PDA

View Full Version : Olivetti M24 - No POST



ChrisCwmbran
June 4th, 2012, 07:01 AM
Hi guys,

Need some advice. My Olivetti M24 has been connected to the mains power and it's monitor today for the first time in many years. Doing the testing, the machine is fitted with it's HDD controller card, but the cradle with the HDD and FDD are removed (and obiously disconnected).

When powered on the machine flashes it's keyboard LEDs for a few seconds and then beeps but produces no display at all. The monitor is the monochrome M24 monitor. The two dials on the monitor are both turned fully clockwise as the manual says.

Interestingly, pressing control+alt+del does seem to cause it to reboot.

Ideas please? Thanks in anticipation!

Stone
June 4th, 2012, 07:11 AM
If it's just one beep that means it booted successfully, is looking for a boot device and the monitor is likely the problem.

ChrisCwmbran
June 4th, 2012, 07:28 AM
I know someone on here (Chuck(G)?? I think) has posted details of how to connect a 6300 to a VGA monitor but I can't find the posting anywhere.

Sadly whilst I've done tonnes of stuff inside PCs I've always shied away from monitors.

Any advice on how to diagnose this fault would be appeciated.

Does anyone have a pinout for the graphics card of the M24 please?

Stone
June 4th, 2012, 07:51 AM
You're right again. :-) Guess we'll have to wait till Chuck finds this thread.

The main problem with the message search feature is that if you indicate a member's post to search for it only returns threads in which that member was the OP. That feature really sucks when you're trying to locate a post from someone who wasn't the OP in the thread, like you're probably trying to do now.

Chuck(G)
June 4th, 2012, 08:58 AM
Chuck's been at the dentist. He found the thread.

For a quick check on a VGA (LCD works fine) monitor, build this cable (http://discoremoto.cheapnet.it/pier3079/video_converter.pdf). I just whacked the D-sub connector and cable from an old VGA CRT monitor that was headed for recycling and soldered a DB25M connector to the end.

In my case, I could open the M24 mono monitor up and see the filament of the CRT glowing, so I knew that it was getting power. A pass over the CRT PCB turned up a cold solder joint.

However, many people wanted to fit an EGA or VGA card (8 bit) to the 6300/M24, so they disabled the video adapter. You can't remove it as it ties the motherboard to the bus converter, so one way was to pull one or two PAL chips (look for empty sockets on the video board). Fortunately, with the help of some folks here I was able to reverse-engineer the content of one of them and my machine is now happily belching video using the onboard video.

I'm going to assume that if your video board is intact, the motherboard video switches are probably set right. But it wouldn't hurt to check them.

ChrisCwmbran
June 4th, 2012, 10:03 AM
Thanks for the advice so far.

I'm a little confused. Reading the AT&T PC 6300 service manual, it appears I have a P4 Display Controller Board, and they talk about a plug being substituted for the 74LS00 chip in position 6H and show a 14 pin DIL chip with strapping between certain pins to disable the onboard video. The chip in that position on my card is a 20 pin DIL chip, and has a label reading PL68 on it.

To further confuse, the settings for DIPSW-0 and DIPSW-1 appear to be for ROMs prior to 1.21 or those version 1.43 onwards. My BIOS chips are labeled as version 1.36! Typical of me to have the odd version that fits neither!!!

Chuck(G)
June 4th, 2012, 10:12 AM
The service manual is really talking about a P2 kit, as far as I can tell. I have the same P4 board that you do and PL68 is the correct chip for that location. There was a 20 pin plug offered that shunted some of the lines to Vcc or gnd that disabled the board. But you could just as easily pull the PL68 and ignore the shunt on the P4 board. Alternatively, you could change a couple of jumpers.

Make up the VGA adapter cable and see if you get anything. You're going to like having color anyway.

I've got the 1.43 BIOS and some board upgrades as well.

ChrisCwmbran
June 4th, 2012, 10:51 AM
I've got the 1.43 BIOS and some board upgrades as well.

I'd love to upgrade my machine to 1.43 but having just bought an oscilloscope for working on my old computers, I don't have the funds for an eprom burner at the moment. What are these board upgrades you talk about?

Compgeke
June 4th, 2012, 03:21 PM
The board is the display board, P2 is a old revision, P4 seems to be the most common. I myself have 2 PC-6300's and both have the P4 revision board.

I would also check to see if the monitor is getting voltage. With a multi meter (or a DC voltage reader) and check the voltages from the wires running to the display board. The orange and grey at P7 is 14v I believe, and if that has no power reading at all then the monitor won't be receiving power.

Chuck(G)
June 4th, 2012, 03:38 PM
I'd love to upgrade my machine to 1.43 but having just bought an oscilloscope for working on my old computers, I don't have the funds for an eprom burner at the moment. What are these board upgrades you talk about?

There are some cuts and jumpers to traces, a new keyboard controller piggyback, etc. I don't know if they're worth the effort, however.

Service manual says +15 on the connector. But I'd still build the VGA cable--it's very simple.

ChrisCwmbran
June 5th, 2012, 08:34 AM
When using the VGA cable, how does the motherboard need to be jumpered? I've made up a cable I think is correct but my LCD monitor displays "Input not supported" when it's on.

Chuck(G)
June 5th, 2012, 08:39 AM
Try a different VGA monitor. Mine works fine with an older 22" ASUS LCD--some newer ones don't have the low res range. I didn't change the jumpering from that indicated in the serivice manual.

At least you know that you've got video.

ChrisCwmbran
June 5th, 2012, 08:57 AM
Can you just verify how I tell which side of the connectors in the diagram are which? i.e. with the top picture of the 25 pin connector is the diagram of the pin side or the other side?

ChrisCwmbran
June 5th, 2012, 10:50 AM
Ok my head hurts now. I've spent all day tinkering with this annoying little connector, and got no where.

I decided to do it using one of these http://www.maplin.co.uk/9-way-d-type-to-25-wayd-type-box-1670

And to do the connections within it using Solid Core Wire (1/0.6) - this may be a mistake (I'm reatively new to soldering and stuff) but it's a bugger to work with such inflexible wire in such a small space. The connectors have what they call solder cups for each wire to go into and they were very fiddly too.

Because Im confused about which side of the connector each image is showing it's hard to be 100% sure I've done it right!

I'm trying to unstress a bit before having another look - but at the moment it's very hard to tell if my adapter is wrong, or the video card of the machine is defective. I had until now assumed that because of the single short beep that the machine at least thinks its video card does work!!

Chuck(G)
June 5th, 2012, 11:36 AM
Many, if not all, solder-cup connectors have, on one side of the connector or the other, pin numbers molded into the surrounding plastic. However, if you're looking at a DB25 male connector at the solder-cup side (not the pin side) with the 13-pin side on top, the rightmost pin is pin 1; the leftmost, pin 13. The numbers on the second row start again at the right with 14, and continue to 25 on the left.

For the three-row DE-15HD VGA male connector, the pins are numbered similarly with pin 1 the rightmost top row position on the solder-cup side and runs to the left at pin 5. The middle row goes from pin 6 on the right to pin 10 on the left. Finally, the bottom row is pin 11-15 in the same manner.

This might help for the VGA side (http://pinouts.ru/Video/VGA15_pinout.shtml)

And This diagram (http://discoremoto.cheapnet.it/pier3079/video_pinout.pdf) identifies the signal names at the M24 connector.

ChrisCwmbran
July 12th, 2012, 09:06 AM
Thanks for the advice Chuck. We made an adapter and double and trebble checked the connections for correctness and continuity.

I connected it to an AOC LM720A monitor (http://www.discoazul.com/tft-17-aoc-lm720a_en.html) which supports 640x350@70Hz, 740x400@70Hz, 640x480@60/75Hz, 800x600@60/75Hz, 1024x768@60/70/75Hz, 1280x1024@60/75Hz resolutions.

Nothing is displayed.

If I were to use my oscilloscope to check the red, green, blue, h sync and v sync outputs on the M24, which appear to be the only ones used, what should I expect to see? I presume I'm going to see 5v ish pk to pk signals, and one of the syncs to be about 60hz.

Chuck(G)
July 12th, 2012, 09:25 AM
I'd have to go back to mine, but I think the horizontal sync will be about 32KHz. Chances are, if you see both horizontal and vertical sync, you'd probably see video also.

Do you know what revision of the video card that you have? Are there any empty sockets or sockets filled with jumper blocks instead of ICs? (Both those approaches were often used to disable the video controller so that a different one (e.g. VGA or EGA) could be used.

ChrisCwmbran
July 12th, 2012, 09:36 AM
96619662

It appears to be a P4 revision.

I was checking the dipswitches on the motherboard, and as I mentioned earlier in the thread it is a little bit odd. The motherboard has dipswitch tables for bios upto version 1.21 and for bios 1.43 onwards. My bios is a version 1.36 according to the stickers which isn't covered by either table. The monitor type is determined by two individual switches, so we tried all four options.

barythrin
July 12th, 2012, 09:39 AM
Sounds like you're perhaps past this already but have you tried a multisync capable monitor? Your input not supported generally means it does see the voltage but can't display that resolution or hertz range.

ChrisCwmbran
July 12th, 2012, 09:47 AM
I can try to hook out one of my older CRT based VGA monitors. I think I have a nice short depth Viewsonic 17" or 19" that is very flexible input wise.

barythrin
July 12th, 2012, 09:59 AM
Yeah I would definitely try that. From what I recall (obviously I didn't play with many due to price) early LCD monitors weren't very compatible ironically with many signal types. They seemed almost like a winchip equivalent of monitor, supporting only popular resolutions and scan rates commonly used in Windows.

Chuck(G)
July 12th, 2012, 11:19 AM
Well, as you've show, you own a 'scope so you can easily determine the H and V sync output frequencies probably faster than I can type this. I run a 19" ASUS LCD monitor on mine and it works fine.

ChrisCwmbran
July 12th, 2012, 12:07 PM
I probably phrased the question wrongly there really sorry. I understand about the frequency of pulses on the horizonal and negative sync pins, and I understand the concept of a negative sync on horizonal. I really was hoping to understand a bit about the nature of a valid signal for the RGB components, and how the data on them is encoded (perhaps).

I've also been trying to find a suitable explanation of how sync works on a composite signal too i.e. where a monochrome composite video signal is carried over a phone cable. I guess it's part of trying to understand what a scope can and can't be used to test.

Chuck(G)
July 12th, 2012, 12:26 PM
The M24 video is simple RGB+I and the VGA adapter ignores the I signal. TTL levels on everything. One of these days, I'll work out a little circuit to use the I output and clamp the outputs to 1V p-p. But it works as-is for now.

So to use a scope, set 1 or 2 v/div, sync on auto, and adjust sweep rate so that you can see two pulses. You then can see the time between the leading (or falling) edges of any two consecutive pulses. The frequency is simply the reciprocal of that.

For example, let's suppose you've cranked the sweep rate to 5 milliseconds/div and you see two pulses. Looking between two consecutive leading pulse edges, you see 3 divisions plus a bit or about 16 msec. between the pulses. The frequency is then 1000 msec./second divided 16 msec. or about 62 Hz.

ChrisCwmbran
July 13th, 2012, 08:24 AM
Ok the settings I found are as follows:


DIPSW1 J5 ON J6 ON Pin 1 (H Sync) 4.9v pk to pk, 25.9kHz
Pin 2 (V Sync) 4v pk to pk, 60Hz
Pin 4 (Red) 3.6v pk to pk... randomish pulses
Pin 5 (Green) 3.6v pk to pk... randomish pulses
Pin 6 (Blue) 3.6v pk to pk... randomish pulses

DIPSW1 J5 OFF J6 ON Pin 1 (H Sync) Signal about 0.16v
Pin 2 (V Sync) Signal about 0.16v
Pin 4 (Red) 3.6v pk to pk... randomish pulses
Pin 5 (Green) 3.6v pk to pk... randomish pulses
Pin 6 (Blue) 3.6v pk to pk... randomish pulses

DIPSW1 J5 ON J6 OFF Pin 1 (H Sync) Signal about 0.16v
Pin 2 (V Sync) Signal about 0.16v
Pin 4 (Red) 3.6v pk to pk... randomish pulses
Pin 5 (Green) 3.6v pk to pk... randomish pulses
Pin 6 (Blue) 3.6v pk to pk... randomish pulses

DIPSW1 J5 OFF J6 OFF Pin 1 (H Sync) 4.9v pk to pk, 25.9kHz
Pin 2 (V Sync) 4v pk to pk, 60Hz
Pin 4 (Red) 3.6v pk to pk... randomish pulses
Pin 5 (Green) 3.6v pk to pk... randomish pulses
Pin 6 (Blue) 3.6v pk to pk... randomish pulses

I also tested them with my adapter cable and found a problem with the HSync pin not making a good connection. I havent had a chance to correct that and try it on a monitor yet.

Aren't those 0.16v signals a bit odd?

Chuck(G)
July 13th, 2012, 09:05 AM
0.16V falls into the category of "noise"--electrically insignificant. Those randomish pulses on RGB are what one would expect to see--they're the individual pixel dots. So, something is coming out of the card.

I'll do some readings on my own setup and see how things compare. 25.9KHz does work out for 60Hz vertical to 431 lines--reasonable for a 640 400 display. I may be lucky with my VGA LCD being able to sync to that.

However, the DIP switch settings for 5 and 6 don't make a lot of sense. Are you power-cycling between changes?

ChrisCwmbran
July 13th, 2012, 02:04 PM
When you say power cycling between changes I assume you mean each time we change the dip switches in which case the answer is yes.

We didn't power cycle between taking readings from individual pins though.

Chuck(G)
July 14th, 2012, 09:51 PM
A quick check with mine yields a horizontal sync of about 25.7KHz, 4.2V positive-going, 4usec pulse.

Close enough. Pick a good monitor--apparently not all VGA LCD monitors will sync down to 25KHz.

ChrisCwmbran
July 15th, 2012, 06:32 AM
Ok, thanks for taking the time to test that Chuck. I'll do some more testing :)

My mum is really excited to see the M24 running again! I get asked everyday. Back in the 1980s she wrote her thesis on one back when she did he Music degree!

ChrisCwmbran
July 22nd, 2012, 09:44 AM
Whilst waiting to be able to pickup a suitable VGA monitor to test this I thought I'd have a look in the back of the M24s own monitor and see if I can see any problems.

What the hell is this? If you look at the last picture it doesn't look too good!!!!

971997209721

Chuck(G)
July 22nd, 2012, 10:07 AM
Well, the resistor is a 6.2 ohm, 10 watt unit. These tend to get very hot (that's why they're called power resistors). But yours looks to have issues with corrosion.

Can't say much about Olivetti's monitors, other than I think they were made by Hantarex Italy. Is that the monochrome monitor?

ChrisCwmbran
July 22nd, 2012, 10:08 AM
Yes. I think its the green screen version rather than the amber one too.

Is the 8.2 ohm version of this correct?

http://www.maplin.co.uk/10-watt-wirewound-resistor-2181

Chuck(G)
July 22nd, 2012, 10:25 AM
Yes it'll work. I'll take a look inside of my own green-screen 6300 monitor to see if mine is in a similar condition.

Chuck(G)
July 22nd, 2012, 09:11 PM
Looked at the innards of my (monochrome green) monitor. It's nothing at all like yours--to start with, all the component silk-screening has components starting with 1, so I see R23, C10, etc. Not R903.

No big power resistor, either. I wonder if that monster in yours is a cheap attempt to drop the +15 supply to +5 for logic. If so, there are other, more efficient ways to do the job. Any identifiers as to who the maker might be and the model number? Sometimes on a bit of paper, silk screening on the PCB, or etched on the foil side of the PCB.

My monitor looks like this one:

http://i.ytimg.com/vi/mUCh46_MzZU/0.jpg

ChrisCwmbran
July 23rd, 2012, 01:06 AM
No labels on mine.... or at least there is one internal label but it is completely faded.

Searching in google for "Olivetti M24 monitor" the first 5 images are from a site called "olivetti.isgreat.org" that appears to be gone, however they appear to be of the same monitor I have, and one of them is a sticker that says it is made by "Goldstar".

I think it says its an "MBN-21318" but its quite hard to be certain:

9734

ChrisCwmbran
July 23rd, 2012, 05:49 AM
Appologies for the poor photo, but I think the important bit is quite visible!

9735

Chuck(G)
July 23rd, 2012, 07:40 AM
So you're good now?

It's curious that the AT&T 6300 used an Italian monitor, but the M24 seems to have used a Korean one. Maybe someone has an idea of why...

ChrisCwmbran
July 23rd, 2012, 07:47 AM
Yeah I don't forsee any problems now.

Thanks for everyones help. I'll post some better pics of the machine when it is all in one piece and booting DOS nicely!

I would still like a set of newer ROMs for this machines and full details of any expansions/mods that are worth doing if anyone can help! :)

Chuck(G)
July 23rd, 2012, 08:08 AM
I still have the installation instructions for 1.43, but it also included at least one PAL as well as the ROMs.

ChrisCwmbran
July 23rd, 2012, 08:43 AM
Ah ok :) I guess it's not hugely urgent anyway! It's not as if I have some kind of BIOS related issue I need to solve!

It's nice to see it come to life again though!

Thanks again! :)

Chuck(G)
July 23rd, 2012, 08:47 AM
Great!

So now to what use will you put your M24?

ChrisCwmbran
July 23rd, 2012, 09:10 AM
Well once I have more of the collection running I'm hoping to start taking some of the machines to UK shows.

In the mean time Im expecting my mum to spend some time in front of it tinkering with the thesis she wrote on an M24 many years ago :)

Once I've reinstalled the floppy and the HDD I'm looking forward to seeing what is actually on it!

Then as you know probably too well I have a Kaypro that needs my attention - and about 100 more machines :P

ChrisCwmbran
July 23rd, 2012, 01:27 PM
Seems I'm not entirely out of the woods yet. Decided to leave it on a while to make sure its ok, and I'm seeing some flickering on and off of the screen, and flickering brightness.

The display itself doesnt seem to jump or deform at all.

It's just variation in brightness.

I'll try to check the stability of the 15v monitor supply whilst it's under load tomorrow.

ChrisCwmbran
July 23rd, 2012, 02:39 PM
I am so fustrated with this computer it is unbelievable. Just when I thought I was past the hardest bit!!!!!

As it stands the display "brightness" on the monitor is altering randomly in a flickering way over time, and it has taken to displaying what looks lie a 40 collumn post screen rather than 80, where part of it is off the screen!

Because of the way it is done it is hard to see how I can take readings of the 15v supply to the monitor whilst the powersupply is still connected to the monitor - there seems to be nowhere I can find that I can actually get a probe to take a reading.

Chuck(G)
July 23rd, 2012, 03:01 PM
Pop the top cover off the M24--you'll see that the +15 feeds to the video controller via a pair of wires (one orange, one black, IIRC). You can take your reading where the thing plugs into the video controller.

ChrisCwmbran
July 23rd, 2012, 03:06 PM
If the 15v rail is losing voltage am I right in thinking that would in this case cause variation in brightness?

Is it possible that if the voltage being supplied is lower than 15v that the guns in the tube would effectively project over a larger area making it look like a 40 collumn image partially off the screen?

Chuck(G)
July 23rd, 2012, 04:19 PM
Well, yes, although if the voltage was fluctuating, you'd also get "bloom" (focus and size variation). Doesn't sound as if that's the case. I'd tackle the obvious--get some contact cleaner (I like DeOxIt, but pick your poison) and squirt some into both the contrast and brightness control pots, agitate, repeat and see if that doesn't make a difference.

ChrisCwmbran
July 28th, 2012, 09:47 AM
Had the monitor apart and used some contact cleaner on the contrast and brightness control pots. Haven't tried it yet though.

Next to the flyback transformer there is a radial electrolitic capacitor (25v iirc, size not visible at the moment) that is bulged and burst.

I'll remove the board tomorrow and desolder the offending capacitor and get replacements in for Tuesday. At the moment Im guessing its a 220uF because all the other identical size and make caps on the board are 220uF 25v.

ChrisCwmbran
August 2nd, 2012, 09:51 AM
It was a 220uF 25v capacitor.

Having replaced it the screen display is now the right size again although the first character on each line of text is squashed and the brightness of that character is somewhat higher. If I remember rightly this symptom can be altered using one of the various trim pots on the monitor board.. perhaps horizontal position?

Would you expect changing a capacitor to require these pots to be altered a bit?

Do you think I should replace the other capacitors in the monitor too?

Chuck(G)
August 2nd, 2012, 02:14 PM
Yes, you can use a pot to change the thing. It might be labeled "Horiz Pos." or "Horiz Phase" or sometimes just "Phase".

If the capacitor affected the voltage situation in the horizontal section, yes, absolutely.

Are any other capacitors bulging or leaking? If not, I say if it isn't broken, don't fix it.

ChrisCwmbran
August 4th, 2012, 09:54 AM
Ok all the other caps look good. Cannot recreate the screen symptom now - it is displaying fine. There clearly is a bad connection somewhere because if you tap the monitor housing sometimes the screen flickers off. I'll try to find out what is causing this.

The HDD controller fitted to the machine is an LCS-6210C Rev C and the HDD is a Shinwa Digital D220. I'm unfamiliar with both these items and can find very little information especially about the HDD at the moment.

The HDD spins up, and other than being quite loud, there is click when the machine is turned on, and that is a small locking mechanism retreating from a gear wheel on the front right hand side of the drive. This gear wheel doesnt seem to move at all, and I hear nothing that suggests the drive attempted to seek.

The machine reports a drive error.

There is one bank of jumpers on the controller card, none of which have any jumpers on them at all.

If the jumpers are the same as a revision G version of the card this is for 306 cylinders and 4 heads as drive 0 and drive 1 disabled.

Any ideas please?

ChrisCwmbran
August 4th, 2012, 03:10 PM
I found a document that says a Shinwa Digital D220 is a 20Mb 614cyl 4head Spt 17 drive.

Assuming the revision C and G jumpers for the controller card are the same I tried both settings for a 614cyl 4head drive and still get no seek sound at all.

Chuck(G)
August 4th, 2012, 04:12 PM
I have no experience with Shinwa drives--they seem to have been marketed in the UK exclusively. Sorry, can't help you there.

modem7
August 4th, 2012, 05:32 PM
The machine reports a drive error.
What is the actual error message?

ChrisCwmbran
August 5th, 2012, 03:40 AM
All it says is "Primary boot strap: DISK READ ERROR.".

Still no indication of any seek activity.

modem7
August 5th, 2012, 04:21 AM
All it says is "Primary boot strap: DISK READ ERROR.".
Still no indication of any seek activity.
I see the error message of "Primary Boot-Strap DISK READ ERROR" in the M24 BIOS, not in the LCS-6210C's BIOS.

It's like the M24 is not 'seeing' the LCS-6210C. If the M24 was seeing the LCS-6210C, it would hand control to it at POST time, and if then, the LCS-6210C detected a problem, the LCS-6210C would issue errors like "POWER ON TEST ERROR - BAD CONTROLLER" and "1701" (both of which appear in the LCS-6210C BIOS).

ChrisCwmbran
August 5th, 2012, 04:27 AM
Ok - thanks for that. If I recall correctly one of the dipswitch settings is for whether the HDD controller card has a bios.

I'll try changing that.

ChrisCwmbran
August 5th, 2012, 05:59 AM
That jumper was set correctly anyway. Ive tried the LCS-6210C in a couple of different slots too with no luck.

Beginning to wonder if perhaps the controller card is dud.

Stone
August 5th, 2012, 06:58 AM
SpeedStor performs controller diagnostics.

Chuck(G)
August 5th, 2012, 07:11 AM
I doubt it. Controller cards rarely just go bad.

Let's make sure. DIPSW-1 should read (postions 12345678 ): 10010001 (1=on, 0=off). That means that if you take the HDC out, the BIOS shouldn't be looking for one.

ChrisCwmbran
August 5th, 2012, 12:33 PM
Can you double check that please Chuck? Doesnt bits 7 and 8 being 0 and 1 mean 2 floppy drives?

Chuck(G)
August 5th, 2012, 03:01 PM
You don't have 2 floppies? Oh, okay. I use a hardhard, so I assumed that everyone was using that.

Sw 1 - on (48 TPI)
Sw 2 - off (fast start up for floppy)
Sw 3 - off (no HD or one with its own BIOS)
Sw 4 - on (reserved)
Sw 5,6 - off, off (80x25 monochrome display)
Sw 7,8 - on,on (1 floppy drive)

The gist of this is that the BIOS won't look for a hard drive, unless it has its own BIOS. So if you're getting a hard disk error, it's coming from the HD controller BIOS. If yo boot from floppy (I assume that you can), you should be able to see the BIOS as absolute address C8000.

ChrisCwmbran
August 6th, 2012, 03:18 AM
With it set for one floppy, the message "Primary boot strap: DISK READ ERROR" still appears when the drive and controller are completely removed from the machine, so it cannot be a message from the controller.

Looking at the controller I noticed that two of the chips which I have ringed in red on the photo below are the opposite way round to all the other chips on the controller card. Both are socketed. Is this a bit odd?

9881

Chuck(G)
August 6th, 2012, 07:49 AM
Yes, it's odd--and lousy engineering practice if it's what was intended.

Let's check the big chip the HDC9226. I've got the datasheet here in front of me and it says that +5 should be on pin 24 and ground should be pin 12. If the chip really is reversed, it's likely toast.

ChrisCwmbran
August 6th, 2012, 08:02 AM
Won't be able to test it electrically for an hour or so. Looking at the printing on the circuit board itself those two chips do have the little knocth marked at the other end from all the other chips though, so it does look like the orientation of them is as designed.

Chuck(G)
August 6th, 2012, 08:29 AM
Can you boot the M24 from floppy?

ChrisCwmbran
August 6th, 2012, 08:41 AM
I have a disk that it boots off, but its origin is unknown. My tweening machine should be up and running properly tomorrow. I've seen it boot through to time/date prompts though.

What do yo have in mind?

The disk images I have came are .img files. What is the best utility to write them with?

Chuck(G)
August 6th, 2012, 09:04 AM
Running DOS DEBUG to see what exactly you have in memory instead of guessing about it.

ChrisCwmbran
August 7th, 2012, 06:43 AM
Right I have the M24 powered up here running AT&T DOS 3.3, and debug loads and works.

What do you need me to do?

Chuck(G)
August 7th, 2012, 07:46 AM
With your hard disk card inserted and DOS booted, do the following and tell me what you see (your keyboard input is in boldface):

A>DEBUG
-DC800:0
...some output...
-Q
A>

This will tell us if your HD BIOS is even present in memory.

ChrisCwmbran
August 7th, 2012, 09:07 AM
Sorry Chuck, bit of a delay. I wrote a 5.25" DSDD floppy disk with a disk image of AT&T DOS 3.3 using a modern PC with a 1.2Mb 5.25" FDD in it.

To start with it was booting in the Olivetti. Now it won't boot in the Olivetti but will in the modern PC.

I had a bit of a struggle finding a program that would write img files to the 5.25" FDD in Windows. Win32image and Winimage both refused to do so, so I ended up using rawdisk.

I'm trying to find a geniune copy of DOS amongst all my stuff instead but not luck atm.

Its a bit annoying. My machines are either too old to be a good tweener or too new. I have nothing in the 386 - Pentium 3 range now :(

Im beginning to wonder if DSDD disks are only reliable if written in a DSDD drive.

Stone
August 7th, 2012, 09:17 AM
Im beginning to wonder if DSDD disks are only reliable if written in a DSDD drive.Why wonder? It's a well known fact. :-)

You can probably overcome this limitation if you format the DS/DD disk in a 360 drive and then write the data to it in a 1.2 drive.

Chuck(G)
August 7th, 2012, 11:07 AM
IOW, "All 5 inch drives are not created equal". AFAIK, you could install a 3" drive and use some "720K" disks instead.

Compgeke
August 7th, 2012, 03:48 PM
Just curious, would you need ROM 1.43 to install a 720 KB drive?

Chuck(G)
August 7th, 2012, 03:54 PM
Not as far as I know. The M24/6300 supported 96 tpi drives right from the get-go. A 96tpi 5" drive in DD mode is virtually indistinguishable from a 3" 720K drive.

ChrisCwmbran
August 13th, 2012, 02:13 PM
I finally got my tweener machine together. I fitted the FDD from the M24 into it just long enough to write some system disks. The media I have is DS-DD, and the drive is a Fuji 5451A02.

I boot the tweener from its HDD which is running DOS 6.22, and at the command prompt I tell it to format the 5.25" floppy drive. It says "Formatting 360k" but when it completes the format it says its only done it as a 180k and accordingly I struggle to write 360k disk images to it!

Any idea why this should be? Defective floppy? Or defective head on the drive perhaps?

Chuck(G)
August 13th, 2012, 02:54 PM
It happens that if an error is detected on side 1, cylinder 0, that DOS will, instead of discarding the floppy entirely, attempt to use it as a 180K floppy. No, you don't get any notice that this is happening. Try some other media, and format it explicity under 6.22:

FORMAT A: /F:360 /U

And see what happens.

Stone
August 13th, 2012, 03:01 PM
IIRC, whenever I made a floppy from an image I didn't format the disk first as the image restore program did that as part of it's included process.

ChrisCwmbran
August 14th, 2012, 05:59 AM
Something is definitely odd about this Fuji drive.

I've tried several floppies from this batch, and it formats them as 180k disks.

They are definitely double sided, and contain data and programs before I reformat them.

It will NOT run the programs on the disks, yet if I put them in the 1.2Mb drive, it sees them as 360k disks, and can run the programs on them.

Stone
August 14th, 2012, 07:11 AM
There is no information available on that Fuji drive.

I think you should use another 360k drive before you waste any more time.

ChrisCwmbran
August 14th, 2012, 07:57 AM
Does this kind of modification really work well:

http://www.classiccmp.org/dunfield/img54306/speed300.htm

Im short on HH 360K 5.25" drives atm.

Chuck(G)
August 14th, 2012, 08:00 AM
Have you tried cleaning the heads on the Fujitsu drive (in this case, the top head)?

ChrisCwmbran
August 14th, 2012, 08:01 AM
Have you tried cleaning the heads on the Fujitsu drive (in this case, the top head)?

Yup, usual thing with isopropyl alcohol. No improvement.

Stone
August 14th, 2012, 08:09 AM
You should really use a cleaning diskette, especially in stubborn cases like this, although even that may not suffice. I still think you need to use another drive altogether.

ChrisCwmbran
August 14th, 2012, 08:18 AM
You should really use a cleaning diskette, especially in stubborn cases like this, although even that may not suffice. I still think you need to use another drive altogether.

I've bid on an IBM half height 360K FDD on Ebay which ends in a couple of hours. Failing that I could borrow one from my 5155 or try a full height unit from a 5160, but in both cases they'd be from machines I haven't looked at yet.

Stone
August 14th, 2012, 08:44 AM
This is only for a test. You do not need to install it (permanently).

You have a 5160 you haven't looked at yet? Now's a good time! :-)

MikeS
August 14th, 2012, 09:06 AM
Does this kind of modification really work well:

http://www.classiccmp.org/dunfield/img54306/speed300.htm

Im short on HH 360K 5.25" drives atm.Yes; a lot of HD drives have jumpers that accomplish the same thing.

But of course you will then have a 720K (80 track, 96TPI) drive; I suspect that's not what you're looking for.

Chuck(G)
August 14th, 2012, 11:55 AM
Yes; a lot of HD drives have jumpers that accomplish the same thing.

But of course you will then have a 720K (80 track, 96TPI) drive; I suspect that's not what you're looking for.

I haven't tried it, but the M24 does have a switch for 720K drives. Exactly what happens when it's thrown, I have no idea. However, what's certain is that you won't be able to create disks readable by your 5150.

A Teac FD55B* (*meaning options for latch placement, etc.) is a very good drive to use in your M24. I've also used Toshiba and Matsushita drives with good results in the 6300.

On the other hand, if you don't have a stock 360K floppies with data on them, you might consider just sticking a 1.44M drive into the M24 and deal with 720K DSDD floppies.

ChrisCwmbran
September 1st, 2012, 01:52 PM
Right I finally have working 5.25" 360K floppy drive in the machine (thanks Malc), and IBM-DOS 3.3 (thanks again Malc), which I have backed up etc etc.

So I can boot to DOS and run DEBUG.

What is the first test you need me to do Chuck?

Chuck(G)
September 1st, 2012, 02:32 PM
If memory serves, we were trying to run down what gives with your hard disk.

So--let's have a look as described in my post #70 on this thing (http://www.vintage-computer.com/vcforum/showthread.php?31098-Olivetti-M24-No-POST&p=236677#post236677).

ChrisCwmbran
September 1st, 2012, 02:41 PM
Ok!

10136

If I press D and Enter a few more times I start to see:

10137

I've tried g=C800:5, g=C800:0, g=e800:5, and g=e800:0 and all just respond with the cursor moving down to the next line, and the machine locking up.

Chuck(G)
September 1st, 2012, 02:48 PM
You don't want to use the g= command unless you know what's at the place you're going to. G= tells the CPU to jump to that location. If there's garbage there, well, you know what happens.

D is the dump command. What looks very odd to me is that contents on the second display appear to duplicate every 16 bytes. That says that there's a disconnect between bit 4 of your controller BIOS PROM and the bus.

If the EPROM on the controller board is socketed, examine it carefully--make sure that no pins are bent or missing.

ChrisCwmbran
September 2nd, 2012, 02:38 AM
Examined the BIOS chip carefully. No sign of any problems. Removed it. Cleaned pins carefully and reinserted it. Still looks the same in debug.

Got another Western Digital MFM controller. Tested post with no drives connected.

New line during POST: "Adapter ROM at C800:0000" and after a few seconds it adds " 1701" to the line so that is much more promising.

I guess now its time to try it with one of the drives.

ChrisCwmbran
September 2nd, 2012, 03:37 AM
Connected ST225.

Booted from floppy. Machine can't see C:. Tried Fdisk but it reports no fixed disks present.

Into Debug.exe. Ran G=C800:5 and did low level format of drive.

Fdisk now works. Formatted drive. 20Mb with no errors.

Reformatted with system tracks. (forgot about sys but it's been a long time).

Machine now boots from HDD no problems.

Made a DOS directory and set about copying the contents of both disks into it.

"Parity error on Systemboard". + lots of screen garbage.

Never rains....lol

Ideas please?

Stone
September 2nd, 2012, 04:04 AM
Sounds like a bad memory chip.

Chuck(G)
September 2nd, 2012, 08:18 AM
Do you have the system diagnostics disk? There are images wandering around the web if you google a bit.

ChrisCwmbran
September 2nd, 2012, 08:33 AM
I have the disk image. My only working 360k FDD is in the M24 atm, so I'll have to swap it back to write the image.

Will report back as soon as I have had a chance to do this.

Thanks again for everyones input.

ChrisCwmbran
September 7th, 2012, 11:24 AM
Ive been running the memory test within "customer.exe" today. Ran it half a dozen times and had no sign of the memory error.

I'm off to Scotland for a week now so it will have to wait until I get back.

Thinking about it, when I was getting the parity error the machine was hooked up only a few feet from a dish washer that was running at the time. I presume the dishwasher is unlikely to have caused any electronic interference?

Chuck(G)
September 7th, 2012, 11:30 AM
Well, if they're sharing the same circuit, it's entirely possible. Dishwashers exhibit nasty surges in power consumption. Even if they're not sharing the same circuit, but if the wiring is close enough, surge currents can inductively couple to another line. I know there in the UK that you often use "ring mains"; in the USA, we're more of a star topology--i.e. branch circuits extend from a single distribution panel with low current (typically 15-20 amp at 120VAC) circuit breakers. In my house, the dishwasher has its own circuit, but that's probably not true in the vast majority of older homes.

patscc
September 7th, 2012, 11:33 AM
More mundane, but does the dishwasher vibrate a lot, mechanically speaking ? Maybe it's mechanical vibration affecting an intermittent on the M24.
patscc

ChrisCwmbran
September 20th, 2012, 03:30 AM
Since getting back from Scotland, I've been running the memory diagnostic on the M24 quite a lot of times and it has not detected a problem.

Accordingly I've been putting a bit of software on the machine - the task which caused it to fall down with the parity error before, and again nothing untoward has happened.

With the machine fully reassembled now, it appears to be working 100%, so thanks to everyone who helped in this thread!

Now to decided which of the other troubled machines in the collection needs attention!

Thanks again everyone!

Trixter
September 20th, 2012, 07:42 AM
Congrats on getting your M24 working again! Now that it's functional, if you're curious to run M24 400-line-specific software on it, head to the AT&T 6300 Shrine's downloads area (https://sites.google.com/site/att6300shrine/Home/downloads) and snag ATT_6300-Enabled_Software.zip (https://docs.google.com/viewer?a=v&pid=sites&srcid=ZGVmYXVsdGRvbWFpbnxhdHQ2MzAwc2hyaW5lfGd4OjU4 YmJjMTQ0OTcxOGI4NWE) which contains a lot of software that has special support for the 6300/M24 640x400 extended CGA mode.

zinamo
October 13th, 2014, 01:52 AM
Hello everyone,
I have found this thread about the Olivetti M24, well I have the same computer but different issue.. I have read all this thread but coulnd't find a solution.. if someone is familiar with this kind of computers, here is the link of my thread: http://www.vintage-computer.com/vcforum/showthread.php?44719-Help-with-Olivetti-M24-AT-amp-T-6300