Subject: Latest Socket 775 fastener farce (a two part story)
By: Ray (IP: 86.128.22.*)
Written on: 07-07-2008 03:13
There are pictures of the new heatsink, fan assembly, fastener & fastener holder here (with notes):-
The demerits of the original Socket 775 fastener system have been widely discussed so there\'s no point repeating it here.
As anyone who has ever tried to remove a Socket 775 heatsink, it can be hit & miss. (Why the Socket 775 fastener system was ever designed in place of the original foolproof socket 478 cradle-plus-fan cage is simple ... penny pinching!
In many cases removal of a Socket 775 heatsink results in damage to the fastener which must then be replaced with a new item - if you can get one (you can - see Ebay UK!).
Anyway, although it was hard to remove the old Socket 775 fastener it was at least relatively easy to do, if you knew how & had the right tools, without damaging it. That was because the fastener was a fairly floppy fit in the metal legs of the heatsink which means that the fastener is at least accessible from almost every angle.
Enter the latest Socket 775 fastener system!
The fastener itself is the same but the metal legs of the heatsink have been done away with. Now, the fan assembly and the fastener holder are moulded into one unit .... and the fastener is almost entirely shrouded by the plastic moulded fastener holder. That makes access to the fastener itself virtually impossible, even using the right tools.
OK, I said it\'s virtually impossible to remove the fasteners from this new plastic holder but it can be done. My first attempt took about 30 minutes PER FASTENER! After that I re-fitted & retried and got the technique right so that I can now remove a fastener in about 10 minutes.
BUT in my humble opinion, 99% of people will just give up and use a hacksaw etc to remove a damaged fastener, possibly damaging the holder & fan itself. Great Job Socket 775 Fastener Designers!
This next part is a lot more worrying, if that\'s possible. It concerns the accuracy of this new design.
I lapped the heatsink - I always do that as a matter of course - and it is MUCH easier to do than the original Socket 775 heatsinks because the heatsink can now be detached from the fan assembly -come- fastener holders. This means that the heatsink can be lapped just like any old heatsink with nothing to get in the way.
(Previously, the original Socket 775 heatsink was difficult to lap because the metal legs holding the fasteners are in the way. Again, that was awkward but not impossible, with the right tools.)
Having lapped the heatsink & reassembled it to the fan I thought about usina a new (for me) thermal compound. It is so thin that it is supplied in a bottle and has to be painted on using the supplied brush. I painted the entire surface of the CPU. All checked & fitted, I fired up the PC, checked the BIOS temps then did some thermal burn-in testing. All good.
Then I took the heatsink off the CPU and ... here\'s the worrying part ... 30% of the CPU had NOT COME INTO CONTACT WITH THE HEATSINK. That\'s not a problem PER SE but the 30% non-contact was in an area of roughly an arc from the center of the Integrated Hear Spreader (IHS) and one edge of the IHS. If this area is direclty above one of the parts of the CPU core, then it can only dissipate heat up through the IHS then sideways along the IHS to other parts of the IHS that ARE in contact with the heatsink ... Very Bad Idea.
WHY does this happen? I believe it is because the plastic fastener and the plastic fastener holder built into the fan is not rigid enough and not accurate to cope with the very large force applied when the assembly is fixed to the motherboard. Note that the Socket 775 heatsink mounting system has always resulted in an alarming degree of deflection of the motherboard, in my experience, but has never before resulted in this kind of poor CPU-to-heatsink contact.
I predict great demand for types of heatsinks with the old, original, metal legs ...
Other info. The new heatsinks are:-
* Very low profile, roughly half the height of older designs.
* All-aluminium compared with quality, older copper-core designs. I tried an old, extremely efficient copper-core HSF and on the same CPU the temperature hardly varied at all, in all tests, at approx 72% less RPMs (read ... quieter!)
* Wider as both the base and the fins themselves extend beyond older designs. The base is wider because there is no restriction imposed by the metal legs of the original design. The fins can be made wider because the fasteners are now attached to the top of the fan assembly rather than the metal legs of the older design. There\'s not much in it, but the extra wide design does mean more surface area for the CPU to lose heat into via the IHS AND more surface area for the fan to do it\'s job of cooling more efficiently.
* able to lose heat well but do not absorb/transfer heat well since the heatsink material is 100% aluminium - NOT COPPER CORE!
As ever in the heatsink world, efficient is beautiful, big & efficient is best!