Thoroughbred on KT133

- OCinside.de PC Forum
https://www.ocinside.de

-- English
https://www.forum-inside.de/forums.cgi?forum=14

--- Thoroughbred on KT133
https://www.forum-inside.de/topic.cgi?forum=14&topic=29

-- Veröffentlicht durch Skligmund am 9:47 am 26. Mai 2003

I forgot you had a KT7 not KT7A or E... It probablty won't work, unless your bios gives you the option.....

-- Veröffentlicht durch Lupson am 12:59 am 25. Mai 2003

I eventually found the problem and solved it. It turned out none of my cuts were actually faulty, I borrowed a multimeter and discovered that the problem was a tiny scratch I accidently had inflicted upon the 3rd L3 bridge. With my L3-bridges cut in that way the multiplier was set to 18.5x or something like that which simply was too much for my system to post with. I reconnected the faulty bridge with a carbon pen - and voila! - it posted and worked like a charm and has done ever since, at 105*16.5x.

Regarding 133 Mhz FSB, I doubt my motherboard (KT7-RAID) can handle anything above maybe 107 Mhz, though maybe Ill try that out sometime when Im not depending on my rig to for work etc.

-- Veröffentlicht durch Skligmund am 12:47 am 25. Mai 2003

Ground the 4th L3 Bridge to the CPU. In other words, cut it too deep and then paint over it agasin. Although your multipliers will drop 4X, you get 133FSB and the ability to close the 5th L3 bridge. It works, I had actually done it by accident, but found out later it was actually the work-around for doing what I was trying to do, get 133FSB.

-- Veröffentlicht durch Lupson am 18:11 am 22. Jan. 2003

paulus: Read through that thread. Wether to belive it or not? I don't know, at least the guy is very open about it and gives plausible explanations for most things he has done.

I still think my processor somehow must have shorted a bridge somewhere. Hopefully I'll get my thumb out of my **** and go visit a friend who's got a multimeter and has a collage degree in electronics. I mean, it DID work for 4 days. I'll be damned if I cannot get it to work again. Or maybe I should get some conductive ink, reconnect everything to the way it was from the beginning, and then do the AJ27 resistor mod instead of desperatly carving deeper and deeper into the CPU.

-- Veröffentlicht durch paulus am 15:48 am 22. Jan. 2003

I just pass this on to you because it fits in this thread however I must admit that I find this hard to believe.

http://forum.oc-forums.com/vb/showthread.php?s=ebabbf0242bbb38b9ab4605e94222a35&threadid=158672

-- Veröffentlicht durch invar am 12:45 am 22. Jan. 2003

my knowledge of this stuff is pretty much limited to what i have found on this forum and then making an educated guess based on that.  sorry lupson.

i know it's kind of late to say, but i think you should have gone the AJ27 resistor mod method instead of cutting/connecting the bridges on the CPU.

-- Veröffentlicht durch paulus am 20:22 am 21. Jan. 2003

Sorry to hear that it didn't work. Closing the 5 th L3 bridge using a tiny copper wire is possible but not the way you tried to do it. (You will never get good contact by applying Scotch tape.).
There is an alternative to the silver conductive ink method:
You must form a small loop with a piece of wire and put it around 2 pins on the downside of the cpu. I have never done this myself so I don't know any details but I am sure you will find good explanations of this method in this or another forum.

-- Veröffentlicht durch Lupson am 17:39 am 21. Jan. 2003

Thanks for your replies guys. Since I have no conductive silver ink at the moment I guess I'll have too wait until I can get hold of some if I am to try paulus suggestion. Im not very familiar in the internal workings of the different bridges and similiar things, but isn't the fifth L3 bridge there for the multiplier control? Are you sure you don't mean the new fifth (On t-breds) L1 bridge?

BTW, would it work if I took a tiny thread of copper cabling wire and bent it to an U. Then cleaned all contact surfaces with acetone and then placed it the same way I would have painted the U. Then I'd secure the wire with a little piece of sticky transparent plastic tape. (Scotch tape in english?) ?

I read through some of the other posts, you seem to be a knowledgeable group of guys. Would it work measuring the resistance between the bridges with a multimeter? If one bridge somehow has reconnected itself or has grounded into the metal below the bridges, would it be possible to determine which bridge is faulty then? I guess a really high resistance between the two "dots" of a bridge would mean a properly cut bridge, while a low resistance would mean connection.

Well, enough for now - thank you again for your kind help!

EDIT: Impatient as I am, I tried to use a wire to connect the 5th L3 bridge, but it didn't post. I tried resetting CMOS and using the reset button trick but neither helped. Then I removed the wire and tried to boot, but again - no post. Tried the same procedure with the 5th L1 bridge but without getting any posts. Inserted my old processor and like the last times, it required a press to the reset button before posting.

(Geändert von Lupson um 19:02 am Jan. 21, 2003)

-- Veröffentlicht durch paulus am 17:13 am 21. Jan. 2003

If invar's guess is true you would have to simulate the open switch in the resistor mod with a closed 5 th L3 bridge.

So my suggestion is: Close the 5 th L3 bridge using silver conductive ink with a U shaped connection around the gap you made with you knife. Insert the cpu and let it boot with the 8.5 Multiplier. Power down and remove some of the conductive ink so that the 5 th L3 bridge is open again. Reinsert the cpu and hopefully see 1650 Mhz on startup.
But don't try this right away! It would be safer if invar and or Tassadar gave some comments on this "method".

-- Veröffentlicht durch invar am 10:25 am 21. Jan. 2003

the aj27 resistor mod has a switch on it for a reason, and that because if the power cord is disconnected from the power supply (or power is not provided to the motherboard), then you have to flip the switch to off (thus undoing the aj27 mod) and then POST the system.  then turn off the computer, and flip the switch, and then repost.

it's documented here:
https://www.ocinside.de/go_d.html?https://www.forum-inside.de/topic.cgi?forum=14&topic=22

i would guess that you are experiencing something similar to that even though you did a different mod.

-- Veröffentlicht durch Lupson am 22:33 am 19. Jan. 2003

Damn, new problems.

My new XP 1800+ worked fine until yesterday evening. When I shut down windows for the night, it froze on the "Shutting down windows screen" (WinXP). Well, no big deal I thought and shut off the computer by holding in the power button for four seconds.

And this morning - no post. It behaved exactly as when it didn't work before (se earlier in this thread) i.e. hdd spin up and stuff but no post.

Well, I reinserted my old processor and it booted up nicely. I reexamined my XP-processor and played around a bit in the gaps with my knife. I reinserted it (the XP) again, but no luck - no post.

I tried pressing the reset button many times and resetting CMOS and so on, but no luck.

Anyone got any ideas? I don't think I've burned the CPU, afterall, it worked just fine until the power off yesterday. Could the way I turned off my computer yesterday have caused the timing problem? I dunno..

Lastly. Could one of my self-cut bridges somehow have reconnected due to heat, heatsink pressure or something? I messed around a bit with my cutting knife, but that didn't help either. Also, I accidently cut a tiny scratch over the middle L3 bridge (the only one that's supposed to be connected). It worked fine for a couple of days, could this bridge somehow have been broken?

Well well...Im grateful for any tips or suggestions.

/Lupson

-- Veröffentlicht durch invar am 0:33 am 18. Jan. 2003

lupson, if you read carefully through the thread i referred to, you should've noted the following:

the pin->resistor->pin mod is not necessary if you have cut the L1 bridges and have forced the multipliers through use of the L3 bridges.

if you do not use the resistor mod, however, then you have to cut the L1 bridge and force the multiplier through the L3 bridge.

i personally opted for an unlocked Tbred processor (connected L1 bridges) and went for the pin->resistor->pin mod.

your friend went for the bridge method.

that's why i suggested to you to cut the L1 bridges, because it didn't sound like you were trying to do the resistor mod, nor did it sound like you wanted to do it.

glad you got it to work now though. :-)


edit: yikes, now that i read through my link carefully again, i think i might've linked you to the wrong post about it or something.... crap, i can't find the other one.  oh well, seems it worked anyway, cutting the L1s.

(Geändert von invar um 0:37 am Jan. 18, 2003)

-- Veröffentlicht durch Lupson am 17:55 am 16. Jan. 2003

paulus: I wasn't sure wether my cutting had been successful so I examined all my cuts with a magnifying glass and continued to cut at the bridges that I wasn't 100% sure were cut. This applies to both L1 and L3 bridges. No other bridges were touched. I also flashed my BIOS to the latest A9 version. All L1-bridges were cut and just the middle L3 bridge were left intact to obtain 16.5x multi.

/Lupson

-- Veröffentlicht durch paulus am 23:48 am 15. Jan. 2003

Lupson did you continue to cut the L1 bridges or did you cut the L3 bridges again?
Did you do anything else besides the cutting of the bridges?

-- Veröffentlicht durch darkcrawler am 2:39 am 15. Jan. 2003

nice:)

-- Veröffentlicht durch Lupson am 0:03 am 15. Jan. 2003

I got it to work!! I cut ALL L1 bridges. Still didn't work. Then I continued cutting the bridges even more just to make sure they were properly cut, and on my next try - voila! A nice clean post.

Checked temprature in BIOS, okey. Into windows and yes allright:



Glory! Now I'll start tweaking FSB and Bios settings to see how fast I can get it to run :)

Thanks everyone!

/Lupson

-- Veröffentlicht durch Lupson am 14:48 am 14. Jan. 2003

Thanks for your answer Invar, though I didn't get any answers from the thread you were refering to.

By reading this and others topics regarding the KT7 with AMD XP issue I've concluded that it has worked for everyone whom had tried it, so I guess it'll work for me when I get the right bridges cut properly...well, as long as I havn't destroyed my CPU in the process... :-|

A friend of mine has promised me to try my CPU in his rig (real 266 FSB rig) tomorrow or Thursday so I'd thought I'd await the results there until I'll continue cutting bridges and stuff.

I'll post my results here when I know more, thanks again for your help guys!
/Lupson

-- Veröffentlicht durch invar am 6:40 am 14. Jan. 2003

https://www.ocinside.de/go_d.html?https://www.forum-inside.de/topic.cgi?forum=14&topic=12

yep, you need to cut L1

-- Veröffentlicht durch Lupson am 15:36 am 11. Jan. 2003

Thanks for your answer. I got hold of my friend with the working rig. He had had to cut all his L1 bridges to get it to work so I'll try that now. I hope I didn't destroy the processor, if I did, at least it was cheap (~80 euros), so it's no big disaster..

/Lupson

-- Veröffentlicht durch Tassadar am 22:07 am 10. Jan. 2003

I suggest you to try your processor on another board, if it don't work, try to rejoin the bridge you cut, using conductive paint or silimar.

If after this your processor don't work, very probably you've broke it.

-- Veröffentlicht durch Lupson am 17:35 am 10. Jan. 2003

Hi, thought this thread would suit my question

I also have a KT7-RAID with A9 BIOS. I bought a T-bred 1800+ which came unlocked from factory (L1 bridges closed). A friend of mine had done some bridge cutting on his rig (also KT7-RAID) to get 16.5x multiplier. It works fine for him. He hasn't mentioned the pin->resistor->pin mod.

Anyway, I cut the bridges at L3 to get 16.5x multiplier, but my system doesn't even post (black screen only, hdd:s and fans spin up). If i reinsert my old processor everything works fine. I've tried resetting CMOS before installing the 1800+ and I've tried setting the BIOS to 7.5x multi and then changing processors. Nothing works. I've also tried the resetting trick.

I'm not sure if my L3 cutting were been successful. Have no tools to check resistance. Could that be the reason it doesn't work? Maybe I've shorted a bridge with the material underneath cutting too deep? But if the L1 is unlocked, shouldn't the BIOS be determining the multiplier?

I havn't gotten hold of my friend to check with him. I'd be very grateful for some tips regarding this issue.

Regards Lupson

-- Veröffentlicht durch invar am 9:55 am 5. Jan. 2003

sorry folks, i kind of got sidetracked for a few days.

the reset button trick isn't really relevant in my case.  today, i received a new (used) 256MB DIMM from my friend in exchange for my 2 128MB DIMMs.  the system runs perfectly stable now with the new single DIMM so now I am back at 512MB.

i do get the reset problem as well, meaning my first totally cold boot after having reconnected the power cord must have the switch set off.  then, i turn it off, and while the power cord is still in, i flip the switch.  turn the system back on, and it works fine.

it sounds like you need to do that ben.

-- Veröffentlicht durch paulus am 19:07 am 4. Jan. 2003

Exactly. But since invar did not answer yet I assume it didn't help in his case.

-- Veröffentlicht durch Tassadar am 1:11 am 3. Jan. 2003

Do you mean press reset button when cumputer is on and not posting?

-- Veröffentlicht durch paulus am 4:53 am 1. Jan. 2003

Did you try the "reset button trick"?

-- Veröffentlicht durch Ben am 14:48 am 28. Dez. 2002

Hi, I've got an Abit Kt7 with a duron 800, a few days ago i bough an Athlon XP 1700+ thoroughbred stepping 'RIRGA'. Fistly I updated the motherboard bios to 'A9' version, as Tassadar adviced me, but when I plug in the Athlon XP the computer don't post, I press the power button and screen is only blank screen, i mean, anything  . I don't know why it happends because it is supposed that XP's works in my board, an abit KT7 with A9 bios, anyone can help me?

Thanks

-- Veröffentlicht durch invar am 23:26 am 23. Dez. 2002

i think the problem will arise when you get more than 2 DIMMs.  luckily, i have found someone willing to trade his 1 256MB module for my 2 128MB modules.  so i'll be up to 512MB pretty soon.  phew!

-- Veröffentlicht durch Tassadar am 3:56 am 23. Dez. 2002

Not everything are good news :(, I got A9 bios on my KT7-raid and works fine at 133 mhz. I've got 2 different 256 mb modules running at 133 mhz.

-- Veröffentlicht durch invar am 11:54 am 21. Dez. 2002

DAMNIT...

okay, you know those timing issues?  Yes, I think they exist, although I can't be sure because I don't have another CPU that I can plug in right now... or rather I don't want to pull it out anymore.

I thought the system was running stable.  I was wrong.. oh yes, I was very wrong.

I tried to run several 3D applications, including 3DMark 2001 SE, which resulted in the system crashing.  I couldn't figure out why for the longest time until finally, I reset my BIOS configuration to the "default safe settings" option.  This let me run all my 3D things just fine.  So I went through the first screen, and set my memory to PC133.  Immediately, it was unstable again.

What in the world??  PC133 worked fine on the other BIOS.  Okay.  So I try toying with all the timing settings.  No go.  Now, let me first tell you that I have 3 DIMMs in my system.  I have 2x128MB from Mushkin PC150 2-3-2 stick, and 1x256MB from Crucial PC133 2-2-2 stick, bringing my total to 512MB.  The system currently refuses to run stable with all 3 sticks in at PC133 using any timing settings at all.  It will only run stable at PC100.  However, if I remove one or two of the memory sticks (it can be ANY of them), the system runs perfectly stable.  I am sure of it.

I tried every BIOS version from YP Beta and up.  This means I have tried YP, ZT, 3C, 3R, 4A, 7N, 64, A9.  The results I have come up with is the following:

YP, ZT, and 3C detect the CPU as an Athlon Processor and thus SSE is disabled.  The memory can run stable... meaning PC133 works just fine.

3R, 4A, 7N, 64 detect the CPU as an Unknow Processor and SSE is enabled as identified by 3DMark 2001SE and WCPUID.  The memory does NOT run stable at PC133.  Yes, it actually says Unknow, not Unknown.
r
A9 detects the CPU as an Athlon XP Processor and SSE is enabled as identified by the same as above.  The memory does NOT run stable at PC133.

Without plugging in my old Thunderbird 900, I can not say for sure.  However, I firmly believe that the Athlon XP Processor does indeed cause a timing issue of some sort with the Abit KT7 when using all 3 DIMM slots.

Currently, I have removed one of my DIMM modules in order to keep my system running stabley.  This whole situation was a nightmare.  Suffice it to say that it used up a good 7 hours of my time as a result... it screwed up my windows boot... uggh.

-- Veröffentlicht durch invar am 2:57 am 21. Dez. 2002

Am currently running at 1500MHz (15x100).  It worked!!

Thanks. :-D

By the way, I too get the strange power cord thing.  First boot, I have to have the switch flipped off and boot at 7x.  Then once the power cord is still in, I shut it off, flip the switch, and it boots at 15x every time afterwards until I yank the power cord again.  Very strange.

(Geändert von invar um 4:25 am Dez. 21, 2002)

-- Veröffentlicht durch darkcrawler am 12:23 am 20. Dez. 2002

:respekt: tassadar


it seems that the good old kt7 series is not to kill:lol:

-- Veröffentlicht durch invar am 8:15 am 20. Dez. 2002

I guess I'll go out and purchase what I need tomorrow.  I'm just worried about the whole soldering thing.  So... how does this work, it just adds 8x to them multiplier that you have set, right?  In that case, before I shut down my system to do the soldering, I should set the multiplier down to something low?

-- Veröffentlicht durch Tassadar am 1:47 am 20. Dez. 2002

At once I've got mine working at 1,59 volts, 1800 mhz (18*100) and 44 degrees with a a cooler master alluminium heatsink.

What are you waiting for? :D

hehehe

-- Veröffentlicht durch invar am 0:00 am 20. Dez. 2002

Recently, I got a 1700+ and this was all before I knew anything about Thoroughbreds and Palominos, etc.  I too have a KT7 and wanted higher multipliers than 12.5x.

I was surprised by the fact that the processor came unlocked.  Currently, I'm running on an old beta YP version of the KT7 BIOS.  It's the only version of the BIOS that allows me to run my FSB at 112.  Thus, I'm at 112 * 12.5 = 1400 right now for a 1700+ processor.  I have no idea if it's a Thoroughbred A or B, but I'd like to run it faster.  

After having read through that multi-mod thing, I'm very tempted to try it.

Any thoughts?

Thanks!

-- Veröffentlicht durch Tassadar am 23:27 am 18. Dez. 2002

Thanks Paulus :), I've tried to enter 'KT133 thoroughbred' in google and the second and third results are post of mine in those forums, hehehe.

The multi-mod is described here:

https://www.forum-inside.de/topic.cgi?forum=14&topic=22

And here you have a link to a post that i started on one spanish forums that i visit everyddays, where i have tell everything i i've made since i started searching info about xp on kt133:

http://www.hardlimit.com/forum/showthread.php?s=&threadid=7014

(in spanish)

And of course, many thanks to pzed22, without him anything of this could have been possible, hehe

Oh, i forgot to say.... as far as i'm concerned it's perfectly possible to use tbred's B on KT133 and even bartons, always when the bios support the processor microcode, but it's perfectly possible (I suppose my bios support tbreds 'b' and i bet also bartons)
My bios is latest official bios from abit, none of unofficial or modified ones

(Geändert von Tassadar um 23:34 am Dez. 18, 2002)

-- Veröffentlicht durch paulus am 19:43 am 18. Dez. 2002

Congratulations Tassador! I have been looking quite some time for a post in forum of somebody who made a T-bred run in a KT7.
As far as I can tell you are the first who made it!

Just one question: What kind of multiplicator mod did you make?

-- Veröffentlicht durch Tassadar am 2:58 am 18. Dez. 2002

Today I've make the multiplicator mod and it works!! :D

the multipliers that works with the mod are:

7 ->  15
8 -> 16
8,5 -> 16,5
9 -> 17
9,5 -> 18

The other ones don't work (system don't post), and it's not problem caused by CPU, because I've post at 2018 mhz (18*112).

-- Veröffentlicht durch Tassadar am 22:24 am 16. Dez. 2002

This morning I bought a 1800+ thoroughbred, and it's working at 1,32 vcore and 1313 mhz.

The board now don't allow me to make it faster, so tomorrow I will make the multi-mod (Called resistor mod) I hope It will work and make me able to set up yo 20,5 x

Someone Knowks if the multimod will work?

The CPU is factory-unlocked althought have L1's and last L3 disconected, would I need to join someone to make it work?

-- Veröffentlicht durch BUG am 12:00 am 13. Dez. 2002

The TBred 1700+ will works on some Abit KT7A (Rev1.0) with Rev1.3 bios. I think you must flash also the Bios from Rev1.3. The differents KT7A / KT7 is only the Northbridge. Bios, layout and so one ist the ~same. Flash the bios from ABit KT7A Rev1.3 to let work TBreds on your Abit KT7.


...but no warranty for working correctly. ;)

cu
BUG

(Geändert von BUG um 12:01 am Dez. 13, 2002)

-- Veröffentlicht durch Tassadar am 23:21 am 12. Dez. 2002

Oh, shit, hehe, I've been waiting for 3 mounths for buying a 1700+ tbred and now that at last are avaiable you say me it won't work on my board.

¿Has anyone test any 1700 or 1800+ tbred on an kt133 board? mine is ABIT Kt7 raid with latest bios

Thank u very much for your answer, darkcrawler

-- Veröffentlicht durch darkcrawler am 12:14 am 12. Dez. 2002

-- Veröffentlicht durch Tassadar am 5:40 am 11. Dez. 2002

pzed22, you said the resistor mod for 2200+ is different from 1500-2100+, but I have a doubt: If a buy the 1700 or 1800+ with tbred core, it will be a 1700 or 1800+, not 2200, then what resistor mod would i need?

-- Veröffentlicht durch Tassadar am 1:26 am 5. Dez. 2002

I have still not buy the XP, as i'm concerned 1700 - 1800 + with thoroughbred core will arrive at shops here (Spain) in one or 2 weeks, and I wanna buy one.
Has anyone test this processors on a KT133??

-- Veröffentlicht durch Tassadar am 1:12 am 25. Okt. 2002

I don't want to buy a 2200+, of course. I want to buy a 1600+ or 1700+ with thoroughbred core, whose should cost as a similar rated palomino.

Then, spending the money of a 1600 or 1700+ I would have a thorouhgbred.

Questions are:

- Are now avaiable the 1700+ with tbred core on shops or the only tbred now are 2200+? (I live in Spain)
- What's the minimum tbred? 1700+?
- Would work a 1700+ tbred on my board?
- Would work this resirtor mod?
- Should I unlock by the 5 L1 or only the last L3 ???

-- Veröffentlicht durch pzed22 am 12:28 am 23. Okt. 2002

It should work fine, but a 2200+ requires a different resistor mod.
The resistor mod picture I drew in the other thread is for 1500+ to 2100+ only.

-- Veröffentlicht durch Miky am 11:52 am 21. Okt. 2002

It's interesting, try it and let us know how did you succeed :)

-- Veröffentlicht durch Tassadar am 2:47 am 19. Okt. 2002

In theory and basing in some publications, thoroughbred 'A' can reach similar speeds than a Palomino aroia-agoia.

The question is than using a good copper heatsink i think I will be capable of get more from a tbred A than from a aroia-agoia palomino, becouse the 2200+ tbred works at a default of 1,65 against 1,75 of palomino 2100+

-- Veröffentlicht durch Miky am 12:35 am 18. Okt. 2002

I don't know, but I'll be happy if it works.
Now I've palomino 1600+ (1400@1760).
How far you think you can go with Thoroughbred 'A' Core?

-- Veröffentlicht durch Tassadar am 3:15 am 18. Okt. 2002

I've got an Abit KT7 raid with bios A9, i tried an XP 1700+ palomino core, and works ok.

Now I want to replace my thunderbird 1400 for a new processor, an XP, my idea is to wait a bit and get a 1600+ or 1700+ but with Thoroughbred 'A' Core, with the aim of getting a better overclocking.

Will this processor work on my board?? I think yes, becouse is exactly as a Palomino core, but smaller die process

What do you think about it? Am I wrong or it can work?