Sorry got no clue. Never used the XTIDE "serial" thing.
In some full null modem cables the CD pin was connected to DSR on each side for only 7 wires but on some very old ones the CD was straight to CD for 8 wires IIRC.
If CD pin is not connected at all that could be part of the problem....
Are you sure your model has true 16550 compatible serial port? iirc that is a requirement for using the "serial" part of XTIDE.
Going mostly from memory and a quick check of some pinouts. Probably wrong but I know CD pin has to be connected somehow when using older stuff... 😁
added: is your model 30 an original plain one with just floppies or does it have an HD ?

HI, how are you? Thank you for your answer.

The CD pin is not connected. I am using the cisco RS-232 cable (planar, lightblue one) with DB9 and DB25 female adaptors, which should be a full handshaking null-modem cable, but with the CD pin not connected.

Horun wrote on 2022-09-07, 02:23:

Are you sure your model has true 16550 compatible serial port? iirc that is a requirement for using the "serial" part of XTIDE.
Going mostly from memory and a quick check of some pinouts. Probably wrong but I know CD pin has to be connected somehow when using older stuff... 😁
added: is your model 30 an original plain one with just floppies or does it have an HD ?

The ps/2 model 30 has an NS8250A UART with no FIFO buffer. This one comes with a floppy and a hard drive, but none of them are working.

I am trying to figure out how to boot the ps/2, since I made an assembly program to diagnose some weird MCGA onboard display adapter issue (which is in another post here in vogons), and I didn't figure out how to put the program in memory through ROM BASIC. So, my best shot is to boot into DOS and use debug to diagnose the MCGA issue.

Best Regards,
Ariel

Hi! This brings back some memories.

A Null-Modem is a Null Modem, right?

The 8250 UART was practically broken.
Sure, it could be made working. Software from the 1980s had workarounds built-in. Maybe carefully written assembler routines, don't know.
I remember the utility Ultrafast Filetransfer Operation , which could do 57600 or 115200 Baud on 8250/16450 era hardware over a 3-wire connection.
However, I don't know if XTIDE has such a workaround, too. It's from the 21th century, after all, long after the 16550A had become standard.

Maybe, the 8250 in that PS/2 can be upgraded? Without breaking BIOS compatibility?
The 16550A was pin compatible to the 16450, at least.
Not sure about the 8250, though. 🤷‍♂️

Here's another post of mine.
Sorry, I'm on a mobile device right now. Typing is cumbersome right now.

Re: Can I connect my IBM AT to a modern computer running dosbox?

Here are two virtual drives for DOS that connect over null modem.
One mentions explicit PS/2 compatibility.

Re: 386 , what can I do with it?

Good luck ! 🤞🙂

Edit: 9600 Baud might be a bit too high for a 8250 already!
Even for a 16450 with a functioning 1 Byte FiFo, it's almost borderline.
Please try 4800 Baud, too, if you can. Or 2400 Baud, if the processor can't catch up with it.

Edit: I know it sounds a bit contradictory.
Thing is, if the processor is fast enough, it can catch the incoming data in time, even without the help of a FiFo.
A fast 286 (16MHz up), a 386 or 486 processor might be able to archive better throughput on an 8250 or 16450.
So higher baud rates become possible.
However, on a slow 8086 or 286 (say 6MHz), the problem of missing characters might occur.

Edit: ^That's related to interrupts and polling, too.
Depending on how the system is running, a slow machine might be able to handle things.
A single DOS communications program that's running alone
and has full control over the PC might be able to do 115200 Baud on a Turbo XT.
Especially with the 7-wire connection and handshaking.
However, if multitasking comes into play, things look differently.
Other programs now will try to do claim interrupts, will steal processor time etc.
That's when buffering comes in handy. That's when the 16550A (or AF, C, AFN etc) is needed.

Edit: I've just checked the net if the 8250 could be directly replaced by a 16550A.
That's when Wikipedia came up with this article.

"Replacement of the factory-installed 8250 UART was a common upgrade for owners of IBM PC, XT, and compatible computers when high-speed modems became available.

Above 9600 baud, owners discovered that the serial ports of the computers were not able to handle a continuous flow of data without losing characters.

Exchange of the 8250 (having only a one-byte received data buffer) with a 16550, and occasionally patching or setting system software
to be aware of the FIFO feature of the new chip, improved the reliability and stability of high-speed connections. "

Source: https://en.wikipedia.org/wiki/16550_UART

That being said, I did encounter communications problems in FileMaven, too, when I worked with a 286 at 10 MHz.
At 19200 and up, it often "hang" after a while.
The 1988 machine had an 16450 FiFo on the motherboard, which I replaced by a socket a few years ago.

Found something..

"Several VLSI gate arrays on the system board consolidate the functions of parallel, serial, PS/2 mouse and keyboard, floppy and hard disk, and on-board graphics."

Source : https://en.wikipedia.org/wiki/IBM_PS/2_Model_30

Now that's bad. How comes that IBM installed an outdated 8250 design in a new product in '87 and up?
How could the company be that incompetent?
It blows my mind. That's ridiculous.

Hm. There are three workarounds that come to mind:

a) use a slow baud rate, use full handshaking
b) install a NEC V30; the speed gain might be just enough for 9600; the mod is reversible
c) use an ISA card with a new serial port or mod an old serial card to have an 16550A etc.
Multi-i/o cards for ISA were cheap and common in 486 days.

Edit: Just double checked the specs. Maybe the system has a 16550/A, after all ?
As I said earlier, how could IBM have been so incompetent not to use it in 1987.

"The corrected -A version was released in 1987 by National Semiconductor.[1]
It is frequently used to implement the serial port for IBM PC compatible personal computers,
where it is often connected to an RS-232 interface for modems, serial mice, printers, and similar peripherals.
It was the first serial chip used in the IBM PS/2 line, which were introduced in 1987."

[..]

"The 16550A and newer is pin-compatible with the 16450,
but the Microsoft diagnostics program (MSD) supplied with MS-DOS 6.x,
Windows 9x, Windows Me, and Windows 2000 often report the 16450 chip as an 8250 chip."

https://en.wikipedia.org/wiki/16550_UART

So please use NSSI or Dr. Hardware, each a good diagnostics program.
It will probe in a much more faithful fashion.
Or, SI (System Information) from PC-Tools, maybe.

https://www.navsoft.cz/products.htm
Dr. Hardware Sysinfo (DOS) - Versions Archive

"I am using the cisco RS-232 cable (planar, lightblue one) with DB9 and DB25 female adaptors, which should be a full handshaking null-modem cable, but with the CD pin not connected"

I have found that some cisco serial cables are a different pinouts to normal serial cables. They are pinned to work with there own equipment. Also a null modem cable is different to standard 232, you could try source another cable?

That's a good idea, thanks!
Wrongly wired cables are often the cause of trouble.

Though guys, the pin is named DCD, strictly speaking.
Don't mean to be nitpicking, though. 😉

The term DCD (Data Carrier Detect) is also used by radio modems (TNCs).
Here, it relates to the detection of a valid radio signal if the squelch is open.

See https://mijndingen.nl/true-open-squelch-dcd/

Edit: Searched the web for DCD once more.
This one seems interesting :

"Much like the Ring Indicator signal, on a PC's serial port, changes to the DCD signal state can generate a hardware interrupt
that can be captured by the processor any time the DCD signal changes state, preventing the PC from needing to constantly poll the pin. "

Source: https://en.wikipedia.org/wiki/Data_Carrier_Detect

So that makes sense to use on a piece of software with "XT" in its name (XTIDE).
If the pin is missing, it could be that the communication doesn't work properly.

Hi Guys, how are you?

Thank you for the answers. I double checked and it has a 8250A UART controller. I made a null modem cable with full handshaking and DCD line connected to DSR, with the same results.
The serial drive stuck always in the same place. I tried several images with the same results. All of them stuck always in the same place (different places between them, but always the same for each one). I even tried 2400 bps, with the same results. it stucks in the same place.

I am trying to get a serial port ISA adapter. I will let you know.

Regards,
Ariel

original_meusli wrote on 2022-09-08, 21:00:

I have found that some cisco serial cables are a different pinouts to normal serial cables. They are pinned to work with there own equipment. Also a null modem cable is different to standard 232, you could try source another cable?

Yes, you are right. The light blue cables are straight ones, which could act as a null modem cable when connected to female DB9 and/or DB25 adapters in both ends. However the DCD pin is not connected. Take a look at this: https://www.cisco.com/c/en/us/td/docs/securit … Xhw/pinouts.pdf

Regards,
Ariel

Jo22 wrote on 2022-09-08, 20:23:

Hm. There are three workarounds that come to mind: […]
Show full quote

Hm. There are three workarounds that come to mind:

a) use a slow baud rate, use full handshaking
b) install a NEC V30; the speed gain might be just enough for 9600; the mod is reversible
c) use an ISA card with a new serial port or mod an old serial card to have an 16550A etc.
Multi-i/o cards for ISA were cheap and common in 486 days.

Hi! Thank you for the detailed answer.
I tried 2400 bps which is the lowest speed possible in xtide serial drive, and I get the same results.
I don't have a NEC V30. However this is a 8086-2 which is clocked at 8 Mhz.

Jo22 wrote on 2022-09-08, 20:23:

Edit: Just double checked the specs. Maybe the system has a 16550/A, after all ? As I said earlier, how could IBM have been so i […]
Show full quote

Edit: Just double checked the specs. Maybe the system has a 16550/A, after all ?
As I said earlier, how could IBM have been so incompetent not to use it in 1987.

"The corrected -A version was released in 1987 by National Semiconductor.[1]
It is frequently used to implement the serial port for IBM PC compatible personal computers,
where it is often connected to an RS-232 interface for modems, serial mice, printers, and similar peripherals.
It was the first serial chip used in the IBM PS/2 line, which were introduced in 1987."

Yes, I read that, but the IC is labeled as NS8250AY.

I am still unable to boot the machine, so I cannot run any DOS software. I just have access to ROM BASIC, but it is very difficult to put my assembly program in memory through it.

Regards,
Ariel