• Please review our updated Terms and Rules here

On again, off again

The machine is relatively stable at this point; it boots, it keeps the CMOS settings, there are no parity errors and the machine passes CheckIt diagnostics for memory, video, keyboard, IO ports, etc.

The video card is an IBM EGA adapter with a memory expansion card, driving a 5151 monochrome display. The display driver loads from the config.sys file, and I don't know any more than what it reports on boot. (see attached pic).

Video is stable and displays all attributes except for high-intensity video, which seems to be missing. I visited MZD and looked at all the documentation: I checked the BIOS date (4/30/89), the motherboard DIP switches, the EGA card dip switches, the EGA card jumper pin positions; nothing was inappropriate for an EGA card with a monochrome display.

I'm thinking I must be mis-configuring something, but I don't know what.

?????

-CH-
 
What I was missing was the proper adjustment of the monitor contrast and brightness controls :crazy:

Easy fix, though. BTW thanks to Daniel Marks and Peter Summers for MFALLSCR.COM.

-CH-
 

Attachments

  • DSCN0998.JPG
    DSCN0998.JPG
    249.6 KB · Views: 1
  • DSCN1000.JPG
    DSCN1000.JPG
    227.3 KB · Views: 1
  • DSCN1005.JPG
    DSCN1005.JPG
    211.9 KB · Views: 1
  • DSCN1006.JPG
    DSCN1006.JPG
    215.9 KB · Views: 1
Last edited:
The video card is an IBM EGA adapter with a memory expansion card, driving a 5151 monochrome display.
Video is stable and displays all attributes except for high-intensity video, which seems to be missing. I visited MZD and looked at all the documentation: I checked the BIOS date (4/30/89), the motherboard DIP switches, the EGA card dip switches, the EGA card jumper pin positions; nothing was inappropriate for an EGA card with a monochrome display.

I'm thinking I must be mis-configuring something, but I don't know what.
Possibly a result of:

From a magazine article on the IBM EGA at [here], "Most programs that run on the Monochrome Adapter should run on the EGA/Monochrome Display combination with no problems. However, some of the attributes (high intensity, underlining, and reverse video) are different, so programs that use the same attributes for monochrome and color may not look the same."
 
Thank you for your reply.

On other machines I always use a screensaver -usually based on a routine I found in a book, one that involves writing a few lines of code - but since I had none on this machine I had been turning down the brightness and contrast when I was away from the machine for a length of time. When I came back I would turn them back up, but I didn't realize that if the brightness is turned up too high it obviates the mediation of the contrast setting. A lower brightness setting and a higher contrast setting brought out the missing attributes.

I really like the velvety green of the 5151's high-persistence phosphor but as you may surmise I'm not familiar with the 5151's personality. Another lesson learned.

-CH-
 
Communication with IBM ProPrinter 4201 confirmed.

The Everex EV-125 card, substituted for the RAMPAGE 286, has added conventional and extended memory, as well as two serial ports, bringing the total to four. On its board it also has ports for a parallel and a game interface, but there is no (extension) cable and apparently these are not enabled. Nor do I know where or how to enable them.

I have not found an EMM.SYS that will enable the use of any of the extended memory.

-CH-
 
Next to address are the serial ports. Anecdotal evidence: the MS serial mouse is recognized on COM2, The Zoom V32bis external modem is not able to be initialized on any COM port.

The BIOS, MSD and CheckIt all agree that there are four COM ports, at different addresses and with the proper IRQs. CheckIt has the most complete suite of tests and will perform with or without a loopback adapter. The ports all pass without the loopback adapter. With a loopback it's a different story.

First, it matters (to CheckIt) what is being looped. My first try used a jumper between D-sub sockets 2 and 3, RXD and TXD. The CheckIt software noted partial success but several failures. I tried another, more extensive loopback, bridging 2:3, 1:4:6 and 7:8, tying DCD and DSR to ground and RTS to CTS. This eliminated all the failures except for "Test modem status" (expected F0h, received B0h). I'm stumped as to what signal / line that test is examining. DSR?

I know TouchStone Software, the authors of V3 which I am using, have since been bought out by Phoenix and subsequently disbanded operations. They used to offer a loopback tester, as well as the spiral disk used for testing floppies. (ASKY, who made the spiral disk, isn't around any more, either.) I am without the documentation for the program and don't know how their loopback was wired, so I'm hoping someone can fill in some missing information.

Thanks,

-CH-
 
The BIOS, MSD and CheckIt all agree that there are four COM ports, at different addresses and with the proper IRQs. CheckIt has the most complete suite of tests and will perform with or without a loopback adapter. The ports all pass without the loopback adapter. With a loopback it's a different story.
Yes, it will perform but is is not worthwhile without... as you have already discovered. IOW, you need the loopback or you are just spinning your wheels. Just because a program doesn't refuse to run doesn't mean it's going to provide you with anything useful.


I am without the documentation for the program and don't know how their loopback was wired, so I'm hoping someone can fill in some missing information.
This has what you want:

https://www.lammertbies.nl/comm/cable/rs-232
 
I happen to have the drivers for the EV-158A (RAM 10000) board, that includes EMS.SYS. Do you think that would help?

Worth a shot, thanks. I would think the algorithm would be the same just because the mfr would probably want to re-use most of the architecture. What info is available on the Internet leaves a lot of questions unanswered, however. Without the OEM docs it's a little like trying to do a jigsaw puzzle in the dark.

Let's run it up the flagpole and see who salutes.

-CH-
 
Next to address are the serial ports. Anecdotal evidence: the MS serial mouse is recognized on COM2, The Zoom V32bis external modem is not able to be initialized on any COM port.

The BIOS, MSD and CheckIt all agree that there are four COM ports, at different addresses and with the proper IRQs. CheckIt has the most complete suite of tests and will perform with or without a loopback adapter. The ports all pass without the loopback adapter. With a loopback it's a different story.

First, it matters (to CheckIt) what is being looped. My first try used a jumper between D-sub sockets 2 and 3, RXD and TXD. The CheckIt software noted partial success but several failures. I tried another, more extensive loopback, bridging 2:3, 1:4:6 and 7:8, tying DCD and DSR to ground and RTS to CTS. This eliminated all the failures except for "Test modem status" (expected F0h, received B0h). I'm stumped as to what signal / line that test is examining. DSR?

I know TouchStone Software, the authors of V3 which I am using, have since been bought out by Phoenix and subsequently disbanded operations. They used to offer a loopback tester, as well as the spiral disk used for testing floppies. (ASKY, who made the spiral disk, isn't around any more, either.) I am without the documentation for the program and don't know how their loopback was wired, so I'm hoping someone can fill in some missing information.

Thanks,

-CH-

The problem was neither with the port hardware or software, nor with the modem, but with the cable which was a null-modem configuration but unmarked as such. I switched to a straight-through ribbon cable and the problems disappeared.

So you might say the problem was with the wetware...

-CH-
 
I happen to have the drivers for the EV-158A (RAM 10000) board, that includes EMS.SYS. Do you think that would help?

The EV-158A drivers work with the EV-125 board. Thanks for the suggestion.

Thus far the following issues have been resolved:

-CMOS configuration after dead backup battery replacement. Configured AMI BIOS to have no HD, allowing XT-IDE board to control HD
-Sudden halt to system after a minute or two of operation. Attributed to failing tantalum capacitors on an IBM IO board (removed, repaired, not reinstalled)
-Display lacking brightness attributes. After adjusting brightness and contrast controls high-level brightness is distinguishable
-Serial port conflict. SIIG multi-I/O and Everex cards were both configured the same; with no info on EV card the SIIG was changed to COM3 and COM4, resolving the conflict
-Serial port loopback test failed. Found that CheckIt utility requires extra jumpers on DB9 in order to complete tests successfully.
-Modem installation problems. Traced to an unmarked null-modem cable, replaced
-Parallel port installation. No problems driving IBM Proprinter, except the Proprinter ribbons are all dry. Constructed alternate ribbon inker covered under another thread
-RAMPAGE memory card parity errors. Unable to resolve, replaced card with Everex card. Plans for Arduino RAM tester investigated
-Everex EV-125 card mystery settings. Referencing Stason documentation and some experimentation allowed 2MB Everex RAM + IO card to install 128K conventional memory plus expanded / extended memory. Drivers as noted above

A very hearty thank-you to all who contributed diagnostic advice and suggestions. Next I plan to integrate a CalComp drawing tablet and a 1522 SCSI card. The E-series plotter comes last.

-CH-
 
On to the topic of SCSI: Added an Adaptec 1522A (has floppy controller & port) for the purpose of communicating with an external Iomega ZIP drive. I configured the 1522A according to the factory defaults except that I disabled the SCSI boot, BIOS, and floppy controller. Adaptec EZSCSI installation recognized the 1522 and installed ASPI2DOS.SYS, which announced the presence of the adapter on the next reboot. I subsequently attached the ZIP disk, and realized I should have attached it before the driver installation. I reinstalled and this time the ZIP was recognized as well and ASPIDISK.SYS.

I could not locate the drive through DOS, however; there was no letter assigned to it. Since then I have made dozens of iterations varying everything I can think of: I swapped ZIP drives, re-enabled the BIOS (but not the BIOS boot or the floppy controller), stepped through the CONFIG.SYS and AUTOEXEC.BAT on bootup, varying what was enabled; I have run Iomega GUEST.EXE instead of the Adaptec device drivers. Nothing reports an error; everything says it has concluded successfully, except for GUEST.EXE, which seems to hang the machine.

I don't know if I have a hardware or a software problem at this point, but I thought I would ask if there is any possibility of a conflict with the XT-IDE BIOS.

Thanks for your replies.

-CH-
 
Perhaps there's a version problem. I run an XT clone with a Trantor MA13B and external ZIP. I use MA13B.SYS and TSCSI.SYS; works fine under DOS 6.22.
 
Perhaps there's a version problem. I run an XT clone with a Trantor MA13B and external ZIP. I use MA13B.SYS and TSCSI.SYS; works fine under DOS 6.22.

I have run a similar setup on a '386 machine, using most of the same hardware. I was beginning to suspect the adapter, but - mirabile dictu - I resolved the issue; a configuration problem. The answer came from one of the many sources I investigated online. It turned out that adding "LASTDRIVE=J" to the top of the CONFIG.SYS file resolved the conflict. I don't know how many drives DOS 6.22 is prepared to address by default, but apparently fewer than this.

A: and B: are floppies, 16 GB Bigfoot is partitioned into three 2GB partitions which are C:, D: and E:. The Everex RAM drive was being assigned F:, but I had disabled that in my attempt to figure out what was going wrong. ASPIDISK.SYS was announcing it had installed the ZIP as F: but when I would try to access it I would get a "general fault" error.

Next I tried running IOMEGA's GUEST.EXE from a floppy (B:). Once again GUEST hung the system, until I found this nugget (below) and edited the GUEST.INI file to reference C:\SCSI\ASPI2DOS.SYS, the driver the Adaptec install wanted to use for the 1522A. You can see that the default INI does not have a value for that, which would explain the system hang. Too bad there was no error generated, though.

"GUEST Locks up the System
If GUEST locks up the system when it is loading from either the
AUTOEXEC.BAT file or the MS-DOS prompt, GUEST might be scanning for
different SCSI adapters from the one you are using. To solve the
problem, edit the GUEST.INI file and insert a semicolon (;) as the
first character in the lines containing ASPI managers that are not
being used. (This makes the software treat these lines as remarks.)
The GUEST.EXE program uses the GUEST.INI file to load ASPI managers.
Below is the default GUEST.INI file:
[----Scan for existing ASPI managers---]
SCAN=ON
[----Load ASPI managers----]
ASPI=ASPIPPA3.SYS SCAN /INFO SL360=NO SMC=NO
ASPI=ASPI1616.SYS SCAN /INFO
ASPI=ASPIPC16.SYS SCAN /INFO
ASPI=ASPIPC8.SYS SCAN /INFO
ASPI=ASPIPC2.SYS SCAN /INFO
ASPI=ASPIPC4.SYS SCAN /INFO"

With that modification GUEST ran, albeit with everything else in CONFIG.SYS disabled, and I had a drive labeled F:. I was unable to read the disk I inserted but got an error stating "invalid media", not general fault. I switched to a disk I had used and knew was good, and got a directory listing. So far so good.

Next I tried allowing CONFIG.SYS to identify the drive. Once again the ZIP disk was recognized and assigned label F: (the RAM drive still disabled). Finally I un-commented the RAM drive reference and it installed as G:.

Nothin' to it.

Final version of CONFIG.SYS:

LASTDRIVE=J
DEVICE=C:\SCSI\ASPI2DOS.SYS /D
DEVICE=C:\SCSI\ASPIDISK.SYS /D
DEVICE=C:\DOS\HIMEM.SYS /INT15=256 /VERBOSE
DOS=HIGH
DEVICE=C:\DOS\SETVER.EXE
FILES=30
DEVICE=C:\DOS\EGA.SYS
DEVICE=C:\EVRX\EDISK.SYS /E

MEM /F reports:

Free Conventional Memory:

Segment Total
------- -----------------
00C67 80 (0K)
00F40 96 (0K)
00F46 88,992 (87K)
02500 502,784 (491K)

Total Free: 591,952 (578K)

No upper memory available

The Everex 125 card was able to backfill conventional memory but I have not figured out how to backfill upper memory. When it creates the RAM drive extended memory is used. I haven't been able to LOADHIGH, in other words.

-CH-
 
Not sure if I need to load the Adaptec BIOS; have to experiment and see if it is necessary for operation of the ZIP disk only. Slows down boot and occupies 16K of conventional memory.

Meanwhile: I've been trying everything I could think of to configure COM3 and COM4, which are on the SIIG multi-IO controller. Using CkeckIt and a loopback tester these ports had tested OK last week, but I never have been able to drive a MS serial mouse or modem off either one. Finally after eliminating every other possibility I went back and retested.

!5$@!^$

Time for Plan B.

-CH-

Problem.jpg
 
Plan D, actually; this is the fourth piece of equipment deployed in an attempt to get serial communications straightened out. And it's not completely straight yet.

My first thought was to replace the SIIG SI-1132 controller with a DTC 2280 that I had acquired recently, but that proved to be a dead end. DTC must have made at least three of these models and probably more than one board revision. None of the documentation I could find matched the board I had in front of me; even the silkscreen on the back didn't agree with the front of the board.

Next idea was to disable the SIIG COM ports but still use it to control the disks while another card would provide two serial ports. Easier said than done; there are a million variations on that 8-bit serial board, most not identifiable and undocumented, many hard-wired COM1 and COM2 as well.

In the box of spares I found an Everex 170A (no clock) "Magic IO" board. It was obtained second-hand and had only one UART and one serial channel, as well as one parallel, on board. It took me a while to find the documentation but eventually I remembered I had a printed copy of my own, carefully hidden inside an IBM manual.

There were three empty sockets on the board, but I had a spare NS 16550 and a 1488 and 1489 sitting around, left over from an unfortunate ESD I once inflicted on an Amiga, so I cobbled together the second channel. On the SIIG controller I disabled the two misbehaving serial ports, leaving only the parallel and game ports enabled besides the disks. The SIIG parallel port I configured as EPP/ECP at 278H and IRQ5, because I configured the EV-170 as a vanilla LPT.

In a perfect world I would have plugged everything back together and all would be well. Not with this machine: Either it is haunted or it really craves attention. Once again neither the mouse nor the modem could be driven from COM3 or COM4, my new additions.

MSD showed all four COM ports, three with 8250s and one 16550AF. Addresses were correct and 2 & 4 shared IRQ3 while 1 & 3 shared IRQ4. But when I ran CheckIt again, testing 3 and 4 with the special loopback adapter, COM4 - the 16550 - failed 5 of the tests. Rats! I double-checked with Everex's MagicIO software utility; it merely confirmed that Serial Port 2 (COM4) was "bad".

But if MSD could recognize the 16550 while the port could not communicate with the outside world, that suggested to me that the 16550 might not be at fault, but the 1488 and 89 line drivers, which I knew from experience are very sensitive to ESD. I have a little pass / fail IC tester so I thought it would be worth a shot to see if my hunch was true.

Before I opened up the computer yet again I retrieved my tester and the extra ICs (88/89) to see if the tester would recognize them. It did not. So I have three working serial ports, COM1, COM2 and COM3, and two parallel ports, LPT1 and LPT2, and no clear path what to do next.

If someone knows of a way to test the 1488 and 1489 out of circuit, please let me know.

In the meantime, I found that the reason the serial mouse was not being recognized on port 3 or 4 was that, apart from port 4 being defective, Microsoft's MOUSE.COM was only checking COM1 and COM2. I switched to CUTEMOUSE.EXE and it immediately found the mouse on COM3. It also observed that I had an EGA adapter and suggested another version if I wanted "EGA RIL". I don't know what RIL stands for, either.

Lastly, I disabled the Adaptec 1522A BIOS load to see if it affected the loading of the SCSI drivers or the use of the ZIP drive. It did not, so I'm leaving it disabled.

-CH-
 
In the meantime, I found that the reason the serial mouse was not being recognized on port 3 or 4 was that, apart from port 4 being defective, Microsoft's MOUSE.COM was only checking COM1 and COM2. I switched to CUTEMOUSE.EXE and it immediately found the mouse on COM3. It also observed that I had an EGA adapter and suggested another version if I wanted "EGA RIL". I don't know what RIL stands for, either.

RIL = Register Interface Library, a feature of EGA.

1488 / 89 line drivers operate on +15VDC & -15VDC; I do not have a power supply capable of generating those voltages.

Thanks to Modem7 for suggesting http://minuszerodegrees.net/software/SERTEST.zip

-CH-
 
Using SERTEST.EXE and the prescribed loopback adapter I ran tests against the Everex EV-170A card. Tests revealed a problem with both serial channels' IO, so I removed the card as well as the SIIG controller and substituted a DTK PTI-227B multi-IO controller with 1xP and 2xS ports. The change reduced the component count, and COM3 tested clean in the new setup, but there were a few glitches on COM4. I guess it's a tough life if you're a serial port.

In the meantime I took the opportunity to install / investigate some of the WinWorld software. It was fun to see GEM Desktop again after 35+ years, and really interesting to finally see DisplayWrite running in its initial configuration. Thanks to the guys at WinWorld for maintaining that valuable resource.

The objective, ultimately, is to install a DOS program capable of receiving input from the CalComp tablet and outputting to the Roland CRX-400AB plotter. To that end I installed AutoCAD R11/DOS on the 5170. It runs but lacks configuration and (once again) I have no documentation.

A +&- 15VDC power supply is in the planning stages; need to acquire a center-tapped transformer with 120VAC input and the appropriate winding. Still don't know if the 148x chips were at fault but I will someday provided I live long enough.

-CH-
 
36VCT transformers are not uncommon, exempli gratia--and give you 18-0-18, which will give you more than enough headroom for a 15V linear regulator (e.g. LM7815).

The trick about commonly-available transformers is to think in terms of 9V multiples. So, 9, 18 and 36 volt units are quite common.
 
Back
Top