• Please review our updated Terms and Rules here

Olivetti M24/AT&T6300 no boot - keyboard leds blink

O

Oscar

Experienced Member
Joined
Sep 11, 2003
Messages
206
Location
Switzerland
Hi,

I think I've read all the M24 posts but could not quite get to my problem.

I bought a M24 (as I really liked them back then) and:
- had to unplug the hard drive as it clearly had crashed, rattling noise.
- on powerup: no video (but you hear the monitor come on), no beep, keyboard LEDs light up for half a second, then start rapidly blinking forever. Until you press reset and it all starts over.

I reseated chips, checked voltages, cleaned contacts to the riser board. Removed hard disk adapter card too. Nothing changes.

The nicad battery looked good and still gave 3.7 volts, I removed it just to be safe.
--> Am I correct in thinking that removing it will not hinder the boot process? It's just a RTC?

The dip switches and jumpers are all normal as per the manual (except the point below) and the seller told me he had not messed with it, it had booted right up the day before to the point where the hard disk obviously did not work. So sudden death.
--> Only question: dipSW 1, #7: this should be ON for mono 80x25, but it's set to OFF for 80x25 color according to the tech manual. That was how the system worked before, and changing it does not change my problem. Should it be on or off, really? For a mono monitor with - I guess - grey scale?

I remember once reading 'rapidly blinking LEDs' means stumbling at the very first hurdle in the boot process, but could not find anything to confirm that anymore.
--> Any hints on where to start looking? What is blinking the LEDs, the CPU or something lower-level?

Any hints welcome!

Regards,

Oscar
 
Rapidly blinking keyboard LED's indicate the keyboard has yet to be initialized and programmed by the BIOS. This happens prior to POST so yes it's stuck very early in the startup process.
I've seen this before several times. Has your battery at all leaked or are there any traces of leakage from an older battery? Every single time I've seen this the battery has leaked. We've all assumed at this point that something is damaged when it leaks but none of us have ever found out what, so the number of machines like yours keep piling up. I've lost an M24 and it's rebranded PC6300 variant this way. :(
 
I've replaced a leaky NiCd in a 6300 with a plain old lithium coin cell and a Schottky diode. It works fine.

There are schematics available for these things, so it's not as if a malfunctioning motherboard is a mystery. Schematics and a scope should narrow things down pretty quickly.
 
Chuck(G) said:
I've replaced a leaky NiCd in a 6300 with a plain old lithium coin cell and a Schottky diode. It works fine.

There are schematics available for these things, so it's not as if a malfunctioning motherboard is a mystery. Schematics and a scope should narrow things down pretty quickly.
Click to expand...

I was never able to get anything useful out of those, my scope and a logic analyzer. Just a lot more questions that nobody wanted to answer. :/
 
They're in storage. The machine is too complete to write it off for such a mysterious failure.
(Sorry oscar, I think I hijacked your thread. ;) )
 
Oscar, you might try re-inserting socketed chips. But do have a cloe look at tge area around the battery. Fortunately, the battery will usually leak downwards--away from the board. Also check any plug-in modules--some M24/6300 have module updates for the keyboard controller. For what it's worth, note that the machine will not boot if the display controller board isn't installed.
 
Oscar, while the system was rebranded for the states, the 6300 and M24 are nearly identical. I would grab the service manual at https://docs.google.com/viewer?a=v&...xhdHQ2MzAwc2hyaW5lfGd4OjJjMTRiYmM3MTAwZjY1NjA and see if that helps. (If the link doesn't work, google "6300 shrine", go to downloads, and the service manual is there.) In particular, chapter five has "troubleshooting without system test".

Unfortunately, that manual lists that for the following conditions:

No Display
LED OK
Fan OK
No Beep

...that the motherboard be replaced :-(
 
All,

Thanks for the answers. I read before that battery leakage leads to this symptom and is tough to solve - but I could not find a description of how to try and solve it.

The battery still gives 3.7v and looks completely clean. battery area2.jpgChecking the battery area on the PCB, it looks completely clean too. Except for some marginal gunk around the bottom battery pin, which I thought it was more like flux residu or something. See picture (I replaced the battery with some pin headers).

But the solder hole did look a little bit ragged. hmmmm. Board connections (bottom pin goes to GND, top two pins go ro R36) are as expected at least.


Question: I *assume* the machine would normally boot with the nicad battery removed, it would just complain about real time clock settings perhaps? I.e., I can debug the board without a battery connected, right?


>> some M24/6300 have module updates for the keyboard controller

You mean a new ROM in the keyboard, you suspect that this has to do with the keyboard itself rather than the main board? Maybe I should check the keyboard's 2716.


>> I would grab the service manual

I went through it - but like you said, for this very basic problem it does not offer much insight.


>> For what it's worth, note that the machine will not boot if the display controller board isn't installed.

Thanks, that's important to know. Dead display controller = rapidly blinking lights, or boot failure later on in the process?


Kind regards,

Oscar.
 
My battery was soldered in directly to the board--no pins as you show. Regardless, it's not needed for successful booting.

By replacement keyboard controller, I mean that there was a small PCB with an 8048, EPROM, and some random logic that replaced the original.

On a quick check, see if your BIOS EPROMs are intact. Bit rot has been known to occur.
 
Thanks!

One more question: is the source code of the M24 BIOS, or the keyboard MCU ROM, available somewhere?

I did find:
- M24 theory of operation (did not see it mentioned before, very useful, more so than the service manual)
- 6300 Plus BIOS sources (although a 286 machine, it looks like a lot of the code came from the M24/6300)

Regards,

Oscar.
 
Hello, there is an internal diagnostics in the M24 BIOS which is done before it starts to display something. Unfortunately this seems not to be mentioned in the normal service manuals or Olivetti pocket guides but only in the circuit diagrams which were issued only Olivetti internal for the electronics workshops to fix the boards. To see it's activity one has to plug a device with 8 LEDs or two digits hexadecimal display to the parallel port. (To drive the LEDs it would make sense to put a 74LS240 buffer between parallel port and LEDs!)

Anyhow, you also should check the graphics card of your M24. There was a modifying task to disable that card and using an ISA VGA instead. If the original owner did that modification on the graphics card to use another graphics card, the integrated OGC is completely disabled and won't display anything. The task to modify it is to remove one of the PAL chips on the graphics card and short some contacts in that socket. This modification is documented on some websites and Olivetti service manuals. But be careful, before trying to power on the M24 with an add on graphics card, please check if that modification has been done. If you start the machine with an extra graphics card and that modification IS NOT DONE then you will DAMAGE your M24. So be carefull!
 
On my 6300, the previous owner simply disabled the graphics card by pulling a PAL from the display board (and then losing it). See my blog on replacing the PAL with a GAL.
In any case, the service manual shows how to disable it, if it's not disabled already. A hint at a previously present video adapter would be one of the rear slot covers missing.

The parallel port is already capable of sinking 24 ma per line and is supposedly capable of operating for a short time with a short-circuited cable so an external buffer (and power supply for it) isn't necessary. I'd put a 270ohn resistor in series with each LED, however, to protect the LEDs. Most bog-standard commodity LEDs will show light in the sub-milliamp range.
 
Thanks! Yes, I saw the mention of the LPT output, when I get back home I'll cobble something together.

The case of this particular M24 is simpler than most: it worked one day, then did not the day after when I picked it up from the seller. So I am sure that nobody messed with settings. It'll be a good old-fashioned (but likely not too successful I admit) bughunt with the logic probe now that I've got the schematics and theory of operation.

Regards,

Oscar.
 
You must log in or register to reply here.
Back
Top