VOGONS


AMD K7 "goldfinger" cloning thread

Topic actions

Reply 220 of 234, by NostalgicAslinger

User metadata
Rank Newbie
Rank
Newbie
Mel427 wrote on 2021-08-27, 20:59:

thanks will try this, but i thought the stock clock rate for thunderbird cpu´s is 1.75v? http://www.bytes-and-more.de/tech/amd_athlon_tech.htm

I have it unvervolted for a longer life/less heat, because the 1GHz version is very rare and expensive at this days.

Yes, 1,75V is the stock voltage for the 900-1000 MHz T-Bird Slot A. The 650-850MHz T-Bird Slot A version has a stock voltage of 1,7V.

Reply 221 of 234, by the_patchelor

User metadata
Rank Newbie
Rank
Newbie

Using a 800Mhz Orion (as it has a 1000Mhz core) with GFD.

1Ghz is enough 😉

(1100 tested but I don't wanna stress the old hardware too much)

CFF180D2-2560-4DBB-B485-B44FE05B2E77.jpeg
Filename
CFF180D2-2560-4DBB-B485-B44FE05B2E77.jpeg
File size
212.36 KiB
Views
243 views
File license
Fair use/fair dealing exception
8F52A813-2F34-4033-A63C-24D6523E5994.jpeg
Filename
8F52A813-2F34-4033-A63C-24D6523E5994.jpeg
File size
103.92 KiB
Views
243 views
File license
Fair use/fair dealing exception

Reply 222 of 234, by bloodem

User metadata
Rank Oldbie
Rank
Oldbie

Nice! Did you have to force a different divider? 😀

2 x PGA132 / 5 x Socket 3 / 9 x Socket 7 / 8 x SS7 / 12 x Socket 8 / 11 x Slot 1 / 3 x Slot A
5 x Socket 370 / 8 x Socket A / 2 x Socket 478 / 2 x Socket 754 / 3 x Socket 939 / 4 x LGA775 / 1 x LGA1155
Current rig: Ryzen 5 3600X
Backup rig: Core i7 7700k

Reply 224 of 234, by bloodem

User metadata
Rank Oldbie
Rank
Oldbie
the_patchelor wrote on 2021-08-29, 17:23:

switched cache to 2/5 (soldering iron...)

Ugh, forgot to reply. 😀
I think you meant 1/3, right? Because the 800 MHz Athlon already has the 2/5 divider by default, and the cache will generally not work at 400 MHz and definitely not beyond that.

Anyway, in other news, I came across a weird motherboard: the Acorp 7VIA71A ver 1.1 (with the VIA KX133 chipset)
I tested various CPUs with the Goldfinger device, including one of my modified 700 MHz CPUs with cache divider switched from 1/2 to 1/3 (and this CPU works great on AMD Irongate 75x motherboards at up to 1 GHz).
The behavior is very strange: the motherboard does not POST ("- - - -" on the POST card) when setting the CPU at 850 / 900 / 950 MHz.
However It will POST just fine at 700 / 750 / 800 / 1000 MHz 😁

Has anyone encountered something like this?
I mean, it's great that the board does post with a 1 GHz overclock and it's fast and quite stable (albeit, not as stable as my AMD 75x based boards, but faster than them), but I'm wondering what exactly might be causing this.

2 x PGA132 / 5 x Socket 3 / 9 x Socket 7 / 8 x SS7 / 12 x Socket 8 / 11 x Slot 1 / 3 x Slot A
5 x Socket 370 / 8 x Socket A / 2 x Socket 478 / 2 x Socket 754 / 3 x Socket 939 / 4 x LGA775 / 1 x LGA1155
Current rig: Ryzen 5 3600X
Backup rig: Core i7 7700k

Reply 225 of 234, by snufkin

User metadata
Rank Member
Rank
Member
bloodem wrote on 2021-09-24, 06:58:

Has anyone encountered something like this?
I mean, it's great that the board does post with a 1 GHz overclock and it's fast and quite stable (albeit, not as stable as my AMD 75x based boards, but faster than them), but I'm wondering what exactly might be causing this.

Which (if any) goldfinger device are you using? I'm wondering if this might be a rare (as in first documented) example of a passive one actually causing a problem, but could only test that by seeing if the board booted after setting the CPU at 850 using either a powered goldfinger or by changing the resistors on the CPU PCB.

Reply 226 of 234, by bloodem

User metadata
Rank Oldbie
Rank
Oldbie
snufkin wrote on 2021-09-24, 09:03:

Which (if any) goldfinger device are you using? I'm wondering if this might be a rare (as in first documented) example of a passive one actually causing a problem, but could only test that by seeing if the board booted after setting the CPU at 850 using either a powered goldfinger or by changing the resistors on the CPU PCB.

I'm actually using powered Goldfinger devices (don't have any passive device) - see attachments.
I've tested multiple golfingers (all powered) and the result on that specific Acorp motherboard is the same: does not boot when the Goldinger is configured for 850/900/950 MHz. On the other hand, on a Gigabyte 7iXE board (AMD Irongate 751), all frequencies work properly.

Yeah, I actually was thinking of changing the resistors for multiplier/voltage on the CPU PCB - especially because the CPU die is actually a 900 MHz part (even though it was sold by AMD as 700 MHz). I'm guessing that this would work, my assumption is that there's some weird incompatibility with the Goldfinger device. Although it's strange, since the motherboard should not matter much in this equation. I would understand if these multipliers would fail on ALL motherboards... but, no.

Attachments

  • 20210924_121652.jpg
    Filename
    20210924_121652.jpg
    File size
    601.24 KiB
    Views
    149 views
    File license
    Fair use/fair dealing exception
  • 20210924_121711.jpg
    Filename
    20210924_121711.jpg
    File size
    911.65 KiB
    Views
    149 views
    File license
    Fair use/fair dealing exception

2 x PGA132 / 5 x Socket 3 / 9 x Socket 7 / 8 x SS7 / 12 x Socket 8 / 11 x Slot 1 / 3 x Slot A
5 x Socket 370 / 8 x Socket A / 2 x Socket 478 / 2 x Socket 754 / 3 x Socket 939 / 4 x LGA775 / 1 x LGA1155
Current rig: Ryzen 5 3600X
Backup rig: Core i7 7700k

Reply 227 of 234, by Kimseo

User metadata
Rank Newbie
Rank
Newbie

Hi guys, I am newbie here. I am old hardware collector from Czech Republic. I started to make my collection 4 years ago. My first computer was ATARI 800XE so in my collection there are mainly 8-bit and 16-bit computers. Last year I turned to PC platform because I like 90's computers and the 90's feel a lot. In my collection there are ESCOM, Autocont, Palladium, ASI, and other PC brands. I also have small collection of processors. I decided to build some top PC based on AMD K7 processor. I also found informations about overclocking tool GoldFinger made by Ninja Micro. Unfortunately I didn't find any on eBay. Today I found your forum thread about new production. I would like to ask you if there is a possibility to buy any for me and my friends around me. #gex85
Thank you for reply. Enjoy your old hardware.
KimSeo

Reply 228 of 234, by feipoa

User metadata
Rank l33t++
Rank
l33t++

Do we have any generalities/statistics as to how far which Athlon CPUs will overclock to, e.g. a 600 MHz Orion generally does 750 MHz without issue, or a 850 MHz T-bird does 1000 MHz? Or maybe someone has noticed that they cannot overclock their 950 to 1000 MHz with stable operation.

After I just witnessed somebody paying $260 for a broken 1 GHz slot A, T-bird, I remembered that I also have a dead 1 GHz T-bird. I was thinking I should overclock a 900 T-bird to 1000 MHz and stick it in the casing for my dead 1 GHz T-bird, that is assuming it tests stable. In this market, I figure this should double it's value after a few years.

Ultimate 486 Benchmark | Ultimate 686 Benchmark | Cyrix 5x86 Enhancements | 486 Overkill Graphics | Worlds Fastest 486

Reply 229 of 234, by bloodem

User metadata
Rank Oldbie
Rank
Oldbie
feipoa wrote on 2021-10-14, 10:39:

Do we have any generalities/statistics as to how far which Athlon CPUs will overclock to, e.g. a 600 MHz Orion generally does 750 MHz without issue, or a 850 MHz T-bird does 1000 MHz? Or maybe someone has noticed that they cannot overclock their 950 to 1000 MHz with stable operation.

There aren't any rules. The best way to know how far you can push it is to open it up and check the CPU die. Many times the CPU die will actually be for a higher tier CPU. Like, you could have a 600 MHz Pluto that is actually a 900 MHz part under the hood.
In that case, all you'd have to do is swap one (maybe two) resistors on the CPU PCB in order to switch the cache divider from 1/2 to 1/3 and there's a good chance that you could reach 1 GHz, with very little voltage increase.

When it comes to Slot A Thunderbirds, I have 3 x 700 MHz, and they are all very poor overclockers, but it's not enough data to draw a conclusion 😀.

feipoa wrote on 2021-10-14, 10:39:

[...]I was thinking I should overclock a 900 T-bird to 1000 MHz and stick it in the casing for my dead 1 GHz T-bird, that is assuming it tests stable. In this market, I figure this should double it's value after a few years.

That... wouldn't really be fair. Unless the CPU die is actually a 1 GHz CPU die, it will never be a 'real' 1 GHz processor. And, even so, it would still be a bit shady. 😁
People pay a lot of money for the 1 GHz Orion parts because of their history and the fact that they were the first CPUs to hit that magical mark.

Last edited by bloodem on 2021-10-14, 11:13. Edited 1 time in total.

2 x PGA132 / 5 x Socket 3 / 9 x Socket 7 / 8 x SS7 / 12 x Socket 8 / 11 x Slot 1 / 3 x Slot A
5 x Socket 370 / 8 x Socket A / 2 x Socket 478 / 2 x Socket 754 / 3 x Socket 939 / 4 x LGA775 / 1 x LGA1155
Current rig: Ryzen 5 3600X
Backup rig: Core i7 7700k

Reply 230 of 234, by Doornkaat

User metadata
Rank Oldbie
Rank
Oldbie
feipoa wrote on 2021-10-14, 10:39:

After I just witnessed somebody paying $260 for a broken 1 GHz slot A, T-bird, I remembered that I also have a dead 1 GHz T-bird. I was thinking I should overclock a 900 T-bird to 1000 MHz and stick it in the casing for my dead 1 GHz T-bird, that is assuming it tests stable. In this market, I figure this should double it's value after a few years.

The chip is marked. I recon anyone willing to pay collectors' prices will notice.

Reply 231 of 234, by feipoa

User metadata
Rank l33t++
Rank
l33t++

I wouldn't be selling it, but putting it in my slot A system running at 1 GHz. Currently, I have it at 950 MHz. According to CPUShack (CPU-World), there's no physical difference between the 900 and 1000 MHz T-bird; just binning.... but if the chips are marked under the cover, I'll have to make a note of this and put a sticker on the case as such. I didn't think they were marked, but I haven't checked.

Ultimate 486 Benchmark | Ultimate 686 Benchmark | Cyrix 5x86 Enhancements | 486 Overkill Graphics | Worlds Fastest 486

Reply 232 of 234, by bloodem

User metadata
Rank Oldbie
Rank
Oldbie
feipoa wrote on 2021-10-14, 11:18:

I wouldn't be selling it, but putting it in my slot A system running at 1 GHz. Currently, I have it at 950 MHz. According to CPUShack (CPU-World), there's no physical difference between the 900 and 1000 MHz T-bird; just binning.... but if the chips are marked under the cover, I'll have to make a note of this and put a sticker on the case as such. I didn't think they were marked, but I haven't checked.

The CPU die is marked, see the pictures I added in my post here.

2 x PGA132 / 5 x Socket 3 / 9 x Socket 7 / 8 x SS7 / 12 x Socket 8 / 11 x Slot 1 / 3 x Slot A
5 x Socket 370 / 8 x Socket A / 2 x Socket 478 / 2 x Socket 754 / 3 x Socket 939 / 4 x LGA775 / 1 x LGA1155
Current rig: Ryzen 5 3600X
Backup rig: Core i7 7700k

Reply 233 of 234, by feipoa

User metadata
Rank l33t++
Rank
l33t++

That looks to be an Orion, but if the T-birds are also marked like this, then there isn't much point in swapping the case over to the 900. In such case, I'll just leave the 900 casing on the 900 and clock it to 1000 MHz, that is, assuming it can handle 1000 MHz.

Ultimate 486 Benchmark | Ultimate 686 Benchmark | Cyrix 5x86 Enhancements | 486 Overkill Graphics | Worlds Fastest 486

Reply 234 of 234, by bloodem

User metadata
Rank Oldbie
Rank
Oldbie
feipoa wrote on 2021-10-14, 12:02:

That looks to be an Orion, but if the T-birds are also marked like this, then there isn't much point in swapping the case over to the 900. In such case, I'll just leave the 900 casing on the 900 and clock it to 1000 MHz, that is, assuming it can handle 1000 MHz.

It's a Pluto. Orions were only 900 MHz/1GHz parts. Anyway, not important, just pointing it out. 😀
Yes, don't have any "naked" picture of my Slot A Thunderbirds, but I do remember their die was marked too.

2 x PGA132 / 5 x Socket 3 / 9 x Socket 7 / 8 x SS7 / 12 x Socket 8 / 11 x Slot 1 / 3 x Slot A
5 x Socket 370 / 8 x Socket A / 2 x Socket 478 / 2 x Socket 754 / 3 x Socket 939 / 4 x LGA775 / 1 x LGA1155
Current rig: Ryzen 5 3600X
Backup rig: Core i7 7700k