it even has ps2 ports! nice!

It works! I ran speedsys, and get these (probably very ordinary) results -

Overall score = 24.93 (I seem to get this value on most of my 486 / dx2-66 set ups)
Memory bandwidth = 76.21 MB/s
Data cache L1 (8 kb) = 52.30 MB/s
Data cache L2 (256 kb) = 37.50 MB/s
Memory throughput = 23.05 MB/s

I have a CPU in front of me that says -

Intel
OVERDRIVE
DX20DPR66
C5020122
SZ935 V4.0
INTEL (m)(c) '89 '92

I think I will attempt to install it, and see what happens...

You can't play Dark Forces without any tunes and the sound of blaster fire! So in goes a CT1600 (Soundblaster Pro 2), and a Roland MPU 401 AT card + daughterboard card - in this instance, it's an SCB-7.

I'm surprised at the responsiveness of the PS/2 mouse in Dark Forces. It's really rather good. Frame rate looks respectable. I've no idea what the value is, it just looks playable.

It's time to check out the 486 Overdrive CPU. I notice that it has V4.0 written on it. I guess that means 4 volts. Currently, the mobo is set to 5 volts, for a typical Intel 486 dx2-66 chip. I will change the mobo's jumpers to a 4 volt setting...Edit: I just saw this webpage -

http://datasheets.chipdb.org/Intel/x86/486/Intel486.htm

And it says for the SZ935 CPU (the Overdrive CPU), that the V4.0 means Version, not Volts! The Overdrive CPU is 5 volts, just like a typical 486 dx2-66 CPU. Lucky I saw that, before switching on the power!

Benchmarks -

Using 486 dx2-66 CPU

Doom -timedemo demo3 (with the screen set to max size, and sound left on) = 2134 / 3019 = 24.73 fps
Quake -timedemo demo1 (res 320x200, with soundblaster) = 6.9 fps

Using Overdrive CPU

Instead of running Doom and Quake first, I decided to run speedsys. The score I get is *exactly* the same I got with the dx2-66 CPU! There's something wrong, isn't there? (I get the score of 24.93 again.) Speedsys tells me this info about the Overdrive CPU -

Processor: Intel 486DX2 tm 66mhz
CPUID (TFMS) : 0435 Codename : P24 (0.80 um)

Hang on a minute...do I need to set the mobo's jumpers to "P24T"? There is some info on 486 Overdrive CPUs (including the SZ935 CPU) on that webpage link above, in this post.

retro games 100 wrote:

It's time to check out the 486 Overdrive CPU. I notice that it has V4.0 written on it. I guess that means 4 volts. Currently, […]
Show full quote

It's time to check out the 486 Overdrive CPU. I notice that it has V4.0 written on it. I guess that means 4 volts. Currently, the mobo is set to 5 volts, for a typical Intel 486 dx2-66 chip. I will change the mobo's jumpers to a 4 volt setting...Edit: I just saw this webpage -

http://datasheets.chipdb.org/Intel/x86/486/Intel486.htm

And it says for the SZ935 CPU (the Overdrive CPU), that the V4.0 means Version, not Volts! The Overdrive CPU is 5 volts, just like a typical 486 dx2-66 CPU. Lucky I saw that, before switching on the power!

Benchmarks -

Using 486 dx2-66 CPU

Doom -timedemo demo3 (with the screen set to max size, and sound left on) = 2134 / 3019 = 24.73 fps
Quake -timedemo demo1 (res 320x200, with soundblaster) = 6.9 fps

Using Overdrive CPU

Instead of running Doom and Quake first, I decided to run speedsys. The score I get is *exactly* the same I got with the dx2-66 CPU! There's something wrong, isn't there? (I get the score of 24.93 again.) Speedsys tells me this info about the Overdrive CPU -

Processor: Intel 486DX2 tm 66mhz
CPUID (TFMS) : 0435 Codename : P24 (0.80 um)

Hang on a minute...do I need to set the mobo's jumpers to "P24T"? There is some info on 486 Overdrive CPUs (including the SZ935 CPU) on that webpage link above, in this post.

P24T is the intel Pentium overdrive

No need to change any jumpers, just have everything set to a 33mhz, 486DX or something similar

I've found a good webpage about overdrive CPUs. It's here -

http://www.cpushack.net/UpgradeProcessors.html

About 20% down this page, it says -

486 ODP Overdrive

Some lowcost motherboards had a 486 SX soldered onto the board but still provided an 'Overdrive Socket' that these chips plugged into.

(end)

So, it looks like the 486 (non Pentium) Overdrive CPU was just a basic way to upgrade a crappy mobo.

BTW, is it possible to overclock a 486? How about I try overclocking this Overdrive CPU? It doesn't matter if I fry it. But I really don't want to damage the mobo. The CPU can go up in flames, so long as the mobo can live another day.

I think it's physically possible to overclock 486DX4's at least, as I seem to remember having to do complex jumper configuration to set the clock speed. Whether or not the CPU and/or mobo can handle it is another story.

That's a very nice Socket 3 board you got there! Very similar to the venerable PVI-486SP3, some features are even better:
It sports four SIMM slots instead of two, and apparently it has a configurable PCI bus divider (Jumper W20), which greatly increases the chances to run a 50 MHz FSB! 😉
What I don't like so much about this board is its conventional "AT-style"-layout (with a heatsink installed, the CPU socket blocks two ISA and one PCI slot for long cards) and the fact that it omits the VLB slot that is normally present on SIS 496/497 boards.

Your benchmark scores are perfectly normal for this kind of system.
Overclocking the 486DX2 Overdrive is very easy with this board: Increase the FSB to 40 MHz (by altering the W15/W16 jumpers) and your CPU will run at 80 MHz (make sure there is a heatsink on the CPU if you want to avoid an unstable system).
This will increase the overall performance by ~30%.

Oh yes! 😀 Simply by removing jumper W15, the CPU speed now reads: 80mhz on the BIOS POST screen. I then reran Speedsys -

Memory bandwidth is now 85.37 (was 76.21; 12% improvement)
Overall score is now 30.05 (was 24.93; 20.53% improvement)

(The other 3 memory related scores have not increased.)

Doom = 2134 / 2543 = 29.37 fps (was 24.73; 18.76% improvement)
Quake = 8.1 fps (was 6.9; 17.39% improvement)

Thanks - that was a bit of fun!

I bet this board came from one of these 😀 :

Or this:

Hmmm, maybe my 30% estimate was a bit optimistic. 😅 I guessed values from my Intel DX4 which supports write-back caching and has the double amount of L1 cache (16 KB).

I don't know which BIOS settings your board offers, but if it has the typical SIS 496/497 Award BIOS, there may be room for improvement, especially with the L2 cache and memory timings.

I've just checked the BIOS POST screen, and also examined the BIOS set up area by pressing the F2 keyboard button on boot up. The board uses the PhoenixBIOS version 4.04 (other Micronics boards also use this BIOS), and this BIOS set up area seems rather "bare bones" - there are no options to mess about with anything "interesting" related to cache and timings, etc.

Also, I was thinking of replacing the dx2-66 Overdrive CPU, and putting in something a bit faster. I had a look about, to see what I had to hand, and found an AMD DX4-100 CPU. It was inside a Soyo board; a 4SA2. It was then that I realised it had the 496 chipset. Does that mean the chip underneath that Soyo sticker (see photo) is the "matching" 497 chip? It's interesting, because this board has the "missing" VLB slot that the Micronics board does not have. And the cache option on this board is incredible! It seems to allow up to 1 meg of cache! Unfortunately, the cache jumpers are currently configured for its minimum setting, which is just 256k.

This Soyo board also sports 4 SIMM slots, and a convenient coin battery too. No PS/2 ports however.

I removed the AMD DX4-100 CPU from the Soyo board, and put it in the Micronics board. Two surprises: 1) Looking at the Stason webpage for this board, I notice that the CPU voltage jumpers need to be changed from 5V to 3.3V. (From a typical dx2-66 5V setting, to an AMD dx4-100 3V setting.) I must admit that I forgot to do this, and checked this jumper area after an unsuccessful boot up attempt. When I examined the voltage jumper J4, it was set to 3.3V anyway! The previous owner may have set it to 3.3V for one of those rare dx2-66 CPUs that are for 3.3V. However, it's strange that both of my 5V dx2-66 CPUs worked OK with a 3.3V setting!

Surprise 2) Again, looking at the Stason webpage, there appears to be no jumper configuration difference between an Intel dx2 CPU, and an AMD dx4 CPU. Not one. (However, I think it's important to point out that the Stason webpage refers to Rev. 3A, and my board is Rev A5.)

Well, there's no Surprise number 3), because it didn't come as a huge surprise when I switched on power and got no POST. I could either mess about with the Soyo board, and double-check that the DX4 CPU is OK, or try to discover why the DX4 CPU is not working in the Micronics board. The problem is that if I ring Micronics tech support, there may be no one there. 😉

retro games 100 wrote:

I removed the AMD DX4-100 CPU from the Soyo board, and put it in the Micronics board. Two surprises: 1) Looking at the Stason webpage for this board, I notice that the CPU voltage jumpers need to be changed from 5V to 3.3V. (From a typical dx2-66 5V setting, to an AMD dx4-100 3V setting.) I must admit that I forgot to do this, and checked this jumper area after an unsuccessful boot up attempt. When I examined the voltage jumper J4, it was set to 3.3V anyway! The previous owner may have set it to 3.3V for one of those rare dx2-66 CPUs that are for 3.3V. However, it's strange that both of my 5V dx2-66 CPUs worked OK with a 3.3V setting!

Surprise 2) Again, looking at the Stason webpage, there appears to be no jumper configuration difference between an Intel dx2 CPU, and an AMD dx4 CPU. Not one. (However, I think it's important to point out that the Stason webpage refers to Rev. 3A, and my board is Rev A5.)

Well, there's no Surprise number 3), because it didn't come as a huge surprise when I switched on power and got no POST. I could either mess about with the Soyo board, and double-check that the DX4 CPU is OK, or try to discover why the DX4 CPU is not working in the Micronics board. The problem is that if I ring Micronics tech support, there may be no one there. 😉

Try phoning Gateway for fun. 😀

Since they still have alot of support files on their site for download for the Micronics boards they used in ther pc's up to the mid 90's...maybe they still have some "responsability"

Ah! Your reference to Gateway was inspirational! 😀 Thanks a lot for that clue. I briefly looked through Gateway's support section, but I couldn't find any information about "Gateway 2000" products, or Micronics motherboards. However, a quick google for: gateway 2000 micronics, and I found this excellent webpage -

http://webpages.charter.net/dperr/micronics/gw2k_faq.htm

Reading this webpage, I now wonder if the Micronics board has been "hard wired" to be a 5V CPU only board. Now *that* would explain why the jumpers on the J4 voltage setting are all set to Open (which is actually the 3.3V setting!). If this J4 voltage setting is being ignored by the motherboard, then why waste jumper caps on these pins? So, I think it doesn't matter what jumper setting you choose for J4, I think this setting will be ignored, and the mobo will simply use 5V for all installed CPUs. So, if you want to use an AMD DX4-100 (3V CPU), you will need to use a "voltage regulator".

On this webpage, it says -

...there are the 486 style processors which have an operating voltage of 3.3 or 3.45 volts. These processors require a voltage regulator to operate in a 5 volt motherboard. A voltage regulator is a tiny circuit board installed in the motherboard socket with a 3 volt processor plugged in piggy back style on top.

This webpage also points out that -

All of the Micronics 486 motherboards known to have been used by Gateway were designed for 5 volt processors.

I think that explains why the AMD CPU will not POST. I looks like I have accidentally sent 5V through the AMD CPU. I wonder if I have fried it?

I've never fried a 486 cpu when usinbg the wrong voltage but might have been lucky 😉

If you need a matching old school, goofy Gateway 2000 wallpaper for this system..look here

http://www.flickr.com/photos/textfiles/sets/7 … 57616506646434/

Hehe! Thanks. I viewed them all as a slideshow. My favourite is this one, because I think it sucks the most -

I reinserted the Intel Overdrive dx2-66 CPU, and set the mobo's CPU jumper speed setting to 40mhz, giving me a real clock speed of 80mhz. As I can't use my AMD CPU without a voltage regulator (or VRM as they are called), I will stop testing this Micronics board today. (I have got one VRM somewhere, but it's in the Attic o' doom, and so I'd rather test the Soyo board with the AMD CPU, to see if I fried it.) On to the Soyo board then...

This thing seems better than the Micronics board. It seems like a "proper" SiS 496/497 board. Here are some of its BIOS features -

Award Modular BIOS v4.51PG
BIOS datestamp of 07/30/97 (that's the latest 486 BIOS I've ever seen!)
Inside its set up area, there are memory timing and cache options

The mobo POSTs OK, and so the AMD CPU lives to see another game of Dark Forces. Lucky chip, eh?