VOGONS


First post, by Sphere478

User metadata
Rank l33t++
Rank
l33t++

I’m trying to put together a catalog of the different configurations found under k6 cpus in the hopes that another mod may have gone unnoticed.

First I need to figure out how many different configurations there were, and once we have images of them we may be able to learn something more.
-

-

-
K6-2 333amz

4E5CAAA8-4344-40CB-92B4-7E7C6BCA933A.jpeg
Filename
4E5CAAA8-4344-40CB-92B4-7E7C6BCA933A.jpeg
File size
1.33 MiB
Views
460 views
File license
Public domain
36CE371F-3C0F-4944-AAD3-D962B696FCEF.jpeg
Filename
36CE371F-3C0F-4944-AAD3-D962B696FCEF.jpeg
File size
1.65 MiB
Views
460 views
File license
Public domain

Not expecting that any cache is hidden here, but I am wondering what the vacant pads are for? Perhaps non implemented features of some sort?

-
350 k6-2
file.php?id=132080&mode=view

-
k6-2 500

20220317_020450.jpg
Filename
20220317_020450.jpg
File size
1.53 MiB
Views
429 views
File license
Public domain

-

K6-III
file.php?id=125388&mode=view
I am suspecting that those SMDs may be configuration pads for turning the III into a 128kb version or a plain 2 with 0kb
But so far no one has tried

-

K62+/3+
We seem to know quite intimately now that there was only one difference between the 2+ and the 3+
file.php?id=132047&mode=view and that we discovered the function of this resistor which makes me curious if there may be some late model k6 + cpus that were made into non plus cpus file.php?id=132046&mode=view

Sphere's PCB projects.
-
Sphere’s socket 5/7 cpu collection.
-
SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)

Reply 1 of 4, by shamino

User metadata
Rank l33t
Rank
l33t

I'm suspicious of the lower voltage and/or higher clocked K6-2 examples if they also have late date codes. The 1.9V CPU you opened is visibly different inside though.
There's also a 2.0V rated CPU that's marked for something like 475MHz I think.
I wish these CPUs were still dirt cheap in quantity.

Did anyone ever find a delidded picture of an actual K6-3+? Just curious if the mod that's been developed actually matches how AMD configured the "official" ones.

Reply 2 of 4, by Sphere478

User metadata
Rank l33t++
Rank
l33t++
shamino wrote on 2022-03-17, 01:49:
I'm suspicious of the lower voltage and/or higher clocked K6-2 examples if they also have late date codes. The 1.9V CPU you ope […]
Show full quote

I'm suspicious of the lower voltage and/or higher clocked K6-2 examples if they also have late date codes. The 1.9V CPU you opened is visibly different inside though.
There's also a 2.0V rated CPU that's marked for something like 475MHz I think.
I wish these CPUs were still dirt cheap in quantity.

Did anyone ever find a delidded picture of an actual K6-3+? Just curious if the mod that's been developed actually matches how AMD configured the "official" ones.

The cracked one I opened, the shape of that core suggests that there is no cache in there. What the empty smd spaces are for I’ve no idea.

Yeah, fritz came along and had a de lidded 3+ and that was the last piece of the puzzle, soon as we all saw it it was obvious what needed to be done. This was alp figured out in the same thread, the thread was just re named.
SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod

Sphere's PCB projects.
-
Sphere’s socket 5/7 cpu collection.
-
SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)

Reply 3 of 4, by snufkin

User metadata
Rank Oldbie
Rank
Oldbie

On the K6-iii, might be worth looking at the pads in the top corners:

k6iii_possibleConfig.jpg
Filename
k6iii_possibleConfig.jpg
File size
38.06 KiB
Views
370 views
File license
Fair use/fair dealing exception

They're marked slightly differently, in a similar way to the k6-2+.

My guess is that the pads marks with a single tiny dot are possibly Vcore, unmarked are Gnd, 2 tiny dots might be Vio (but they never seem to have components fitted), don't know about the 3 tiny dot one. The slightly larger dots for config pads. So most components are capacitors for Vcore, with a couple of disguised 0-ohm links to pull the config pads high or low. Would need to go through them with a broken chip and a multimeter to see what's connected to what (e.g. check resistance between all pads mark with two tiny dots).

Reply 4 of 4, by Sphere478

User metadata
Rank l33t++
Rank
l33t++
snufkin wrote on 2022-03-17, 13:19:
On the K6-iii, might be worth looking at the pads in the top corners: k6iii_possibleConfig.jpg […]
Show full quote

On the K6-iii, might be worth looking at the pads in the top corners:
k6iii_possibleConfig.jpg

They're marked slightly differently, in a similar way to the k6-2+.

My guess is that the pads marks with a single tiny dot are possibly Vcore, unmarked are Gnd, 2 tiny dots might be Vio (but they never seem to have components fitted), don't know about the 3 tiny dot one. The slightly larger dots for config pads. So most components are capacitors for Vcore, with a couple of disguised 0-ohm links to pull the config pads high or low. Would need to go through them with a broken chip and a multimeter to see what's connected to what (e.g. check resistance between all pads mark with two tiny dots).

Yeah, I’m betting that those are the 0,128,256 kb cache config pads. But so far as I know AMD never released a 128k version…. But they may have released a 0kb version….🤔

Sphere's PCB projects.
-
Sphere’s socket 5/7 cpu collection.
-
SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)