• Please review our updated Terms and Rules here

Make-it-486 in an IBM 5160 XT -- speed rush! ;)

DeathAdderSF

Experienced Member
Joined
Jul 2, 2016
Messages
315
Location
USA
Hello, all.

I recently picked up a Make-it-486 CPU upgrade for 286 machines. While its permanent home is going to be in my Tandy 1000 TX, curiousity got the better of me and I decided to give it a quick spin in my IBM 5160 XT. I accomplished this by way of the Orchid Tiny Turbo, which if you're not already aware is a 286 upgrade card for the XT. What's neat about the Tiny Turbo is that it provides a toggle switch to flip between 286 and 8088 processors (you install your original 8088 chip on the card itself), which is very helpful if -- like me -- you are using hardware and / or software in the XT that specifically requires the 4.77Mhz speed to run properly.

The Tiny Turbo comes with a Siemens 286-N processor which, according to Norton's System Index benchmark, provides 6.6 times the processing power of the 5160's base 8088.
With the Make-it-486 installed in place of that 286-N, the power goes up to 7.5 times the 8088.
But with the cache enabled on the Make-it-486 -- thanks to Cloudschatze for providing the cache utility -- the system ultimately becomes 14.3 times quicker than the base 8088.

Ultimately I think I'm going to go with a Harris CS80C286-20. That should offer more than enough horsepower to this venerable machine.

I thought this was a pretty cool, fun experiment. And I thought to post my results here in case anyone else feels like turbo-charging their 5160 in a similar manner.
 
A fun experiment, but a quick note that Norton SI is not a good test of relative performance since all it tests is the speed of the MUL opcode. For a more rounded view, try running TOPBENCH on your franken486 (load it, and select Benchmark->Realtime) and see what system it most closely matches in the database.
 
It would be interesting to benchmark the Makei-It-486 piggybacked off a TinyTurbo route against a Intel Inboard 386 PC upgraded with a 486. Miy InBoard uses a Cyrix Cx486DRx2 which tops out at 33Mhz as it clock-doubles the 16MHz bus, but is limited to go further by the InBoard stock oscillator and a 1KB cache. The nice thing about the InBoard route is gaining upwards of 5MB RAM (I have the 4MB RAM piggyback card sitting on the 1MB InBoard), most of which can be used as extended, not expanded memory, and thus able to install and run Windows 3.1 in protected mode well on a 5160 base However, given the current going prices and rarity of InBoards these days, your path sounds far more affordable.

Mike
 
Topbench performed on my machine (IBM 5160 + Orchid Tiny Turbo w/Make-it-486 installed) reported the following...

MemoryTest=977
OpcodeTest=1192
VidramTest=1895
MemEATest=1597
3DGameTest=1092
Score=11

I haven't really had the time to throw any software at the machine to really see the performance firsthand. All I've done so far is play a quick game of Golden Axe -- in CGA. :huh:
 
If only one of these Orchid Tiny Turbo's were made for an 8086 system like my Model 25, I might, just might, get to see if Doom could run on it, haha, with a 486 upgrade.
 
Whoa, that's really low for the 486 installed -- did you enable the cache?

Yeah, that's with the cache enabled.

When I did the initial tests using Norton's System Index, it showed the Make-it-486 as 14.3 times XT speed in the 5160. But with the Make-it-486 installed in my Tandy 1000 TX (and cache enabled), System Index reported 25 times faster than XT. (I haven't run Topbench on the Tandy with the Make-it-486 installed, yet.)

My only guess is that the Tiny Turbo has a bottleneck. It wasn't, after all, originally designed to take anything over a 286.
 
If the board had its own private memory, that might make it worth the original purchase price. But these things were outclassed by the Taiwanese replacement motherboards, which turned in better performance with fewer compatibility issues for less money.

So these things that fit inside a 5160 or 5170 are mostly curiosities.
 
Back
Top