I see there is no cache in this machine so it's kinda meh... I had a Hyundai 386DX25 that was cache-less as well, didn't find much value to it tbh , a normal DX40+128kb is a much better option.

Another case of fake cache?

Damn, this is an outstanding machine!

Real connoissoirs appreciate this:
- C&T chipset (by far the best one at that time)
- Full AMI bios setup. You can even fiddle with the chipset registers.
- 5150 class heavy high quality case.
- who knows about computer history knows ICL, their quality and their importance

For this box I would ditch a dozen of these mass produced late 386/40 poor quality nonames with cloned CPUs and crappy chipsets like UMC, SIS etc.

P.S. Turbo switches are pointless on all computers except PC/XT clones. They are just a marketing gag which tricks people to think "turbo=faster", which is nonsense.
The purpose of these switches was and is to make the machine compatible with 4.77 MHz 8088, for allowing early PC games which used loops instead of timers to run at correct speed. Obviously, this only can work at 8088; even on an 8086 that won't result in "correct" speed. For this reason, I observed the turbo button never being released on all computers past 8088.

oeuvre wrote:

Another case of fake cache?

nope , these old DX25 systems simply did not have the option for cache. My Hyundai one was from 1988 or something - pretty expensive back then I imagine

keropi wrote:

oeuvre wrote:

Another case of fake cache?

nope , these old DX25 systems simply did not have the option for cache. My Hyundai one was from 1988 or something - pretty expensive back then I imagine

That depended on the mobo.
For example, back then in the second half of the 1980s, a friend of mine got a 16MHz 386 mobo with cache as a promotional gift directly from Intel.
Mobos with cache were outrageously expensive at the beginning, and for this reason were practically nonexistent except of the ultra highend segment.
Cached mobos became common in the mass market only since the 33MHz 386.

The mass of what is surviving today is low-end cheap stuff of questionable quality which I never would have considered buying back then.
It is very hard nowadays to find high quality vintage PCs like that ICL.

I was actually working for ICL about 4 months in 1998 when I was still a university student. At least in 1998 their Pentium 2 computers had a lot of quality issues which might have been caused by them being merged with Fujitsu.

I worked for ICL from 1983 to 1992.
From 1983-1989 I worked as a componenet level electronics repair engineer at the Central Spares & Repairs workshop in Stanier Street, Newcastle Under Lyme, Staffordshire (1983-1986) then the workshop moved to Bradwell Staffordshire until 1989 when I moved to Kidsgrove Staffordshire to work as a programmer.

I worked on a variety of systems at ICL from an old 'Key Edit' system that used core plane memory and had a wierd 9 bit processor (3x octal = 1 byte) to the ICL Perq that was really a very powerful graphics computer for it's time and had a microcoded processor board made from discrete logic chips. The Perq was the first itme I ever saw a mouse (but actually it was a puck) and a GUI. That was in 1984. Other systems I repaird were an 8086 based ICL PC that ran CP/M and was kinda like an early IBM clone. Later there was an ICL PC that you could switch between ICL and IBM modes and also I repaired the weird ICL / Sinclair QL clone thing - One Per Desk http://www.old-computers.com/museum/computer.asp?c=275 I even had one of them at the house for a while as they gave some away to the engineers. You could play Sinclair QL games on it.

Rich

best regards
Rich

You should check out the speed of your DRAM, maybe there's no cache because it DOESN'T NEED cache.

Try to put the best ram you have (70-60ns sticks) and lower all the timings the lowest the bios allows you and then run some benchmark on it

Personally I like this system, I'd keep it if I had it (but that's probably because I have ""only"" two SX25 and a DX40)

dicky96 wrote:

I worked for ICL from 1983 to 1992.
From 1983-1989 I worked as a componenet level electronics repair engineer at the Central Spares & Repairs workshop in Stanier Street, Newcastle Under Lyme, Staffordshire (1983-1986) then the workshop moved to Bradwell Staffordshire until 1989 when I moved to Kidsgrove Staffordshire to work as a programmer.

Rich

Hi Dicky96,

I checked on Vogons and read that you had worked for iCL. I have this beautiful i386 SBC from iCL which I would like to bring to life using a PICMG backplane. I have some issues with it and was wondering if I could contact you directly to explain the problem I have. On the passive backplane I've installed a controller for hard- and floppy drive as well as a graphics card and an Adlib sound card. As a harddrive I use a industrial grade compact flash card of 256MB. The systems boots OK. I get to boot from a disk and manage to install DOS 5.0 on it. However, on the next boot I get the following message: BIOS(14): Backplane ID incorrect. I removed the battery and believed to have cleared the BIOS settings. The problem remains. Are there some hidden BIOS settings for backplanes on which the SBC is mounted? I don't really understand what provokes this error message. Looking forward to your feedback. Sorry for using this post for my problem. I didn't know how else to send you this message.

keropi wrote:

I see there is no cache in this machine so it's kinda meh... I had a Hyundai 386DX25 that was cache-less as well, didn't find much value to it tbh , a normal DX40+128kb is a much better option.

Yeah but at that point why wouldn't you go for a 486SX/DX instead? To me a slower 386 is the best way to run AT software.. I would keep it, it probably slows down quite a bit when de-turboed.