Vipersan wrote on 2022-10-05, 13:45:

Not really made any progress on this Elonex as yet.. I still think there is a problem with Keyboard communication. What I have w […]
Show full quote

Not really made any progress on this Elonex as yet..
I still think there is a problem with Keyboard communication.
What I have worked out so far is that the VIA PLCC package VT82C42V is the keyboard and mouse controller...and that PS2 keyboard pin 1 (DATA) is directly connected to PIN2 of the VT82C42 (Test 0)
and the PS2 keyboard pin5 (clock) is connected directly to pin30 (P10)
Both of which check out fine with a meter in contiuity mode.
I also know this PLCC chip can operate in PS2 or in AT mode...though I'm not sure just how this is achieved..
bios control ?
I think the chip can also lock out mouse and keyboard though I have yet to find a specific pin to do this...if I could I could trace it to a jumper perhaps ?
..again this could be controlled electronicaly perhaps by bios on the many ports on the chip.
All this assumes the VT82C42 is actually ok...and only if I knew for certain it was damaged would I attempt replacement as it isn't socketed.
Corrupt Bios is still an option IMO as the bios may possibly set up the mode of operation of the VT82C42 at boot....PS2 or AT.
The mystery continues sadly.
rgds

The difference between ps/2 mode and AT mode is whether the keyboard data gets inverted or buffered. I've attached the datasheet for your KBC, look on page 5 in the app notes there are schematics of how it would be connected.

Vipersan wrote on 2022-10-02, 09:59:

...and the Bin dump of the Bios. Full label Award software inc 1994 ISA/PCI 486 Aries […]
Show full quote

...and the Bin dump of the Bios.
Full label
Award software inc 1994
ISA/PCI 486 Aries

I can help with the BIOS. Thanks for the BIOS dump btw.

I first opened the BIOS image with my hexeditor and did a check on the checksums: all OK!
Then I opened the BIOS with Award’s Modbin tool and all the screens looked fine as well. 😀

Finally I ran the BIOS in 86Box, using the emulated Asus PVI-486AP4 machine that uses the same i420EX chipset. This work flawlessly! Here are some screenshots:

And here is one of the CMOS Setup screens:

So I didn’t find any corruption and I’m convinced the BIOS is fine.
I only found the usual 11/1994 Award v4.50G BIOS bugs like the Y2K bug and the 2GB Harddisk display limit bug. But these can be easily fixed.

Jan

Thanks for helping guys...
If the Bios is OK and looks like it is...then this problem has to be a hardware issue on the motherboard..
I still think this is keyboard related.
Possibly the Keyboard controller is faulty...or is a link or jumper to allow PS2 to function.
I have tried running without the cache stick fitted...also a single stick of ram...
I'm assuming the CPU is ok ...on account is recognised correctly during the partial boot...and doesn't actually crash..
The cursor continues to blink ...but no response from keyboard...so I cannot enter bios.

CPU is Identified..
Next should be memory test but this never happens...but I do get on screen message 'press DEL to enter setup'
..this I cannot do of course...as the keyboard is unresponsive.

Did you review the schematics I posted? There is a 7406 that could also be faulty in the keyboard clock/data signal paths

maxtherabbit wrote on 2022-10-06, 14:35:

Did you review the schematics I posted? There is a 7406 that could also be faulty in the keyboard clock/data signal paths

I've downloaded the PDF to study later when I get home..
Currently at work fixing a sickly Toshiba DVR combi.
That takes priority is I get paid for that one ..if I fix it...🤣

What I can say is..that this motherbord doesn't have a 7406 on it...but does have a 74LS05 ..which is also a hex inverter...
I'm assuming it does the same job.

its stopping at 30, checking size of installed ram? would explain rapid beeps. As I said 2 weeks ago:

"Start by removing all SIMMs and reinserting just one."

rasz_pl wrote on 2022-10-06, 22:13:

its stopping at 30, checking size of installed ram? would explain rapid beeps. As I said 2 weeks ago:

"Start by removing all SIMMs and reinserting just one."

thanks for that advice rasz...and for checking out my boot avi.
already done the single simm check in simm slot 1
I doubt both simms would be faulty ...though I'll try and find some replacements anyway..
definate possiblity there are trace breaks between cpu and simm slots...though I can't see any corrosion.
Lots of trace checking to be done I guess.

OK ..
found a working replacement for the 486-DX2 66
No change..
Also found a couple of 72 pin ex compaq HYM532220W-70 simms..
Tried 1 in slot 1 no change...
so tried both fitted slots 1 and 2 populated..
Still no change.
So if it is a ram issue..I suspect traces or worse the chipset..
The fact that this Elonex was put away working in the 90s according to it's previous owner...and had a faulty psu now fixed...imo hard to see how this is chipset failure ?
A multiple coincidence of failure ?
If there was evidence of bad corrosion...I would be more inclined to think trace damage...but this motherboard is in very nice condition physically.

http://mrbios.com/techsupport/award/postcodes.htm
Late Award BIOS (4.5x-non PnP)
30 Size Base and Extended Memory. From 256-640K and that above 1 Mb.
this is where you stop. Bad keyboard or kbr controller would not stop boot process here.
>Tried 1 in slot 1 no change...
>so tried both fitted slots 1 and 2 populated..

try all the simms you have just in slot2

Curiously I dont see "0C Initialise keyboard; Set NUM LOCK status.". Your post card skips from 0B straight to BF and 0D.
does Numlock illuminate on your keyboard? What happens when you press Numlock while sitting on the frozen screen, does it make keyboard led blink?
Do you have an iphone/ipad or another phone/camera able to record at high refresh/slow motion? would be helpful to record _just_ post codes in best way possible, that means in good lighting too so the camera can use fast shutter.

Have you tried Babasha babasha suggestion?
>There interesting JP1 jumper “manufacture… and security” check it, maybe it for keyboardless boot/oerations

Ill look into the bios to see what could cause doing fast beeps at code 30.

Elonex got a lot of contracts in education in the UK... if this is an ex-schools or higher ed machine, then there may also be chassis intrusion lockouts. There may also be physical issues with the keyboard and mouse sockets, either from having random objects jammed into them, or from glue intrusion when it's original keyboard and mouse were glued in place, where it happened to make contact fine with those, but it's coating a pin just wrong for what you're using now. So I'd say to work over the sockets with a pick to scrape all pins and contacts clean.

Possibly finding an IT bod/sysop active in schools through the mid 90s will get you someone who is very familiar with these machines.

I located JP1 which has 4 pins ...
None of wich have jumpers on them...so nothing to lose I tried linking 1 and 2 ...it didn't help..
Then tried linking 3 and 4 ...again no help there.
The other options are to link 1 and 3 or 2 and 4....but since I dont know what the pins do I'm hesitant to try this as it would be counter intuitive.
I could try linking both at once 1 and 2 also 3 and 4 I guess.
In reply to Raz suggestions ...
I have no way to slow motion capture The post card display...
I wonder if I can reduce the capture speed in software perhaps ?
and finally ...
Although I get all 3 leds num caps and scroll illuminating briefly when powered up from mains off...(indicating the KB is getting power..
All keys remain unresponsive thereafter...so no ...pressing the num lock key does not illuminate the corresponding LED either before or after the rapid beeps.
I have yet to try simms in just slot 2 but I'll do that soon.
I have to be cautious as they simm catches are plastic and brittle with age.
I have ordered a replacement VT82C42V (PLCC44) from little diode...just in case and a 74LS05D....which I hope will be a straight swap for the 74LS05M on the motherboard.
Neither chip will be an easy swap ..but it's nice to have options.
I do appreciate any further help I get on this as I dont just give up easily until I exhaust possibilities.
I contacted it's previous owner again..who informed me that this Elonex was used in a work environment until he claimed it before being binned at NORWEB UK.
but he only used it and had no ideas beyond that as to what was connected or if it booted independatly or more likely ..networked in some way.

...A thought just ocurred to me..
This may sound stupid but hear me out.
There is a riser on this mobo...which takes PCI at the top..ISA at the bottom and a shared position in the middle which could hold either PCI or ISA ...making 3 slot positions at the rear ...and 3 blanking plates.
Now...the middle blanking plate is missing so obviously this had been removed in the past in order to fit a card..
My guess would have been a network card since it was used in a work environment.
There is no card fitted there now of course...
Could it be that this Elonex is waiting for a response from that missing card which will never come...?
We know that some of these network cards could be fitted with a rom not unlike how some XT-IDE can function.
..and that rom intercepted the boot process and this machine is waiting for that rom and of course will never see it.
again..this setup would have been proprietory at NORWEB...
That being remotely possible ?....could there be a way around this ?

its too early in the boot for a network
it doesnt appear to be initializing keyboard in any way whatsoever
keyboard controller error would either freeze it earlier or not matter beyond generating "keyboard error press any key" later 😀
no idea why 0c post code is missing, but 0F is in there and this also has something about initializing keyboard

rasz_pl wrote on 2022-10-08, 13:26:

its too early in the boot for a network it doesnt appear to be initializing keyboard in any way whatsoever keyboard controller e […]
Show full quote

its too early in the boot for a network
it doesnt appear to be initializing keyboard in any way whatsoever
keyboard controller error would either freeze it earlier or not matter beyond generating "keyboard error press any key" later 😀
no idea why 0c post code is missing, but 0F is in there and this also has something about initializing keyboard

Interesting Raz ..
So it is possible the VIA keyboard controller chip is faulty ..or indeed the hex inverter.
I'm glad I sourced and ordered them now.
But of course the error could be in one of the chips which make up the chipset...
If this is the case ..I'm stuffed.

no, the other way around, its almost impossible to be caused by bad keyboard controller

So sad to hear that..
Looks like I may be out of options then..and this Elonex may never be resurrected.
That's soooo sad imo.

So ..I had to at least try one simm in slot 2
First tried no memory at all...and as expected ...the post didn't even start..
Long slow single beeps ..continuosly with regular intervals.
Single Simm in slot 1 ..the POST starts and ends as previously with the rapid beeping or as previously described ...a screech.
Single Simm in slot 2
Exactly as above with slot 1
Obviously the presence of ram is acknowledged.
I found this which presumably is in sequence..

AWARD BIOS
Code
(hex) Name
C0 Turn Off Chipset Cache
01 Processor Test 1
02 Processor Test 2
03 Initialize Chips
04 Test Memory Refresh Toggle
05 Blank video Initialize keyboard
06 Reserved
07 Test CMOS Interface and Battery Status
BE Chipset Default Initializatiion
C1 Memory presence
C5 Early Shadow
C6 Cache presence test
08 Setup low memory
09 Early Cache
0A Setup Interrupt Vector Table
0B Test CMOS RAM Checksum
0C Initialize Keyboard
0D Initialize Video Interface
0E Test Video Memory

..and so the post sequence must be at least reaching C1 ...and C6 will be ignored if this refers to the cpus external cache..as this is not fitted...
..and appears to be failing/hanging at 0C...but has passed the earlier step 05 Blank video Initialize keyboard.
I'm out of ideas now..
but perhaps someone will have a moment of inspiration.

could the POST be failing at OB...just before keyboard initialisation ?
do remember the Dallas RTC was empty ...and although I have now added an external button cell...CMOS will have been dumped.
Maybe there was something other than RTC info in there...
or maybe I'm clutching at straws...🤣

Vipersan wrote on 2022-10-08, 15:44:

..and so the post sequence must be at least reaching C1 ...and C6 will be ignored if this refers to the cpus external cache..as this is not fitted...
..and appears to be failing/hanging at 0C...but has passed the earlier step 05 Blank video Initialize keyboard.

your post card stops at 30, you see video output (initialized at 0D)

Vipersan wrote on 2022-10-08, 15:44:

could the POST be failing at OB...just before keyboard initialisation ?

no

Vipersan wrote on 2022-10-08, 15:44:

do remember the Dallas RTC was empty ...and although I have now added an external button cell...CMOS will have been dumped.
Maybe there was something other than RTC info in there...
or maybe I'm clutching at straws...🤣

resetting cmos is standard procedure, you can do it again with JP2 CMOS CLR
things to try:
-take off VGA ram expansion, no keyboard connected, no riser connected

-what is the glued jumper bodged with green wires?
-what is J2? dual jumpers with no description on the silk screen?

If anything gets wedgy with the KBC and RAM count involved I always wanna say "Gate A20 line problem"

BitWrangler wrote on 2022-10-08, 17:16:

If anything gets wedgy with the KBC and RAM count involved I always wanna say "Gate A20 line problem"

damn, you are right
that glued jumper bodged with green wires is wired on one side to +5V on a cap, and on the other side mysteriously very close to pin 44 of VT82C42. Maybe this is a hacked up Case open/security mechanism - jumper removed kbc gets no power?

Vipersan very carefully measure voltage between ground (metal bits around connectors will do) and pin 44 (VCC) of VT82C42v. Its the first pin to the right from the white dot https://pdf1.alldatasheet.net/datasheet-pdf/v … TC/VT82C42.html page 8 has pinout
then measure resistance between pin 44 and wherever that green wire is soldered to nearby

EDIT: sadly breaking kbc altogether stops boot at "05 Blank video; initialise keyboard". Commenting out all the A20 code doesnt do anything and 86box happily counts memory.

new plan. Vipersan measure continuity between green wire near VT82C42v and all of its pins looking for a short
https://pdf1.alldatasheet.net/datasheet-pdf/v … TC/VT82C42.html page 8 has pinout
it can still be a super secret lockout, and emulated 86box keyboard controller is to dumb do understand command or by shear coincidence returns value letting it boot