PDA

View Full Version : IBM 5155 - F6000 ROM - error



rtwo
March 9th, 2010, 01:33 AM
Hello,
I recently bought myself an almost mint looking 5155 (15 euro's / is about 20 USD)
It has 6 cards in it, one floppy drive and a 20Meg hard disc.
When I boot up, memory starts counting till "640kb Ok" then I get an F6000 ROM error message followed bij press F1 to resume.
After pressing F1, I can boot from floppy or hard disc and everything seems to work just fine, only that F6000 ROM error message at (re)boot.
I search the net, but could not find what this error means, and whats more important, how can I fix it.
Can one off you help me out on this??

cosam
March 9th, 2010, 01:43 AM
I believe the BASIC ROM hangs out at address 0xF6000, so the message would suggest there's something wrong with this ROM. Unless you want to use the ROM BASIC, or programs which access it, it's little more than an annoyance. If you want to fix it, you may try reseating (http://wiki.vintage-computer.com/index.php/Reseat) the ROM (assuming it's socketed) or sourcing a replacement.

rtwo
March 9th, 2010, 01:51 AM
Thank you for your quick reply.
How can I tell what ROM is the ROM BASIC?, is there a board layout so I can check the correct ROM?

rtwo
March 9th, 2010, 10:32 PM
Tried reseating all socketed roms, but no luck, still got the F6000 ROM error. Removed the hard drive to see if it would go to basic, but the floppy led stays on, and no basic is loaded.
Is there anything else I can try??
Can anyone point me to a site where they have replacement parts??

modem7
March 10th, 2010, 12:33 AM
A 5155 (IBM PC Portable) contains a 5160 (IBM XT) motherboard. On the motherboard are two ROM BIOS chips, U18 and U19.

The "F6000 ROM" error means that when the POST (power on self test) in ROM BIOS chip U18 did a checksum of U19, the 8-bit checksum was not the expected value of zero. The most probable cause is that U19 is faulty (faulty ROMs are relatively common).


Is there anything else I can try??
You've reseated the ROMs. I can't think of anything else for you to try.


Can anyone point me to a site where they have replacement parts??
U18 and U19 are IBM specific parts and not sold any more, but there is an alternative:

The part number on U19 will either be one of the following: 5000027 / 6359116 / 62X0854 / 62X0853 / 62x0819 / 68x4370

With that part number, a look at web page http://www.vintage-computer.com/vcforum/showthread.php?9693-IBM-PC-XT-5160-BIOS-versions will reveal which BIOS you have (either 11/08/82 or 01/10/86 or 05/09/86).

There are images of those BIOS' at http://members.dodo.com.au/~slappanel555/bios.htm (in the "5160 IBM-XT" section).

That means that someone with an EPROM programmer, and a 27256 EPROM (rated at 200ns or faster), can create a replacement U19 for you, then post it to you.
Maybe someone on these forums will do that freely or for cost. I certainly can do that if you are in Australia. Where are you (country/region)?

per
March 10th, 2010, 12:40 AM
F6000 ROM error means that the sum of the bytes from address F6000 to F7FFF doesn't equal zero. It basically means that U19 doesn't work properly, either because it's broken or not properly seated. I expect the first one as you've already reseated all the chips.

It's not that hard to fix it, all you need is an 8KB EPROM (chip ID: 2764 or 27C64), and an EPROM burner. After having burned the proper first 8KB of the BASIC code into the EPROM, simply replace U19 with it. I guess somebody here can help you with the burner, but you problably have to get the EPROM yourself.

*Edit*
If you got either the third or fourth revision of the BIOS (according to the site mentioned above), then you will have to get a 32KB EPROM instead, as said above. The chip ID for those are 27256 or 27C256. a 32KB EPROM will also work for the first/second revision of the BIOS, but it's a waste of space since only 8KB are actually being used.

rtwo
March 10th, 2010, 01:09 AM
Part number on the U19 is "6359116" its a 64/256K motherboard.
@modem7: I am from the Netherlands, the other side of earth :) but if you can help me out, that would be great, we can work something out.
If I need to burn an EPROM, is it than also possible to upgrade to a newer version BIOS, by burning 2 EPROM's, the U18 and U19 and replace both of them?, or would my motherboard not accept newer BIOS versions.

per
March 10th, 2010, 01:27 AM
If I need to burn an EPROM, is it than also possible to upgrade to a newer version BIOS, by burning 2 EPROM's, the U18 and U19 and replace both of them?, or would my motherboard not accept newer BIOS versions.
Yes, that's possible.

cosam
March 10th, 2010, 01:31 AM
I should be able to burn these for you if you can provide the chips and point me to the correct images (and cover return postage).

rtwo
March 10th, 2010, 01:47 AM
I should be able to burn these for you if you can provide the chips and point me to the correct images (and cover return postage).

You are also from the Netherlands??, I am from Rotterdam, where are you coming from, maybe whe can use PM to discuss this further.

modem7
March 10th, 2010, 02:27 AM
I should be able to burn these for you if you can provide the chips and point me to the correct images (and cover return postage).
Per per's statement, any of the 5160 BIOS images can be used.
So with rtwo's desire for a newer BIOS, use the last released BIOS.
It's at http://members.dodo.com.au/~slappanel555/bios/BIOS_5160_09MAY86.zip

Two 27256 (or 27C256) rated at 200ns, or faster, are required. One for U18 and one for U19.


@modem7: I am from the Netherlands, the other side of earth :) but if you can help me out, that would be great, we can work something out.
If you can't work something out with cosam, then let me know.

cosam
March 10th, 2010, 02:57 AM
You are also from the Netherlands??, I am from Rotterdam, where are you coming from, maybe whe can use PM to discuss this further.

I'm based in Den Haag, but I get around a bit. Drop me a PM (Dutch is fine) and we'll see if we can arrange something.


Per per's statement, any of the 5160 BIOS images can be used.
So with rtwo's desire for a newer BIOS, use the last released BIOS.
It's at http://members.dodo.com.au/~slappanel555/bios/BIOS_5160_09MAY86.zip

Two 27256 (or 27C256) rated at 200ns, or faster, are required. One for U18 and one for U19.

OK, just don't want to leave anything open to interpretation and end up burning the wrong images to someone else's EPROMs ;-)

MikeS
March 10th, 2010, 06:53 AM
I'm based in Den Haag, but I get around a bit. Drop me a PM (Dutch is fine) and we'll see if we can arrange something.



OK, just don't want to leave anything open to interpretation and end up burning the wrong images to someone else's EPROMs ;-)That's what the 'E' is for ;-)

cosam
March 10th, 2010, 07:06 AM
That's what the 'E' is for ;-)

Heh, indeed! Although not having the kit to "E" them myself, it could make for a long turnaround...

MikeS
March 10th, 2010, 07:11 AM
Heh, indeed! Although not having the kit to "E" them myself, it could make for a long turnaround...Especially this time of year, when the days are short and the sun's low on the horizon...

rtwo
March 12th, 2010, 02:53 AM
I am having a hard time finding 27(c)256 eproms rated 200ns or faster.
I can however get my hands on new 27C512-120 ( 64k*8 ) eproms
Is it possible to burn the images on the 512 eproms?, I know its a waste of space, but at least I can use the new version Bios, without errors at POST.

cosam
March 12th, 2010, 04:34 AM
Conrad have '256s as fast as 45ns (Componenten > Componenten actief > IC's > Geheugen IC's > Opslag-ICs, EPROM). Be aware that some of them are OTP (one-time programmable) with which there are no second chances.

As far as I can see, '512s should work too. I'm pretty sure I can program them for you, as long as they don't require 13.5V Vpp. You would need to tie the extra address line (pin 1) to ground somehow, unless this is already done on the motherboard. If I'm burning them for you anyway, I can tack a wire on there no problem.

per
March 12th, 2010, 05:28 AM
You would need to tie the extra address line (pin 1) to ground somehow, unless this is already done on the motherboard. If I'm burning them for you anyway, I can tack a wire on there no problem.

Pin 1 is actually wiered to XA14 (Internal address line 14), together with pin 27. This is to keep compability with both 8KB (A14 on pin 1, pin 27 unused) and 256KB (A14 on pin 27, pin 1 unused) EPROMs.

I have a hypothesis that the 256-640KB models may have the Pin1 line disconected, however, I'm not able to verify it at this point. To keep safe, assume the above pharagraph to be the case.

One solution is to burn the images in both the lower halfs and upper halfs (mirrored) of the EPROMs. By this, the state of pin 1 doesn't actually have any effect on the data returned.

cosam
March 12th, 2010, 06:29 AM
One solution is to burn the images in both the lower halfs and upper halfs (mirrored) of the EPROMs. By this, the state of pin 1 doesn't actually have any effect on the data returned.
Yep, I can do that. I was just worried what may happen if that pin didn't go anywhere at all and what may happen if that input was left floating.

per
March 12th, 2010, 08:29 AM
I was just worried what may happen if that pin didn't go anywhere at all and what may happen if that input was left floating.

You're not the only one there. I have been thinking about that for some time, and based on my level of self-education on that field, I would say that:

In a standard TTL chip, V+ are applied to an output if it's on (through a resistor. Open-collector means that the resistor and V+ are external).
If an output is off, a transisor drains the current from the V+ source to ground, and reduces the output to 0V.

Two TTL Inputs mainly consists of a dual-emitter tetrode transistor. It worls like, if (the drain on the emitters is big enough [I'm unsure if the current actually has something to say here, so please correct me if I'm wrong] and) the voltage is signifficantly lower than V+, it will activate the output.



The Thing when nothing is connected to an input or output is; Current can't flow. No current means no voltage. If we now consider the TTL input, the transistor will not be able to activate, and both inputs will appear on.

Intergrated chips should have pull-resistors on the lines just to prefix the defaults, but if that's not the case, I actually think that an unconected input will apear high there too.

per
March 12th, 2010, 01:41 PM
BTW, I just found out that the ROM's used in the XT actually is 250ns ones, not 200ns. This means that 250ns EPROMs can in fact be used in addition to the faster kinds.

rtwo
March 27th, 2010, 11:20 AM
Received my eproms from Cosam today, thank you for burning the images for me.
My 5155 starts up without any errors and with the newer bios with a quicker memory count.
Cosam, you made me a very happy person. Now I have 2 IBM's working 100%, now have to fix my other two.
Thanks again..

cosam
March 27th, 2010, 12:52 PM
Glad to hear they did the trick. If only all our problems were so easy to diagnose and fix!

Be sure to let us know how it goes with your other machines.

k2x4b524[
March 27th, 2010, 02:04 PM
I'm in the US, so it is possible to make the Memory count faster by using a faster rom? I'd like more details on this one..

modem7
March 27th, 2010, 06:14 PM
I'm in the US, so it is possible to make the Memory count faster by using a faster rom? I'd like more details on this one..
On a 5160 motherboard, the timings used for reading the ROMs are fixed. Therefore getting ROMs faster than that specified by IBM will not affect performance.

The reason why rtwo saw a faster RAM count was because he upgraded to a newer version of the BIOS, one that in this case, performs a faster RAM check compared to previous versions (the RAM is checked on a word basis rather than a byte basis).

k2x4b524[
March 27th, 2010, 06:56 PM
ohh i gotcha, would explain why my 1984 xt counts faster than my 1986 xt, the 1984 has the newest roms in it..

modem7
March 27th, 2010, 07:14 PM
ohh i gotcha, would explain why my 1984 xt counts faster than my 1986 xt, the 1984 has the newest roms in it..
I think you may have crossed your "1984" and "1986".

The differences among the 5160 ROM revisions are listed at http://www.vintage-computer.com/vcforum/showthread.php?9693

k2x4b524[
March 27th, 2010, 08:07 PM
nope, not crossed anything, the tag on the speaker on one say mid '86, and is DEFINETLY slower. Doesn't leave out the possibility that the board in there ISNT the original