• Please review our updated Terms and Rules here

replacement for 80286

Don't forget that the 386SX was a "budget" notion. Why waste money on that, when you can sell the "full bodied" version? I don't see why a 386SX couldn't be soldered to a PGA carrier board, if that was really needed. I may even have seen such a beast.
 
The main problem is Intel never released the SX in a physically compatible form factor; it *only* came in a surface mount flat pack, so you need an adapter board to put it in a 286 socket.
In the early days that was the main reason I let this idea go: you really need a decent PCB. The CPU in those days was too expensive for just a hobby project. And PCBs were not cheap then as well. Nowadays it could be a nice hobby project but I have too much other project running. And what should I do with the machine once finished? But I will place it on my list "To do in a very far future...".
But in case somebody has some plans of his own, please let me know!
 
There were kits with the higher clock oscillator and updated BIOS to go along with the new CPU. Note that the 10 MHz AT variant from IBM was a bit unreliable so pushing that far might be unwise. It wasn't 5170 but I think was the industrial system.

Do you have any sources for the existence of the 10MHz AT variant? I've never heard of this before, except for the existence of some prototypes. I have seen IBM made "Turbo" AT motherboards listed on TH99, but I assumed those must be errors.
 
Do you have any sources for the existence of the 10MHz AT variant? I've never heard of this before, except for the existence of some prototypes. I have seen IBM made "Turbo" AT motherboards listed on TH99, but I assumed those must be errors.

Code:
7552-040 h IBM PC/AT System Unit (10MHz, 512KB)
Hardened, for use without keyboard or display via Remote Terminal Interface
7552-540 h IBM PC/AT System Unit (10MHz, 512KB)
Hardened, 20MB fixed disk for use without keyboard or display via Remote
Terminal Interface

Taken from page 46 of the Sep 88 Product Reference. I have never seen one and what I heard about it indicates some software didn't work on it which is why it went with a terminal interface.
 
Code:
7552-040 h IBM PC/AT System Unit (10MHz, 512KB)
Hardened, for use without keyboard or display via Remote Terminal Interface
7552-540 h IBM PC/AT System Unit (10MHz, 512KB)
Hardened, 20MB fixed disk for use without keyboard or display via Remote
Terminal Interface

Taken from page 46 of the Sep 88 Product Reference. I have never seen one and what I heard about it indicates some software didn't work on it which is why it went with a terminal interface.

I read a little about IBM 7552 in a copy of a document I found on Google Books. It sounds like it is an MCA/ISA hybrid, so it may not have an AT motherboard afterall.

https://books.google.co.jp/books?id...edir_esc=y#v=onepage&q=ibm 7552 10mhz&f=false

A VCfed member actually bought one and took photos of it too:

http://www.vcfed.org/forum/archive/index.php/t-22672.html
 
Last edited:
Network World has the earliest description of the 7552 online from 1986 https://books.google.com/books?id=n...ChDoATAAegQIChAB#v=onepage&q=ibm 7552&f=false

I doubt it had a prerelease version of MCA but as a passive backplane system, adding support for MCA should be relatively easy. I can't find the 7552 related IBM announcements but I think the Microchannel industrial line up wasn't introduced until 1989.

What's clear from the first document I linked to is that this system is not using a 5170 motherboard, and it's not even completely AT compatible as claimed because some signals are missing from the ISA slots.
 
It certainly doesn't sound any more like a "10 mhz variant of the AT" then the PS/2 Model 50 is. Or maybe more charitably, the PS/2 Model 30/286?

Funny how IBM was apparently allergic to using the 12mhz 80286, I'm pretty sure those were in wide circulation by 1987. Certainly by 1988 when the 30/286 came out.
 
That's really cool looking. Do you have a machine to test that on?

hello - yes an unassembled IBM AT. Some time ago i did try the MakeIt486 chip in the IBM-at, and at very slow speed did boot Windows 95. ..very slow.. The memory is on the ISA bus and the overall speed not high.
/cimonvg
 
But i would not say this card has close to no chips glued on :D.

Heh. I'm going to hazard a guess that the logic is to double the host CPU clock so the 386sx runs at closer to its rated clock speed. (And also to support the FPU socket?) I do vaguely remember once reading a review of 386sx upgrade boards (PC Magazine?) that featured at least one board that didn't do that. It's amazing to contemplate what it'd be like to have an original 6mhz AT with one of those installed.
 
Heh. I'm going to hazard a guess that the logic is to double the host CPU clock so the 386sx runs at closer to its rated clock speed. (And also to support the FPU socket?) I do vaguely remember once reading a review of 386sx upgrade boards (PC Magazine?) that featured at least one board that didn't do that. It's amazing to contemplate what it'd be like to have an original 6mhz AT with one of those installed.

Hello
i did some tests.
AMD286-6mhz - to compare against:

DSCN9159.JPG
;Data collected by: TOPBENCH | Benchmark and detection stub | Version 0.97
;This file contains fingerprinting information about your computer. Please
;email this file to trixter@oldskool.org with a subject line of "Benchmark" to
;help test these routines and seed the TOPBENCH database.

[UIDF102A099F]
MemoryTest=924
OpcodeTest=398
VidramTest=779
MemEATest=515
3DGameTest=386
Score=16
CPU=Intel 80286
CPUspeed=8 MHz
BIOSinfo=R(C)1985-1989, American Megatrends Inc,.All Rights Reserved.1346 Oakbrook Drive, Suite-120, Norcross, GA-30093. Phone-(404)-263-8181. (04/30/89, rev. 0)
MachineModel=01FC
BIOSdate=19890430
BIOSCRC16=F102
VideoSystem=VGA
VideoAdapter=VGA, Paradise PVGA1A, 8-Bit-Video, 8-Bit-ROMfixed-sync, 512kb Vi
Machine=AT clone

and the MakeIt486 (no cache enabled):

DSCN9151.JPG
;Data collected by: TOPBENCH | Benchmark and detection stub | Version 0.97
;This file contains fingerprinting information about your computer. Please
;email this file to trixter@oldskool.org with a subject line of "Benchmark" to
;help test these routines and seed the TOPBENCH database.

[UIDF1026E7]
MemoryTest=769
OpcodeTest=372
VidramTest=787
MemEATest=491
3DGameTest=358
Score=18
CPU=Cyrix/TI 486-class processor
CPUspeed=33 MHz
BIOSinfo=R(C)1985-1989, American Megatrends Inc,.All Rights Reserved.1346 Oakbrook Drive, Suite-120, Norcross, GA-30093. Phone-(404)-263-8181. (04/30/89, rev. 0)
MachineModel=01FC
BIOSdate=19890430
BIOSCRC16=F102
VideoSystem=VGA
VideoAdapter=VGA, Paradise PVGA1A, 8-Bit-Video, 8-Bit-ROMfixed-sync, 512kb Vi
Machine=AT clone

and at last the "new" Nokia-module:
DSCN9164.JPG
;Data collected by: TOPBENCH | Benchmark and detection stub | Version 0.97
;This file contains fingerprinting information about your computer. Please
;email this file to trixter@oldskool.org with a subject line of "Benchmark" to
;help test these routines and seed the TOPBENCH database.

[UIDF102920]
MemoryTest=1048
OpcodeTest=396
VidramTest=771
MemEATest=527
3DGameTest=401
Score=16
CPU=AMD Am386SX
CPUspeed=8 MHz
BIOSinfo=R(C)1985-1989, American Megatrends Inc,.All Rights Reserved.1346 Oakbrook Drive, Suite-120, Norcross, GA-30093. Phone-(404)-263-8181. (04/30/89, rev. 0)
MachineModel=01FC
BIOSdate=19890430
BIOSCRC16=F102
VideoSystem=VGA
VideoAdapter=VGA, Paradise PVGA1A, 8-Bit-Video, 8-Bit-ROMfixed-sync, 512kb Vi
Machine=AT clone

..i had to build some socket extensions , but after i firm pressure the Nokia-module did connect to the motherboard ;)
DSCN9160.JPG

The speed of this "new" (a call it new - i have never used it before :D ) nokia-module is not doubled. But the feeling , when used in PCDOS is better then the MakeIt486 (without cache enabled) !!
/cimonvg
 
Wow, the scores for that Make-it 486 module are astoundingly bad.

I have to admit I'm jealous of your 8mhz 386sx now. :)

That is only because the cache is turned off. Superior performing chip doesn't do much work if spending all the time waiting on motherboard memory.
 
That is only because the cache is turned off. Superior performing chip doesn't do much work if spending all the time waiting on motherboard memory.

Certainly true, but I guess I expected at least a *little* more of a goose from having the actual instruction-executing hardware running at 33mhz, since most instructions on the 286/386sx take a significant number of cycles. Maybe the instruction mix in Topbench just happens to be particularly bus-bound.

How does it score with the cache enabled?
 
Certainly true, but I guess I expected at least a *little* more of a goose from having the actual instruction-executing hardware running at 33mhz, since most instructions on the 286/386sx take a significant number of cycles. Maybe the instruction mix in Topbench just happens to be particularly bus-bound.

How does it score with the cache enabled?

hello
in the Topbench database , there is a 80286 Bridgeboard with MakeIt 486, doing Score=23
but i will post my score (if/when i get the correct makeit486 driver and the makeit486 on the motherboard)
/cimonvg
 
hello
in the Topbench database , there is a 80286 Bridgeboard with MakeIt 486, doing Score=23
but i will post my score (if/when i get the correct makeit486 driver and the makeit486 on the motherboard)
/cimonvg

That seems rather low, too. I also get a score of 23 on my Compaq DeskPro 286/12. Could be they ran without the cache enabled as well? BTW, your 80286 is 8 MHz as evidenced by the 16 MHz crystal and benchmark results, not 6 MHz as you labeled it. Not a big deal, I wasn't sure you were aware of it or not.
 
Back
Top