• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

[DEAD] Celeron 633 $17 @ Directron, free S/H

Page 5 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
I'll try and stop by the local electronics store today and pick up the magnetic wire.

As it is, I may be sunk anyway. Too ka quick look at the pin last night and got the wrong location for the loose pin. It seems to sit at the SMI location, not the VCC site. A little research leads me to believe that this is tied to the system's sleep mode. If so, that may be the reason the system hangs upon booting as well.

In case this is the problem, two more questions:

1. If the pin sets into the socket, irregardless of how well connected it is to the chip, will it serve its function, or will it need to be resoldered on tightly to the chip?

2. Is there an alternative to soldering if it does need to be re-tightened?

Getting the feeling that it may be a lost cause, but I just want to check before shopping for the wire.

Thanks as always!
 
Hey everyone, just wanted to give you all an update. I used the link that mindless1 provided about modding a PPGA slocket to FC-PGA to tried it out. I had to insulate one pin and solder a piece of wire from two points on the rear of the slocket. I had no idea if the cheapo generic slocket would take the mod or not. So I said what the heck, I slapped it into the Tekram and /emeril voice "BAM"! it POSTED!!! I was like, well d@mn, that mod does work. 😀

Thanks again everyone and especially mindless1 for the linkage. 😉 Don't forget to set your slockets to 1.7V, mine would only go down to 1.8V but I don't think a little more juice is going to hurt anything. 🙂
 
Originally posted by: AllPurposeNothing
I'll try and stop by the local electronics store today and pick up the magnetic wire.

As it is, I may be sunk anyway. Too ka quick look at the pin last night and got the wrong location for the loose pin. It seems to sit at the SMI location, not the VCC site. A little research leads me to believe that this is tied to the system's sleep mode. If so, that may be the reason the system hangs upon booting as well.

In case this is the problem, two more questions:

1. If the pin sets into the socket, irregardless of how well connected it is to the chip, will it serve its function, or will it need to be resoldered on tightly to the chip?

2. Is there an alternative to soldering if it does need to be re-tightened?

Getting the feeling that it may be a lost cause, but I just want to check before shopping for the wire.

Thanks as always!
Click to expand...
Since you only need about 1" of wire, you might ask politely if they're throwing anything away, or check a repair shop, since a few inches of that wire is worth < $0.01 but a whole spool might easily cost $5-10 or more. If you have an old "wall wart" AC/DC adapter lying around, so long it's not the switching type (most aren't) it would have plenty of wire in it too, around the transformer, under the outer wrapper... you could even dismantle the wall-wart using the "brute-furce" (hammer) method :Q the switching type also has a tiny transformer but less wire, harder to cannibalize.

I don't know if the CPU will work OK without the SMI pin or not. When you wrote that it "hangs upon booting", do you mean that it IS displaying video, the POST screen, or not? If it's not POSTING, no video at all, focus on the voltage increase, a no-video situation is exactly what would happen with a voltage that far too low.

IF it turns out that this loose pin is a problem, there "might" be a few things to try later, after increasing the voltage... they are involved and I have no idea if they would work, but involve things like using APM, not ACPI power management mode, or removal of the STPCLK signal (the pin directly beneith that one on the pic), but do not worry about these things yet, I would focus on getting a scrap piece of wire and trying it AS-IS. I think it's most likely that without that pin, the CPU simply wouldn't be able to enter the system management state, "sleep mode", which means it would use more power at idle, but it's likely not a significant problem, not with the CPU using a relatively low amount of power to begin with, and considering that many other more demanding (Athlon) boards don't allow the CPU to enter the sleep-state at all without special BIOS feature or user-hacks. The SMI signal is pulled-low when active, meaning without the pin, the mode simply can't be enabled, IF I understand it's function. This leads me to believe the CPU will function 100% normally in "processing/use".

The pin being connected to the socket properly means nothing if it's not contacting the CPU, but if it's attached at all, even by a tiny patch of metal, it should still have sufficient electrical contact to the CPU. This is a logic signal, doesn't need to 'sink current, so I would suspect that any attachment at all is enough. I don't have high hopes for you being able to solder it, I imagine it would be extremely hard to do so with a pin in the middle of other pins as that one is. Spot-welding it might almost be easier, but this is drifting into theory, in practice it isn't a reasonable solution either.

I would not yet assume it's a lost cause, you still have yet to do the one most important thing, raise the voltage.

On the other hand, if you wanted to check function of the CPU with the pin missing, all you would have to do is use same wire as used for wrapping the pins, but only ~5mm long piece, inserted into the BSEL0 and BSEL1 socket holes before inserting the CPU. You might need an extra mm or two in the hole where the broken-off BSEL0 pin was, to be sure it's touching the contact in the hole, which would make the board run at 66MHz FSB again. Just remember that when looking at the top of the socket, it's a reverse, mirror-image of the pin-pictures I've linked. You still might not be able to assume the wire was making good contact in the BSEL0 socket-hole though... if it DID work, you would know the CPU was still fine, but if it didn't work, there would be uncertainty about the contact.

On the other hand, I would just skip doing that and go for the voltage-increase first.
If all else fails, send me your CPU and I'll see what I can do with it.
 
Hi...
I'd really liek to buy this, but I'm not sure if it will work for me. I have an old BX board, can't remember who made it, and it has a P2 233... plugged in vertically. Can i use this chip, and if so, what else would I need? I do remember that the mobo will take about an 800mhz chip after a bios update. Thanx for any help.
 
legoleg, you may need that bios update and either a FCPGA compatible slotket or a modified PPGA slotket, AND possibly that slotket may need support manually setting 1.8V if your board can't go < 1.8V, or moddin' the slotket for 1.8V too. Of course there are no guarantees with old boards, nor with overclocking if you got an old chip, but "one way or the other" you'd have at least 633MHz, providing that BIOS really does support 'em or at least doesn't halt the system. A Google search for your particular board make/model/revision might provide detail.
 
legoleg, I have an old abit BX motherboard, and using the Abit slotket III I have a fc-pga Celeron 1GHz CPU running. All just plugs and plays, not messing around with anything.
 
Originally posted by: AllPurposeNothing
I'll try and stop by the local electronics store today and pick up the magnetic wire.
Click to expand...

Why not use a wire from a floppy ribbon cable? It's what's typically used for pin mods on Athlon XPs.
 
Confucious say "Clumsy man with big stubby fingers should not overclock cpu."

Took a st*b (can't believe the forum won't allow this word!) at modding the voltage. Pulled apart an old AC adapter, took off the wire, wrapped it around the pins and....nada.

Went back and re-read mindless1's instructions, realized the AC adapter used multistrand wire, and that he has said not to use this wire.

Found another AC Adapter, opened it, and found the solid wire. Wrapped the wire around the pins, twisted twice to tighten around both pins, and snapped one of the pins off of the cpu.

Screamed incessantly, took the hammer I used to open the AC Adapter, and broke all of the pins on the CPU.

Swore I would never attempt a mod again, and the next morning ordered another similar CPU. (What can I say...I'm a glutton for punishment!)

And now, I'm toying with ripping apart my son's XBox for a mod.

I'm turning into my Dad....lord help me!
 
mindless1-

I have that effect on people. 🙂

I know I should probably give up, but I'm one of those people who will try anything twice.

Don't know a lick about pattern making or sewing and somehow put together my son's Halloween costume (he wanted to be Sonic the Hedgehog). Read a book and have been performing minor VCR repairs for friends for years. Hardly knowing a thing about soldering or wiring, I modded a Sega Master system controller (I believe) into a Vectrex controller some years ago.

A night's rest gave me a clear head and I'm ready to learn from my mistakes and try again.

To be honest, I"m really hoping that the next CPU ends up being 1.7v with cC0 stepping but there are no promises. But I"ll learn from my mistakes and hopefully succeed this time.

As for the XBox, I'm not crazy enough to dig into the insides myself. I"ll either software mod though the Agent Under Fire loophole or pay someone local to install the modchip and new HD.

Anyway, I'll let you know how I progress when the new CPU arrives.

Thanks for all of your help, BTW. It has been extremely informative up to now.

 
AllPurposeNothing,
if the pin you'd broken off was the VSS, you could've wrapped the wire around the VID3 and VID0 pins, it would've had the same effect. If you'd broken off the VID3 pin, and also had previously broken off the VID2 pin as per the pic I linked, the CPU would'd been set to 1.45V, which "might've" ran at 633MHz. if you had not yet broken off the VID2 pin, it would've still ran at it's default 1.65V.

To anyone with a soldering iron, board or slotket with voltage adjustment, your CPU was still worth something.
 
mindless1-

Yes, it was the VSS pin that broke. I had already broken the VID2 pin off. I was just tired and frustrated (slept maybe a total of 20 hours the past six days) and figured it was better taking it out on the chip than the rest of the computer or the wall. Had I known I could have used the other pin, at the time, I"m not sure my reaction would have been much different.

As a fair assessment, the other pins aren't broken...just very bent. But I believe the cpu itself bent and split so it doesn't make much difference. If you think the chip has any possible life, you or anyone is welcome to it...probably not worth the effort now.

However, if the mod doesn't work the second time, anyone is free to have the second cpu. I promise to have a better head about me when working with it. Promise!

 
Got a SL3VS chip (cB0 stepping) from Directron and now just awaiting the slocket which has FSB jumpers but no voltage. Before I start thinking about modding the voltage, has anyone who got this stepping gotten it running at 950 MHz, and if so at what voltage?

Also, as for the voltage mods, I'm assuming insulating a pin is an alternative to breaking it off, and since I'd at least like to be able to use this chip at 633 MHz I want to keep the option of going back. How does one insulate a pin?
 
A long time ago I had an SL3** cB0 Celery 566 @ 952, needed 1.85V.

To insulate the pin, cut a ~ 3mm piece of insulation off of an old ATA33 or floppy cable, slip it over the CPU pin, then use a tiny drill bit or jeweler's screwdriver, etc, to CAREFULLY ream out the socket hole just large enough.
 
LS52X from celeron 766. best cumine o/cer EVAR. (cD0)
you can get cD0 from 800->1100, but those are all 100fsb, while 766 is still 66fsb

use ata/100 wire
 
Well, here we go again....

Just got the second CPU from Directron, and once again got a 1.65 v, cB0 stepping one.

Before I go modding this, I think I may take your advice from way back and try modding the 566 instead. I'll put this aside as my back up in case the 566 mod doesn't work. That way, at least I'll have some minimal speed increase for the time and money.

So here's my new question:

With the 566 Celeron chip, I'll insulate the BSEL0 pin with nail polish or the ATA33 trick to get the 100 FSB (I think given my track record this is a better idea that snapping it).

Am I going to set this chip at 1.85 as well, or would a lesser voltage work? My understanding from an earlier post would be 1.85, but I just want to make sure.

Any help is appreciated.

Thanks in advance.
 
AllPurposeNothing,
your 566 @ 850 might not need 1.85V, but I think it will need at least 1.8V. I'd go ahead and try 1.85V first, just to be sure it'll do 850 stable. The odds are very good that it can do 850. THEN, after you have it confirmed working right at 850, you can consider IT the backup CPU, and proceed to mod the other one if desired.
 
Originally posted by: mindless1
You can't run Coppermine/FCPGA Celerons in Dual/SMP mode, the last Celery to allow that was the PPGA Celeron 533 (and slower models of Covington and Mendocino).

The last Celery I bought from Directron (after they were already old tech) was a 700 or 733Mhz. It would not run stable at > 1GHz no matter what Vcore I tried. This doesn't bode well for old Celerons from Directron, but it's a LOT easier to get one running at < 1Ghz (950MHz). I'd think it's a likely, reasonable goal but no guarantees of course..
Click to expand...

That's not really a surprise, considering the recall of the 1.13Ghz PIII chips. Intel had to do another rev (cD0 stepping) before they could get > 1Ghz clock speeds stable. These chips are all of course pre-cD0-stepping, most likely.

 
From what I vaguely remember, the biggest difference between steppings was from cB0 to cC0, where the cC0 actually added another layer to the core, then cD0 was just a further tweak of the cC0 redesign. My experience with cC0 was that it's hit-or-miss how high they'll go, with some not doing much over 1GHz but others quite happy till around 1.2GHz.

A few years ago I set up a cC0 Celeron 700@1050 for my Aunt, she's been using it trouble-free all day, every day since. I also still have a cC0 Celeron 600@1200, is a rare example of a CPU that can run "double-speed". From what I recall the cD0 step was a lot more likely to hit 1.2GHz, maybe even 1.3-1.4, but that's pushing it, and by the time the cD0 came out I wasn't messing with Coppermine Celerons anymore so I have less first-hand experience with the cD0.
 
I have a old slotket with a celeron 400 MHZ and works fine, When I pop the celeron 633 in I get nothing.

IS this because my slotket?
Here are the jumpers that are on the slotket
66/100/133 HMZ
intel or cyrix jumpers

NO volt jumpers
Also says ver 1.0.

I just wanna know so I can either get a newer slotket or a new socket 370 board with FCPGA....
 
Originally posted by: DJMiX
I have a old slotket with a celeron 400 MHZ and works fine, When I pop the celeron 633 in I get nothing.

IS this because my slotket?
Click to expand...

Yes. That 400Mhz Celeron is PPGA, which isn't pin-compatible with the FCPGA pinout. You should be able to mod it, though (i believe someone (mindless1, most likely, as he appears to be the main man for this info) posted a bit of info and a link or two earlier in the thread). As for a replacement, a new slotket will probably cost you somewhere between $5 and $10, or a new/refurb S370 board between $20 and $40.

I snagged a refurbed P6STMT from Newegg, and it's actually impressing the hell out of me, especially for $29. I wouldn't play anything newer than Quake II on this Alice (American McGee's), but for everything else, it's surprisingly stable (she sits and runs SETI@home 24/7, only gets rebooted when the power fails or i'm doing something like adding the router card i just got). Not very zippy (the performance is just slightly better than Konoko's (P-II 450 @ 504Mhz, 128MB SDRAM, FIC VB-601), but for less cache-intensive things, she seems to perform better (boots into Win98 in under 40 seconds, something even the other Alice (Resident Evil, Duron 900 @ 1Ghz, 256MB SDRAM, Biostar M7VKQ) fails to manage).

And for those who missed the discussion last month, I build systems around female themes. Specifically, women in games and/or movies, using sounds and icons taken from the games/movies, cases fitting the image I have of them (and when i get my Dremel fixed/replaced, custom-cut logos, such as the Bloodrayne logo for Rayne, and the Triop logo for SHODAN (my main squeeze, speced in the sig)).

And while I'm at it, I need to shout out a big "THANKS!" to mindless1 for posting the deal in the first place. $17 for a decent CPU was just too good to pass up, and even though I didn't have anything other than the RAM (old Outpost 64MB deal), I figured I'd pick one up just in case. The other parts (board and case) all but fell into my lap after that, but I'd've never even given them a second look if I didn't already have the CPU. So, thanks, mindless1! You triggered a killer deal for me (complete system for under $90) and put me one step further along in building my digital harem. 😀
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top