PDA

View Full Version : XT startup questions



clh333
March 23rd, 2017, 07:37 AM
Going through a 5160 before applying power: Cleaned out the power supply (see attached) with compressed air, used compressed air followed by ChuckG's method to clean the motherboard. Tested the 5V - to - ground paths on the motherboard for resistance; all around 475 Ohms. So far so good, but a couple of questions:

I would like to test the power supply before reattaching it to the board. Last year Modem7 was kind enough to observe that the PS needs a load resistor applied to two of its pins before it will start. The problem is that I can't remember which two pins - well, one of them is ground, obviously, but what's the other one? 12V? 5V? Power Good? I looked for the old post but could not find it, so I have to ask a stupid question - twice!

Also: this board is a 256 / 640 configuration. Banks 0 and 1 are populated, 2 and 3 are not. MZD says these other banks use a different memory IC. I have AST SixPack Plus boards I could use instead of populating the motherboard, but am wondering if there is an advantage to one configuration over another.

Finally: Checked the switch settings and have a question about switches 7 and 8. Both are on (see attached), which indicates 1 floppy drive. However, there are two drives on the machine: one 5.25 and one 3.5. According to MZD the setting should be 7 off, 8 on when two floppies are present. Am I missing something?

Thank you for your assistance.

-CH-

Attachments:

37249 37250 37251 37252

Stone
March 23rd, 2017, 07:53 AM
You can just connect the PSU to the motherboard or a hard drive to load the PSU.

Motherboard RAM should be somewhat faster than expansion card RAM but on an XT... does it really pay to quibble??? :-)

No, you are not missing anything RE: the switches. Your observations are entirely correct. :-)

clh333
March 23rd, 2017, 10:20 AM
I'll try an HD. Didn't want to accidentally fry the board.

Take a look at this 3.5 Toshiba FDD; dust bunnies around the head! Who was the slob who used this?

37262

-CH-

Chuck(G)
March 23rd, 2017, 10:58 AM
Far more common than you'd otherwise think. One of the problems with the conventional PC design is that it relies on a negative case pressure, which allows all sorts of junk to be sucked in through the floppy slots.

The better approach would have been a positive case pressure with a filter on the PSU fan intake. I used to do this, particularly with big tower systems that sat on the floor. The floppy drives stayed clean, as did the innards of the machine in general.

clh333
March 24th, 2017, 05:34 AM
Spent most of the past 36 hours cleaning components and came to testing the power supply. Following the diagnostic procedure outlined at MinusZeroDegrees.net I tested the minimum diagnostic configuration. The Power supply would start, click and stop all within a half-second, with either P8+P9 or P9 only plugged into the motherboard. However, when I attached the power supply to the IBM HD (looks like the Astrodome: the biggest drive I've ever seen) the power supply started and ran the fan continuously. MZD conclusion: highly likely a bad tantalum capacitor on the 5V rails.

As Dorothy said to Toto, "We're not in Kansas anymore". Now the question is threefold: Is there a diagnostic procedure from here? Is there a list of the requisite caps? Is there a supplier anyone can recommend for obtaining the parts? I'm assuming you would replace them all at this point, no?

Thanks for your reply.

-CH-

Stone
March 24th, 2017, 06:02 AM
Unfortunately while there's only a few tantalums on the 12V rails there are plenty of them on the 5V rails so you've got your work cut out for you.

Malc
March 24th, 2017, 07:18 AM
Check C68 it's the only one on the -5V line and 13 on the +5V line, Have a good look at them using a magnifying glass, They don't always show signs of damage though, Did you do the resistance check.

Stone
March 24th, 2017, 07:54 AM
Here they are:

http://www.minuszerodegrees.net/5160/misc/5160_5volt_caps_plus.jpg

clh333
March 24th, 2017, 08:53 AM
Unfortunately while there's only a few tantalums on the 12V rails there are plenty of them on the 5V rails so you've got your work cut out for you.

Thank you for not saying "I told you so..."

-CH-

fatwizard
March 24th, 2017, 09:11 AM
You might also keep in mind that if that old hard drive was originally parked, it won't be after a spin test (if it's still operational, that is). That means the heads are likely resting on track zero now, and damage to track zero is usually the end of an old drive. Try to not subject it to shock. I LOVE those old, giant, fragile hard drives, and eBay sellers are forever connecting them to a power supply and they will announce, "spins up, no further testing done".

clh333
March 24th, 2017, 09:15 AM
Check C68 it's the only one on the -5V line and 13 on the +5V line, Have a good look at them using a magnifying glass, They don't always show signs of damage though, Did you do the resistance check.

I ran a resistance check before applying power, but only checked the three +5V pins. Each showed resistance around 475 Ohms, so I figured no worries. After your suggestion I went back and checked again: +5V OK, -5V infinite resistance, + 12V and -12V each around 100 MOhm, Power Good (P8 pin 1) infinite resistance. C68 does not appear abnormal in any way but it's right up against a slot. C10, 12, 14 and 16 are populated with 3-legged tantalum caps, C11, 13 and 15 have pcb holes that have been filled when wave soldering was performed; no caps there.

I'll have to do more research to understand the power-good circuit. Thank you for your suggestion.

-CH-

Malc
March 24th, 2017, 09:15 AM
... I LOVE those old, giant, fragile hard drives, and eBay sellers are forever connecting them to a power supply and they will announce, "spins up, no further testing done".

All mine ' Spin up ' don't do anything else though, Only good for door stops or to use as a load.

clh333
March 24th, 2017, 09:24 AM
You might also keep in mind that if that old hard drive was originally parked, it won't be after a spin test (if it's still operational, that is). That means the heads are likely resting on track zero now, and damage to track zero is usually the end of an old drive. Try to not subject it to shock. I LOVE those old, giant, fragile hard drives, and eBay sellers are forever connecting them to a power supply and they will announce, "spins up, no further testing done".

And no returns. Last year I purchased one to go with another (the first) XT I was rehabilitating. The eBay seller guaranteed it to work and to be loaded with DOS. By the time it made the trip from California it would not boot, however, quite possibly from having been x-rayed to the bejeezus belt on account of being suspected to be a bomb. Fortunately I was able to stumble though a low-level format and reload it. Last time I did one of those was 1985 - I was a little rusty.

This IBM drive looks like it ought to have its own ZIP code. Thanks for the warning. I'll be gentle.

-CH-

modem7
March 24th, 2017, 03:24 PM
Last year Modem7 was kind enough to observe that the PS needs a load resistor applied to two of its pins before it will start.
It wasn't me.


However, when I attached the power supply to the IBM HD (looks like the Astrodome: the biggest drive I've ever seen) the power supply started and ran the fan continuously.
That's a good sign. So, with the power supply powering the hard drive, measure the +5V, -5V, +12V, -12V, and POWER GOOD lines, per [here (http://minuszerodegrees.net/5150_5160/MDC/no_beeps_2.htm)].
If they measure as expected, then you have good confidence that the power supply is functional.


Following the diagnostic procedure outlined at MinusZeroDegrees.net I tested the minimum diagnostic configuration. The Power supply would start, click and stop all within a half-second, with either P8+P9 or P9 only plugged into the motherboard.

It is true to say that usually, the 5160 motherboard by itself is enough of a load for the power supply. I use "usually" rather than "definitely" because there was one case seen in these forums where (for whatever reason) the 5160 motherboard by itself was not enough of a load.

Your resistance measurements of the motherboard suggest (repeat: suggest) no shorts, nor partial shorts. Does the power supply and motherboard start if you connect up the power supply to both the motherboard and hard drive ?

Stone
March 24th, 2017, 03:41 PM
It is true to say that usually, the 5160 motherboard by itself is enough of a load for the power supply. I use "usually" rather than "definitely" because there was one case seen in these forums where (for whatever reason) the 5160 motherboard by itself was not enough of a load.That is what is commonly referred to as an outlier ​and can usually be ignored.

clh333
March 25th, 2017, 03:31 AM
That is what is commonly referred to as an outlier ​and can usually be ignored.

I'm afraid I don't know enough to make assumptions. I just keep asking questions. Thank you all for your suggestions.

-CH-

clh333
March 25th, 2017, 04:23 AM
...measure the +5V, -5V, +12V, -12V, and POWER GOOD lines, per [here (http://minuszerodegrees.net/5150_5160/MDC/no_beeps_2.htm)].
If they measure as expected, then you have good confidence that the power supply is functional.

With both the HD and the motherboard connected the power supply once again started and ran. Measurements taken at the suggested pins (relative to ground):
P9, Pin 6: +5.17V
P9, Pin 5: +5.17V
P9, Pin 4: +5.17V
P9, Pin 3, -5.04V
P8, Pin 4: +12.22V
P8, Pin 3: -12.18V (transposed? must be: your reference has pin 4 at -12V.)
P8, Pin 1: +5.11V


... Does the power supply and motherboard start if you connect up the power supply to both the motherboard and hard drive ?

The power supply started and ran for more than ten seconds, it seemed to me, when the speaker emitted one long tone and then shortly thereafter two shorter, higher-pitched tones. After what seemed to be a 30-second to one-minute interval two short tones were repeated. After that, no more sounds until shutoff.

On the strength of this test I will reassemble the unit and see if I can get a display. Thank you very much for your suggestion.

-CH-

clh333
March 25th, 2017, 04:33 AM
It is true to say that usually, the 5160 motherboard by itself is enough of a load for the power supply. I use "usually" rather than "definitely" because there was one case seen in these forums where (for whatever reason) the 5160 motherboard by itself was not enough of a load.

Previously I tried applying a load to the power supply using the 1.5 Ohm resistor shown in the attachment below. Last year it was successful in getting the other (earlier version) XT power supply to run independent of any other connection. This year, with this one, it was not. Pretty close to a dead short, I obviously did not apply this load for more than a few seconds.

-CH-

37289

clh333
March 25th, 2017, 12:48 PM
Reassembled the XT as it was when it was shipped to me. Attached my original XT's keyboard and monitor. Stuck in a floppy drive of suspect quality, crossed my fingers and flipped the switch.

Display showed the usual memory count, without complaint, and then booted into MS DOS 6.22 - from the hard disk!

This machine was sold as "for parts only" so I am feeling quite lucky. Thanks to all who contributed...

-CH-

Pictures attached.

37291 37292 37293

modem7
March 25th, 2017, 06:19 PM
P8, Pin 4: +12.22V
P8, Pin 3: -12.18V (transposed? must be: your reference has pin 4 at -12V.)

My reference is from the IBM manual, and I just now confirmed it as correct by measurement of one of my power supplies. If the 12V lines have been transposed in your power supply, there are nasty implications for the 12V filtering capacitors on the 5160 motherboard, and for any expansion cards that use 12V.


Previously I tried applying a load to the power supply using the 1.5 Ohm resistor shown in the attachment below. Last year it was successful in getting the other (earlier version) XT power supply to run independent of any other connection. This year, with this one, it was not. Pretty close to a dead short, I obviously did not apply this load for more than a few seconds.
Different make/model power supplies have different minimum loading requirements. For example, some also require a load on the +12V line.

We know that in the history of IBM 5150 and 5170 production, that the 'guts' of the 5150's and 5170's power supplies changed. It may have been the same for the IBM 5160 (resulting in varying minimum loading requirements).

clh333
March 26th, 2017, 03:55 AM
My reference is from the IBM manual, and I just now confirmed it as correct by measurement of one of my power supplies. If the 12V lines have been transposed in your power supply, there are nasty implications for the 12V filtering capacitors on the 5160 motherboard, and for any expansion cards that use 12V.)

"transposed" as in "had the probes of my auto-ranging multimeter transposed", i.e. red on ground and black on the 12V pins. Was the proper orientation for P9 but P8 is the reverse. Checked and verified just now.


Different make/model power supplies have different minimum loading requirements. For example, some also require a load on the +12V line.

We know that in the history of IBM 5150 and 5170 production, that the 'guts' of the 5150's and 5170's power supplies changed. It may have been the same for the IBM 5160 (resulting in varying minimum loading requirements).

The internals of this later power supply are enough different that it supports that hypothesis. Would make sense to test not just for load but load on all important circuits.

-CH-

Malc
March 26th, 2017, 04:28 AM
Out of interest what's the wattage of that PSU ?, The original 5160 PSU was 130W, When i first looked at your Pic i thought 157W PSU as fitted in the XT 286's, But on second look i can see differences to the outside housing.

clh333
March 26th, 2017, 09:12 AM
Out of interest what's the wattage of that PSU ?, The original 5160 PSU was 130W, When i first looked at your Pic i thought 157W PSU as fitted in the XT 286's, But on second look i can see differences to the outside housing.

130 Watt.

martinw
March 27th, 2017, 12:29 AM
Far more common than you'd otherwise think. One of the problems with the conventional PC design is that it relies on a negative case pressure, which allows all sorts of junk to be sucked in through the floppy slots.

The better approach would have been a positive case pressure with a filter on the PSU fan intake. I used to do this, particularly with big tower systems that sat on the floor. The floppy drives stayed clean, as did the innards of the machine in general.

That's very interesting Chuck, I've just replaced the fan in my 5160 and it's providing a positive case pressure so it must be the wrong way around, thank you for that. I realised as I have a tape backup unit in mine and air is blowing out of it and the PC smells a bit more than it did before :)

clh333
March 27th, 2017, 07:51 AM
You might also keep in mind that if that old hard drive was originally parked, it won't be after a spin test (if it's still operational, that is). That means the heads are likely resting on track zero now, and damage to track zero is usually the end of an old drive. Try to not subject it to shock. I LOVE those old, giant, fragile hard drives, and eBay sellers are forever connecting them to a power supply and they will announce, "spins up, no further testing done".

Have been running the XT a bit and notice there is a little bearing noise coming from the HD, AFAICT. Sometimes better, sometimes worse, sometimes there, sometimes not. Is this a maintenance issue or a harbinger of doom?

-CH-

Stone
March 27th, 2017, 08:09 AM
Is this a maintenance issue or a harbinger of doom?Probably neither. I've had drives with bearing noise run for years and years with no problems. More often they die of media failure or rot than mechanical issues.

lyonadmiral
March 27th, 2017, 09:15 AM
...with a filter on the PSU fan intake. I used to do this....

What would you use?

lyonadmiral
March 27th, 2017, 09:17 AM
You might also keep in mind that if that old hard drive was originally parked, it won't be after a spin test (if it's still operational, that is). That means the heads are likely resting on track zero now, and damage to track zero is usually the end of an old drive. Try to not subject it to shock. I LOVE those old, giant, fragile hard drives, and eBay sellers are forever connecting them to a power supply and they will announce, "spins up, no further testing done".

That is why I am always very afraid to power up any MFM/DLL drive without having it connected in someway I can park the heads... :/

lyonadmiral
March 27th, 2017, 09:20 AM
Have been running the XT a bit and notice there is a little bearing noise coming from the HD, AFAICT. Sometimes better, sometimes worse, sometimes there, sometimes not. Is this a maintenance issue or a harbinger of doom?

-CH-

I have a 5161 with both that WD equivalent black drive you have along with the brass golden Seagate. Both sound like World War 3 is going on inside the case but every few months I run them through Advanced Diagnostics and come back with flying colors every time. Go Big Blue.

clh333
March 27th, 2017, 11:47 AM
"Expect the worst," they say, "you'll never be disappointed." If the Astrodome bites the dust we'll deal with it later. Until then, no quarter.

Speaking of drives: My newer XT has the IBM FDC which does not support the 1.2 or 1.4 floppy. I saw on the MZD site the Multi Unique FDC, used as an illustration of one that would support higher density. Couldn't find anything like it on eBay; lots of controllers but few 8-bit and fewer still with FDC. None that I could readily associate with DSHD.

I may have to wait for a while to find one but could someone suggest a few that I might be looking for? I didn't see anything from Adaptec or Western Digital, but again my knowledge of the field is limited.

Thank you for your replies.

-CH-

Malc
March 27th, 2017, 12:03 PM
A 16-bit floppy/IDE controller will work in an XT, Disable the IDE and use the floppy with a HD floppy bios, But you'll need somewhere to home the bios.

clh333
March 28th, 2017, 10:13 AM
Replaced the Intel 8088 with NEC v20. Also added Intel 8087. Added SixPack Plus memory card to bring system memory to 640K total, 512 on motherboard. Have cables and connectors for parallel and game ports but haven't tested any ports yet...

Verified that Drive A and Drive B are operational. I haven't run any diagnostics yet so I am guessing they are both DSDD, i.e. 360 and 720K. The 3.5 floppy has an unusual interface: It is a card plugged into the back of a conventional 3.5 drive that adapts the pins to card-edge. The IBM cable therefore is a dual card-edge connector with the usual twist. There is also a 3-pin jumper labeled A/B on this interface card, I guess to allow configuration (in case the cable does not have a twist?) Obtained a SIIG 16-bit FDC capable of driving 1.2 ands 1.4 drives; hasn't arrived yet.

The lower density drives are something of a problem. Three problems, actually; I have spare 360K 5.25 drives (not all of them work but that's another issue) but I don't have another 720K IBM-compatible drive, and I don't know when I will see one. 720s were supplanted pretty quickly by 1.4s and I never saved one from any of my old stuff. When working with the old Macintosh I recently acquired I ran across the same problem: The "writer" of the disk is a 1.4 and while the disk itself may be a 720k and the "reader" is a 720K, and sometimes the reader may not be able to read what the writer writes. I found that it helps to format the 720K disk on the 720K drive first, then let the 1.4 write to it.

That brings me to the second problem: The IBM 20Mb HD ("the Astrodome") has command.com (plus the hidden files) on it; just enough to boot, but no utilities. No FORMAT command, in other words. I really need to start bringing over some software utilities, and eventually some applications, but I don't know if I should format the drive first, and if so whether to go as far as a low-level format, before bringing over more software.

Finally, although there are a lot of applications available to me, many or most are on higher density disks. I'll either have to split these distributions and reassemble them on the HD or find a way to transfer them by serial cable.

If anyone has any ideas I'd love to hear them.

Thanks,

-CH-

Chuck(G)
March 28th, 2017, 10:29 AM
The 3.5 floppy has an unusual interface:

Not unusual at all for the time. When 3.5" drives were offered by OEMs for the 5160 and 5170, they included an adapter bracket with both power and signal adapters, sometimes (e.g. Sony) on the same card.

Any decent 1.44M drive should be able to format and write 720K media. I do it all the time. In fact, I find that some 1.44M drives are better at it than native 720K ones.

Do a high-level format of the hard drive (i.e. FORMAT /S) first. If it goes through with no bad sectors detected, leave things alone. That is, don't try to fix what's not broken.

Stone
March 28th, 2017, 10:35 AM
Three problems, actually; I have spare 360K 5.25 drives (not all of them work but that's another issue) but I don't have another 720K IBM-compatible drive, and I don't know when I will see one. 720s were supplanted pretty quickly by 1.4s and I never saved one from any of my old stuff.IIRC, you can use a 1.44 MB drive as a 720 K drive quite easily. Maybe as easily as just plugging it in normally. :-)



Finally, although there are a lot of applications available to me, many or most are on higher density disks. I'll either have to split these distributions and reassemble them on the HD or find a way to transfer them by serial cable.Or parallel cable which is way faster. But if not, a serial cable should suffice quite nicely.

Or, set up a parallel port network (which also requires a parallel transfer cable). I like to use UNET-II which is a low overhead parallel port network.

clh333
March 29th, 2017, 11:58 AM
Dusted off the tweener and attached a 360K FD, with which to make a copy of the IBM Advanced Diagnostics disk. Thanks again to MinusZeroDegrees for making the tools available.

Ran the series of tests and everything passed (except serial port, which is on the AST card and has nothing attached to it, and the parallel port which is on the display adapter and also is bare). Relieved to know that procs, memory and keyboard checked out. Ran all HD tests except for writing on sector 305. Surface scan did not report any errors but maybe I need to check a log to find out? Anyway I did not modify anything.

I'm still considering a low-level format. MZD states that the IBM controller defaults to a 3-to-1 interleave which it also states is too conservative and recommends a 2-to-1. I didn't see an option for divining the interleave on this drive. Did I miss something (as usual)? It just seems to make sense to be sure about correct interleave and verified surface before loading the drive.

Side note: It's been years since I used DISKCOPY; I had forgotten it existed. But I used it to create the ADD disk from an .IMG file, and I'm wondering if this is the prescribed way of writing all of the disk images that are on WINWORLD and other sites. If so I really have missed the boat.

Thanks again for your input,

-CH-

Stone
March 29th, 2017, 12:43 PM
I'm still considering a low-level format. MZD states that the IBM controller defaults to a 3-to-1 interleave which it also states is too conservative and recommends a 2-to-1. I didn't see an option for divining the interleave on this drive. Did I miss something (as usual)? It just seems to make sense to be sure about correct interleave and verified surface before loading the drive.I hope you meant defining, not divining above. :-)

You can use SpinRite. It will do a non-destructive LLF and set the interleave to one of your choice at the same time.



Side note: It's been years since I used DISKCOPY; I had forgotten it existed. But I used it to create the ADD disk from an .IMG file, and I'm wondering if this is the prescribed way of writing all of the disk images that are on WINWORLD and other sites. If so I really have missed the boat.If you are referring to the DISKCOPY by DR I don't think it can handle all IMG files; maybe only the one's it's created. But it certainly can't hurt to try it.

clh333
March 29th, 2017, 02:48 PM
Divine (v): To discover or perceive.

Diskcopy was in a ZIP file that I obtained at MZD from a link on this page: http://www.minuszerodegrees.net/ibm_xebec/ibm_xebec_llf_add.htm. I don't know if that's the same utility ("external command") I used to use with DR DOS. Ran it from the tweener C: drive: I gave it a source and a target; it formatted the disk and then transferred the image as a one-sided, bootable, 5.2-inch disk. File size was maybe 170K. Thanks to whoever wrote that one.

-CH-

Chuck(G)
March 29th, 2017, 02:53 PM
...and there's always RAWRITE for DOS. You can find it all over the place. The last time I saw it, it was on a Red Hat Linux CD under the "dosutils" directory.

Stone
March 29th, 2017, 03:34 PM
Diskcopy was in a ZIP file that I obtained at MZD from a link on this page: http://www.minuszerodegrees.net/ibm_xebec/ibm_xebec_llf_add.htm. I don't know if that's the same utility ("external command") I used to use with DR DOS. Ran it from the tweener C: drive: I gave it a source and a target; it formatted the disk and then transferred the image as a one-sided, bootable, 5.2-inch disk. File size was maybe 170K. Thanks to whoever wrote that one.Yes, that's the DR program.

clh333
April 1st, 2017, 02:21 AM
Obtained SpinRite from WinWorld and ran its tests on the HD. After a surface quick scan found no problems I ran the analysis: it found a 6:1 interleave and recommended a 5:1. I opted for the deep pattern analysis and a reformat. Twelve hours later it finished and I shut the machine down for the night.

I was sure something would fail this stress test but started right up this morning without complaint. Everything fails sometime but I guess I can quit worrying about this for now.

Next step is to get the 5154 functioning. And send the people at Gibson Research some money for a license! Still in business, they are up to version 6. This is a product worth supporting. Thanks for recommending it.

-CH-

Stone
April 1st, 2017, 04:57 AM
Next step is to get the 5154 functioning. And send the people at Gibson Research some money for a license! Still in business, they are up to version 6. This is a product worth supporting. Thanks for recommending it. I don't know anything about version 6 as I only use version 2 because it's for MFM/RLL drives.

clh333
April 2nd, 2017, 03:17 AM
Retrieved the 5154 yesterday; long story, longer trip. Now about to embark on diagnosis and implementation.

First step is to fit and test an EGA card. Obtained the Paradise Basic EGA card illustrated below from an eBay reseller. It came with the switches and jumpers set as illustrated. According to the reference I found here http://stason.org/TULARC/pc/graphics-cards/U-Z/WESTERN-DIGITAL-CORPORATION-EGA-Monochrome-BASIC-E.html these are the correct settings for an EGA display. I am assuming here that for the switches UP is OFF and DOWN is ON.

Can anyone confirm any or all of this?

According to what I have read you do not want to connect a monochrome monitor to an EGA card unless it is set for a monochrome monitor - which would be a different setting (in the table under MONOCHROME). With two unknown pieces of hardware to test, I would want to connect my known-good 5151 to confirm the EGA adapter is operational before attaching the 5154 to test it.

Is there anything else I should consider or beware of?

As always, thank you for your suggestions.

-CH-

Attachments:

37505 37506

Stone
April 2nd, 2017, 04:27 AM
Just set it for EGA and plug it in.

If you've got a CGA monitor you can try that one first to reassure yourself.

clh333
April 2nd, 2017, 05:53 AM
Paradise card when operated in monochrome mode and attached to 5151 display worked properly after I remembered to set motherboard switches 5 and 6 properly.

Removed the card and reconfigured as illustrated in previous post, attached 5154 monitor and powered up. Power on LED lights up but no image displayed on the monitor as the computer posts or boots.

With the computer shut down but the monitor powered up, and with the brightness and contrast turned up as far as they will go, the image below is displayed on the monitor. I had to adjust the vertical 1 and vertical 2 knobs on the rear of the console to get this image; previously it was stretched farther vertically. One obvious problem is the irregularity of the horizontal sweep. A possibly related second problem is the fact that as I adjusted vertical 1 through its rotation the display jumped markedly: variations in brightness and regularity of sweep were noticeable.

I don't know anyone in my area with the expertise to repair this unit so I am thinking I will try it myself. As always, I'll welcome any advice from other VCF members.

Thanks for your replies.

-CH-

Edit, correction:

I went back and restarted the machine, this time letting the CRT warm up first. I had forgotten that these things aren't "instant on". After a warmup I turned down the brightness and contrast until the no-signal pattern was just invisible, and restarted the machine. I was able to make out the characters of the post and boot sequence on the display. They were very faint but discernible. I take this as a hopeful sign, although I still don't know what the problems are or how to fix them.

Attachment:

37511

Stone
April 2nd, 2017, 06:06 AM
You still can't be sure whether it's the card or the monitor although I'd bet on the monitor.

It is entirely possible, however, for a multifunction card to be defective in one mode and work perfectly in another.

OTOH, I have one Multisync monitor that works correctly in TTL mode but has no GREEN in Analog mode. :-)

Stone
April 2nd, 2017, 06:10 AM
Edit, correction:

I went back and restarted the machine, this time letting the CRT warm up first. I had forgotten that these things aren't "instant on". After a warmup I turned down the brightness and contrast until the no-signal pattern was just invisible, and restarted the machine. I was able to make out the characters of the post and boot sequence on the display. They were very faint but discernible. I take this as a hopeful sign, although I still don't know what the problems are or how to fix them.There are brightness and focus trims either on the back of the monitor or possibly inside. Normally you need a long reach plastic tool to get to it if it's on the outside of the case. If it's on the inside, however, it's a normal screw adjuster.

clh333
April 3rd, 2017, 02:19 PM
Set aside the 5154 for now in order to work on other issues. Spent time loading software, and one of the items I wanted to load was the AST clock's utility. I had added and configured the other ports on the card and had tested them with the IBM advanced diagnostics. The only test for the card however was to invoke it and see the results. The results were disappointing: I got "34/00/50" for a date and "99:99:99:99" for the time, even after loading the "correct" time and date with the setclock utility.

I checked the manual to be sure the jumper and switch configuration was right; I checked the voltage in the battery I had just replaced - it was 2.75V so I put in a fresh 2032 which measured 3.4V. Made no difference. I could set the time and date manually but the settings would not persist through a reboot.

Exasperated, I retrieved another AST Six Pack card; this one was older and untested. I put a fresh battery in that and configured it according to the manual, and plugged it in. After booting the machine ran the astclock utility and the time displayed was wrong, but sensible; something - something - '80. I re-ran the setclock utility, setting the correct date and time, and this time the date and time persisted through a warm boot and a cold boot.

However, the astclock utility reported the time as 04/03/99, even though the OS knew the date to be 04/03/2017. I tested this by creating a directory and observing the date: sure enough, the date was 2017. My conclusion is that the "Y2K" bug is with the AST software, not with the clock chip.

Looking at the (screwy) board I notice that it is a later version than the one in MZD's documentation. The battery tray is in a different location and there are some small electrolytics around the tray. One of them is leaning and maybe bulging as well. On a hunch I will try replacing them and see if the clock will remember how to tell time.

If I am unsuccessful in this repair I will have about 54 4164 RAM chips, 36 of which I just purchased to populate the board. The XT has Bank 0 and Bank 1 populated but 2 and 3 are bare. These chips could be repurposed for that, could they not?

-CH-

Pictures attached:

37545 37546 37547 37548 37549

Stone
April 3rd, 2017, 03:09 PM
Do you know that a 256K XT board can be modded with just one soldered jumper to take 640K? This allows for 256K chips in banks zero and one. It's quite easy.

clh333
April 4th, 2017, 02:23 AM
Do you know that a 256K XT board can be modded with just one soldered jumper to take 640K? This allows for 256K chips in banks zero and one. It's quite easy.

?

This XT has a 256-640 board. Banks 0 and 1 have 41256 chips in them, totaling 512k on board. I thought the remaining two banks required 4164s: 2x64=128; 512+128=640. So I thought I might transfer the 4164 that I bought for the AST card onto the motherboard.

-CH-

37569

ibmapc
April 4th, 2017, 05:46 PM
You are correct. That memory configuration will get you to 640k as long as you set the dip switches correctly.

bhtooefr
April 5th, 2017, 02:00 AM
Motherboard RAM should be somewhat faster than expansion card RAM but on an XT... does it really pay to quibble??? :-)

That's true for some late AT clones (especially 16, 20, or 25 MHz ones, which run a /2 or /3 on the ISA bus typically), but for a PC or an XT, that's not true - memory accesses take 4 cycles whether it's motherboard or expansion RAM.

And, in an AT, it's actually possible for expansion RAM to be faster than motherboard RAM, I believe - motherboard RAM takes 3 cycles, whereas 16-bit expansion RAM boards asserting the 0WS signal takes 2 cycles. (8-bit expansion RAM, OTOH, will of course be slower, and for even an 8-bit transfer, will take 4 cycles if it asserts 0WS, 6 cycles if it doesn't). The XT 286, as well as many clones, however, have zero wait state motherboard RAM, so the slots at least can't be faster than the motherboard.

I believe the PS/2 Model 50 is another case where expansion RAM can be faster than motherboard RAM, but I'd need that verified. The Model 50Z went to zero wait state motherboard RAM, though.

clh333
April 5th, 2017, 03:32 AM
Thanks to all for their replies.

I removed the suspect caps and found them within spec so I am going to consider the AST board supernumary and remove the memory. I bought the board from an eBay e-waste seller ("untested") but knowing how these things go I bought a second one soon after. The second one is now in the XT; it was in working condition and fully populated with memory, included a parallel port cable and cost less, too. There's a lesson in there somewhere...

Now to work on the ports: printer, serial and SCSI.

-CH-

Stone
April 5th, 2017, 04:18 AM
I have an extra NOS 256K board that can give DOS more than 640K:

http://www.vcfed.org/forum/showthread.php?56857-256K-RAM-Expansion-card

clh333
April 7th, 2017, 10:15 AM
The parallel and serial ports appear to be in good order: the parallel port writes to a ProPrinter and the serial port operates a V32bis external modem.

However, the SCSI port is not yet configured and I am not able to find the reason why. The adapter is a IOMEGA 8-bit adapter, PC800, with a 25-pin connector on the outside and a 50-pin connector on the inside. I installed the OAD DOS drivers and utilities from a distribution disk I got with the adapter back in the early '90s. As the documentation suggested I modified the CONFIG.SYS and AUTOEXEC.BAT files with the appropriate entries. The first entry runs a configuration executable and rebuilds the driver on the fly. There is a lot of disk thrashing going on.

I attached a ZIP 100 Plus and cable to the external connector and powered everything up. MS DOS 6.22 loaded but the executable, which got as far as examining the SCSI chain and found the adapter at ID 0, could not find any other devices on the chain. The ZIP drive is either 4 or 5 depending on a switch setting.

The ZIP 100 Plus is a dual-port device, usable either from a parallel port or SCSI (but the user must take care to mate parallel with parallel or SCSI with SCSI, otherwise bad things can happen.) The devices are marked "autodetect" on the cable. I thought there might be a problem with this technology, so I got an older ZIP 100 that was SCSI only and for good measure swapped the cable, too. But I got the same results on booting: no device found.

One possibility occurs to me that the ZIP drive, which appeared on the scene in the early 90's as I recall, might not have been around when this driver utility was written. Has anyone had success using a ZIP drive with an IBM XT, and if so, what did you do differently?

Thanks for your replies.

-CH-

Screen shot:

37610

modem7
April 7th, 2017, 01:55 PM
Has anyone had success using a ZIP drive with an IBM XT, and if so, what did you do differently?
Parallel port Zip100 drive (Z100P and Z100P2), with palmZIP driver.
See [here (http://minuszerodegrees.net/transfer/zip/zip_drive.htm)].