Strange AMD K6 233MHz CPU

[kc8eyt]

Back in the late 90's while working as a computer tech I built many K6 systems so I am pretty familiar with the CPUs. I recently purchased a Micronics Twister motherboard on ePay and it included the K6 233MHz CPU shown in the pictures below:



I've never run across a K6 (or any AMD CPU for that matter) that was blank. As you can see there is no silkscreen or engraved markings as is typically found with K6 era CPUs.

At first I thought perhaps the lettering had been removed with some sort of solvent. However the heatsink/fan combo has a sticker on it identifying the chip, its voltage, and what appears to be a serial number. It also appears that the heatsink/fan is original and has never been removed since installed (it was a little effort to get the CPU separated from the heatsink). I've never seen a dedicated AMD heatsink/fan combo like this with information about the underlying chip printed on the spinning sticker.

Anyone care to venture what I have here? An engineering sample? The original seller only remembers obtaining the board sometime around the early 2000's as a spare.

The chip does indeed work. I pulled it to put an Intel 233MMX CPU in its place.

 

[Agent Orange]

Back in the late 90's while working as a computer tech I built many K6 systems so I am pretty familiar with the CPUs. I recently purchased a Micronics Twister motherboard on ePay and it included the K6 233MHz CPU shown in the pictures below:

View attachment 67159View attachment 67160

I've never run across a K6 (or any AMD CPU for that matter) that was blank. As you can see there is no silkscreen or engraved markings as is typically found with K6 era CPUs.

At first I thought perhaps the lettering had been removed with some sort of solvent. However the heatsink/fan combo has a sticker on it identifying the chip, its voltage, and what appears to be a serial number. It also appears that the heatsink/fan is original and has never been removed since installed (it was a little effort to get the CPU separated from the heatsink). I've never seen a dedicated AMD heatsink/fan combo like this with information about the underlying chip printed on the spinning sticker.

Anyone care to venture what I have here? An engineering sample? The original seller only remembers obtaining the board sometime around the early 2000's as a spare.

The chip does indeed work. I pulled it to put an Intel 233MMX CPU in its place.

My quick guess would be an OEM version or it just slipped through without being marked.

 

[kc8eyt]

My quick guess would be an OEM version or it just slipped through without being marked.

I never thought of it being OEM, hmmm. However, a few years back I upgraded some Compaq systems that came from the factory with K6 233 CPUs and they had the engraved markings as is typically found. This is the first time I've ever run across an AMD chip in this condition and a matching heatsink/fan identifying the unmarked chip.

 

[the3dfxdude]

You could fire it up and get the cpuid information. Then check that online. If it matches what's online, then you know what spin it really is. What's the serial number?

 

[kc8eyt]

You could fire it up and get the cpuid information. Then check that online. If it matches what's online, then you know what spin it really is. What's the serial number?

Good idea. I'll reinstall it later today and get that information.

 

[Unknown_K]

I have a Cyrix 5x86/100GP that was blank on the top except for the white circle that points to pin 1. The underside had the usual date, company, and model number. I figured it was just an engineering sample or something.

 

[Timo W.]

There's nothing special about that CPU. If you look at the fan, you'll see that it belongs to the CPU. There was no need to mark the CPU as all the data is on the fan already and the CPU was not meant to be sold without the fan. Can't say whether these were made for OEMs or are retail boxed ones, but certainly nothing unusual here.

It's not an engineering sample. No idea why people always think that when they find an "odd" CPU...

 

[jlang]

I purchased one in the late 90's.
They were available through normal retail.

joe

 

[kc8eyt]

There's nothing special about that CPU. If you look at the fan, you'll see that it belongs to the CPU. There was no need to mark the CPU as all the data is on the fan already and the CPU was not meant to be sold without the fan. Can't say whether these were made for OEMs or are retail boxed ones, but certainly nothing unusual here.

It's not an engineering sample. No idea why people always think that when they find an "odd" CPU...

How many CPUs have you personally seen without any markings or indications of what brand/speed/voltage/etc they are? This the very first one in 30+ years of tech work that I have seen.

From 92 to 2000 I worked various roles from computer tech to system administrator. From 2000 to 2018 I was a CompTIA/CCNA instructor for the local school district. Every two years my classroom was replenished with 50 new computers. My classroom also received all of the district's out of service computers for use as lab/legacy/show-tell purposes (min 500+ systems to harvest). The local community also brought in their computer systems for cleaning/anti-virus/etc which included pulling the heatsink to replace the thermal paste. So I can confidently say that in all my time in IT I have literally worked with 1000's (perhaps even 10's of 1000's) of CPUs. AMD/Intel/Texas Instruments/Zilog/IBM/Cyrix/IDT Winchip/... you name it. First CPU ever with no markings. That is why I commented on it.

To say that a fan needs to always accompany a CPU to identify the CPU makes absolutely no sense to me. Bearings in fans, especially the P5/socket7 generation, lasted maybe two to three years tops in daily operation before the warble of death.

So yes, certainly something unusual here.

 

[the3dfxdude]

How many CPUs have you personally seen without any markings or indications of what brand/speed/voltage/etc they are? This the very first one in 30+ years of tech work that I have seen.

I believe I've seen one of these before as well. I distinctly remember so, because I noticed the writing on the fan identified the CPU, and told myself, this fan must always go with the CPU. I guess they figured the fan would never go bad before the CPU was obsolete. It was either a late 486 (with 586 inside) or later pentium-class processor, so knowing the settings was important, because you had to set jumpers for these to the correct voltage... etc and that will be the only way knowing. It wouldn't surprise me if it was a K5/K6. In fact, I just tried looking for it a couple days ago because of your post. I could not find it. I must of gave it away or threw it out eventually.

The cpuid or serial number is going to probably help explain if it was sold in stores, oem, or maybe an engineering sample. There is usually pretty good information on cpu runs on the net.

 

[jlang]

AMD glued the heatsink to the CPU at the factory. No point silk screening the surface you intend
to glue. You don't want ink in the thermal path either.
They used a long life ball bearing fan so it should last the life of the system.

Here is an article from 1997 describing the product. (sorry no pictures)
https://www.cnet.com/news/amd-bundles-k6-cuts-prices

joe

 

[kc8eyt]

AMD glued the heatsink to the CPU at the factory. No point silk screening the surface you intend
to glue. You don't want ink in the thermal path either.
They used a long life ball bearing fan so it should last the life of the system.

Here is an article from 1997 describing the product. (sorry no pictures)
https://www.cnet.com/news/amd-bundles-k6-cuts-prices

joe

Thank you for the link to the article. That may explain this chip then. The time frame mentioned in the article for the VAR program falls squarely in the time frame I was building a few K6 computers a day when working at local computer shops (I moonlighted at night after my sysadmin day job). I'm surprised I never ran across one of these clean K6 CPUs then ... all local computer shop owners always bought the cheapest hardware they could find to be competitive. Even though these 233MHz K6 chips were faster at 2D operations they performed poorly at 3D, such as Quake!, but still sold like hot cakes because the systems sold for typically $100 less than their Intel MMX counterparts.

I agree also that this CPU appears to have been glued on. Again, I never ran across a K6 CPU/Heatsink combo already glued together like this when building new systems. So it would appear my mystery is solved. Thank you for the information.