PDA

View Full Version : XT Clone Not Booting Floppy First



TMA-1
January 24th, 2014, 08:08 AM
Hello all,

I've been lurking on this forum for awhile, and have learned a lot. Having a rather mysterious problem though. I searched, but only found one similar post, and no answer that fit my situation.

I have an Eagle PC Spirit luggable clone. It is an 8088 with no CMOS BIOS setup. It originally came with two floppies, but I later substituted a hard drive in the 2nd of two bays. I have never this problem before, but now it seems to want to boot to the hard drive instead of the floppy first.

Once it boots from the hard drive, the floppy is available and works fine.
I've tried two different floppy controllers. Neither seems to take priority from the hard drive.
Both the "straight" connector and the "twisted end" connector of the floppy ribbon produce the same results.
If I remove the hard drive controller, the system boots from the floppy as it should. Put the hard drive back, and it takes priority over the floppy.

I don't remember changing anything since last time I had this machine "out to play", nor do I know of any jumper setting (main board, floppy controller, HD controller, drives) that could alter what I always thought was carved in stone. Floppy first, then hard drive.

Which aging component has gotten senile, the computer or the user?
Ian

Stone
January 24th, 2014, 08:40 AM
I know that WD RLL controller I sent you couldn't have gotten there yet so what HD controller are you currently using?

SomeGuy
January 24th, 2014, 08:54 AM
Yes, that is very odd.

Is the floppy drive fully functional? Can you read, write, and format disks in it OK? If the drive is fully functional, then that suggest an issue relating to the BIOS.

Most PC and XT clones have DIP switches on the motherboard that indicate how many floppy drives are installed. Odd things can happen if they are not set right. I'd start by double checking those.

Is there a long pause at boot up as if it were trying to boot from a floppy, but no activity showing on the drive?

Is the drive actually showing up in dos as "A" and not "B"?

Even if it is, I'd be tempted to use Imagedisk and its utilities to check that it is really the first BIOS/hardware drive regardless of what DOS says (ImageDisk talks directly to the FDC chip). Also, some very early clones were known to use flat cables instead of IBM style twisted cables.

Are you leaving a non system disk in the drive when it boots? Some floppy disk formatters include a floppy boot sector that immediately continues booting from the C: drive rather than just sitting there telling you to insert a system disk.

vwestlife
January 24th, 2014, 09:20 AM
Older 5.25" drives require a terminating resistor to be installed if they are configured as the A: drive. If you swapped the drives, so that what used to be the B: drive is now installed as the A: drive, it might not work correctly unless you swap the terminating resistor to it as well.

TMA-1
January 24th, 2014, 09:25 AM
Hey Stone!

Actually your controller did arrive already just yesterday, so I was getting organized to try it out on this machine. Thank-you very much! The cables are a godsend too, as I was using scrounged ones that were really too short for my layout.

SomeGuy,

DIP switches say one floppy, which is true.
The lights and activity sequence seem correct, ie the hard drive spins up and lights briefly, then the floppy light goes on and I can hear the stepper motor or whatever it is click--it just acts as if the disk in there is not a system disk (but it is), and I've tried more than one.
After booting, the floppy drive is fully available in DOS as drive A, formats a floppy and everything.

I was hoping it wasn't my BIOS going bad. I've already lost the indigenous keyboard to what I think is ROM degradation.
I'll look into Imagedisk.

TMA-1
January 24th, 2014, 09:30 AM
Hi vwestlife,

Actually I already did exactly that swap, hoping it was the floppy drive itself that was at fault. Sad to say it make no difference, (and I don't see any resistor pack on them.) But thanks.
Ian

TMA-1
January 24th, 2014, 09:42 AM
Oh yeah, Stone's question. (Sorry I somehow lost another reply I had made.)

The existing controller is a WD1002A-WX1. I tried experimentally putting the new controller in. The result was the drive lights oscillated endlessly between a hard drive that could not be interpreted with the pre-existing format, and the floppy drive which the system refuses to acknowledge contains an OS.

But again, if there are no HD controllers present, the system merrily boots from the floppy drive. And then smiles innocently. The bastard.
Ian

Stone
January 24th, 2014, 09:57 AM
It might be possible that the HD controller has a device address setting that is in conflict with the floppy address. You can try changing the WX1's address according to the information here:

http://www.dcllabs.net/docs/1002awx1.txt

See if that makes anything work differently.

TMA-1
January 24th, 2014, 10:44 AM
Sorry to say I'd already tried changing jumper 4 for the device address. With the alternate address I get the dreaded 1701 and a frozen system (with the floppy light on solid.) I think the alternate address is to be used when the controller is configured as the 2nd HDC in the system, but I thought it was worth a try.

Way back when I bought my first hard drive (having survived on twin floppies for close to two years), the controller they gave me did have a conflict with the floppy controller. They symptoms were similar, but the hard drive, being brand new, was not then bootable. I had to get them to exchange the HDC for a different one, then the boot from floppy functioned normally. Now that problem seems to have come back, strangely.

I've tried two other MFM controllers being present in the system. They produce 1701 (which is not surprising because I didn't connect the cables), then the floppy light comes on and the heads engage, but will not read. Stays that way. It seems my system now ignores the floppy if it catches even a whiff of a hard drive. And that's with 2 different FDCs.

I'm stumped.
Ian

tezza
January 24th, 2014, 11:03 AM
The lights and activity sequence seem correct, ie the hard drive spins up and lights briefly, then the floppy light goes on and I can hear the stepper motor or whatever it is click--it just acts as if the disk in there is not a system disk (but it is), and I've tried more than one.
After booting, the floppy drive is fully available in DOS as drive A, formats a floppy and everything.

Just checking. Have you made a system disk in that particular drive since it started having problems, then tried to boot it? I'm thinking that if not, this problem could just be one of drive alignment. If it's out it will likely be ok with it's own formatted disks but not others.

When you say "and everything." have you given it a good workout? Have you run a diagnostic program on the drive to see if seek/read/write is AOK?

Tez

TMA-1
January 24th, 2014, 11:23 AM
Another worthy angle to try, but when SomeGuy suggested I try formatting a floppy to test the drive itself, I later SYS'd it and left that one in, so it's fresh as a summer rose. As long as I remove any HDC from the system, and she boots from that, or any, floppy just fine.

You guys are tryin' hard and I thank you all, by the way.

SomeGuy
January 24th, 2014, 11:47 AM
The lights and activity sequence seem correct, ie the hard drive spins up and lights briefly, then the floppy light goes on and I can hear the stepper motor or whatever it is click--it just acts as if the disk in there is not a system disk (but it is),


But again, if there are no HD controllers present, the system merrily boots from the floppy drive.
O_o That is just weird. Probably some kind of conflict, but not sure what off hand.

It seems like the BIOS knows that it should boot the floppy, and the hardware is, indeed responding. But somehow it cant read from the disk at boot when the controller is present.

I would expect a ROM conflict to either not recognize the hard drive or hang at startup. With an interrupt conflict or interrupt controller failure I would expect one or the other to not work at all, and those don't use the same interrupt anyway.

Perhaps a power supply issue? Is there any difference between cold and warm boot?

Further testing the floppy drive with the controller present might reveal some additional information. ImageDisk should reveal any interrupt or I/O problems and you can also use it to check the rotation speed.

Perhaps inspect the boot sectors to make sure there isn't some odd old boot sector virus floating around?

TMA-1
January 24th, 2014, 01:44 PM
Power supply was an interesting vein, as I had power supply problems in the relatively recent past. Without going into detail, the solution at that time turned out to be a "V Adj" potentiometer. :huh: A rather embarrassing episode, actually. Anyway, I tried turning up the juice another slight notch, but it has made no difference to the current issue.

A warm boot and a cold boot act exactly the same. Actually, on the boots from floppy with no HDC, the hard drive itself was still in place and powered-up, so I don't think it's a power supply issue this time.

[...]

Okay, that took some doing. My 8088 is not exactly web capable. ;) Found ImageDisk. Here is the report from TESTFDC. The 128 byte sector test produces an "error (0) NoAdrMark". Is this significant? (The drive and disks are standard 360's.)

Report on FDC capabilities, issued 1/01/1980 0:42:54:
Single-Density at 250 kbps ............................ Passed
Single-Density at 300 kbps ............................ Not tested
Single-Density at 500 kbps ............................ Not tested
Double-Density at 250 kbps ............................ Passed
Double-Density at 300 kbps ............................ Not tested
Double-Density at 500 kbps ............................ Not tested
Double-Density at 250 kbps / 128 byte sectors ......... Failed
Double-Density at 300 kbps / 128 byte sectors ......... Not tested
Double-Density at 500 kbps / 128 byte sectors ......... Not tested

(This is with the hard drive present and booted from.)
Ian

TMA-1
January 24th, 2014, 08:10 PM
Hey.

I just took another look in my box of cards and pulled out an 8-bit DTC MFM controller. I'd missed it the first time because it's a longer card, but it is still only 8 bits. Guess what? Yep, she works. With that controller present and the drive attached, the floppy boots first.

It seems my Eagle PC "Spirit" clone simply does not like Western-Digital controllers. All the other controllers I was using were WD, including the one that my current hard drive is paired with. But, as I recall, I actually formatted and populated that drive in another machine, an IBM 5160, and later transplanted the controller and drive combination into the Eagle. (Funny how you can do that with the old stuff--complete brain transplant with no device driver issues.)

This echos the experience I had in '84 or so, when I had to return the original controller for a substitute one. Exactly the same scenario, actually, but at that time I was "system halted" because the hard drive was a virgin install, and I couldn't populate it unless I could boot the floppy. I got around it accidentally this time, and never noticed the boot priority flaw until now.

So, it's a good news/bad news sort of deal. The good news is it works, the bad news is I can't use the controller I had intended in this machine. But, I can do some dancing and make it all work out.

Thanks again for all the help guys. Strange that the behavior is present with two different FDCs and a variety of WD HDCs. If anyone wants to pursue information about interrupts or just what the fool motherboard is doing with her children, I will try to oblige.

SomeGuy
January 25th, 2014, 06:28 AM
Ok, that's great! Do you think this card will suffice, or will you still pursue trying to get a different card in there?

My next suggestion was going to be try the cards in a different machine to see if they still do anything like that (doesn't sound like you have one handy), and try a different model of controller if possible.

Anyway, yes, it sounds like you may have just found one of those early freaky compatiblity issues. That that Eagle has a very early BIOS and it could easily have some bugs in it. I can imagine that as the WD rom initializes it could happen to set some bits somewhere that your Eagle bios expects to be set differently.

At a glance, there doesn't seem to be much information about the Eagle PCs out there. Cool early PC clone BTW. If you dump the ROM (If you don't have an EPROM programmer you can probably dump it with DEBUG (http://www.mess.org/dumping/dump_bios_using_debug)), I'm sure Modem7 would like to add it to http://www.minuszerodegrees.net/ . Also, you don't happen to have the Eagle OEM of DOS do you?

Stone
January 25th, 2014, 07:59 AM
I can imagine that as the WD rom initializes it could happen to set some bits somewhere that your Eagle bios expects to be set differently.The WD ROM can be disabled by removing jumper W3.

TMA-1
January 25th, 2014, 08:11 AM
Knowing that my Eagle is allergic to Western-Digital will allow me to change things around some and come up with a satisfactory layout. I also do also have a 5160 I got for 50 bucks, including a 20M hard drive. (My wife isn't pleased with that purchase, but I couldn't resist.) The WD RLL controller will go in there to expand the drive to 30M. I have another card in mind for the Eagle.

As I recall, Eagle went out of business after settling out of court with IBM for the famous "undisclosed" amount of money. The BIOS was found to be a little too close to IBM's. There's an Eagle for sale on eBay right now (not mine), item number 190857597851

Don't have a burner, but I will look into dumping the ROM for that collection.

And yes, if the disks have survived storage I do have the original OEM DOS, (version 2.0 I believe it is), as well as the early "competitor" CP/M-86. When I bought my machine they were discontinuing including that OS because it was becoming clear that DOS was the winner, but I managed to scoop up a copy. I have the complete set of 4 small hard binder manuals.

SomeGuy
January 25th, 2014, 09:45 AM
The WD ROM can be disabled by removing jumper W3.
Right, if the FDD still fails to boot with the ROM disabled, that would point back to a hardware problem or conflict on the boards rather than just BIOS. Of course the hard drive won't boot without that.


a 20M hard drive... The WD RLL controller will go in there to expand the drive to 30M.
If that is a ST225, that probably won't work. With other drives, it is hit and miss and depends on the type of recording surface they use.


And yes, if the disks have survived storage I do have the original OEM DOS, (version 2.0 I believe it is), as well as the early "competitor" CP/M-86. When I bought my machine they were discontinuing including that OS because it was becoming clear that DOS was the winner, but I managed to scoop up a copy. I have the complete set of 4 small hard binder manuals.
Cool. A lot of these early OEM versions have disappeared because people think they are identical to the mainstream versions, but often they include extra features or tools specific to that machine. As an example, a Kaypro 2000 cant properly use PC-DOS 2.x or another OEM's MS-DOS 2.11 on its 720k 3.5" floppy drive, but it uses the Kaypro OEM MS-DOS 2.11 just fine.

In your case, the "Eagle PC" series is supposedly mostly IBM PC hardware compatible, but the same can't be said for the ealier Eagle 1600 MS-DOS machines.

Oh, and:

The 128 byte sector test produces an "error (0) NoAdrMark". Is this significant?
That's normal. Very, very few floppy controllers support that. But yours appears to support Single Density, which is good, as not all do.

TMA-1
January 25th, 2014, 12:07 PM
Right, if the FDD still fails to boot with the ROM disabled, that would point back to a hardware problem or conflict on the boards rather than just BIOS. Of course the hard drive won't boot without that.

For the record, I tried it. (This is the WD1002A-WX1 MFM controller I had been using.) W3 open,... floppy boots, hard drive unrecognized.

TMA-1
January 25th, 2014, 10:16 PM
If that is a ST225, that probably won't work. With other drives, it is hit and miss and depends on the type of recording surface they use.

Hmm. I seem to be lucky today then. The (flawless) ST225 is formatted at 32MB in the 5160, which is busy populating it from the Eagle. It seems healthy enough. The Eagle also formatted a MiniScribe 3425 (20M MFM) at about 30M (there were some defects specified and others found), and also seems happy with the result. I guess time will tell. No data is at risk if one or both bite me on the arse.

Anybody need 8-bit MFM controllers? I have a mitt full for some reason, and a couple of 16-bit ones too. Oh, I remember now--I was buying them from a local surplus store for $0.50 each, looking for an RLL controller, because I thought my Adaptec didn't support the full size of the ST-157R I had then. I later found out it was my ancient low-level formatting program; when I used the card's own BIOS routine, it worked fine.

SomeGuy
January 25th, 2014, 11:01 PM
RUN SPINRITE! The problems on inner tracks may not always show up on an initial format. If you run Spinrite on maximum pattern testing and it says everything is OK, then you lucked out, got a better than normal drive, and you are good to go. Otherwise you will see bad sectors pop up out of nowhere starting about 1/2 or 2/3s in to the disk.

k2x4b524[
January 25th, 2014, 11:24 PM
he's right, plunking a standard MFM onto an RLL is NOT a good idea, but you can go RLL to MFM just fine.

TMA-1
January 26th, 2014, 01:05 PM
Okay, thanks for the warnings. You guys are awesome. I'll experiment some more, which after all, is ALL I'm doing with this obsolete tech. Just because. :D
[Edit] Hmmm, SPINRITE 6 seems to hang my 5160. (DOS 6.22)

modem7
January 26th, 2014, 01:27 PM
[Here (http://www.vintage-computer.com/vcforum/showthread.php?19800-Hard-Drive-Process-Opinions-Wanted)] is an earlier thread (one of a few) on the subject of 'MFM' drives on 'RLL' controllers.

Stone
January 26th, 2014, 02:00 PM
Okay, thanks for the warnings. You guys are awesome. I'll experiment some more, which after all, is ALL I'm doing with this obsolete tech. Just because. :D
[Edit] Hmmm, SPINRITE 6 seems to hang my 5160. (DOS 6.22)Does SpinRite 6 even do MFM/RLL drives? Or run under DOS?

You need SpinRite 2.0

http://dl.dropbox.com/u/14062994/SPINII20.zip

SpidersWeb
January 26th, 2014, 02:24 PM
I'm curious about the 'its bad' opinion - is this from personal experiences?
I haven't had problems with many models of MFM drive being used as RLL.

Edit: just reading the thread modem7 linked to, which has good stuff in it.

Stone
January 26th, 2014, 02:44 PM
I used to use my Maxtor tanks e.g., XT-1140, XT-2190 exclusively on RLL controllers.

SomeGuy
January 26th, 2014, 03:41 PM
Although it was a long time ago, I tested several ST225 20MB drives, and a number of different ~40MB MFM drives with RLL and they failed as described. Formatted with few errors but then Sipinrite barfed on the last third or so of the drive. For a while I had tried to "make do" using one such drive by reducing the number of tracks (a significant amount of space) to avoid the weak areas, but still ran in to intermittent data errors.

On the other hand I have an IMI/Corvus 5 MB drive and Tandon 10MB drive that have been happily formatted to greater capacity with RLL since the early 90s and still work fine.

The thing is, the way RLL encoding works, sectors full of "00"s are no challenge for the weaker surface to hold. Once you put actual data in, it all goes to heck.

Hmm, I supposed they could have varied the surface type during the production of the ST225 and nobody would have noticed unless they cracked it open. But I doubt that.

Also, yea, I think the newer versions of Spinrite have 386 instructions in them.

SpidersWeb
January 26th, 2014, 05:07 PM
I did some sniffing around old news groups. Good reading. Apparently early model ST-225's had a problem with RLL, but later models did not. Also, apparently, if you change the circuit board on an old ST-225 and replace it with an ST-238R board - RLL was reliable afterwards. How legit that information is, I'm not sure, but it's been good fun having a read through.

Posts also show NEC D5126 was pretty good with RLL, and that is a model I've used a few times (with data, and no data loss). I hope groups.google.com doesn't clear their cache, so much fun reading old conversations from the 80's and 90's. A lot of the MFM/RLL drives I use - tend to be from the mid-late 80's with a few exceptions - so that could explain why I haven't had issue.

TMA-1
January 27th, 2014, 09:43 AM
Thanks Stone for pointing me to SPINRITE. It just finished it's thorough pattern testing of the ST225, and apart from 4 questionable sectors (at random spots on the disk) it removed from service, it seems happy enough with the drive in general. So it would seem correct that some 225s' are okay. There's a maintenance tag on the drive stating "repaired and tested 08/89", but I don't know the production date. ("Repaired", drive, really?) Serial number is 2651737.

I will re-install the Miniscribe 3425 and give it a whirl too.

Stone
January 27th, 2014, 09:58 AM
Sounds like you ran it on level 4. I can imagine that took overnight, or longer. :-) I usually run it on level 3 and IIRC that takes quite long enough on an XT.

TMA-1
January 27th, 2014, 03:47 PM
Yep, I let it do the complete scrub-down. I didn't trouble to note exactly, but it was the better part of 24 hours.

Now the Miniscribe 3425 has proven interesting. I formatted it with a recently acquired Adaptec RLL controller. No apparent problems,... SPINRITE is chewing on it now. But before running the current test, I tried to reformat it using the Western-Digital controller, since if it proves reliable at 30M I would use it in the 5160. Instant error as soon as the format starts. It goes thru the dialog, but after the final "Y" to cut it loose, it throws the dish on the floor without even tasting. Perhaps reliability of MFM "rated" drives running RLL encoding is also a function of the controller being used. It will be interesting to see if SPINRITE finds the Adaptec-Miniscribe marriage sound.

SpidersWeb
January 27th, 2014, 04:34 PM
That also happens when you have the data cable upside down or not connected properly, something I seem to by accident on a regular basis at my test bench.
Same symptoms - finds the drive ok, goes through the formatter, then when you go to start it bails out. (Just another possibility)

TMA-1
January 27th, 2014, 05:57 PM
Tactful diplomacy. :) I can't check now because I detached it again. I didn't think I had them on wrong, but I guess that goes without saying doesn't it? You never think you have it wrong, else you'd fix it. When the current experiments have finished I'll try it again. I'm sure you're right--I had a senior moment.

SpidersWeb
January 27th, 2014, 06:15 PM
Worth a shot to check it out :) Just came to mind because that happened to me on Sunday when I was testing a pile of spares.

TMA-1
January 28th, 2014, 09:33 AM
It's working now on that controller. I think what happened is that the data cable had pulled out from the controller, and I didn't notice because it was the 2nd drive, and the primary data cable was hiding the connection from view.

The Miniscribe passed full testing on the Adaptec controller too. 2 for 2. Busy re-assembling the 5160 with two 20-come-30 drives.