It is soldered and can be removed using a razor blade and a small propane torch. But this requires some practice and courage.

I was wondering the same thing. I've delidded a bunch of Northwood and Prescott 478 CPUs. All the Prescotts were soldered, and the Northwoods were not. I've got P4EE 3.2 Gallitan 478(SL7AA) coming in the next few days. I was kinda hoping it was soldered. So I wouldn't need to delid it. Not that it's that big of a deal really(I accidentally discovered an easy and safe way to do it). I've just found it to be all but fruitless with the soldered Prescotts(so I'd assume it's the same with soldered Gallatins). Temps after delid with Conductonaut were pretty much identical to before with the Prescotts(maybe 1-2°C cooler if that). Waste of time. With the Northwoods however, it's a must. Massive gains to be had there(like 5-7°C cooler easily, maybe a bit more even).

Anywho. This is how I do it. Cut the silicone with a razor blade first. Then use the straight edge of a table top or bench top to support the lid with the PCB part hanging over. Hold it in place and whack downward on the PCB part with a flat block of wood. I use a 6" piece of 2x4. Don't get too crazy with it. Just keep whacking on it with light blows until the solder gives way. Works ever time. And I haven't killed one yet doing it. Way easier and safer than applying heat IMO. It's similar to the "stick it in a vise and beat on it until it pops off" method you've heard of for delidding non-soldered CPUs. It sounds rough. But if done right you've got nothing to worry about. Piece o' cake.

conenubi701 wrote:

So I just scored my second SL7CH today and will be testing/comparing the OC headroom on both of my chips. Whichever can't overclock as high would be de-lidded for testing purposes (and also to not kill the "good" sl7ch).

Part of me wants to kill you for even thinking of destroying a working rare chip for "testing" stupid crap such as this. You are no better than that arsehole who buys Voodoo 5's to destroy them. 😒

D'oh, I realize that I replied to an old thread bumped by some resident idiot delidder who just signed up. Welcome! I hope you will soon join OP in the special hell where they take your scalp off and pour some liquid metal on your exposed brain to improve cooling. 😊

dr_st wrote:

conenubi701 wrote:

So I just scored my second SL7CH today and will be testing/comparing the OC headroom on both of my chips. Whichever can't overclock as high would be de-lidded for testing purposes (and also to not kill the "good" sl7ch).

Part of me wants to kill you for even thinking of destroying a working rare chip for "testing" stupid crap such as this. You are no better than that arsehole who buys Voodoo 5's to destroy them. 😒

D'oh, I realize that I replied to an old thread bumped by some resident idiot delidder who just signed up. Welcome! I hope you will soon join OP in the special hell where they take your scalp off and pour some liquid metal on your exposed brain to improve cooling. 😊

*TRIGGERED*

Chill man, i know how you feel, but he bought it, it is his property, he can do whatever he wants.

I do know that many Northwoods and Prescotts were able to be delidded

Both high-end Northwoods and Prescotts use indium. It's pointless to apply liquid metal, it won't help you to overclock in any way.

The Serpent Rider wrote:

Both high-end Northwoods and Prescotts use indium.

What's considered a "high-end" Northwood? Just the 3.2 and 3.4? Or just the 800 FSBs. Because I've delidded two 3.06 533 FSB Northwoods and neither of those were soldered. And they were the highest of the high-end when they were released.

Anywho, I literally can't believe how cool this Gallatin runs. It's undoubtedly soldered.

"Idiot delidder"...I've been called a lot of things...looks like I've got one more to add to the list. 🤣

Even a lowly 2GHz Celeron which heatspreader I turned into a keychain had soldered on silicon. I would think most of the 478 parts are soldered.

I think all 800mhz Northwoods are soldered. Some late 3.06ghz D0 stepping CPUs probably too.

If my memory serves me right, all M0 Gallatins are solder bonded. Including those with the L3 cache disabled which were sold as regular Pentiums. All C1 Northwoods that I've delidded had a thermal compound. I don't seem to have delidded any D1 Northwoods yet, but something tells me there's no difference to C1.

Tiido wrote:

Even a lowly 2GHz Celeron which heatspreader I turned into a keychain had soldered on silicon. I would think most of the 478 parts are soldered.

You might think that. But you might be wrong. I'm thinking more than half of them aren't soldered. Maybe not a lot more than half. Maybe not even half. I don't know. But probably A LOT of them aren't.

BTW, do you remember which 2GHz Celeron it was? Willamette or Northwood? If Northwood, which sSpec number and stepping? Because the Northwood 2GHz Celeron SL6VR D1 stepping I delidded wasn't soldered.

The Serpent Rider wrote:

I think all 800mhz Northwoods are soldered. Some late 3.06ghz D0 stepping CPUs probably too.

There are no D0 stepping 3.06GHz Northwoods AFAIK. Just C1 and D1. There's a Mobile 3.06GHz SL7DT D0, but it's a Prescott. And AFAIK all Prescotts are soldered.

ph4nt0m wrote:

All C1 Northwoods that I've delidded had a thermal compound. I don't seem to have delidded any D1 Northwoods yet, but something tells me there's no difference to C1.

Well, I can authoritatively tell you that there is no difference then. Because I've delidded three D1 Northwoods in total. And they all had paste two of them had paste, the other one was soldered.

So far the Socket 478s I've delidded are:

Willamette
1 x Pentium 4 1.8(SL5UK) 1.8GHZ/256/400 D0 = Paste

Northwood
1 x Celeron 2.0(SL6VR) 2.0GHZ/128/400 D1 = Paste
2 x Pentium 4 HT 3.06(SL6PG) 3.06GHZ/512/533 D1 = Paste(on both one), Solder(on the other)
1 x Pentium 4 HT 3.2(SL6WG) 3.2GHZ/512/800 D1 = Solder

Gallatin
1 x Pentium 4 Extreme Edition 3.2(SL7AA) 3.20GHZ/2M/800 M0 = Solder
1 x Pentium 4 Extreme Edition 3.4(SL7CH) 3.40GHZ/2M/800 M0 = Solder

Prescott
1 x Celeron D 335(SL7DM) 2.8GHZ/256/533 D0 = Solder
1 x Pentium 4 HT 3.0E(SL7E4) 3.0GHZ/1M/800 D0 = Solder
1 x Pentium 4 HT 3.0E(SL7PM) 3.0GHZ/1M/800 E0 = Solder
1 x Pentium 4 HT 3.2E(SL7KC) 3.2GHZ/1M/800 D0 = Solder
1 x Mobile Pentium 4 HT 548(SL7X5) 3.33GHZ/1M/533 E0 = Solder
2 x Pentium 4 HT 550(SL7PP) 3.4GHZ/1M/800 E0 = Solder
1 x Pentium 4 HT 550(SL8K4) 3.4GHZ/1M/800 G1 = Solder

Coming soon maybe...eventually...probably not...

Prescott
1 x Pentium 4 HT 3.4E(SL7Q8) 3.4GHZ/1M/800 E0

UPDATE 2019-4-8: Corrections made for accuracy. I could have sworn I'd delidded both of my SL6PG Northwoods. Long story short...I hadn't...only the second one I bought got delidded. So I just delidded the first one to see what the actual story is with it. And to my complete shock, awe, surprise and amazement...IT WAS SOLDERED! 😲 Apparently some Northwoods ARE soldered! How/what/where/why? I haven't a clue. Two otherwise identical Northwood processors. One soldered...one not...WTF? 😕

Anyway. I'm tearing the lids off all of them now. Not to find out if they're soldered or not. I don't really care if they are or aren't. Since I'm going to be running them all bare die from now on. 😁

I'll edit in the results of those other delids when they're finished...

EDIT: Most likely finished with all I'm gonna do. We'll see on those last two...

UPDATE 2019-6-10: I ran across another soldered Northwood. That SL6WG D1 stepping. I'm not likely going to be purchasing any more s478 P4s or Celerons. I would like to get my hands on an SL7CH. But the prices are still a bit loony for what you get with those. Considering a 3.4 Prescott with a decent OC kicks its ass on pretty much all fronts. And can be had for ~1/4 the price...or less! 🤣

Also, since I don't have a setup to run the SL7Q8 bare die, I don't plan on delidding that one. I might do it anyway though. Who knows. But I can almost guarantee it's soldered...being a Prescott...and with the temps it runs.

UPDATE 2019-8-15: I ended up grabbing an SL7CH after all. Since I killed the SL7AA...😵. Too much HSF mounting pressure while running it bare die. Resulting in some extreme(pardon the pun) compression against the inner "ledge" of the socket, pinching some traces in the CPU PCB. Causing them to short...and let the smoke out. 😲 Also burned up a low-side MOSFET on my mobo in the process. All kinds o' fun!!! 🤣

Anyway...picked up another mobo as a replacement which came with an SL5UK. I delidded that and added it to the list above. Probably not going to delid the SL7CH. Then again...you never know what I'll do. 😊

EDIT: I delidded the SL7CH...and killed it too. 😒 Not with excessive mounting pressure...with like 1.68V. Seriously? That's the first s478 CPU I've managed to kill with voltage. The rest have held up just fine with 1.8V+. 😕

Unfortunately I no longer remember what chip was it, but it was the weakest chip I had in the pile of socket 478 CPUs I had, otherwise I wouldn't have tried to satisfy my curiosity. All markings have worn off over the years (5+) so exact identity of it remains a mystery...