I'm there most weekends - and every time, there are different piles of discarded computer hardware. A lot of it is not interesting to me (a lot of office PCs from Dell and HP). But there are real retro gems to be found!
Yes, very sad to pull such a board from the scrap. If I would have gone through the entire pile, I'm sure I would have uncovered more hardware of that calibre.
might be related, but i worked with pentium 3's in an industrial setting where they ran Siemens PLC stuff (Step 7 under Windows 2000 mostly) and C0 was common on high heat, high vibration installations where the BIOS was unable to verify cache size but was still detecting cache presence. if cache was disable entirely it could boot, but often we would send the cartridge or CPU (which was thermal glued to a small heatsink/fan combo unit from i think AVC) out for repair and the turn around could be as little as a day. C0 was also used by early Xeons to indicate an issue with initializing L2 cache which was a separate IC on the PCB. this might be different though since pins for Slot 1 and Slot 2 differ and might have different POST codes. this initialization step was though through the BIOS because you had to have BIOS support for ECC cache. this differed from the Oracle/Sun servers that ran the database and application end of things where the C0 was usually an add-in controller that had failed to boot somehow and you needed to identify and remove the device (disk controller with bad battery backed cache, misconfigured memory or video, anything where you had to use an unmasked interrupt to actually boot).
It may be more convenient sometimes or safer against loss to not desolder a 2-pad component fully but to rotate or shift it such that one pad is connected and the other isn't, in order to measure it. Especially if you move to more hot air soldering.
you're really good with the SMD stuff. Seeing proper SMD soldering is always ASMR. lol I just love seeing the solder flow exactly where it's supposed to go
@@bitsundbolts I did too, but not SMD. lol I've never done SMD. I did get a rework station (it helps with 3d printing sometimes, but I don't use as much as I thought I would). I'm tempted to take apart an intel laptop that died 2 months after I bought it during covid and practice on that for shits and giggles. (it's why I never followed through on the warranty and it's the last intel I ever bought).
It's quite funny how some dies are in horrible condition and they still work! I have a Duron that looks like a total disaster and I have a feeling that there is some piece of metal sticking out from the chipped silicone die. Stull works though without issues... Go figure.
I had a Thunderbird 1.2GHz (socket A) back in 2000 or so - the die was extremely chewed up by the crappy CPU cooler (very small die, very large pressure spring) but it ran like a champ. Could fry eggs on it.
I've learned a lot! Thank you!!! Also for the insights of a scapyard in the desert (it kinda hurts to see this!). Here are some unqualified suggestions to your POST-problem: do all pins of the cpu work? Looks really scratched up. One other Idea: is there a possibility for a "flexible" resistor? Maybe the one resistor that is different on PIII 650 and 1000 have to be different. Worst case is defitely a broken coppermine die.
Thanks for watching and I'm happy to hear you found value in that video. My thought process is similar to what you mentioned. I did actually check the golden pins. They are indeed scratched badly. On one of those pins, I removed a small piece that was getting too close to the neighbouring pin. Other than that, they look ok, even though they're in bad shape. The similarities to the working PIII 1000 are interesting though.
are those non labeled resistors by any chance connected to pins bsel1 and bsel 2 on the cartridge connector if yes those two resistors were only responsible for changing between 133 and 100 or 66 bus and high values make sense since on 133mhz both of those pins should be high impedance and that would make it a pentium 3 with 133 mhz bus of unknown multipler and speed but you could decode that by multiplier pins too biggest problem you may have with cpu that trashed might be core delamination so i would try pressure on the core substrate while booting
The front side bus is selected on the board. I haven't looked at those pins you mentioned. I'll have a look. Maybe you're right and there is an issue with the connection between the substrate and the PCB of the cartridge.
@@bitsundbolts it would be great to have a late super slot socket adapter with all the jumpers for voltage and bus like ms6905 master so you could reverse engenier voltage and bus jumpers, it also have that ic that got mangled on your cpu but multiplier i think is burn into the die, so that would have to came from intel technical specyfication for slot 1 but i would start with poking the cpu core substrate to see if it doesn't have broken balls or look closer what that ic do because it might be related to cache
I think it's a PIII 667/256/133 Coppermine SL3XL. Because i found pictures with the same modelnumber, PB 746509-001 and the same PCB layout, on a russian auction website.
It was a generic Dell/HP board that was screwed to a desktop case. I usually do not bother looking at those because it's hard to get information on those - or even find mods, BIOS updates, and user stories.
I wish there was something different that could be done about scrapping old hardware. Anything retro is worth 100x more as-is on Ebay than whatever they'll get from scrap but I guess most people, including the employees, just don't realize it.
The people working at the scrapyard don't care what they have in front of them. They separate different materials: cases, plastic, power supplies, electronics... They don't know what's valuable and what's really scrap (80-90% is Dell & HP stuff), then you get servers, switches, and other enterprise hardware. All there for scrap to get the valuable metals back.
@@bitsundbolts I work at one of those scrapyards and sort the electronics, I do care what's going though and I recover some of it. I just picked up a nice 386 MB yesterday. There is so much and you can't save it all. I have recovered some really nice equipment and have even sent some to UA-camrs.
Oh, I didn't want to generalise. When I said people don't care, I specifically meant the scrapyard I go to. It must be nice to find some exotic and rare items. I'm sure the people you sent those items to were very happy and thankful!
My second computer was a P3 running at 800MHz, but it was mounted in a socked. What's the difference between a socketed CPU and one that's mounted on a cartidge? I remember my friend's dad upgrading his P2 computer using a cartridge, but I never bothered asking why his computer is set up the way it was, as I was pretty young.
Many cartridge CPUs had level 2 chips which was the reason to move to this design. Then there were Celeron and Pentium III that had the level 2 cache integrated into the die and therefore no longer needed the cartridge design. Those CPUs could fit in a socket again (socket 370). To make them compatible with slot 1 boards, the adapter card was a neat solution to allow users to continue using their boards with new CPUs. I don't think there's any performance difference when you compare a Coppermine 800 for socket 370 or slot 1.
Yes, I had my fair share of issues with a cartridge CPU not making proper contact with the slot. I tried the CPU in this video multiple times to be sure it's not a connection issue. Also, the same behaviour on a different board. I don't think I have an issue with the contact. It's most likely something on the cartridge.
Hi, there are some socket 370 pentium 3 1000eb 1.7v and 1.75v. Do you think the last could pull less amp from motherboard vrm because of the "higher" voltage?
Hm, I am no expert, but I would expect that at the same frequency, the CPU with 1.75 volts would pull the same current from the VRM. That would lead to a higher power consumption and probably slightly higher temperatures.
Good job, I don't know if you have another motherboard that natively supports P3 1ghz like a P3B-F or equivalent, it would be ideal for testing, otherwise in the last resort just doing a reballing.
While looking at the footage, I had the same thought that it must be a pain trying to fix it! But that board stayed at the scrapyard. There are for sure a lot more issues than just the socket.
@@bitsundbolts I'm fairly certain it's just not possible. The contacts on modern CPUs are way too sensitive and too close to their physical limits with not much room to spare for tinkering. The system probably wouldn't be stable or may even pose a fire hazard. Thankfully it doesn't matter though, most of these boards aren't worth preserving, let alone saving, anyway. 😅
I fixed a single pin once on an X58 board? At least this is what I remember. The boards with triple channel DDR3 support for i7 920/930/950 etc. you could just put a Xeon 6-core/12-thread server CPU in there and get a half-way decent system (about 6 to 7 years ago)...
@@bitsundbolts Impressive! Luckily, I haven't had to resort to this so far, and there's a good chance I would just make it worse. 😅 I suppose it also depends on how important/mission-critical the machine is. If it's just a redundant or hobbyist machine and just one pin, one might get away with it. Coincidentally, I still have an LGA1366 board (P6T Deluxe v2 with an i7 9somethingsomething) in my basement. With its 3 PCIE slots, it still worked well for ETH mining, as long as that was a thing. 😁
Haha, right. Now they're on proof-of-scheibenkleister. I also have an X58 in my pile of hardware to be tested. I think it's also an ASUS/ or Gigabyte - I need to check that. It was just one pin and I tried to flip PCs at that time. Didn't work well. A lot of work, no financial reward. Almost like UA-cam 😂
I am surprised there are scrap yards in Dubai but I probably shouldn't be because they are everywhere.I buy a lot of motherboards, cpus and graphics cards from scrap yards in Vietnam. So many socket 7 boards, Slot 1s and the odd 486 and 386. Some need a lot of work.
I'm pretty sure that's a 1.13ghz SL4HH Cpu, but is only speculation as it didn't have the other resistor, but I have one and the cpu number on it is 7122A223
Oh oh... Really? I thought the 1.13 GHz was the only Coppermine that requires 1.8 volts. How unlucky that it didn't Post. It would have been a pleasant surprise.
What if some of the resistors are to configure CPU clock speed? It may be getting overclocked and crashing at boot. Maybe try underclocking it by lowering the bus speed?
@@bitsundbolts Also try 133MHz. As I mentioned previously none of my three Slot 1 133MHz Coppermines will POST when underclocked to 100MHz. Perhaps this is the same problem.
Good point. Btw, I could reproduce that with a socket adapter and a Pentium III 1 GHz for socket 370. It's a 133 MHz model that refuses to post at 100 MHz FSB.
So, why not just bridge the pads with either a wire or solder blob? Seems a bit silly to use a 0 ohm resistor in that situation. I know why they do from the factory, easier for pick and place, but when you're doing it, I don't quite understand the point.
It's personal preference. Yes, you could use a wire or a solder blob to bridge the contacts. I just have endless 0 ohm resistors which I can use. Also, if I were to use a solder blob or wire, someone would complain 😅
@@bitsundbolts Given I've seen people complain about astrological events occurring on a school day, I wouldn't put too much stock in it, but fair enough.
Wow! Such a scrapyard i would like to search through 😱 Would you mind to Sell some of your 1000MHz P3 Slot 1 CPUs? Im from germany and would very Like to get some, even if broke of some sorts 🤓❤️
I think I have four Pentium III 1GHz (3X 133 MHz FSB and 1X 100 MHz FSB). If you're interested, please contact me through email and we'll continue there. Thanks
Not living in Germany but in another European country, I wonder where I could find scrapyards like this. How does it work in Germany and how do you get access? Local municipalities here have recycling stations but once something is handed in, nobody can take it out anymore (I guess to prevent employees trading/selling things).
I'm not in Germany. I'm in the United Arab Emirates. I heard many stories from viewers how difficult or impossible it is to get access to a scrapyard (especially in the USA and Europe). I guess the only way is to build friendships with owners or employees of those facilities.
Yes, that is most likely true, but what exact model it is will only be known if we get it to POST. you're correct about the SRAM chips. Coppermine Pentium III do not have cache chips.
Given the amount of physical damage, I think that you may have to "rehot cpu friend!". Seriously, if you can't find any other damage, solder balls under the CPU are the most likely thing, other than the silicon or the board the CPU die is bonded to being damaged. Since those are probably impossible to fix, I would suggest a reflow or even a reball.
That would be nice if it's really such a CPU. You'd be surprised what the percentage of working (or repairable) hardware is I find there. If I exclude hard drives, it's easily above 90%.
Athlon vs Pentium III: the race for 1 GHz. I think socket 370 and slot 1 Pentium III were released at the same time. Slot 1 was declining already and the 1 GHz model was one of the last CPUs for slot 1. So, you probably already moved on to the new socket 370 platform.
@@bitsundbolts AMD and Intel both switched to sockets at the same time. I believe the Athlon Thunderbird came out at the same time as the coppermine. It had the integrated cache. I had one. I don't think it was available in a slot configuration. And yeah I do remember Intel and AMD both hit 1ghz at about the same time. Within a week.
Yes, they indeed released their GHz flagships at about the same time. I was happy, because I had AMD back then. I continued with Athlon XP and 64 X2. Then Core 2 Duo.
@@awilliams1701 @bitsundbolts Yeah I read somewhere that AMD released the 1GHz Athlon two days before the P3 1GHz came out! And another fun fact - AMD didn't have on-die cache at this point so they had to set a 2/5 cache divider to get the cache to work stably
@@bitsundbolts at one point I had an Athlon XP mobile CPU in my desktop. I have no idea how that happened. You'd have to ask my cousin. He gave it to me. I'm like.......how the hell is there a mobile cpu in a desktop? lol And then I switched to Intel core 2 quad, pentium-dual (2 cheaper systems used that), I7 laptop, i7-2600 desktop. And now back to AMD 4700u and 5900x. I think it might be time to replace the 4700u laptop. It has served me well though. AMD has RDNA based CPUs now. I was hoping it would play KSP2..........except that now KSP2 is dead. Funny thing. The 4700u loads KSP2. The menus are buttery smooth. And then when you load an actual game it crashes.
none of the markings on the board are much help, there is a whole series of PIII's that it could be, but if you can manually set the multiplier, try setting it up as a 450mhz 100mhz 4.5x as that is the lowest spec for that board, the lowest 133 fsb is 533 under clocking it has more chance to get it to boot than over clocking it.
I'll try that. There are some CPUs that follow the settings of the motherboard for the multiplier. Most slot 1 CPUs have a fixed multiplier - so, the setting on the board will be ignored. I'll try it anyway! Thanks for the suggestion.
@@bitsundbolts it's been a long time since I worked with slot 1 CPU's but there were some boards that you could manually set the clock and multiplier. if I recall some CPU's had an internal clock doubler, others didn't. but my memory might be a bit fuzzy. I think one of the slot 1 SOYO boards would allow us to manually set up the CPU, but can't remember if that was for PII only or if it worked on PIIIs too.
Yes, all of this was there. Clock doubling was introduced with 486 DX2 CPUs and then DX4 with the clock tripled. From socket 3 to 8, most settings were configured using jumpers. Then came fixed clock multipliers and the ability to configure CPUs using the BIOS. I think only a few early slot 1 CPUs have an unlocked multiplier that can be set by the motherboard.
Haha, those are mainly server boards. I don't mind bending some pins on 486 CPUs back into place, but those LGA sockets are on a different level... I don't know if I want to get into something like that. Maybe as a challenge one day, but it should be a board worth saving.
@@bitsundbolts LGA775 (or in this case, LGA771) have relatively large pins, it shouldn't be too hard. The newer LGA sockets have a much greater pin density. Even then it's not too hard to correct a couple of bent pins. For a hard time find an LGA3647 board.
Ты не прав. GD900 среднего качества дешёвая термопаста. Для проверок и экспериментов, отлично подходить. У меня 2 пасты GD900 и MX4. MX уже наношу при окончательной сборке.
@@denyda1754, и в каком месте я не прав. В том что термуха нижнего эшелона и годиться конкретно для проверок на включения? Или для совсем "плюшевых" процов? Я повторюсь и обращу ваше внимание на рофло надпись на тюбике "high performance".
I've had several Athlons where the die has corner damage but the CPU works (and even overclocks) without any issues. No experience of Intel chips though with damaged dies.
😢 omg i can't listen to this anymore! Man! You really need to focus on intonation, REALLY. It's the same! Every sentence has the same ups and downs and it just tears my ear drums apart!
Seeing the ROG Mainboard scraped hurts very much. I had the chance of taking a ROG Commando from a scrapbin. With the exception of the status display in the back not showing all the segments it works. Currenty I'm looking for parts to build a gaming computer like I wanted in my youth but couldn't (because of money). I'm still looking for a Radeon HD4890 with 1GB from Saphire and the waterblocks for the gpu's. Unfortunatly waterblocks for older gpu's or gpu's with waterblocks never comes up for sale. And I have no acces to electronic scrapyard like you have.
I'm sorry to hear that you don't find that GPU and the water blocks for it. I have a dual GPU from AMD, but I don't remember what model - and it might not work. If I find it and I remember, I'll have it in a video or on my website some day.
Foot point. Since it's twisted, maybe something happened internally. There is also a chip in it. Could be that due to the impact, something happened to it.
The POST code stops at C0 because the next step, likely Processor Test 1 (01) if the POST codes are listed in execution order, is failing. I don’t know what that test is trying to verify, but maybe someone else might have some insight…
1.1 GHZ actually, but those are more rare. I have a slotket myself with a 1300 MHZ Tualatin Pentium III/Celeron on it in my CPU collection, works well and is a very fast CPU for the platform.
use a different cpu to turn off internal cache. or maybe the chip you moved the pic needs to be removed to see if any of the traces are messed up and if you are feeling froggy you can swap the chip on to another CPU that does run to see if it works
Why did you say even though it's 133mhz fsb on the P3 1Ghz cpus? I've always seen those as the more desirable ones since they don't overclock much and higher fsb is better than the 100mhz fsb versions.
Correct, however, on an ASUS P2B, the AGP bus is overclocked if I run them at 133 MHz FSB. That was only fixed on later boards like the P3B-F. A 100 MHz FSB GHz CPU would work without issues on those boards. And if I'm not mistaken, the 100 MHz FSB PIII is less common and therefore more valuable.
@bitsundbolts Makes sense, and sounding very familiar. Now that I think about it I may have had a 100Mhz fsb P3 1 ghz in that same board at my job. It started life as a Celeron 300A that I overclocked to 464. I want to say it ended up as a 1ghz cpu but it was sold in the P4 era to an employee.
love your videos thanks for share and doing them for us ;) at some time you should do a bench tour, so we can learn what tools you use for you work ;) if you done such, please link or point me to it ;)
Ah, I haven't done a bench tour. I'm ashamed that I don't even have a dedicated bench for my work. It's also my home office desk 😅. Some day, I hope to change that. Thanks for watching!
Yes, you're right. It does look cracked. I checked and removed it from the processor, but it looks fine. 50 ohms - similar to a resistor on a working CPU.
Another great video. What soldering tips do you prefer to use? Have you ever used a hollow tip? I remember seeing someone talking good about hollow tips on youtube. Thank you for the spreadsheet.
I'm using a simple bevel tip. My soldering iron is the FNIRSI HS-01 and I use the tips that came with it. I move on from chisel to bevel tips. That's a personal preference I believe. In my opinion, the bevel is better in transferring heat to solder joints. I never used a hollow tip, but it looks interesting with space to store solder.
@@bitsundbolts Thanks for the answer. I need to buy new tips for my soldering iron. I'm tired of using big tips for small jobs. I'm gonna buy bevel (I'll see if I find an hollow tip as well to buy). Someday I'll also get a better soldering iron, like the FNIRSI HS-01, but for now I can't justify the expense because it's rare for me to solder. I have a generic Lidl soldering iron with a knob to control the temperature.
Do you mean the number on the substrate? I googled that number, but nothing came up. Maybe it's an internal number for Intel to keep track of production batches, but not meant for the public.
There is an app from Intel to scan and identify those 2d datamatrix markings on their products but at least in my attempts it doesn't seem to work with CPUs as old as the Pentium III
I downloaded the Intel toolkit 3.0, but that application wasn't able to tell me anything. It doesn't recognise the QR-code like matrix 😔. But thanks for the great hint!
it's sad seeing all that hardware being scrapped
I'm there most weekends - and every time, there are different piles of discarded computer hardware. A lot of it is not interesting to me (a lot of office PCs from Dell and HP). But there are real retro gems to be found!
At least it's not on a beach!
Where is there?
Ya, where is this magical land of pc hardware?
I live in the United Arab Emirates.
Seeing that scrapped RVE 10 hurt my soul so much.
Yes, very sad to pull such a board from the scrap. If I would have gone through the entire pile, I'm sure I would have uncovered more hardware of that calibre.
might be related, but i worked with pentium 3's in an industrial setting where they ran Siemens PLC stuff (Step 7 under Windows 2000 mostly) and C0 was common on high heat, high vibration installations where the BIOS was unable to verify cache size but was still detecting cache presence. if cache was disable entirely it could boot, but often we would send the cartridge or CPU (which was thermal glued to a small heatsink/fan combo unit from i think AVC) out for repair and the turn around could be as little as a day.
C0 was also used by early Xeons to indicate an issue with initializing L2 cache which was a separate IC on the PCB. this might be different though since pins for Slot 1 and Slot 2 differ and might have different POST codes. this initialization step was though through the BIOS because you had to have BIOS support for ECC cache.
this differed from the Oracle/Sun servers that ran the database and application end of things where the C0 was usually an add-in controller that had failed to boot somehow and you needed to identify and remove the device (disk controller with bad battery backed cache, misconfigured memory or video, anything where you had to use an unmasked interrupt to actually boot).
It may be more convenient sometimes or safer against loss to not desolder a 2-pad component fully but to rotate or shift it such that one pad is connected and the other isn't, in order to measure it. Especially if you move to more hot air soldering.
I agree. I did lose SMDs before. Usually, I leave one terminal connected to a pad with the component rotated.
Try measuring resistance/capacitance on each pin of the slot connector to a known working slot 1 cpu, see if there is a glaring difference?
Good point. I put that on the list and try that if there's nothing else that will fix it... Because, you know, 121 pins on each side.
you're really good with the SMD stuff. Seeing proper SMD soldering is always ASMR. lol I just love seeing the solder flow exactly where it's supposed to go
Hehe, thanks. I think it's all practice. I did a lot of soldering about 20 years ago. Looks like it's a skill you don't forget.
@@bitsundbolts I did too, but not SMD. lol I've never done SMD. I did get a rework station (it helps with 3d printing sometimes, but I don't use as much as I thought I would). I'm tempted to take apart an intel laptop that died 2 months after I bought it during covid and practice on that for shits and giggles. (it's why I never followed through on the warranty and it's the last intel I ever bought).
Could be possibly a broken solder joint from the CPU to the PCB?
Yes, it could be! I tried to look between the substrate and the PCB. You might be on to something!
awesome, BuB!! love it for the 20 min!! :) i also have an Athlon XP which has chipped corners, works flawlessly
It's quite funny how some dies are in horrible condition and they still work! I have a Duron that looks like a total disaster and I have a feeling that there is some piece of metal sticking out from the chipped silicone die. Stull works though without issues... Go figure.
I had a Thunderbird 1.2GHz (socket A) back in 2000 or so - the die was extremely chewed up by the crappy CPU cooler (very small die, very large pressure spring) but it ran like a champ. Could fry eggs on it.
@@bitsundbolts yeap, this strangeness is fun in retro computing. today it's like this: chipped corners, cool, let;s see if it works :))
Oh My god gtx 690
in scrapyard
With enough time, you'll probably be able to find anything you can imagine over there.
Incredible and awesome!
I would have grabbed that ZOTAC legacy PCI card :)
I've learned a lot! Thank you!!! Also for the insights of a scapyard in the desert (it kinda hurts to see this!). Here are some unqualified suggestions to your POST-problem: do all pins of the cpu work? Looks really scratched up. One other Idea: is there a possibility for a "flexible" resistor? Maybe the one resistor that is different on PIII 650 and 1000 have to be different. Worst case is defitely a broken coppermine die.
Thanks for watching and I'm happy to hear you found value in that video. My thought process is similar to what you mentioned. I did actually check the golden pins. They are indeed scratched badly. On one of those pins, I removed a small piece that was getting too close to the neighbouring pin. Other than that, they look ok, even though they're in bad shape. The similarities to the working PIII 1000 are interesting though.
I could be wrong, but I think a post code tells you what completed. So if I'm right it's failing at the processor test
are those non labeled resistors by any chance connected to pins bsel1 and bsel 2 on the cartridge connector if yes those two resistors were only responsible for changing between 133 and 100 or 66 bus and high values make sense since on 133mhz both of those pins should be high impedance and that would make it a pentium 3 with 133 mhz bus of unknown multipler and speed but you could decode that by multiplier pins too
biggest problem you may have with cpu that trashed might be core delamination so i would try pressure on the core substrate while booting
The front side bus is selected on the board. I haven't looked at those pins you mentioned. I'll have a look. Maybe you're right and there is an issue with the connection between the substrate and the PCB of the cartridge.
@@bitsundbolts it would be great to have a late super slot socket adapter with all the jumpers for voltage and bus like ms6905 master so you could reverse engenier voltage and bus jumpers, it also have that ic that got mangled on your cpu but multiplier i think is burn into the die, so that would have to came from intel technical specyfication for slot 1
but i would start with poking the cpu core substrate to see if it doesn't have broken balls or look closer what that ic do because it might be related to cache
I think it's a PIII 667/256/133 Coppermine SL3XL. Because i found pictures with the same modelnumber, PB 746509-001 and the same PCB layout, on a russian auction website.
At 10:51 you can see the connector looks super filthy, maybe that's the problem. I would try some magic eraser with alcohol to clean it.
I cleaned the contacts again with a magic eraser again, but no change - still stops at C0.
I guess I would go for the chip with the bent legs, because you know it had drama.
I hope you wrote those discrete component values down so someone else doesn't have to dismantle a spare to get that information
Yes, I did. I'll add those details in a separate post on my website in the near future.
The dual cpu motherboard would have been cool, I like looking through the footage of the scrapped motherboards.
It was a generic Dell/HP board that was screwed to a desktop case. I usually do not bother looking at those because it's hard to get information on those - or even find mods, BIOS updates, and user stories.
I wish there was something different that could be done about scrapping old hardware. Anything retro is worth 100x more as-is on Ebay than whatever they'll get from scrap but I guess most people, including the employees, just don't realize it.
The people working at the scrapyard don't care what they have in front of them. They separate different materials: cases, plastic, power supplies, electronics... They don't know what's valuable and what's really scrap (80-90% is Dell & HP stuff), then you get servers, switches, and other enterprise hardware. All there for scrap to get the valuable metals back.
@@bitsundbolts I work at one of those scrapyards and sort the electronics, I do care what's going though and I recover some of it. I just picked up a nice 386 MB yesterday. There is so much and you can't save it all. I have recovered some really nice equipment and have even sent some to UA-camrs.
Oh, I didn't want to generalise. When I said people don't care, I specifically meant the scrapyard I go to. It must be nice to find some exotic and rare items. I'm sure the people you sent those items to were very happy and thankful!
Where do i buy a post code analyzer?? I do similar to you
My second computer was a P3 running at 800MHz, but it was mounted in a socked. What's the difference between a socketed CPU and one that's mounted on a cartidge?
I remember my friend's dad upgrading his P2 computer using a cartridge, but I never bothered asking why his computer is set up the way it was, as I was pretty young.
Many cartridge CPUs had level 2 chips which was the reason to move to this design. Then there were Celeron and Pentium III that had the level 2 cache integrated into the die and therefore no longer needed the cartridge design. Those CPUs could fit in a socket again (socket 370). To make them compatible with slot 1 boards, the adapter card was a neat solution to allow users to continue using their boards with new CPUs. I don't think there's any performance difference when you compare a Coppermine 800 for socket 370 or slot 1.
a lot of these cards had a horrible problem with the board flexing in the CPU not fully contacting to
the board
Yes, I had my fair share of issues with a cartridge CPU not making proper contact with the slot. I tried the CPU in this video multiple times to be sure it's not a connection issue. Also, the same behaviour on a different board. I don't think I have an issue with the contact. It's most likely something on the cartridge.
@@bitsundbolts and i'm guessing it would be really hard to swap card number 3 and mystery cards cup
wow your skill is amazing good job sir
Thank you 🙏
Hi, there are some socket 370 pentium 3 1000eb 1.7v and 1.75v. Do you think the last could pull less amp from motherboard vrm because of the "higher" voltage?
Hm, I am no expert, but I would expect that at the same frequency, the CPU with 1.75 volts would pull the same current from the VRM. That would lead to a higher power consumption and probably slightly higher temperatures.
Good job, I don't know if you have another motherboard that natively supports P3 1ghz like a P3B-F or equivalent, it would be ideal for testing, otherwise in the last resort just doing a reballing.
I did try on exactly that board: P3B-F. I got the same behaviour though. The POST analyser card stops at C0.
0:35 now THAT would be a pin straightening challenge. 😆 don't tell me you didn't consider it! 😉
While looking at the footage, I had the same thought that it must be a pain trying to fix it! But that board stayed at the scrapyard. There are for sure a lot more issues than just the socket.
@@bitsundbolts I'm fairly certain it's just not possible. The contacts on modern CPUs are way too sensitive and too close to their physical limits with not much room to spare for tinkering. The system probably wouldn't be stable or may even pose a fire hazard. Thankfully it doesn't matter though, most of these boards aren't worth preserving, let alone saving, anyway. 😅
I fixed a single pin once on an X58 board? At least this is what I remember. The boards with triple channel DDR3 support for i7 920/930/950 etc. you could just put a Xeon 6-core/12-thread server CPU in there and get a half-way decent system (about 6 to 7 years ago)...
@@bitsundbolts Impressive! Luckily, I haven't had to resort to this so far, and there's a good chance I would just make it worse. 😅 I suppose it also depends on how important/mission-critical the machine is. If it's just a redundant or hobbyist machine and just one pin, one might get away with it.
Coincidentally, I still have an LGA1366 board (P6T Deluxe v2 with an i7 9somethingsomething) in my basement. With its 3 PCIE slots, it still worked well for ETH mining, as long as that was a thing. 😁
Haha, right. Now they're on proof-of-scheibenkleister. I also have an X58 in my pile of hardware to be tested. I think it's also an ASUS/ or Gigabyte - I need to check that. It was just one pin and I tried to flip PCs at that time. Didn't work well. A lot of work, no financial reward. Almost like UA-cam 😂
I am surprised there are scrap yards in Dubai but I probably shouldn't be because they are everywhere.I buy a lot of motherboards, cpus and graphics cards from scrap yards in Vietnam. So many socket 7 boards, Slot 1s and the odd 486 and 386. Some need a lot of work.
I'm pretty sure that's a 1.13ghz SL4HH Cpu, but is only speculation as it didn't have the other resistor, but I have one and the cpu number on it is 7122A223
Oh oh... Really? I thought the 1.13 GHz was the only Coppermine that requires 1.8 volts. How unlucky that it didn't Post. It would have been a pleasant surprise.
did you also save the VRMs?
No, I did not take the VRMs. I guess, that was a mistake.
What if some of the resistors are to configure CPU clock speed? It may be getting overclocked and crashing at boot. Maybe try underclocking it by lowering the bus speed?
I had it set to 100 MHz FSB. Since it's a Coppermine, it should have no problems. But good call, I'll try 66 MHz and see if it changes anything.
@@bitsundbolts Also try 133MHz. As I mentioned previously none of my three Slot 1 133MHz Coppermines will POST when underclocked to 100MHz. Perhaps this is the same problem.
Good point. Btw, I could reproduce that with a socket adapter and a Pentium III 1 GHz for socket 370. It's a 133 MHz model that refuses to post at 100 MHz FSB.
CooL! A P3 SECC processor from 20 years ago. I'm sure many people will want to mod that old dinosaur.
These days, it's just scrap and can be had for cheap. I would never attempt to do mods on items costing hundreds of dollars.
Did you happen to grab the voltage regulator modules from that dual slot 1 board? I could use those :)
Ahh, sorry. No, I didn't. I wasn't interested in the board and didn't think that those VRMs would be useful.
i want to see all the stickers used up by the end of this year!
Haha, that will be a challenge! I need to quit my job and only repair.... about 6 items a day 😂
Love the videos!
Thank you
@@bitsundbolts yw 😄
It is possible that the balls under the BGA chip have broken. Reflow or rebaling can be performed for testing.
That seems to be a common suggestion. I guess I'll attempt to do that.
So, why not just bridge the pads with either a wire or solder blob? Seems a bit silly to use a 0 ohm resistor in that situation. I know why they do from the factory, easier for pick and place, but when you're doing it, I don't quite understand the point.
It's personal preference. Yes, you could use a wire or a solder blob to bridge the contacts. I just have endless 0 ohm resistors which I can use. Also, if I were to use a solder blob or wire, someone would complain 😅
@@bitsundbolts Given I've seen people complain about astrological events occurring on a school day, I wouldn't put too much stock in it, but fair enough.
Try cheching all the legs on the IC or just reflow them.
Some other traces might have a scratch. But it sucks to go over the board in a mm grid heh.
There might be other traces, but I haven't seen any. Maybe I need to take another closer look.
Try it after temporarily removing CPU cache chip.
It's a Coppermine, the cache is on-die, there's no cache chip to remove.
@@Ale.K7 "then decap the chip"
If it would have been a Katmai CPU, it would be possible. But as @Ale.K7 said, the cache is on-die.
Wow! Such a scrapyard i would like to search through 😱
Would you mind to Sell some of your 1000MHz P3 Slot 1 CPUs? Im from germany and would very Like to get some, even if broke of some sorts 🤓❤️
I think I have four Pentium III 1GHz (3X 133 MHz FSB and 1X 100 MHz FSB). If you're interested, please contact me through email and we'll continue there. Thanks
CPU NO BAD
Not living in Germany but in another European country, I wonder where I could find scrapyards like this. How does it work in Germany and how do you get access? Local municipalities here have recycling stations but once something is handed in, nobody can take it out anymore (I guess to prevent employees trading/selling things).
EU regulations forbids access to this. You must be friend with owner.... circular economy in EU visions...
I'm not in Germany. I'm in the United Arab Emirates. I heard many stories from viewers how difficult or impossible it is to get access to a scrapyard (especially in the USA and Europe). I guess the only way is to build friendships with owners or employees of those facilities.
Before I finish the video, I'm gonna guess a Pentium III Coppermine as it doesn't have the two L2 cache SRAM chips?
Yes, that is most likely true, but what exact model it is will only be known if we get it to POST. you're correct about the SRAM chips. Coppermine Pentium III do not have cache chips.
Given the amount of physical damage, I think that you may have to "rehot cpu friend!". Seriously, if you can't find any other damage, solder balls under the CPU are the most likely thing, other than the silicon or the board the CPU die is bonded to being damaged. Since those are probably impossible to fix, I would suggest a reflow or even a reball.
I may attempt a reflow / reball if I can't figure out anything else. It's also some extra practice!
Почему не попробовал определить коэффициент умножения?
П.С. Попробуй запустить на FSB 66 МГц.
Since this is a BGA chip on that PCB, check the solder balls
That's a very common suggestion. It's probably what I'm going to do
Pentium III 1000EB SL4BS C0 rev. but maybe the core not work basically .. that is the reason scrap
That would be nice if it's really such a CPU. You'd be surprised what the percentage of working (or repairable) hardware is I find there. If I exclude hard drives, it's easily above 90%.
odd......I thought by 1ghz they were already done with slot 1.
Athlon vs Pentium III: the race for 1 GHz. I think socket 370 and slot 1 Pentium III were released at the same time. Slot 1 was declining already and the 1 GHz model was one of the last CPUs for slot 1. So, you probably already moved on to the new socket 370 platform.
@@bitsundbolts AMD and Intel both switched to sockets at the same time. I believe the Athlon Thunderbird came out at the same time as the coppermine. It had the integrated cache. I had one. I don't think it was available in a slot configuration. And yeah I do remember Intel and AMD both hit 1ghz at about the same time. Within a week.
Yes, they indeed released their GHz flagships at about the same time. I was happy, because I had AMD back then. I continued with Athlon XP and 64 X2. Then Core 2 Duo.
@@awilliams1701 @bitsundbolts Yeah I read somewhere that AMD released the 1GHz Athlon two days before the P3 1GHz came out! And another fun fact - AMD didn't have on-die cache at this point so they had to set a 2/5 cache divider to get the cache to work stably
@@bitsundbolts at one point I had an Athlon XP mobile CPU in my desktop. I have no idea how that happened. You'd have to ask my cousin. He gave it to me. I'm like.......how the hell is there a mobile cpu in a desktop? lol And then I switched to Intel core 2 quad, pentium-dual (2 cheaper systems used that), I7 laptop, i7-2600 desktop. And now back to AMD 4700u and 5900x. I think it might be time to replace the 4700u laptop. It has served me well though. AMD has RDNA based CPUs now. I was hoping it would play KSP2..........except that now KSP2 is dead. Funny thing. The 4700u loads KSP2. The menus are buttery smooth. And then when you load an actual game it crashes.
Broken solder balls, maybe?
That's probably what I'll try to find out.
none of the markings on the board are much help, there is a whole series of PIII's that it could be, but if you can manually set the multiplier, try setting it up as a 450mhz 100mhz 4.5x as that is the lowest spec for that board, the lowest 133 fsb is 533 under clocking it has more chance to get it to boot than over clocking it.
I'll try that. There are some CPUs that follow the settings of the motherboard for the multiplier. Most slot 1 CPUs have a fixed multiplier - so, the setting on the board will be ignored. I'll try it anyway! Thanks for the suggestion.
@@bitsundbolts it's been a long time since I worked with slot 1 CPU's but there were some boards that you could manually set the clock and multiplier. if I recall some CPU's had an internal clock doubler, others didn't. but my memory might be a bit fuzzy. I think one of the slot 1 SOYO boards would allow us to manually set up the CPU, but can't remember if that was for PII only or if it worked on PIIIs too.
Yes, all of this was there. Clock doubling was introduced with 486 DX2 CPUs and then DX4 with the clock tripled. From socket 3 to 8, most settings were configured using jumpers. Then came fixed clock multipliers and the ability to configure CPUs using the BIOS. I think only a few early slot 1 CPUs have an unlocked multiplier that can be set by the motherboard.
I See Dual lga 775 socket Motherboard
Maybe You Should checkout(And Restore)
Haha, those are mainly server boards. I don't mind bending some pins on 486 CPUs back into place, but those LGA sockets are on a different level... I don't know if I want to get into something like that. Maybe as a challenge one day, but it should be a board worth saving.
@@bitsundbolts LGA775 (or in this case, LGA771) have relatively large pins, it shouldn't be too hard. The newer LGA sockets have a much greater pin density. Even then it's not too hard to correct a couple of bent pins. For a hard time find an LGA3647 board.
1. GD900 is high performance thermal paste??😂 Sure No!
2. Next step to reball Cu-Mine CPU?)
Ты не прав. GD900 среднего качества дешёвая термопаста. Для проверок и экспериментов, отлично подходить. У меня 2 пасты GD900 и MX4. MX уже наношу при окончательной сборке.
@@denyda1754, и в каком месте я не прав. В том что термуха нижнего эшелона и годиться конкретно для проверок на включения? Или для совсем "плюшевых" процов?
Я повторюсь и обращу ваше внимание на рофло надпись на тюбике "high performance".
@@ApostolCV На стене можно написать XYN, от этого стена не станет органом :)
If corner is damaged, most of time, it is dead.
Most of the time if the corner is slightly chipped they still work since the actual transistors are only on the very thing bottom of it.
I've had several Athlons where the die has corner damage but the CPU works (and even overclocks) without any issues. No experience of Intel chips though with damaged dies.
@@RuruFIN (Edit - misread your post) ... I have a duron 900 that specifically survived this.
I did have Intel CPUs with damaged corners and chipped edges. But they usually do not show any POST code.
My 370 866MHz coppermine has damaged corners but it works just fine
😢 omg i can't listen to this anymore! Man! You really need to focus on intonation, REALLY. It's the same! Every sentence has the same ups and downs and it just tears my ear drums apart!
That pile of scrap hurts my soul :'(. Great video, shame the CPU didn't work :(
At least the other two Pentium III 1000 work 😅
@@bitsundbolts Yeah, that was nice! :D
It may need a reball, might be able to float the chip to test before spending the time on a full reball
Seeing the ROG Mainboard scraped hurts very much. I had the chance of taking a ROG Commando from a scrapbin. With the exception of the status display in the back not showing all the segments it works. Currenty I'm looking for parts to build a gaming computer like I wanted in my youth but couldn't (because of money). I'm still looking for a Radeon HD4890 with 1GB from Saphire and the waterblocks for the gpu's. Unfortunatly waterblocks for older gpu's or gpu's with waterblocks never comes up for sale. And I have no acces to electronic scrapyard like you have.
I'm sorry to hear that you don't find that GPU and the water blocks for it. I have a dual GPU from AMD, but I don't remember what model - and it might not work. If I find it and I remember, I'll have it in a video or on my website some day.
@@bitsundbolts I think a lot of these gpu waterblocks are getting salvaged for their copper. They are made from pure copper and and huge and heavy.
That is probably the first thing scrappers are after. It's easy to remove heatsinks. In the process, many destroy the electronics attached to it 😔
You could try to replace the big chip that was twisted in case it's dead
Foot point. Since it's twisted, maybe something happened internally. There is also a chip in it. Could be that due to the impact, something happened to it.
The POST code stops at C0 because the next step, likely Processor Test 1 (01) if the POST codes are listed in execution order, is failing. I don’t know what that test is trying to verify, but maybe someone else might have some insight…
I didn't even know Slot 1 CPUs went up to 1GHz - I thought that was only possible via a slotket o.o
1.1 GHZ actually, but those are more rare. I have a slotket myself with a 1300 MHZ Tualatin Pentium III/Celeron on it in my CPU collection, works well and is a very fast CPU for the platform.
Interesting. Great video, as always.
Thank you
I had dual 1ghz slot 1 p3 in a Tyan Tiger 133 board way back. super responsive system back when xp was the new thing
It executes code so a different motherboard with a non Award BIOS might be able to get it far enough to display the cpu info.
use a different cpu to turn off internal cache. or maybe the chip you moved the pic needs to be removed to see if any of the traces are messed up and if you are feeling froggy you can swap the chip on to another CPU that does run to see if it works
I thought about a chip swap. That IC with the bent pin must have been hit hard. The solder joints and the legs are quite strong.
Look closely at your video! R4 looks like it is cracked.
I checked. It measures 47 ohms in the circuit. I'll have a look at it once more under the microscope, but I believe R4 is ok.
Why did you say even though it's 133mhz fsb on the P3 1Ghz cpus? I've always seen those as the more desirable ones since they don't overclock much and higher fsb is better than the 100mhz fsb versions.
Correct, however, on an ASUS P2B, the AGP bus is overclocked if I run them at 133 MHz FSB. That was only fixed on later boards like the P3B-F. A 100 MHz FSB GHz CPU would work without issues on those boards. And if I'm not mistaken, the 100 MHz FSB PIII is less common and therefore more valuable.
@bitsundbolts Makes sense, and sounding very familiar. Now that I think about it I may have had a 100Mhz fsb P3 1 ghz in that same board at my job. It started life as a Celeron 300A that I overclocked to 464. I want to say it ended up as a 1ghz cpu but it was sold in the P4 era to an employee.
Maybe the balls underneath the cpu or even worse the balls between the core/die and the substrate are broken.
That would be sad. I may be able to do something regarding the connection between substrate and PCB, but not between die and substrate 😞
@@bitsundbolts indeed, probably is the ones in the pcb so you can try
Check the cache chip.
Кэш интегрирован в процессор.
OP neglected to break out his electron microscope to implement your suggestion 🤔
Well it makes more sense that the cache chip is bad since it is giving the code about turning it off.
love your videos thanks for share and doing them for us ;) at some time you should do a bench tour, so we can learn what tools you use for you work ;) if you done such, please link or point me to it ;)
Ah, I haven't done a bench tour. I'm ashamed that I don't even have a dedicated bench for my work. It's also my home office desk 😅. Some day, I hope to change that. Thanks for watching!
At 10:58 the chip on the back looks crooked, maybe have some lose pad.
I checked the chip. The pads look good, but I'll check what this chip is for and maybe swap it with a different one.
R4 looks cracked.
Yes, you're right. It does look cracked. I checked and removed it from the processor, but it looks fine. 50 ohms - similar to a resistor on a working CPU.
Another great video. What soldering tips do you prefer to use? Have you ever used a hollow tip? I remember seeing someone talking good about hollow tips on youtube. Thank you for the spreadsheet.
I'm using a simple bevel tip. My soldering iron is the FNIRSI HS-01 and I use the tips that came with it. I move on from chisel to bevel tips. That's a personal preference I believe. In my opinion, the bevel is better in transferring heat to solder joints. I never used a hollow tip, but it looks interesting with space to store solder.
@@bitsundbolts Thanks for the answer. I need to buy new tips for my soldering iron. I'm tired of using big tips for small jobs. I'm gonna buy bevel (I'll see if I find an hollow tip as well to buy).
Someday I'll also get a better soldering iron, like the FNIRSI HS-01, but for now I can't justify the expense because it's rare for me to solder. I have a generic Lidl soldering iron with a knob to control the temperature.
I guess the CPU dye is cracked.
All cracked dies I had before did not show any POST code. Since we see C0, I assume the CPU die is working.
@@bitsundbolts C0 not rare with cracked corners
Hm, that would be too bad. Then we probably would never know what CPU that is. Let's see if there's something else we can try.
Can the 2D matrix code help identify the chip? Or is that just some unhelpful serial number?
Do you mean the number on the substrate? I googled that number, but nothing came up. Maybe it's an internal number for Intel to keep track of production batches, but not meant for the public.
There is an app from Intel to scan and identify those 2d datamatrix markings on their products but at least in my attempts it doesn't seem to work with CPUs as old as the Pentium III
I downloaded the Intel toolkit 3.0, but that application wasn't able to tell me anything. It doesn't recognise the QR-code like matrix 😔. But thanks for the great hint!
I have a slot 1 that take katmai pentium iii max, it wont take coppermine, what can i do to make it run coppermine?
volt mod presented by BuB in the P2B repairs series and the bios mod look for bios patcher.