GA-586ATV & ECS P5VX-Be:Fixing CPU Table for AMD K6-2

Don't ask how to over-clock.
KachiWachi
The New Guy
Posts: 1451
Joined: Fri Mar 29, 2002 10:32 pm
Location: Pennsylvania, USA

Oops...there was a typo earlier - GA-586ATV. :oops:

Contact Jan anyway...see what he can do for you.

Thanks.
BiosMan
The BIOS Patcher
Posts: 79
Joined: Mon May 20, 2002 9:20 pm
Location: The Netherlands
Contact:

Beta testing on the patched GA-586ATV Ver 1.28 BIOS is completed now and this BIOS can be downloaded via
http://home.hccnet.nl/j.steunebrink/Atv128j2.zip
I will add this link to my k6plus page as well, on the next update.

I found a lot of things to fix. For details see the included patch.txt file.

KenOath,
Let me know if this is the BIOS you need, or if you need a patched GA-586ATS BIOS.

Jan.
The Unofficial AMD K6-2+ / K6-III+ page
http://web.inter.nl.net/hcc/J.Steunebrink/k6plus.htm
KenH
Chip off the ol' block
Posts: 110
Joined: Wed Mar 30, 2005 7:11 pm

BiosMan wrote: KenOath,
Let me know if this is the BIOS you need, or if you need a patched GA-586ATS BIOS.

Jan.
It was the GA586ATS I was originally after, Googles search led
me here, only to find that there was a typo in one of the posts...
However I do have an atv board so thanks for that patch above...
But yeah, still wanting to run a K6 on the ATS board...

You know i've been using a PC-Chips M550 board for 6 years now & was
really hoping I could run more than the 233MMX I had on it, it
died on me 3 weeks ago & now you're looking for beta testers of the
bios for a K6 cpu, :roll: , typical... I noticed you asked for testers on your
bios page...
The M550 board was the fussiest board I had ever used as far as ram
compatibility, very fussy on what works without giving explorer errors
in Win98 when more than 32mb of ram used regardless if 72 or 168pinSD..
However 168pin EDO worked perfect on this board if more than 32mb used,
never had one explorer crash or error using this ram, ever...
Sorry to get off topic, I don't see you here much though...
BiosMan
The BIOS Patcher
Posts: 79
Joined: Mon May 20, 2002 9:20 pm
Location: The Netherlands
Contact:

It took me a bit longer than expected but the GA-586ATS Ver.1.32 BIOS is also patched now. I found about the same things to fix as in the 586ATV BIOS: K6-2CXT, K6-III, K6-2+, K6-III+, and Cyrix MII 4.0x multiplier support. I also fixed the 32GB and 64GB bugs for 128GB HDD support.

KenOath, I send you a PM with the patched BIOS and test instructions.
If anyone else likes to beta test this 586ATS BIOS, please send me an e-mail.

Jan.
The Unofficial AMD K6-2+ / K6-III+ page
http://web.inter.nl.net/hcc/J.Steunebrink/k6plus.htm
KenH
Chip off the ol' block
Posts: 110
Joined: Wed Mar 30, 2005 7:11 pm

BiosMan wrote:It took me a bit longer than expected but the GA-586ATS Ver.1.32 BIOS is also patched now. I found about the same things to fix as in the 586ATV BIOS: K6-2CXT, K6-III, K6-2+, K6-III+, and Cyrix MII 4.0x multiplier support. I also fixed the 32GB and 64GB bugs for 128GB HDD support.

KenOath, I send you a PM with the patched BIOS and test instructions.
If anyone else likes to beta test this 586ATS BIOS, please send me an e-mail.

Jan.
Hey i'm back, never got the PM though...
The place became that dead there for a while that I forgot about it...
I still have the board if you still have the bios handy, no probs if
you don't though...
BiosMan
The BIOS Patcher
Posts: 79
Joined: Mon May 20, 2002 9:20 pm
Location: The Netherlands
Contact:

KenOath,

Thanks for reporting back!
I send you the PM with patched GA-586ATS BIOS again.
Strange that you didn't get it... :?
I used the e-mail address from yr signature.

Jan.
The Unofficial AMD K6-2+ / K6-III+ page
http://web.inter.nl.net/hcc/J.Steunebrink/k6plus.htm
KenH
Chip off the ol' block
Posts: 110
Joined: Wed Mar 30, 2005 7:11 pm

Well, I have 2 revisions of this board, 1A & 2A.
1A has the cache tag ram ( I think it's called ) showing up as 256kb
of asynchronous.
2A has the sram showing as 256kb of pipelined burst.
Tried this bios on both & only managed to run any K6-2 cpu at 233mhz.
I tried a 300afr-66, 333afr-66, 366afr-66, 300/333/366/400
all the same result, showed only "AMD K6/2-233" cpu on screen but
rarely got through the full counting of the ram, usually halted at some
stage of counting the ram, except for the K6-2/366 cpu, got all the way
through counting the ram, but only on a cold start, on a reboot would
freeze the same whilst counting the ram...
also tried a couple of cyrix MII cpu's, all showed up as MII-233.

Also tried a K6/266 cpu but wouldn't fire up, I have 4 of them which I know
they work OK...
I was rather amazed that this board was able to regulate the voltage to
cater for the K6 cpu as i've yet to see an FX chipset board do that...
Is there anything else you'd like me to do/try...
I also tried different timings of EDO & fast page ram on it, thinking it may
have been responsible for the ram count freezing...

Thanks...
cp
BIOS Guru
Posts: 1914
Joined: Mon Oct 21, 2002 9:07 pm
Location: Germany

ahmm..just for your information: all K6 after the K6 233 (e.g. all K6 faster than 233MHz and all K6-2) need a core voltage of 2.2V
and i'm sure that the GA-586ATS with an Intel 430FX chipset only has a linear regulator for the CPU voltage. that means two things:
1. it is not designed to supply 2.2V out of the box. my guess: 3.3V and 2.8V
2. even if you would modify the regulator to supply 2.2V, it would have to burn (really, it DOES!) (5V-2.2V) * CPU current = 2.8V * CPU current. so you'd need a cooler that could handle more heat than the CPU dissipates. additionally most linear regulators are only able to supply around 7-8A current which is not enough for any K6 beyond 200MHz.

so if the RAM count didn't succeed i'd say the regulator or the cpu (probably both) got as hot as hell. and for the K6 266: maybe you already fried them with the 2.8V core voltage as they only need 2.2V
If you email me include [WIMSBIOS] in the subject.
KenH
Chip off the ol' block
Posts: 110
Joined: Wed Mar 30, 2005 7:11 pm

cp wrote:ahmm..just for your information: all K6 after the K6 233 (e.g. all K6 faster than 233MHz and all K6-2) need a core voltage of 2.2V
Yes I know this, & was amazed to see that both revisions 1a & 2a of the
boards had a voltage setting via jumpers that drops to 2.5v, yet iv'e
seen K6/2 cpu's run OK on that...
cp wrote:so if the RAM count didn't succeed i'd say the regulator or the cpu (probably both) got as hot as hell. and for the K6 266: maybe you already fried them with the 2.8V core voltage as they only need 2.2V
No I tested them in the open board layout I have here for testing both
cpu's & ram, they still fired the system up fine...
& the voltage regulators barely warmed up, I was continually checking on
them...

I never came here expecting a bios update allowing for a K6 CPU as I
never thought it possible on this chipset, but since the board clearly had
jumpers on it with cpu options for both 54c & 55c intel cpu's, I figured it'd
be handy if I could at least run a 233mmx on it, but for some reason,
even the latest revision of bios from gigabyte allows the mmx cpu's to
run on it, but all of them detect as a 166 cpu regardless of the jumper
settings, 3 x for the 200mhz, & 1.5 x for the 233mhz..
yet it runs the K6's at 233...

Hell i'm amazed that some VX chipset boards have the voltage regulation
on them from factory allowing for modded bios's to run K6 cpu's, I never
dreamed that an FX chipset would even fire up on a K6 let alone boot
completely, as it did with the 366 in it...

But hey, I now have a bios allowing for a huge hard drive, thanks Jan, I
hope you didn't do this just on my seeking a bios for this board, i'd feel
pretty low if you did...
cp
BIOS Guru
Posts: 1914
Joined: Mon Oct 21, 2002 9:07 pm
Location: Germany

to get the missing multipliers on the intel cpus you have to add pullup resistors to the BF0/1 pins of the cpu socket. either they were not required during the early days of the Socket 7 or they were omitted to save some money.

the regulator can supply 2.5V but has to burn the other 2.5V (if fed with 5V) so the regulator isn't designed to supply high currents. if you'd supply the markings on the regulator it would be easy to help with the cpu selection. the problem that still persists: the regulator has to burn the same amount of current (multiplied with the voltage drop) that the cpu demands which leads us to a power dissipation of 28W (2.8V * 10A) for the regulator when you install a K6-2 2.2V which draws as much as 10A, while the CPU itself would dissipate 22W only.

well...the socket 7 chipsets are all compatible to all socket 7 cpus or i should rather say: the socket 7 cpus are all compatible to the socket 7 chipsets because they were designed to ;) you could run a Pentium 90 or a Cyrix 6x86 PR100 (with 80MHz) in a Socket 7 Aladdin V or VIA MVP3 board. or take a K6-III+ on a journey with the intel 430FX (switching regulator on a VRM module helps with the cpu voltage and current).
If you email me include [WIMSBIOS] in the subject.
KachiWachi
The New Guy
Posts: 1451
Joined: Fri Mar 29, 2002 10:32 pm
Location: Pennsylvania, USA

If you really want to run one of these CPU's, and your on-board regulator cannot supply the required current, you need to look for an "in CPU socket" power supply adapter...such as a PowerLeap.

You also want to look for the + series CPU's (they use less current).
KenOath wrote:Hell i'm amazed that some VX chipset boards have the voltage regulation on them from factory allowing for modded bios's to run K6 cpu's, I never dreamed that an FX chipset would even fire up on a K6 let alone boot completely, as it did with the 366 in it...
Check out my DFI (see my sig).

Thanks.
cp
BIOS Guru
Posts: 1914
Joined: Mon Oct 21, 2002 9:07 pm
Location: Germany

about the Pentium MMX multiplier problem. this is the official reason:
"The Pentium processor with MMX technology differs from the Pentium processor in that its BF0 has an internal pulldown resistor, whereas the Pentium processor's BF0 has an internal pullup resistor. Some motherboard vendors may rely upon their understanding of the Pentium processor, and may not incorporate circuitry necessary to operate the Pentium processor with MMX technology at all required frequencies."

and here is the fix (with pictures):
http://www.ga586hx.de/de/reworks.html

also applicable to any other Socket 5 or Socket 7 board.
If you email me include [WIMSBIOS] in the subject.
KenH
Chip off the ol' block
Posts: 110
Joined: Wed Mar 30, 2005 7:11 pm

cp wrote: if you'd supply the markings on the regulator it would be easy to help with the cpu selection.
On the revision 1a board with a heatsink under it - it's

LT1585
CT
9531

On the rev 2a board with a heatsink under it - it's

EZ1084CT
9549
15632

There's another next to each of them that reads.
Rev 1a
EZ1086CT
9520

Rev2a
LT1587
CT
9519
KachiWachi wrote:If you really want to run one of these CPU's, and your on-board regulator cannot supply the required current, you need to look for an "in CPU socket" power supply adapter...such as a PowerLeap.
Yeah I knew about the powerleaps, I didn't really feel like spending
any money on such an old board though, not when I get given
50+ units for free most months, from a friend who works at a
recycling station...
My collection of PIII systems now is pretty good..
I just like running old systems a speeds that most people won't beleive...
cp
BIOS Guru
Posts: 1914
Joined: Mon Oct 21, 2002 9:07 pm
Location: Germany

rev.1a:
LT1585: 4,6A max. @ CPU voltage
EZ1086CT: 1.5A max. @ chipset/dram voltage

rev.2:
EZ1084CT: 5A max. @ CPU voltage
LT1587: 3A max. @ chipset/dram voltage

AMD K6 166: 6.25A
Cyrix 6x86 PR166 (133MHz) : 6.6A
Cyrix 6x86MX PR200 (166MHz) : 6.7A
Intel Pentium 200 : 4.6A
Intel Pentium MMX 166 : 4.75A
Intel Pentium MMX 200 : 5.7A
(all max. currents)

go figure for yourself.
If you email me include [WIMSBIOS] in the subject.
KenH
Chip off the ol' block
Posts: 110
Joined: Wed Mar 30, 2005 7:11 pm

cp wrote: go figure for yourself.
I never really believed they would run it in the 1st place as it is...
When I'd originally posted here in july last year, I had what looked like
the latest revision of this board, that had jumper settings that I really
thought wasn't possible, so for the sake of trying I did a quick search
& landed on this thread...
That board has long been scrapped for it's bios chip, & I have since
landed another 2 boards of earlier revisions...
Then shortly after, Jan had emailed me stating he'd modded a bios
for the board...
I pulled these 2 boards from the pile I have here, ready to be scrapped
& gave them a go...
Here's the next pile to be scrapped for thier chips, & perhaps capacitors...

Image
Post Reply