crazytiti wrote on 2024-11-27, 16:52:

Hello did you see this guys selling modded K6-2+ 570 on ebay ?
114$ for one already modded and tested !
https://www.ebay.fr/itm/166591461971?_trksid= … c101782.m155878

Why not buy 3 x K6-2+ 570 for $60 and mod them yourself? 😀
PS: The seller accepts $60 offers for a pack of 3.

It say 69 + 22$ shipping cost (i'm in france)
I have linked the ad for the people who do not want to mess with their cpu.

Edit : oh sorry i just understand for the 60$
I may buy 3 just to have one for me. (even i don't have a super 7 motherboard)
And resell the other to some french people...

crazytiti wrote on 2024-11-27, 16:52:

Hello did you see this guys selling modded K6-2+ 570 on ebay ?
114$ for one already modded and tested !
https://www.ebay.fr/itm/166591461971?_trksid= … c101782.m155878

Delete

bloodem wrote on 2024-11-27, 17:56:

crazytiti wrote on 2024-11-27, 16:52:

Hello did you see this guys selling modded K6-2+ 570 on ebay ?
114$ for one already modded and tested !
https://www.ebay.fr/itm/166591461971?_trksid= … c101782.m155878

Why not buy 3 x K6-2+ 570 for $60 and mod them yourself? 😀
PS: The seller accepts $60 offers for a pack of 3.

Weirdly, they don't post to the UK... I have previously enquired

I can post to uk from france, do you want one or two cpu ?

crazytiti wrote on 2024-11-28, 07:00:

I can post to uk from france, do you want one or two cpu ?

already have one coming in the post, thanks. This time won't mod the chip so that I can complete my GA-5AX rev. 4.1 DOS/98 build

Just tried the mod on my K6-2+ 500ACZ and it worked great, stable @ 550Mhz and 2.0V

Question, I just removed the 0 ohm resistor instead of moving it to the side since it was indicated before in this thread that it could work this way, and indeed it did.

Any advantage in reinstalling the resistor on the side position or leaving it open is exactly the same thing?

PS.: I have another K6-2+, a 400Mhz 1.6V, bought it from the seller in US that offers it for $20. Will try to mod it later this week.

Not aware of a disadvantage to doing it that way. No one that I know of has complained. I like to put it to where it should be like AMD did it though. Personally

Sphere478 wrote on 2025-01-22, 20:33:

Not aware of a disadvantage to doing it that way. No one that I know of has complained. I like to put it to where it should be like AMD did it though. Personally

Thank you Sphere, I decided to leave it open on my first unit and it's been working great, so I guess it doesn't make any difference.

I just modded my second CPU today, it's one of those K6-2+ 400ATZ been sold on ebay US for $20. Worked great and I can run it stable at 550Mhz set to 2.0V
Just wanted to share here cause it seems like a great deal, if you live in US.

Pino wrote on 2025-01-30, 22:42:

Thank you Sphere, I decided to leave it open on my first unit and it's been working great, so I guess it doesn't make any differ […]
Show full quote

Sphere478 wrote on 2025-01-22, 20:33:

Not aware of a disadvantage to doing it that way. No one that I know of has complained. I like to put it to where it should be like AMD did it though. Personally

Thank you Sphere, I decided to leave it open on my first unit and it's been working great, so I guess it doesn't make any difference.

I just modded my second CPU today, it's one of those K6-2+ 400ATZ been sold on ebay US for $20. Worked great and I can run it stable at 550Mhz set to 2.0V
Just wanted to share here cause it seems like a great deal, if you live in US.

I'm currently working on a Soyo SY-5BT, to get the VCORE voltages below 2.2V, I have a K6-2+ 400 1.6V, and I want to use it there with multi 6X and FSB 68, it's a secret setting (called Turbo), I would gain a little more than 10 MHz, but if I wanted to I could even get to 500, selecting 83 FSB, I've never made a change like that, also because I only have that CPU, too bad I didn't know it so many years ago, then they were little and you could risk some CPU, the biggest difficulty I don't know if it's removing the lid or the resistor (or capacitor?).

Has anyone here had any failures with this mod?
I dug out 3 I bought ages ago and put in the loft when life got busy, modded them and no joy. Post, but hang upon trying to boot into Windows 98.
As soon as I return them to K6-2+, they work fine. All came from the US seller, selling lots of three. I know I have run this PC with a modded K6-2+ to K6-3 before with no issues, but don't currently have a previous one to try. Trying to decide whether something is wrong with the PC, or these CPU's!
Cheers
Skip

If they were sold as k6-2+ instead of 3 there was a reason for sure... it would be dreaming to think that every will work with all the cache enabled.

Nemo1985 wrote on 2025-02-10, 21:50:

If they were sold as k6-2+ instead of 3 there was a reason for sure... it would be dreaming to think that every will work with all the cache enabled.

My hunch would be this would be especially true with earlier batches of them. It seems likely that the K6-2+ 570s that have been so successful with this mod are later ones. I was much less confident of the outcome when I tried it with a K6-2+ 533, but I was lucky with that one as well. All of the effort we spend to have the fastest slow computer.

it seems nearly every 570 mods fine, but the lower the mhz the more likely to find a dud. I think the 1.6v ones usually are fine though as I think those were kinda higher bin anyway. but, yeah. sometimes it doesn't work. I recommend modding 570s only as they are plentiful, or at leas they were, are they still? the rest are collectable. and I hate to see em modded.

Btw, I've been keeping the list in OP up to date. check it out.

End of the day it's a lottery.
Binning is the same as grading anything.
Even rice has different grades.

As with any grading system, premium grade can be sold as a lower grade if that's what the market is willing to pay for.

You can't sell "animal feed" grade rice as human grade though.
Well, not for long.

With chips, it's the same. The gamble is that yields were good enough that many (not all) premium chips were sold as mid-grade.

Many of you are winning this gamble. Some won't.

Apologies, I should clarify!
This wasn't a moan, just curiosity.
All of the K6-2+s I've worked on have been the 570Mhz ones from that US seller. I had probably done 6 or 9 before that worked perfectly, 100% success rate, and then got 3 bad ones in a batch. Was just curious, really!
Cheers
Skip

Skip94 wrote on 2025-02-11, 06:41:

Apologies, I should clarify! This wasn't a moan, just curiosity. All of the K6-2+s I've worked on have been the 570Mhz ones from […]
Show full quote

Apologies, I should clarify!
This wasn't a moan, just curiosity.
All of the K6-2+s I've worked on have been the 570Mhz ones from that US seller. I had probably done 6 or 9 before that worked perfectly, 100% success rate, and then got 3 bad ones in a batch. Was just curious, really!
Cheers
Skip

Sorry to hear that, that's definitely bad news and some bad luck for you. My working assumption was that the 570s were late enough that they were disabling cache mostly for marketing rather than out of necessity, similar to the Celeron.

Still I don't understand the meaning of the curiosity... as if every k6-2+ MUST be able to become 3+? Why would people think that?
Maybe there are enough cpus able to work a 3+ but at the same time it's possible that some of them had a part of the cache non working properly and for this reason they have been labelled as 2+.
It's the very same thing that you do when overclocking, some cpus are go higher than others, in this case the difference is that the possibility is just yes\no.

As a period correct example, most of the Pentium 166 mmx were able to reach 233, does it means EVERY CPU is able to do so? Obviously not...

Nemo1985 wrote on 2025-02-11, 06:53:

Still I don't understand the meaning of the curiosity... as if every k6-2+ MUST be able to become 3+? Why would people think that?
Maybe there are enough cpus able to work a 3+ but at the same time it's possible that some of them had a part of the cache non working properly and for this reason they have been labelled as 2+.
It's the very same thing that you do when overclocking, some cpus are go higher than others, in this case the difference is that the possibility is just yes\no.

In my case, I modded more than 10 of these, and they all worked perfectly as K6-3+, except for one.
So I can definitely understand Skip94's dilemma; at this point I would also think it's strange to have an entire batch of 570ACZ CPUs that fail to run when enabling the full 256k L2 cache.

Basically, based on our previous experience, these have a 90+% success rate, so having 3 consecutive duds is what constitutes the "curiosity" factor, since it's an atypical behavior.

In layman’s terms: if these mods work almost every time, getting three failures in a row is like flipping a coin and getting heads ten times - it’s not impossible, but it’s weird enough to make you wonder if something else is going on.

bloodem wrote on 2025-02-11, 07:03:

In my case, I modded more than 10 of these, and they all worked perfectly as K6-3+, except for one. So I can definitely understa […]
Show full quote

Nemo1985 wrote on 2025-02-11, 06:53:

Still I don't understand the meaning of the curiosity... as if every k6-2+ MUST be able to become 3+? Why would people think that?
Maybe there are enough cpus able to work a 3+ but at the same time it's possible that some of them had a part of the cache non working properly and for this reason they have been labelled as 2+.
It's the very same thing that you do when overclocking, some cpus are go higher than others, in this case the difference is that the possibility is just yes\no.

In my case, I modded more than 10 of these, and they all worked perfectly as K6-3+, except for one.
So I can definitely understand Skip94's dilemma; at this point I would also think it's strange to have an entire batch of 570ACZ CPUs that fail to run when enabling the full 256k L2 cache.

Basically, based on our previous experience, these have a 90+% success rate, so having 3 consecutive duds is what constitutes the "curiosity" factor, since it's an atypical behavior.

In layman’s terms: if these mods usually work almost every time, getting three failures in a row is like flipping a coin and getting heads ten times—it’s not impossible, but it’s weird enough to make you wonder if something else is going on.

I see your point but we don't know how many cpus have been produced in that batch (are they all coming from the same?) let's say they produced 1000 cpus in that batch? There are 100 cpus will not work, that considering your 90% ratio which seems quite optimistic to be honest.
We are sure that not every k6-2+ cpu can be modded but I never had such doubt personally...