I have an ASUS K7V motherboard that can handle a 1 GHz Orion no problem, but won't work with Thunderbirds above ~750 MHz. This is a well known issue with the K7V and ASUS released the K7V-T to address it. The K7V is a VIA KX133 chipset board and a quick google search tells me that the Gateway Kadoka uses the AMD 750 chipset. So it's not the exact same, but could be caused by a similar issue that affects the K7V.

Most 750 boards supported the 1ghz third, while kx133 won't run 800mhz and above

candle_86 wrote on 2020-05-05, 23:00:

Most 750 boards supported the 1ghz third, while kx133 won't run 800mhz and above

That's what I thought. It's weird to see a 750 board have this issue. If the issue is specific to the KX133 chipset, then it may be something else. Do all the capacitors on the board look good? Maybe some need to be repalced?

The board doesn't appear to have any electrolytic capacitors.

Here's a photo of it.

I'm going to put the bezel into the whittening booth this evening. When I removed the Athlon and Win98 case stickers, you could really see how yellow the case was. Only concern is that the hydrogen peroxide removes the Gateway logo and the SELECT sticker.

feipoa wrote on 2020-05-05, 21:50:

I'm trying to setup a Gateway 2000 Kadoka system, aka Gateway KAD Select 750. I've updated it to the latest BIOS, 0AASNP06, and have started testing various slot A CPUs. My system came with a 750 MHz Thunderbird. I have tested the 1 GHz Orion CPU and it seem so work so far, but when I use the 1 GHz Thunderbird CPU, I get HIMEM errors, Speedsys memory errors, and cachechk gives out a "divide by 0" error. The next fastest Thunderbird CPU I have is the 800 MHz one and it works fine in these tests.

I'm wondering if it is the motherboard that can't cope with the Thunderbird at 1 GHz, or if there's something wrong with my 1 Ghz Thunderbird - perhaps the CPU's onboard cache is bad? The Orion at 1 GHz draws more current than the Thunderbird, so I figured if one of the two chips wasn't going to work, it would be the Orion. Is this a common problem?

While researching this board (or trying to anyway), I did find a couple references to issues with certain memory. It seemed some people had problems with 133 MHz memory (system acting flaky) and at least one person mentioned having trouble with a Kingston module that worked in other machines but wouldn't work in this one. While I doubt it is the problem, it may be worthwhile trying other memory if you haven't already.

The Thunderbird and Orion have a little difference in the bus interface.

Re: AMD K7 Slot A 1000 Orion 512KB Cache working, Slot A 1000 Thunderbird 256KB Cache NOT

The Irongate 750 doesn‘t support the Thunderbird officially, but there existed some workarounds.
Maybe the workaround is power- or clock sensitive, so that the 800MHz TB is running and the 1000MHz TB is not.

feipoa wrote on 2020-05-05, 23:53:

Only concern is that the hydrogen peroxide removes the Gateway logo and the SELECT sticker.

This Gateway logo sucks anyway, so no real loss. The old Gateway 2000 logo with the big 'G' was much cooler.

Gateway sold this system with Thunderbird core CPUs, so they probably felt it was probably official in their book. The 1 GHz Orion consumes around 10 W more than the Thunderbird, so the issue isn't likely to be a power issue. I suppose I should have also pointed out that when I received the Thunderbird chip, one of the fans were ceased. When I removed the sticker from the back, a bunch of grey powder spilled out. I lubed it, but the fan is pretty wobbly. Somebody must have been running the bearings dry until it stopped spinning. By the time someone opened up the case, the CPU could have been running hot for months. I suspect a worn out CPU.

To further this hypothesis, upon a subsequent boot-up, I started smelling an odor from the case, it was coming from the Slot A cartridge. It has a really strong peach smell. I powered off and powered back on, but it was dead - blank screen. Tried again, but the smell was gone and CPU completely dead. I put the 800 MHz Thunderbird back in and had no problem. Any chance that the fault is something on the PCB and is it worth trying to get the cover off?

I agree, the old Gateway2000 logo actually looked like it belonged on a professional system.

Redhill mentioned Slot A 1Ghz Tbird was pretty selective in PSU's (under the FIC SD-11) entry
https://www.redhill.net.au/b/b-00.html

None of my K7 pro (5 boards) work with 1GHz tbird.
My 2 GA-7IXE and my 1 GA-7IX do.

Those 8 Boards are AMD750 equipped. Some of my K7 pros at least are bypass enabled, so are my 7IXE (not my 7IX)
the 7IX has a 99.9MHz FSB, the others are at 100.9MHz (which explains their better results at "stock" speed, I guess that's around when motherboard builders realized the reason ASUS was on top in performance tests was just that they cheated)

All with the same ram (3x128MB PC133CL2 and 3x256MB PC133CL2)
All with the same peripherals (EG365 enermax PSU, compact flash adapter or basic IDE drive)

Which is annoying because the K7 pro has MHz increment overclocking while the gigabytes are 100/105/110/115/120 MHz only, and I get my best results on a 119MHz bus (there is some kind of hard limit, as soon as I touch 120MHz "FSB" it doesn't work)

Could all be so, but will any of this cause the CPU to smoke and die?

For me it's always been simple system freeze after random time between 2minutes and 15 minutes under windows.

feipoa wrote on 2020-05-05, 23:53:

I'm going to put the bezel into the whittening booth this evening. When I removed the Athlon and Win98 case stickers, you could really see how yellow the case was. Only concern is that the hydrogen peroxide removes the Gateway logo and the SELECT sticker.

I retro-brighted the same Gateway front cover from a Pentium III machine of the same era. I did my best to keep the solution off the logos, but some got on them and they survived. So you may be in the clear on that. No guarantees of course. I'm pretty sure you can remove the little Gatway cowbox badge and put it back on after.

Katmai500 wrote on 2020-05-06, 15:50:

feipoa wrote on 2020-05-05, 23:53:

I'm going to put the bezel into the whittening booth this evening. When I removed the Athlon and Win98 case stickers, you could really see how yellow the case was. Only concern is that the hydrogen peroxide removes the Gateway logo and the SELECT sticker.

I retro-brighted the same Gateway front cover from a Pentium III machine of the same era. I did my best to keep the solution off the logos, but some got on them and they survived. So you may be in the clear on that. No guarantees of course. I'm pretty sure you can remove the little Gatway cowbox badge and put it back on after.

The logos survived just fine. I also did the floppy drive. But then the CD-ROM wouldn't match even after the whitening, so got another drive from my bin.

sunaiac wrote on 2020-05-06, 11:49:

For me it's always been simple system freeze after random time between 2minutes and 15 minutes under windows.

That is what I would expect as this has been my experience with 1 GHz on the Slot 1 440BX for certain boards.

Came out looking pretty good. It only takes about 2 hours under the 400W grow bulb.

Looking great! These are one of my favorite OEM case designs of all time.

candle_86 wrote on 2020-05-05, 23:00:

Most 750 boards supported the 1ghz third, while kx133 won't run 800mhz and above

My Epox 7KXA Rev. 0.4 with VIA KX133 runs fine with the 750Mhz T-Bird, also overclocked to 825 MHz (110MHz FSB). No chance to get this board working with a 1 Ghz T-Bird CPU. The same 1 GHz CPU works fine on a Asus K7M 1.04 and Asus K7V-T.