Durandal wrote:

Any suggestions? 😕 😕 😕

Yes, post a picture of that CPU.

IBM motherboards are notorious for rejecting CPUs they were not designed for.
Just for the record, the 133 version of the 5x86 normally requires 3.7V to work properly, although there is a version that can work at 3.6V and is labelled as such. 3.45V is not going to cut it.

Anonymous Coward wrote:

Just for the record, the 133 version of the 5x86 normally requires 3.7V to work properly

That's why I asked for a picture. The only 133 MHz version of that CPU that was sold in retail was for upgrading and should sit on an adapter, thus working with 5V. The ultra-rare normal version of the CPU is easily killed when trying to use it with 5V.

Aside from providing a photo of the CPU (top and bottom), could you provide a high resolution photo (not over 2 MB) of your motherboard? I'm looking for jumpers and the existance of a voltage regulator.

That article was mentioning upgrade paths. Upgrade interposers contained contain a voltage regulator to bring down 5 V, which was typical for an i486-33/66 down to 3.45 V. The sentiment in that Japanese article about the Am5x86 being an easier upgrade is probably due to it not needing its registers setup via software. Also, I have found the Am5x86 more compatible in general with older motherboards.

The Cyrix 5x86-133 requires 3.6 V or 3.7 V, which will be stated on the surface of the chip. Do not run the Cyrix 5x86 or AMD 5x86 at 5 V. While they may power up initially, they will return errors on the screen like you describe. They will overheat at 5 V and ruin.

The simplest most probable upgrade for your system would be an Intel DX4 overdrive, which already contains a voltage regulator. e.g. DX4ODPR100. We could workout a trade if interested.

Yip for a 5x86 to work in a 5v 486 board it should be attached to an interposer to step from 5v down to 3.x volts. You could buy the interposer separately but normally supplied as part of a 5x86 upgrade package such as Kingstone, HyperRace or similar put out.Some of the interposers have jumpers for fine tuning purposes and software is usually supplied.

As above photos will be really helpful here.

This is the first time I've taken off the cooling fan and it looks like this is the 100MHz version, not the 133. So this site is just wrong about the G5FB607B?
http://www.cpushack.com/cyrix-486-cpus/

Regardless, the interposer thing is a good point, I hadn't thought about that. I see them sometimes coupled with CPU upgrades for Japanese PC-98 computers on Yahoo Auctions and sometimes the interposers show up separately for cheap, should I try getting one of those? And what are the chances I've destroyed the CPU by inadvertently running it at the incorrect voltage?

Thanks for all the help guys, I'm pretty new to anything beyond a DX2. Hell I only just built a Pentium MMX Win98SE PC last week for the first time. 🤣

feipoa wrote:

Aside from providing a photo of the CPU (top and bottom), could you provide a high resolution photo (not over 2 MB) of your moth […]
Show full quote

Aside from providing a photo of the CPU (top and bottom), could you provide a high resolution photo (not over 2 MB) of your motherboard? I'm looking for jumpers and the existance of a voltage regulator.

That article was mentioning upgrade paths. Upgrade interposers contained contain a voltage regulator to bring down 5 V, which was typical for an i486-33/66 down to 3.45 V. The sentiment in that Japanese article about the Am5x86 being an easier upgrade is probably due to it not needing its registers setup via software. Also, I have found the Am5x86 more compatible in general with older motherboards.

The Cyrix 5x86-133 requires 3.6 V or 3.7 V, which will be stated on the surface of the chip. Do not run the Cyrix 5x86 or AMD 5x86 at 5 V. While they may power up initially, they will return errors on the screen like you describe. They will overheat at 5 V and ruin.

The simplest most probable upgrade for your system would be an Intel DX4 overdrive, which already contains a voltage regulator. e.g. DX4ODPR100. We could workout a trade if interested.

I just put the damn thing back together! 🤣 I'll get it apart again and take a pic of the board later today. I'm not really looking to upgrade it in the first place, it's meant to be for DOS games up to/around Doom and Doom II which it runs nicely, I just thought it would be fun to mess with a 5x86 while I've got one. If I do get the CPU working I'll probably be selling it anyway.

An interposer could be useful for testing and experimenting in future. Just depends what your intentions are.

Caluser2000 wrote:

Just depends what your intentions are.

To test the CPU and pass it on to someone else, really. I don't need a 5x86 in this system.

You cannot look at that list and assume that all G5FB607B are 5x86-133 chips. Many cx5x86-133 chips which did not qualify at 133 MHz were marked for 100 or 120 MHz operation. That list represents only what a few individuals have encountered and is not exhaustive.

For how long had you run the Cyrix 5x86 at 5 V? If the chip still runs, it is possible that it may require more a smidge more voltage (e.g. 0.1V) than the nominal 3.45 V to run normally now. This is what I've noticed with some of these chips anyway.

feipoa wrote:

You cannot look at that list and assume that all G5FB607B are 5x86-133 chips. Many cx5x86-133 chips which did not qualify at 133 MHz were marked for 100 or 120 MHz operation. That list represents only what a few individuals have encountered and is not exhaustive.

For how long had you run the Cyrix 5x86 at 5 V? If the chip still runs, it is possible that it may require more a smidge more voltage (e.g. 0.1V) than the nominal 3.45 V to run normally now. This is what I've noticed with some of these chips anyway.

You mean they labeled chips with the same identifier that weren't the same? Seems silly to me, couldn't they have given them a different code?

Only maybe 5 minutes all up with lots of off/on and reboots since I was trying different boot floppies and hard drives.

I suspect the dye caps were stamped, then installed first. Then the chips were qualified.

feipoa wrote:

I suspect the dye caps were stamped, then installed first. Then the chips were qualified.

I see. So should I bother getting an interposer?

If you have the desire to upgrade your system using an Am5x86 or Cx5x86, then you'll need an interposer to drop the voltage down to ~3.45 V. Keep in mind that Write-back L1 cache may not work and your L1 cache may only work in write-through mode.

If you don't want to bother with an interposer, but still desire to upgrade the CPU, you can consider an Intel DX4 Overdrive, which has the voltage regulator built onto the unit. e.g. https://www.ebay.com/itm/264400658137 These chips run in L1 write-through mode. All of the Intel DX4-100 chips I've tested have run at 120 MHz if you are wanting more speed. This is assuming you have the ability to set the FSB on your motherboard to 40 MHz.

feipoa wrote:

If you have the desire to upgrade your system using an Am5x86 or Cx5x86, then you'll need an interposer to drop the voltage down to ~3.45 V. Keep in mind that Write-back L1 cache may not work and your L1 cache may only work in write-through mode.

If you don't want to bother with an interposer, but still desire to upgrade the CPU, you can consider an Intel DX4 Overdrive, which has the voltage regulator built onto the unit. These chips run in L1 write-through mode. All of the Intel DX4-100 chips I've tested have run at 120 MHz if you are wanting more speed. This is assuming you have the ability to set the FSB on your motherboard to 40 MHz.

Thanks for the information, I might upgrade to a DX4 100 at some point.

I've decided to auction this 5x86-100 as-is - you should be able to see it on eBay if you search for it in about 2 hours.

Well, what it has going for it is that it should work with a 4x multiplier. This may be useful for boards which only work with a 25 MHz FSB, but I don't think many people are running those.