De-lidded 3770k!!

bremen1874 · 30 Jan 2013 at 16:51

TwsT said:
I de-lidded my chip definitely worth it.

How many additional MHz and/or what reduction in noise?

TwsT · 30 Jan 2013 at 16:55

bremen1874 said:
How many additional MHz and/or what reduction in noise?

no reduction in noise as it's water cooler with the 680.

no real boost in MHz just 20 degrees cooler.

have been running a 4.9 but I've settled it back down to 4.5 @ 1.21 vcore as the MHz gained isn't worth the increase in temps.

http://www.hwbot.org/image/900738.jpg I've got better screenies at home of temps and tests

bremen1874 · 30 Jan 2013 at 16:58

TwsT said:
no reduction in noise as it's water cooler with the 680.

no real boost in MHz just 20 degrees cooler.

have been running a 4.9 but I've settled it back down to 4.5 @ 1.21 vcore as the MHz gained isn't worth the increase in temps.

http://www.hwbot.org/image/900738.jpg I've got better screenies at home of temps and tests

So no benefit at all then?

TwsT · 30 Jan 2013 at 17:14

well yes plenty of benifits it runs at 60 degress underload instead of 80 @ 4.5 and i was able to post a stable 4.9 oc within temps.

I would do it to a chip even if i wasn't overclocking. Especially if was using the useless and noisy stock cooler.

The only risk is damaging the chip I didn't I don't think Intel will honour my warranty if it goes faulty either but since I upgrade so often I doubt it will be an issue for me.

bremen1874 · 30 Jan 2013 at 17:27

Reduced temperatures are a means to an end, not an end in themselves.

IMO 60 degrees Vs 80 degrees makes no odds unless it allows the chip to be pushed further (and you actually do it) and/or it allows for a quieter machine. That's not to say that delidding shouldn't be done, even if it's just out of interest.

BluD · 30 Jan 2013 at 23:36

If the chips staying cooler, wont it last longer?

troops · 30 Jan 2013 at 23:46

I cant bring myself to do it, bottled it

bremen1874 · 30 Jan 2013 at 23:56

BluD said:
If the chips staying cooler, wont it last longer?

Possibly, especially if the improved heat transfer allows for reduced voltages.

Chip longevity isn't a major issue anyway. They all seem to last long enough unless really silly voltages are used.

I have wondered what the resale value of delidded chips will be compared to the unmolested examples.

wazza300 · 31 Jan 2013 at 00:12

i still cant believe intel did this,such a cop out

Jakus · 31 Jan 2013 at 04:48

BluD said:
If the chips staying cooler, wont it last longer?

Yes, massive difference to life of electronics, Heat and thermal shock is the no1 enemy for electronics.

Jakus · 31 Jan 2013 at 04:57

wazza300 said:
i still cant believe intel did this,such a cop out

I can't believe different TIM makes such a difference... all tests prove there is little between all of them.....must be more at play here especially if the heat spreader is just rested back on top, I suspect a problem of bad original contact is more likely

If I went to the trouble I would not be putting the heat spreader back on that's for sure, just make a couple of pads for the clamp plate.

chrisoverson · 31 Jan 2013 at 11:48

An idea I had... rather than de-lidding totally and risking crushing the cpu and require modding of the mounts etc... why not just lap the heatspreader but sand it down a lot more than would normally be required?

I know direct contact is best, but would making the IHS thinner give better transfer of heat? Still protecting the CPU die but dropping temps some more.

RE: the "only worth it if you can push it more" argument... I'm only running mine at 4.4ghz now, but the point is I can definately push it higher if I wanted to. Before de-lidding I got 101 degrees when I attempted 4.7ghz, now I can do it at 67 degrees. I'm not, because I don't like leaving the voltage that high. But just for benchmarking and "because I can" purposes that headroom is there now

.

Jakus · 31 Jan 2013 at 12:09

Or, don't know if anyone tried with a dremel drum sander, simply just sanding through the top edges of the heat spreader should not take much, and simply remove the center whilst leaving the outer frame for the CPU Clamp

chrisoverson · 31 Jan 2013 at 15:26

JAKUS said:
Or, don't know if anyone tried with a dremel drum sander, simply just sanding through the top edges of the heat spreader should not take much, and simply remove the center whilst leaving the outer frame for the CPU Clamp

You could do that I suppose, but you'd permanently ruin a heat spreader by doing so. Also you'd have to get it down to the perfect thickness so that it allows direct contact between CPU die and heatsink, whilst using the remains of the heat spreader to distribute the load.

I suppose you could achieve the same through other means. If we can accurately measure the height of the CPU die, maybe we can find some kind of washer or other material that could be universally used for this, rather than cobbling together one each time.

Edit: Also, that idea still requires removal of the clamp etc.

troops · 31 Jan 2013 at 15:44

Sounds way to drastic to sand down the heat spreader and more likely to cause damage than cutting the rubber seal would.

chrisoverson · 31 Jan 2013 at 16:57

troops said:
Sounds way to drastic to sand down the heat spreader and more likely to cause damage than cutting the rubber seal would.

Myself and some others have already sanded it down by lapping it to make a flat surface.

See mine:

I was just suggesting sanding down further in order to make it thinner, and less of a barrier to heat... if that would work

.

TwsT · 31 Jan 2013 at 17:05

bremen1874 said:
Reduced temperatures are a means to an end, not an end in themselves.

IMO 60 degrees Vs 80 degrees makes no odds unless it allows the chip to be pushed further (and you actually do it) and/or it allows for a quieter machine. That's not to say that delidding shouldn't be done, even if it's just out of interest.

I see what your saying but the less time a cpu spends at high temps the better no matter what other benefits you gain from it.

doyll · 31 Jan 2013 at 17:23

TwsT said:
I see what your saying but the less time a cpu spends at high temps the better no matter what other benefits you gain from it.

+1
40c or 60c may bot be as big a deal but choosing between 60c and 80c, 60c is definitely preferable.

troops · 31 Jan 2013 at 21:05

I didnt mean lapping.

Or, don't know if anyone tried with a dremel drum sander, simply just sanding through the top edges of the heat spreader should not take much, and simply remove the center whilst leaving the outer frame for the CPU Clamp

Sorry my fault for not quoting.

Jakus · 1 Feb 2013 at 01:07

troops said:
Sounds way to drastic to sand down the heat spreader and more likely to cause damage than cutting the rubber seal would.

And it's not drastic forcing a blade between the board-heatspreader

I think as someone raised the point, a few crucial measurements would be needed to ensure the cooler would fit without fouling the clamp assy but regarding the risk of damage, with everything taped up and hand held to absorb vibration, I would say minimal.

Just seems to me if you commit to this route go all the way

Maybe someone could measure the heat spreader thickness, I doubt its much more than .30 mm that's about .012"....could be wrong !