VOGONS


Definition of "Super Socket 7"

Topic actions

Reply 40 of 59, by PcBytes

User metadata
Rank l33t
Rank
l33t
meljor wrote:
Well.... i almost found one! […]
Show full quote

Well.... i almost found one!

I was picking motherboards at the recycler and came across an ECS board with Sis chipset and white isa slots, socket 7.
Since the only board i have that can run Tillamook with cache looks alot like that (Ecs, Sis and lots of white) i bought it, together with lots of other boards.

At home i took another good look at it and noticed what i had missed: it has an agp slot!! It is the ECS P5SD-B revision 1.1 (NOT the P5SD-B+ which is a Via mvp3 board).

I was amazed but very happy to have found a (in my own opinion) non super 7 with an agp slot. It does support sdram and has a 100fsb setting but anything above 83/90fsb seems unstable on these boards (according to the web).

UNFORTUNATELY, before testing i wanted to make sure everything was there as it was missing a little memory lever and the socket was a bit busted up (one ''hook'' for the cooler is damaged, so only a cooler with three holes will clamp on). Then i noticed the more important: one capacitor was gone (no big deal), the heatsink of the vrm was gone (no big deal) and another part was gone next to a rectifier 😢 i have no clue what that was!!

It is very clear that i didn't take a good enough look at this board (too busy picking others) and normally i do not take home such damaged boards. But i am still very happy with this one and still have hope that it will work. But i really need to know what part is missing so i can fix her up.

Does anyone know? Or does anyone here have this board and can take a picture? Or at least look and tell me what part i need? VERY MUCH APPRECIATED! Here some pics, i'm talking about the missing part next to the Mospec S10S40C.

Best case scenario: Fixing this board and hopefully it works like a charm AND runs the Tillamook with cache for the ultimate Tillamook board with agp 😎

Okay so I tracked down an ECS motherboard that uses the same kind of layout for that side you are missing the part on and the missing part might be a 45N03L mosfet.

That and also the BIOS chip Is missing. I do have an BIOS file if you need to flash on another board.

"Enter at your own peril, past the bolted door..."
Main PC: i5 3470, GB B75M-D3H, 16GB RAM, 2x1TB
98SE : P3 650, Soyo SY-6BA+IV, 384MB RAM, 80GB

Reply 41 of 59, by lazibayer

User metadata
Rank Oldbie
Rank
Oldbie
meljor wrote:
Well.... i almost found one! […]
Show full quote

Well.... i almost found one!

I was picking motherboards at the recycler and came across an ECS board with Sis chipset and white isa slots, socket 7.
Since the only board i have that can run Tillamook with cache looks alot like that (Ecs, Sis and lots of white) i bought it, together with lots of other boards.

At home i took another good look at it and noticed what i had missed: it has an agp slot!! It is the ECS P5SD-B revision 1.1 (NOT the P5SD-B+ which is a Via mvp3 board).

I was amazed but very happy to have found a (in my own opinion) non super 7 with an agp slot. It does support sdram and has a 100fsb setting but anything above 83/90fsb seems unstable on these boards (according to the web).

UNFORTUNATELY, before testing i wanted to make sure everything was there as it was missing a little memory lever and the socket was a bit busted up (one ''hook'' for the cooler is damaged, so only a cooler with three holes will clamp on). Then i noticed the more important: one capacitor was gone (no big deal), the heatsink of the vrm was gone (no big deal) and another part was gone next to a rectifier 😢 i have no clue what that was!!

It is very clear that i didn't take a good enough look at this board (too busy picking others) and normally i do not take home such damaged boards. But i am still very happy with this one and still have hope that it will work. But i really need to know what part is missing so i can fix her up.

Does anyone know? Or does anyone here have this board and can take a picture? Or at least look and tell me what part i need? VERY MUCH APPRECIATED! Here some pics, i'm talking about the missing part next to the Mospec S10S40C.

Best case scenario: Fixing this board and hopefully it works like a charm AND runs the Tillamook with cache for the ultimate Tillamook board with agp 😎

I have a Z-shape string that can hold the heatsink to the diagonal tabs on the socket. Do you want it?

Reply 42 of 59, by meljor

User metadata
Rank Oldbie
Rank
Oldbie
PcBytes wrote:
meljor wrote:
Well.... i almost found one! […]
Show full quote

Well.... i almost found one!

I was picking motherboards at the recycler and came across an ECS board with Sis chipset and white isa slots, socket 7.
Since the only board i have that can run Tillamook with cache looks alot like that (Ecs, Sis and lots of white) i bought it, together with lots of other boards.

At home i took another good look at it and noticed what i had missed: it has an agp slot!! It is the ECS P5SD-B revision 1.1 (NOT the P5SD-B+ which is a Via mvp3 board).

I was amazed but very happy to have found a (in my own opinion) non super 7 with an agp slot. It does support sdram and has a 100fsb setting but anything above 83/90fsb seems unstable on these boards (according to the web).

UNFORTUNATELY, before testing i wanted to make sure everything was there as it was missing a little memory lever and the socket was a bit busted up (one ''hook'' for the cooler is damaged, so only a cooler with three holes will clamp on). Then i noticed the more important: one capacitor was gone (no big deal), the heatsink of the vrm was gone (no big deal) and another part was gone next to a rectifier 😢 i have no clue what that was!!

It is very clear that i didn't take a good enough look at this board (too busy picking others) and normally i do not take home such damaged boards. But i am still very happy with this one and still have hope that it will work. But i really need to know what part is missing so i can fix her up.

Does anyone know? Or does anyone here have this board and can take a picture? Or at least look and tell me what part i need? VERY MUCH APPRECIATED! Here some pics, i'm talking about the missing part next to the Mospec S10S40C.

Best case scenario: Fixing this board and hopefully it works like a charm AND runs the Tillamook with cache for the ultimate Tillamook board with agp 😎

Okay so I tracked down an ECS motherboard that uses the same kind of layout for that side you are missing the part on and the missing part might be a 45N03L mosfet.

That and also the BIOS chip Is missing. I do have an BIOS file if you need to flash on another board.

Thank you very much for the help! Someone else (by mail) said it was probably a rectifier, just the same as the one that is still on there.... How can i be sure what to order? I don't want to kill this board. My knowledge of electronics is next to non-existent 😢

asus tx97-e, 233mmx, voodoo1, s3 virge ,sb16
asus p5a, k6-3+ @ 550mhz, voodoo2 12mb sli, gf2 gts, awe32
asus p3b-f, p3-700, voodoo3 3500TV agp, awe64
asus tusl2-c, p3-S 1,4ghz, voodoo5 5500, live!
asus a7n8x DL, barton cpu, 6800ultra, Voodoo3 pci, audigy1

Reply 43 of 59, by meljor

User metadata
Rank Oldbie
Rank
Oldbie
lazibayer wrote:
meljor wrote:
Well.... i almost found one! […]
Show full quote

Well.... i almost found one!

I was picking motherboards at the recycler and came across an ECS board with Sis chipset and white isa slots, socket 7.
Since the only board i have that can run Tillamook with cache looks alot like that (Ecs, Sis and lots of white) i bought it, together with lots of other boards.

At home i took another good look at it and noticed what i had missed: it has an agp slot!! It is the ECS P5SD-B revision 1.1 (NOT the P5SD-B+ which is a Via mvp3 board).

I was amazed but very happy to have found a (in my own opinion) non super 7 with an agp slot. It does support sdram and has a 100fsb setting but anything above 83/90fsb seems unstable on these boards (according to the web).

UNFORTUNATELY, before testing i wanted to make sure everything was there as it was missing a little memory lever and the socket was a bit busted up (one ''hook'' for the cooler is damaged, so only a cooler with three holes will clamp on). Then i noticed the more important: one capacitor was gone (no big deal), the heatsink of the vrm was gone (no big deal) and another part was gone next to a rectifier 😢 i have no clue what that was!!

It is very clear that i didn't take a good enough look at this board (too busy picking others) and normally i do not take home such damaged boards. But i am still very happy with this one and still have hope that it will work. But i really need to know what part is missing so i can fix her up.

Does anyone know? Or does anyone here have this board and can take a picture? Or at least look and tell me what part i need? VERY MUCH APPRECIATED! Here some pics, i'm talking about the missing part next to the Mospec S10S40C.

Best case scenario: Fixing this board and hopefully it works like a charm AND runs the Tillamook with cache for the ultimate Tillamook board with agp 😎

I have a Z-shape string that can hold the heatsink to the diagonal tabs on the socket. Do you want it?

Thanks for the offer but i'm good. I've got a lot of coolers and also one or two like you mention, just like 486 coolers right? But i believe a good socket 370/socket A cooler with the 3 holes at each side of the clamp is perfect for this. I use the same on a Tualatin board that has the same piece snapped off the socket.

The capacitor, heatsink, cpu clamp, hotflashing a bios etc. etc. is not a problem. The only problem i have is the missing part as i have no clue. Really hope to get the right answer or even better: someone here having the same board.

asus tx97-e, 233mmx, voodoo1, s3 virge ,sb16
asus p5a, k6-3+ @ 550mhz, voodoo2 12mb sli, gf2 gts, awe32
asus p3b-f, p3-700, voodoo3 3500TV agp, awe64
asus tusl2-c, p3-S 1,4ghz, voodoo5 5500, live!
asus a7n8x DL, barton cpu, 6800ultra, Voodoo3 pci, audigy1

Reply 44 of 59, by kanecvr

User metadata
Rank Oldbie
Rank
Oldbie
feipoa wrote:

Has anyone found a regular socket 7 board which had an unofficial 100 MHz setting? And did it work?

yes. Lucky Star 5V-1A, using the VIA Apollo VPX chipset. It has unofficial jumper settings for 100MHz. At first I tought it was 95MHz, but turns out it's 100 and PCI runs at CPUCLK/3 automatically when using 83 or 100MHz FSB. It's not stable with L2 cache enabled tough.

infiniteclouds wrote:

How about a S/S7 Board that has AGP, can do 100 FSB, but can also do BELOW 66.. ?

Yes. The above mentioned board supports 7.15, 50, 55 and 60Mhz FSB as well. At 7.15Mhz FSB you get XT speeds when disabling cache or 386 speeds with cache enabled.

Reply 45 of 59, by meljor

User metadata
Rank Oldbie
Rank
Oldbie
kanecvr wrote:
yes. Lucky Star 5V-1A, using the VIA Apollo VPX chipset. It has unofficial jumper settings for 100MHz. At first I tought it was […]
Show full quote
feipoa wrote:

Has anyone found a regular socket 7 board which had an unofficial 100 MHz setting? And did it work?

yes. Lucky Star 5V-1A, using the VIA Apollo VPX chipset. It has unofficial jumper settings for 100MHz. At first I tought it was 95MHz, but turns out it's 100 and PCI runs at CPUCLK/3 automatically when using 83 or 100MHz FSB. It's not stable with L2 cache enabled tough.

infiniteclouds wrote:

How about a S/S7 Board that has AGP, can do 100 FSB, but can also do BELOW 66.. ?

Yes. The above mentioned board supports 7.15, 50, 55 and 60Mhz FSB as well. At 7.15Mhz FSB you get XT speeds when disabling cache or 386 speeds with cache enabled.

What -ns rating does the cache have on that board? And can you tell me please what is on that lucky star 5v-1a board next to the Mospec S10S40C? Is it 2x this same rectifier or has it another value?
I can see the s10s40c on pictures but can not read the other one... Thanks.

asus tx97-e, 233mmx, voodoo1, s3 virge ,sb16
asus p5a, k6-3+ @ 550mhz, voodoo2 12mb sli, gf2 gts, awe32
asus p3b-f, p3-700, voodoo3 3500TV agp, awe64
asus tusl2-c, p3-S 1,4ghz, voodoo5 5500, live!
asus a7n8x DL, barton cpu, 6800ultra, Voodoo3 pci, audigy1

Reply 46 of 59, by kanecvr

User metadata
Rank Oldbie
Rank
Oldbie

7ns as far as I know. It should run at 100Mhz... it just doesn't. Maybe it's the cache controller that doesn't like 100MHz - or the chipset itself (most likely). I've got a pretty good pic here: Lucky Star 5V-1A / 5V-1B review - scaling form XT levels to K6-II levels see if you can read off that.

Reply 47 of 59, by lazibayer

User metadata
Rank Oldbie
Rank
Oldbie
kanecvr wrote:

7ns as far as I know. It should run at 100Mhz... it just doesn't. Maybe it's the cache controller that doesn't like 100MHz - or the chipset itself (most likely). I've got a pretty good pic here: Lucky Star 5V-1A / 5V-1B review - scaling form XT levels to K6-II levels see if you can read off that.

The tag ram seems to be rated as 15ns, which is only good for 66MHz.
I couldn't read the model number on the clock generator; from other pictures on the web it seems to be ICS9169CJ-27. I guess the "tristate" setting gives you 100MHz bus speed.

Attachments

Reply 48 of 59, by meljor

User metadata
Rank Oldbie
Rank
Oldbie
kanecvr wrote:

7ns as far as I know. It should run at 100Mhz... it just doesn't. Maybe it's the cache controller that doesn't like 100MHz - or the chipset itself (most likely). I've got a pretty good pic here: Lucky Star 5V-1A / 5V-1B review - scaling form XT levels to K6-II levels see if you can read off that.

It is a pretty good picture but i still can't read it... could you please have a look and see what it says on there? I think the setup is the same and i could order that part to get my board going.

About cache: My 100mhz ss7 boards all have 5ns cache chips, and overclocking is limited on some of these boards, but even the very best hardly reaches 115fsb with cache still on (p5a/p5a-b/ga5ax).This is on Ali based boards as my Mvp3 boards do even worse or some do not have overclock settings.

I think that 6ns might be able to reach 100mhz stable on some boards but will already be stretching it?

asus tx97-e, 233mmx, voodoo1, s3 virge ,sb16
asus p5a, k6-3+ @ 550mhz, voodoo2 12mb sli, gf2 gts, awe32
asus p3b-f, p3-700, voodoo3 3500TV agp, awe64
asus tusl2-c, p3-S 1,4ghz, voodoo5 5500, live!
asus a7n8x DL, barton cpu, 6800ultra, Voodoo3 pci, audigy1

Reply 49 of 59, by deleted_Rc

User metadata

I always defined a ss7 by their ability to run k6-2 and k6-3 CPU which requires the extra pin on the socket.
The agp support and faster fsb is almost inherent to those mobos atleast on the mid to high end mobos although not always supported it doesn't fall within the criteria to be a ss7 mobo.
Although admitted no s7 mobo has support for agp as far I know (I looked for them)

Reply 50 of 59, by lazibayer

User metadata
Rank Oldbie
Rank
Oldbie
Richo wrote:

I always defined a ss7 by their ability to run k6-2 and k6-3 CPU which requires the extra pin on the socket.
The agp support and faster fsb is almost inherent to those mobos atleast on the mid to high end mobos although not always supported it doesn't fall within the criteria to be a ss7 mobo.
Although admitted no s7 mobo has support for agp as far I know (I looked for them)

K6-2 and K6-3 physically use standard socket 7 interface and even some 430HX based boards, such as T2P4, can run K6-3+.

Reply 51 of 59, by BitWrangler

User metadata
Rank l33t++
Rank
l33t++

Yah, I've run them on TX, VX and HX, and a K6 on an FX, mostly voltage support stops you.

Unicorn herding operations are proceeding, but all the totes of hens teeth and barrels of rocking horse poop give them plenty of hiding spots.

Reply 52 of 59, by mperu99

User metadata
Rank Newbie
Rank
Newbie

THIS IDIOT thinks his ASUS SP97-V 1.02 is a super socket 7 board. Clearly he cannot read. Hell even the SIS5598 chipset only supports 66Mhz (75Mhz officially) (83Mhz unofficially) and that chipset has NO AGP support.
I think because he sees super tx on the chipset heatsink he therefore claims it is super socket 7. He will argue with anyone that tries to reason with him. OH and he said AGP is nothing more than a PCI , although PCI had a major Influence on the AGP port design it is NOT a PCI port. I even sent him the manual that any one can look up and it clearly tells you what this board can do.. but still deny's

The two Major features of SUPER SOCKET 7 was the ability to support processors running at 100+Mhz bus, (like K6-II and K6-III) therefore the ability to run the FSB at 100+Mhz is part of Super 7 , the other aspect was the ability to have AGP support.. normal socket 7 does not have AGP support. This does not mean all super socket 7 boards had AGP slot.

so according to him we are all stupid.

Reply 53 of 59, by mperu99

User metadata
Rank Newbie
Rank
Newbie
lazibayer wrote on 2017-10-27, 19:40:
Richo wrote:

I always defined a ss7 by their ability to run k6-2 and k6-3 CPU which requires the extra pin on the socket.
The agp support and faster fsb is almost inherent to those mobos atleast on the mid to high end mobos although not always supported it doesn't fall within the criteria to be a ss7 mobo.
Although admitted no s7 mobo has support for agp as far I know (I looked for them)

K6-2 and K6-3 physically use standard socket 7 interface and even some 430HX based boards, such as T2P4, can run K6-3+.

Sure you can physically plug them into a socket 7 and as long as you can get the proper CORE and IO voltages you still will be limited to the 66mhz bus , super 7 gets you that 100+Mhz bus , PC 100 vs PC -66 , AGP vs. PCI vga

Reply 54 of 59, by BitWrangler

User metadata
Rank l33t++
Rank
l33t++
kanecvr wrote on 2017-10-25, 21:00:
yes. Lucky Star 5V-1A, using the VIA Apollo VPX chipset. It has unofficial jumper settings for 100MHz. At first I tought it was […]
Show full quote
feipoa wrote:

Has anyone found a regular socket 7 board which had an unofficial 100 MHz setting? And did it work?

yes. Lucky Star 5V-1A, using the VIA Apollo VPX chipset. It has unofficial jumper settings for 100MHz. At first I tought it was 95MHz, but turns out it's 100 and PCI runs at CPUCLK/3 automatically when using 83 or 100MHz FSB. It's not stable with L2 cache enabled tough.

infiniteclouds wrote:

How about a S/S7 Board that has AGP, can do 100 FSB, but can also do BELOW 66.. ?

Yes. The above mentioned board supports 7.15, 50, 55 and 60Mhz FSB as well. At 7.15Mhz FSB you get XT speeds when disabling cache or 386 speeds with cache enabled.

One of these turned up in my stash since I last commented, so tryna make a mental note of this so I can try it out.

Unicorn herding operations are proceeding, but all the totes of hens teeth and barrels of rocking horse poop give them plenty of hiding spots.

Reply 55 of 59, by brassicGamer

User metadata
Rank Oldbie
Rank
Oldbie

Sorry to necro this thread, but I was looking for information specific to the definition of Super7 and hoped this thread might be it. I have done my own digging, however, and found this somewhat useful 'AMD Editorial Background' document that covers definitions of Super7 somewhat comprehensively. The only potentially dissatisfying aspect of it is that the date is late - I would have hoped for something from Q2 1997, but this is dated 28 May 1998. Earlier material, however, such as the K6 press release don't yet mention Super7, just 'Socket 7', so there's something missing in-between, perhaps.

The following items are defined within the document linked above as "Planned Super7 performance and feature-set enhancements":

  • 100-MHz local bus frequency.
  • Accelerated Graphics Port (AGP) support.
  • Support for backside L2 cache and frontside L3 cache.

So, while it seems AMD were keen to provide some guidance and direction for the platform, manufacturers ultimately did their own thing. What's also entertaining about the document is how much they talk about how trashy Slot 1 is compared to SS7, and yet they developed Slot A not so long afterwards. Biggest u-turn in PC history?

Check out my blog and YouTube channel for thoughts, articles, system profiles, and tips.

Reply 56 of 59, by dionb

User metadata
Rank l33t++
Rank
l33t++

Marketeers gonna be marketeers 😉

That's the point though, Super Socket 7 is a marketing term combining a bunch of nice features that don't directly have any thing to do with the socket itself. The only change to the socket is the addition of BF2 multiplier pin, but that was already done prior to the "super" branding...

Reply 57 of 59, by the3dfxdude

User metadata
Rank Oldbie
Rank
Oldbie
brassicGamer wrote on 2021-10-21, 11:26:
Sorry to necro this thread, but I was looking for information specific to the definition of Super7 and hoped this thread might b […]
Show full quote

Sorry to necro this thread, but I was looking for information specific to the definition of Super7 and hoped this thread might be it. I have done my own digging, however, and found this somewhat useful 'AMD Editorial Background' document that covers definitions of Super7 somewhat comprehensively. The only potentially dissatisfying aspect of it is that the date is late - I would have hoped for something from Q2 1997, but this is dated 28 May 1998. Earlier material, however, such as the K6 press release don't yet mention Super7, just 'Socket 7', so there's something missing in-between, perhaps.

The following items are defined within the document linked above as "Planned Super7 performance and feature-set enhancements":

  • 100-MHz local bus frequency.
  • Accelerated Graphics Port (AGP) support.
  • Support for backside L2 cache and frontside L3 cache.

Which if you see the document on the K6 release, those specs wouldn't have benefited from the K6 anyway. K6 was intended for 66mhz bus and really only competing with the pentium mmx. The K6-2 would have benefited from Super Socket 7, and it coincided with the super socket 7 if you look at the timeline.

brassicGamer wrote on 2021-10-21, 11:26:

So, while it seems AMD were keen to provide some guidance and direction for the platform, manufacturers ultimately did their own thing. What's also entertaining about the document is how much they talk about how trashy Slot 1 is compared to SS7, and yet they developed Slot A not so long afterwards. Biggest u-turn in PC history?

Well, sometimes trying to one-up your competition and pointing out their flaws is actually harder to do in practice, even if they were right. The K6-3 showed it was possible to have more integrated cache on die and match Intel on bus speeds, but they did it on an aging architecture. If they decided that Athlon had to be the same, it might have been technically better, but it could have been made the Athlon more expensive and with poor yield in the early days, if the die and process wasn't up to task. Having Athlon be too ambitious could have caused a market failure faster than Intel could say Itanium. But anyway, using K6-3 to pull off technical feats and then using Athlon for the Mhz wars was probably a smart thing to do to keep Intel on their toes, and they pulled off psychological wins, hitting speed records and keeping socket 7 viable beyond what Intel wanted.

Reply 58 of 59, by BitWrangler

User metadata
Rank l33t++
Rank
l33t++

I think it might have been to keep the low end market engaged and give it reasons not to buy celerons while they readied the duron, which is what kicked the K6-2/3 out of the Dresden fab, even as they could have sold them for another couple of years to s7 diehards.

Unicorn herding operations are proceeding, but all the totes of hens teeth and barrels of rocking horse poop give them plenty of hiding spots.

Reply 59 of 59, by the3dfxdude

User metadata
Rank Oldbie
Rank
Oldbie
BitWrangler wrote on 2021-10-21, 18:28:

I think it might have been to keep the low end market engaged and give it reasons not to buy celerons while they readied the duron, which is what kicked the K6-2/3 out of the Dresden fab, even as they could have sold them for another couple of years to s7 diehards.

Even at the time, early 2000's I though socket 7 could have lasted a little longer. I think 133mhz FSB K6-4 was feasible, given how long it lasted in production and the capability some boards could reach, and the k6-3+ was already a quite cool, low power chip, so one more performance revision could have been done. But at some point, you would have had to move on from socket 7 to do an i686 compatible chip anyway, so since intel didn't let amd share a socket anymore, you see they came out with their own (duron), and that meant k6 had to be discontinued in the desktop market.