Early PC 6300 Help

[marcoguy]

Hi all, it's been a while

I'm posting today because a friend of a friend was cleaning out his attic, one thing lead to another, and I'm now the proud owner of two working 6300s, plus a monochrome monitor, two keyboards, and the original docs and software :D

Cracking open the earlier of the two units was a little bittersweet, there was some damage to the frame in one corner that required me bending it back into place so I could slide the bottom cover off, and when I did I found that some spiders made a home in the underside of this system. After a little cleaning, however, I was overall pleased to find that the only component that had suffered damage was the internal speaker, which had detached from the motherboard. Not even any corrosion on the RTC battery. I was more pleased to find that the BIOS in this thing is 1.0, dated 5/84. That makes this a very early system, as according to Wikipedia AT&T didn't launch the 6300 until June of that year. This leads me to my first question: Is this BIOS revision particularly rare, or of value to anyone here? I can't seem to find any info about the 6300's BIOS online, other than the fact that many systems shipped with 1.21, and that the last revision was 1.43. Does the 6300's BIOS differ at all from that of the M24? Other than the BIOS, I can't find any indications that this machine is a particularly early one. The RAM was upgraded to 640k, and it was upgraded with a long dead ST-225 (which definitely wasn't original as the drive labels on the front are "A" and "B," suggesting this was originally a dual disk drive system). There are also three 16-bit connectors on the bus extender on this older system rather than two on the newer one, but I don't know if that is of any significance. Anyway, if someone has info on anything to look for in this system unit that would differentiate an early 6300 from a late one other than the BIOS revision, that would be much appreciated so I can get a better idea of the system I have.

The more recent of the two systems has a BIOS version 1.1 which, to my knowledge, is relatively early as well. This system is pretty standard, but it's causing me an issue. I built one of James Pearce's amazing ISA compact flash adapters a while back, and was able to successfully flash the 8k IDE_XT ROM and use this card in my Deskpro 286. However, when I insert this board into either 6300, they are able to detect an option ROM at C800:0000, which is correct, but the XT-IDE BIOS isn't loaded by the system. I just get "No fixed disk present" and it attempts to boot from the floppy drive. I've inspected the jumper settings and, using the manual fir reference, configured one of the systems to "use external hard drive code" (which I take to mean the system will assume there is an HDC installed with an on-board BIOS), but no dice. I won't be able to access either system for a day or two, but I could take a few pics of a boot up to show the issue when I get the chance. Until then, I'd appreciate if someone had some insight into how the 6300 handles option ROMs, and if it differs from that of a standard PC compatible.

One last bit of info, one of the keyboards isn't an original AT&T. Its a Key Tronic KB5151, and it appears to have been sold specifically for the 6300, as it's compatible both electrically and in protocol. I wasn't aware of any 3rd party keyboards for the 6300, and I found it interesting as this KB has a superior layout compared to the original AT&T keyboard IMO.

Well, there you have it. These are my two latest acquisitions, and some of my favorites. The PC 6300 is my favorite XT compatible, and I'm very fortunate to now have two in my possession. I haven't yet decided what the fate of these systems will be, I might keep one and sell one, or maybe set one up with the original AT&T peripherals, and the other with a VGA monitor/adapter and the 3rd party KB. Either way, I'm glad to finally have a 6300 in my collection. I look forward to running Windows 1.03 with the AT&T drivers. I appreciate any help that anyone has to offer, thank you.

 

[Compgeke]

The more recent of the two systems has a BIOS version 1....I built one of James Pearce's amazing ISA compact flash adapters a while back, and was able to successfully flash the 8k IDE_XT ROM and use this card in my Deskpro 286. However, when I insert this board into either 6300, they are able to detect an option ROM at C800:0000, which is correct, but the XT-IDE BIOS isn't loaded by the system. I just get "No fixed disk present" and it attempts to boot from the floppy drive...Until then, I'd appreciate if someone had some insight into how the 6300 handles option ROMs, and if it differs from that of a standard PC compatible.

One of the first things you're going to want to do is find a copy of the PC-6300 BIOS 1.43. I had to do that in order to use a Future Domain TMC-850 in my PC-6300 and even then there were occasional issues of it wanting to boot from floppy first. If you have a programmer and a few spare ROMs the images are easy to find. If not, asking around here usually turns up someone willing to write 'em for a few bucks.

 

[NeXT]

Not even any corrosion on the RTC battery.

Remove it. I don't care if it even checks good you MUST remove it.
Myself included, we've been discovering that if the battery leaks, no matter how clean the board looks after the machine seemingly drops dead. We still can't figure out what causes it.

 

[marcoguy]

One of the first things you're going to want to do is find a copy of the PC-6300 BIOS 1.43. I had to do that in order to use a Future Domain TMC-850 in my PC-6300 and even then there were occasional issues of it wanting to boot from floppy first. If you have a programmer and a few spare ROMs the images are easy to find. If not, asking around here usually turns up someone willing to write 'em for a few bucks.

Well, its a good thing I've already got a copy of the 1.43 BIOS coming in the mail I'm hoping the BIOS upgrade will fix the way it handles expansion cards with on-board ROMs, and based on your experience with the TMC-850 it seems that's true to some degree.

EDIT: Don't worry NeXT, that was the first thing I did after opening the case

 

[Stone]

Remove it. I don't care if it even checks good you MUST remove it.
Myself included, we've been discovering that if the battery leaks, no matter how clean the board looks after the machine seemingly drops dead. We still can't figure out what causes it.

Probably the same scenario that many boards appear to succumb to after battery leakage. The leakage gets to the inner layers (that are not visible) of the board and destroys some traces there. I have several different boards with battery leakage where all the visible surface traces are continuous or have been successfully repaired to become continuous but still will not function. Apparently there is internal damage that is neither visible nor accessible.

 

[Chuck(G)]

I replaced the leaking 6300 NiCd battery with a plain old coin cell in series with a Schottky diode. Works fine. One thing works to your advantage--if the battery leaks in a 6300 that's been stored in a "working" orientation (i.e. not on its side), the goop from the battery will tend to drip away from the inverted motherboard. As documented in other threads, there are modifications to the bus converter board that can make for better operation.

IIRC, there were some issues with the XTIDE board and the 6300. Check your data by using DEBUG to dump the area starting at C800:0.

 

[marcoguy]

I've always loved the 6300's downward-facing motherboard for that among other reasons. A nice piece of Italian engineering

Chuck, is this the bus converter mod you're talking about? If so, I might have to see about implementing that in one of my systems. It seems like it would be pretty handy to have.

I'll be home later today, and can dump the XTIDE BIOS with debug then to see if the data is intact/being read properly. However, from the research I've done it seems the 6300 BIOS 1.43 should help with the issue.

 

[1ST1]

Please note that with BIOS 1.43 the meanings of the dip switches will change. You have to resetup your harddisk dip switches.

Seagate ST225 is not the original harddisk when Olivetti shipped out M24 with HDU kit. Usually it was Olivetti Lexikon 5520 harddisk, which has the same CHS parameters as ST225.

 

[Chuck(G)]

My own 6300 has an 80MB "hard card" drive, and, although I've reported on using the XTIDE and issues, I've largely forgotten about it. Something about using the "Chuck Mod" on the XTIDE requiring special code because 16-bit I/O instructions on the 6300 fetch the high order byte first, then the low order, instead of what the 8088 BIU does.

Sorry that I can't be of more help, but memory of the events is suffering from wetware bit-rot.

 

[Krille]

My own 6300 has an 80MB "hard card" drive, and, although I've reported on using the XTIDE and issues, I've largely forgotten about it. Something about using the "Chuck Mod" on the XTIDE requiring special code because 16-bit I/O instructions on the 6300 fetch the high order byte first, then the low order, instead of what the 8088 BIU does.

Sorry that I can't be of more help, but memory of the events is suffering from wetware bit-rot.

I think this is what you are referring to.

AFAIK, there shouldn't be any problems as long as you're using the "XT-CF 8-bit PIO mode". It's the default (and also the slowest) transfer mode on all XT builds with XT-CF support and compatibility with these machines is the very reason for that.

 

[Oscar]

PC 6300 / M24 Battery damage

PC 6300 / M24 Battery damage

Originally Posted by NeXT

Remove it. I don't care if it even checks good you MUST remove it.
Myself included, we've been discovering that if the battery leaks, no matter how clean the board looks after the machine seemingly drops dead. We still can't figure out what causes it.

Probably the same scenario that many boards appear to succumb to after battery leakage. The leakage gets to the inner layers (that are not visible) of the board and destroys some traces there. I have several different boards with battery leakage where all the visible surface traces are continuous or have been successfully repaired to become continuous but still will not function. Apparently there is internal damage that is neither visible nor accessible.

I have a battery goop damaged 6300 as well. Actually, the Olivetti M24 version of it. It has the classical symptoms of this problem.
There's no visible damage, it must be a matter of eating away the inner layers.

Very unlikely, but does anyone have the PCB layout sheet showing what traces pass by the battery area?

Getting the M24/6300 going is one of my to-do projects.
Assuming the battery goop does not travel very far inside the board, it must be one or two bad traces at most.

Regards,

Oscar.

 

[marcoguy]

Alright, time for an update. I've finally got one of the 6300s up and running with BIOS 1.43. However, I'm still having some issues with the XT-IDE BIOS. I've since built another one of James' boards, this time his Lo-Tech 8-bit IDE adapter. I've tested it on a 486 computer of mine and it works fine with the BIOS provided on James' site. However, when plugged into the 6300 with the same hard drive, the BIOS now loads thanks to the AT&T BIOS 1.43, but no drive is detected. I've tried flashing the XT-IDE BIOS with the controller setting set to PIO mode, but no luck. Has anyone else had this issue with the 6300 and an XT-IDE card?

 

[NeXT]

I have a battery goop damaged 6300 as well. Actually, the Olivetti M24 version of it. It has the classical symptoms of this problem.
There's no visible damage, it must be a matter of eating away the inner layers.

Very unlikely, but does anyone have the PCB layout sheet showing what traces pass by the battery area?

Getting the M24/6300 going is one of my to-do projects.
Assuming the battery goop does not travel very far inside the board, it must be one or two bad traces at most.

Regards,

Oscar.

Whatever it is, it's not something super critical or major. When I hung a logic analyzer off the EPROMs and CPU I was finding the Data bus was pushing 1's and 0's around and it would go silent when the RESET was asserted but beyond I could not tell exactly why there was data flying around or what it was coming from. It's not like the systems go 100% brain dead. Even if it's pure garbage it's trying to do something.

 

[Krille]

Alright, time for an update. I've finally got one of the 6300s up and running with BIOS 1.43. However, I'm still having some issues with the XT-IDE BIOS. I've since built another one of James' boards, this time his Lo-Tech 8-bit IDE adapter. I've tested it on a 486 computer of mine and it works fine with the BIOS provided on James' site. However, when plugged into the 6300 with the same hard drive, the BIOS now loads thanks to the AT&T BIOS 1.43, but no drive is detected. I've tried flashing the XT-IDE BIOS with the controller setting set to PIO mode, but no luck. Has anyone else had this issue with the 6300 and an XT-IDE card?

Regular 16-bit I/O doesn't work on these machines. That's why you need to use the XT-CF 8-bit PIO mode. That mode in turn is only supported by ATA-2 compliant drives, ie Compact Flash cards and Microdrives. There are also "real" harddrives that support this transfer mode but they are rare.

 

[marcoguy]

Ah, thanks for the info. I wasn't educated as to the nature of 8-bit PIO, and didn't realize it would restrict me to 8-bit ATA compliant devices. Time to get a CF adapter

 

[Trixter]

I was more pleased to find that the BIOS in this thing is 1.0, dated 5/84. That makes this a very early system, as according to Wikipedia AT&T didn't launch the 6300 until June of that year. This leads me to my first question: Is this BIOS revision particularly rare, or of value to anyone here?

I'm not sure if it is of any value to anyone, but that is indeed exceedingly rare. They switched to the 1.21 BIOS fairly early on, although the 1.43 BIOS is what you want because it enhances compatibility (later versions of Microsoft Word for DOS require the 1.43 BIOS) and allows for add-in video cards. I've never seen the 1.0 BIOS; if you can dump it before replacing it, that might be nice to archive somewhere.

There are also three 16-bit connectors on the bus extender on this older system rather than two on the newer one, but I don't know if that is of any significance. Anyway, if someone has info on anything to look for in this system unit that would differentiate an early 6300 from a late one other than the BIOS revision, that would be much appreciated so I can get a better idea of the system I have.

I don't think I've ever seen that before -- cool. You may want to leave that one alone as stock or something.

I look forward to running Windows 1.03 with the AT&T drivers.

Geoworks Ensemble using the AT&T drivers is also wickedly fast and impressive. However, it locks up if you swap your 8086 for an NEC V30, so you have to keep the 8086 if you want to run it.

 

[Krille]

I guess it should be possible to add a transfer mode specifically for Olivetti machines. Something like this;

	; Reading from drive
	in	ax, dx
	xchg	ah, al
	stosw

	; Writing to drive
	lodsw
	xchg	ah, al
	out	dx, ax

It would be slow but would allow use of normal harddrives.

Am I correct in this line of thinking? What do you Olivetti experts say?

 

[Trixter]

Seems ok to me, but someone should rope Chuck(G) in to verify. Are all communications (not just sector transfers) with the hardware performed in 16-bit words? If not, my concern would be reading the final single byte reliably.

I question the usefulness of doing this though. CF cards are such an improvement over period hard drives (10x or more IOPS, capacity, small size, silent) that I can't imagine anyone preferring anything else...

 

[Krille]

Time for an update.

I went ahead and added a new device type/transfer mode to XTIDE Universal BIOS based on the idea I wrote about in the above post. It's the same code as used in the "XT-CF PIO16" mode except for the xchg instructions added as above. Marcoguy then kindly tested the new build on his machine and the results are very strange.

This image is with the regular "XT-CF PIO16" mode;

This image is with the "XT-CF PIO16RO" mode (RO as in Reversed Order);

As you can see (hopefully, unless the forum destroys these images), the problem is not (only) that the byte order is swapped. It appears that there's a single odd byte inserted somewhere which cause the bytes to be swapped across word boundaries.

If someone could determine exactly what happens when doing 16-bit I/O on these machines (and provided there's no data loss in transfers) then I'm sure we could add a compatible transfer mode to the XTIDE Universal BIOS.

Bottom line is; I need help with this.

 

[Trixter]

I'd love to help but I'm not knowledgeable at that low a level. Would a link to the hardware/tech schematics help? Try this: ftp://ftp.oldskool.org/pub/drivers/...livetti_M21_M24_Theory_of_Operation_Nov84.pdf
Specifically pages 3-2 and 3-3 describe how 16-bit transfers work.

I have the actual 6300 tech ref somewhere; it's the only thing I haven't scanned and put online because it's a beast. But I'll give it my best effort if the above document doesn't help, just let me know.