Can you post a picture ? My guess is Peaktron is a relabeled Vtech or other OEM made adapter. A picture would help identify it better and the FCC ID if it has one....
I have two XT multi cards with battery (for a clock) and you generally need special software to set and access the clock.

I took a couple of photos of the card:

(Getting images on here seems to be quite a pain)

I also took a few other photos.

Standard 640k Juko ST 12MHz board

Longshine 6210D rev A1 st506 controller (looks like the type that steals a track to write its config to the drive)

Juko monochrome display adapter with lots of other ports and even more mystery jumpers and switches

https://stason.org/TULARC/pc/io-cards/P-R/PEA … NTY-PERI-I.html

Looks like your mystery card?

Looks like your mono card..
https://stason.org/TULARC/pc/graphics-cards/I … TIMATE-EGA.html

Nope, wrong jumper header, wrong ram layout. Tularc does not know what that is.

wierd_w wrote on 2024-01-18, 07:44:

https://stason.org/TULARC/pc/io-cards/P-R/PEA … NTY-PERI-I.html

Looks like your mystery card?

It is indeed!

Wow that was fast.

Thank you very much, this helps a lot!

wierd_w wrote on 2024-01-18, 07:44:

https://stason.org/TULARC/pc/io-cards/P-R/PEA … NTY-PERI-I.html […]
Show full quote

https://stason.org/TULARC/pc/io-cards/P-R/PEA … NTY-PERI-I.html

Looks like your mystery card?

Looks like your mono card..
https://stason.org/TULARC/pc/graphics-cards/I … TIMATE-EGA.html

Nope, wrong jumper header, wrong ram layout. Tularc does not know what that is.

It might be a close relative of it. The description of the jumper header makes sense to me; mine just has one extra unused jumper to the left. It looks like the extra jumper is JP1 and the other 9 jumpers are SW1.

The three combinations for configuring the two serial ports and the two combinations for configuring the prallel port looks to be right.

I still cannot see how the RTC is supposed to work, there is one 14 bit counter chip on the section labeled RTC in the diagram.

As for VGAcards, a number of old Oak and Trident TVGA isa cards can work in 8bit slots.

see here
Advise a good VGA card for 8-bit ISA slots?

Thanks, I used to have an 8 bit Paradise Wonder (if its gonna work) lying about which I cannot seem to find. I wanted to use that one. I have a couple of Trident 9000's and an 8900D that I will pop in and see what they do.

st31276a wrote on 2024-01-18, 08:01:

It might be a close relative of it. The description of the jumper header makes sense to me; mine just has one extra unused jumpe […]
Show full quote

wierd_w wrote on 2024-01-18, 07:44:

https://stason.org/TULARC/pc/io-cards/P-R/PEA … NTY-PERI-I.html […]
Show full quote

https://stason.org/TULARC/pc/io-cards/P-R/PEA … NTY-PERI-I.html

Looks like your mystery card?

Looks like your mono card..
https://stason.org/TULARC/pc/graphics-cards/I … TIMATE-EGA.html

Nope, wrong jumper header, wrong ram layout. Tularc does not know what that is.

It might be a close relative of it. The description of the jumper header makes sense to me; mine just has one extra unused jumper to the left. It looks like the extra jumper is JP1 and the other 9 jumpers are SW1.

The three combinations for configuring the two serial ports and the two combinations for configuring the prallel port looks to be right.

I still cannot see how the RTC is supposed to work, there is one 14 bit counter chip on the section labeled RTC in the diagram.

Frequently, these RTC cards worked along with a small program called in autoexec.bat, that just set the system time for you. (There are several such programs to try with this card in places like the sim-tel bbs archive.)

A binary counter with a crystal clock source, is really all you need with such a bespoke product. Count seconds in the bottom 6 bits, count hours on the next 6 bits, etc..

Thoughts on the 'RTC'

It probably stores / counts 'minutes', with the high bits set to store the year.

Remember, this was an era before high precision network time was required for cryotographic functions. Just having the right minute mark was a huge step up from 'not preserving date and time at all' which is what stock XTs did.

#minutes/year will handily fit with room to spare in 17bits.

st31276a wrote on 2024-01-18, 08:09:

Thanks, I used to have an 8 bit Paradise Wonder (if its gonna work) lying about which I cannot seem to find. I wanted to use that one. I have a couple of Trident 9000's and an 8900D that I will pop in and see what they do.

Here i have tested a lot of 16bit graphics cards if they work as 8 bit:
Re: Resurrecting a Lanner AP-40AHD board PC

It turns out, most of them (at least the ones i own) do.

Thank you for the thoughts on the RTC and the pointers to the vga cards.

Turns out I have a 9000C trident that is based on the same reference design, but lacks the legend on the silk screen that describes the jumpering for 8 bit operation.

I will change that one over and try it out later.

I tried the 8900D with the same jumper config, does not work. Tried some more tridents by closing the block of 3 jumpers and fiddling about with the rest of them, does not work.

Then I took the 16 bit Paradise (WD) wonder if it will work, which I actually do not know if it even works or not, and fiddled about with it for a bit and found it to work when mystery switch 4 is off.

I have the pictured 8 bit 3c503 lying about, which I put in because networking. I used this one a decade or two ago in a P120 with Linux and it saturates at about 3.3Mbps in that application. I set the memory jumper to DC000, as C8000 might overlap with the mfm controller. I left the io base at 300, which means it will use irq2. When this thing boots, I want to use it to telnet to another machine from where I can ssh to stuff to see how practical such a vintage terminal can be used in modern day.

Meanwhile, the disk controller does not read its config from the disk and says it knows about zero drives. Attached is a 225 seagate which sounds really nice. It spins right up, maintains speed well and makes its classic pheeguhley sound when it calibrates the heads. The light on it remains off, however. If I am not mistaken, this is the way XT's do it, that light flashes only on disk access if I remember correctly. On 16 bit AT MFM controllers, I have seen that the light on the drive remains on for the selected drive. Please chip in if you know more.

The floppy drive seeks and it tries to boot from it as well. It seems as if it really wants to boot.

I really hoped it would boot from the hard disk, as I do not have any 360k floppies close by at this stage.

"1701" is the error message by the HDD controller BIOS that the hard drive is not properly responding to controller requests. You should make sure that

• The drive select jumper on your ST225 is set to "first drive"
• The (wide) control cable between the hard drive controller and the ST225 is properly conneted. This may be a 3-way cable.
• The (narrow) data cable between the hard drive controller and the ST225 is properly connected. This is a per-disk point-to-point data cable and needs to be connected to the data connector associated with the primary hard drive.

mkarcher wrote on 2024-01-19, 17:23:

"1701" is the error message by the HDD controller BIOS that the hard drive is not properly responding to controller requests. Yo […]
Show full quote

"1701" is the error message by the HDD controller BIOS that the hard drive is not properly responding to controller requests. You should make sure that

• The drive select jumper on your ST225 is set to "first drive"
• The (wide) control cable between the hard drive controller and the ST225 is properly conneted. This may be a 3-way cable.
• The (narrow) data cable between the hard drive controller and the ST225 is properly connected. This is a per-disk point-to-point data cable and needs to be connected to the data connector associated with the primary hard drive.

I assume system was set up to work correctly, as I have not changed anything in it and those who threw it out 25 years ago would not know how to. The controller and drive looks to my forensic eye as if they were sold as a kit and added to the computer about six months to a year after it was first purchased. The control cable is a single drive flat cable (no middle connector, no twist) which means second drive select must be selected, which it was.

Just for the joke, I moved the jumper to first drive, and there it started working! I have no idea why.

The drive has read errors. It tried to boot straight into Brilliant Accounting, which seems to be licensed to a local spares place. The software no longer works right, because general Failure is reading drive C.

It has IBM DOS 4.00 on it and the accounting software. That’s it. 3MB worth of stuff.

I looked around to see how the RTC was accessed, but nothing in config.sys or autoexec.bat seems to do something with time at all. There is a cmosclk.sys file in the dos directory that is 878 bytes in size. The files have dates between 1990 and 1998 so the computer must have known the time somehow while it was used. It now thinks it is 1980. Perhaps the accounting software could read the RTC, but that seems like a long shot.

It worked for 15 minutes, stopped reading correctly, and now says non system disk on startup. I guess the thermal expansion of the platters moved the outer tracks just far enough that they no longer want to read properly.

My best guess is that this system was used for accounting until it became flaky in the late 90’s, then it was thrown out. It sat for 25 years, of which the last 24 was in my attic.

I hope a LLF will re-align the tracks properly. I do hope I can recover some stuff off of it first, so I am letting the disk cool to see if it will play again.

When it cools down it seems to work again for another 15 minutes or so.

st31276a wrote on 2024-01-20, 09:46:

I looked around to see how the RTC was accessed, but nothing in config.sys or autoexec.bat seems to do something with time at all. There is a cmosclk.sys file in the dos directory that is 878 bytes in size.

So there likely is a DEVICE=C:\DOS\CMOSCLK.SYS line in CONFIG.SYS that loads this driver, which is supposed to add support for the RTC chip to DOS. Maybe the driver just copies the current RTC time into the system time, but it might also register as proper clock driver in DOS, which will in turn also store updates to the system time/data into the RTC chip.

I will give that a try when it cools down again, the driver is not referenced in config.sys as I found it. I wonder if this file is standard in pc dos 4 or if it was added specifically for this card.

Nope, it does not read or write this RTC. Perhaps it has switches/parameters it accepts, but I am inclined to think it is not for this peaktron card.

This time the hard disk kept going for longer than 15 minutes, until I saw smoke rising from the case. A tantalum cap next to a memory chip was glowing yellow, emitting light, so I shut it off.

This is strange, usually the 5v tantalums are all right.

It feels like a waste to replace all of them just because one failed, what do the experienced fixers think about this?

Edit: interesting reading https://electronics.stackexchange.com/questio … -in-new-designs

Edit: I found this about cmosclk.sys - looks like somebody copied it from pc dos 5 files to the pc dos 4 on this system so see if it will do something.

CMOSCLK.SYS

PC DOS 5.02 and above

CMOS clock driver.

DEVICE=[drive:][path]CMOSCLK.SYS
Installing CMOSCLK.SYS will cause DOS to always access the CMOS real-time clock instead of the system timer for date/time requests.