystal or DS1285 ?
 
modem7 said:
I did an experiment. I fired up one of my IBM 5170's into DOS (IBM DOS 3.10). As expected, periodic use of the TIME command showed the current time. Also as expected, the MC146818 RTC's clock output pin (CKOUT), pin 21, outputted a 32.768 KHz signal. (Pin 21 on the DS1285 RTC is used for a different purpose.)

I then disabled the 32.768 KHz oscillator signal going into the MC146818 RTC. As expected, the MC146818 RTC's clock output pin no longer showed a signal. Periodic use of the TIME command continued to show the current time.

Therefore, in an AT-class machine, DOS' TIME command reads the DOS clock (clock 'ticking' triggered by an interrupt), not the RTC's clock.

[ When I then warm booted the 5170, I saw the POST display a '163-Time & Date Not Set' error. One of the known triggers for that error in a 5170 is the MC146818 RTC's clock not advancing. I then enabled the 32.768 KHz oscillator signal going into the MC146818. The error did not appear when I then warm booted the 5170. ]

Putting everything together, it appears that the clock in your DS1285 RTC is not advancing at all, computer powered on (DS1285 RTC powered by motherboard) or powered off (DS1285 RTC powered by battery). When the computer is turned on, the POST does a couple of reads of the DS1285 RTC's clock, sees that the clock is not advancing, then displays 'Time of day clock stopped'. So, crystal or DS1285 ?
Click to expand...

Crystal would be easiest to replace, so I will start there.
 
Replaced the crystal with a 6pf (listed on a datasheet as compatible) and still the same error.

Not sure what else it could be other than the DS1285Q itself. I have a few of those coming in the mail now. Not sure if I'm up to resoldering that particular chip though...
 
Last edited:
keenerb said:
Replaced the crystal with a 6pf (listed on a datasheet as compatible) and still the same error.

Not sure what else it could be other than the DS1285Q itself. I have a few of those coming in the mail now. Not sure if I'm up to resoldering that particular chip though...
Click to expand...

Doesn't seem to affect much other than having to hit F1 during boot, I'll probably just live with it. I wonder if I could just remove that F1 check with a hex editor somehow.
 
FOr what it's worth, I ordered a half-dozen random oscillators from ebay and one batch fixed the problem. There were no specs, just listed as "32.768khz oscillator for ARDUINO blah blah blah" but as soon as I soldered this one in, it booted perfectly.
 
keenerb said:
FOr what it's worth, I ordered a half-dozen random oscillators from ebay and one batch fixed the problem. There were no specs, just listed as "32.768khz oscillator for ARDUINO blah blah blah" but as soon as I soldered this one in, it booted perfectly.
Click to expand...

This has been in wrong forum also; it can be moved to later pcs if it's a problem.
 
You must log in or register to reply here.
Back
Top