Why Arn't CPU's Produced "Lapped"?

[Mr Krugga]

Intel should lap the 'K' CPUs and put another £10 each on the price.

Wouldn't be worth it. There definitely aren't that many K CPUs sold compared to their normal versions that are sold to OEMs plus Intel doesn't set UK's prices, they're all the same besides customs/import duties/VATs/retailers margins variables.

Realistically how many of these "K" CPUs is sold worldwide? 100k? 200k? 300k? $10 on top of the price would probably affect the sales in a harsher way than it would attract the enthusiast who will buy the CPU regardless. Lapping would have to be done in factories and that would drive the costs up significantly enough to disregard the idea. Not mentioning the potential returns from damaged CPUs.

 

[mmj_uk]

Please see post #9.

2500k/2600k are CPU's aimed purely at enthusiasts, it wouldn't hurt Intel to skip the step of fusing on a heatspreader for those particular chips.

Intel should lap the 'K' CPUs and put another £10 each on the price.

They don't need to lap, just don't put a crappy uneven heatspreader on in the first place.

 

[Jokester]

Is there a problem with the 2500/2600K chips IHS not being flat again?

 

[Richdog]

2500k/2600k are CPU's aimed purely at enthusiasts.

Correct... see post #21.

 

[tbh]

Proper lapping takes time. Its dirty, produces very very fine metal particles. These can get into the CPU, affect reliability and require more money to clean the CPU. Its an exotic unneeded process!

I believe I remember hearing that these are also highly carcinogenic?

Might be worth wearing a filtration face mask if this is true.

 

[MumboJumbo]

K chips arn't just aimed at enthusiasts. They contain the best graphics processor the 3000hd is it? I forget. Either way put in a H67 board they are the better HTPC chip.

 

[PhillyDee]

I believe I remember hearing that these are also highly carcinogenic?

Might be worth wearing a filtration face mask if this is true.

Its all contained in the fluid / lapping oil though! Lapping oil contains carborundum powder, which really is a gloopy mess, makes your hands manky just by looking at it.

 

[LaurenceGough]

I don't understand either, that air must make a really bad thermal conduction pocket.

My theory is that heat sink manufacturers make the modern heatsinks to fit the latest processors, at least from what I have gathered.

I'm going to stick about 3-4mm ball of paste on mine, a bit more than the rice grain I do normally after researching on youtube, it doesn't spread as much as you think.

And pretty much anything is better than air...

 

[JonJ678]

The electroplated nickel coating decreases thermal resistance. That is, a copper to copper interface leads to a hotter processor than coating them with nickel achieves.

The hope with lapping is that uniform mounting pressure improves temperatures more than losing the nickel layer worsens them.

It's far from certain whether lapping is a good thing. I dont believe it to be cost cutting, as it would be cheaper to skip the electroplating step.

 

[mangobreeder]

as it would be cheaper to skip the electroplating step.

+1 i find it hard to believe than intel would spend a fortune on R&D only to add a coating which the processor does not need?

 

[Lakeland]

Is there any proof that lapping to make the surface flat makes the difference? The thermal paste should fill in and gaps and transfer the heat either way.

I can see it making a difference if there was no paste, but with it..?

 

[Durzel]

Is there any proof that lapping to make the surface flat makes the difference? The thermal paste should fill in and gaps and transfer the heat either way.

I can see it making a difference if there was no paste, but with it..?

t does make a small difference to temperatures. The only way it would make no difference would be if the heat spreader material conducted heat perfectly to the TIM & heatsink/fan. Removing metal (the heatspreader), or even flattening it out and using less TIM as a result, means there is less material for the heat to be conducted through, and therefore a better transfer to the heatsink/fan you've got on it.

2500k/2600k are CPU's aimed purely at enthusiasts, it wouldn't hurt Intel to skip the step of fusing on a heatspreader for those particular chips.

Problem is you, and others, are talking about it as if Intel would have a guy there at the end of the assembly line or whatever spending however long it takes lapping each CPU by hand.

It simply isn't worth the money for Intel to tool up machines to actually do lapping when 99.99% of their customers would see no benefit from it at all whatsoever.

Don't forget CPUs & heatsinks are slapped together on OEM stuff with whatever TIM does the job at the lowest cost.

 

[LaurenceGough]

Good point about the plating, my NH-D14 came with a nickel coating over the copper, I thought it was just to stop the copper from umm, galvanizing? Not sure if that's the right word, but after a year or so the temps can increase a lot.