Printable View
I just realized that laptop Phenom II CPUs don’t use IHS so they are naked. I was wondering if AMD is planing to sell (at least one) FX model with optional IHS or unsoldered HIS?
I miss the days when we could just remove the IHS and mount a waterblock directly on the core:clap:
Please don’t tell me that there will be no temperature difference because there will be some:yepp:
There will be no noticable temperature difference and you will kill your chip amazingly fast.
chew* killed Deneb with 1.45v without IHS I believe.
Any idea what the cause was? You think he applied too much pressure to the core or it was a bad contact between the core and cooling?Quote:
Originally Posted by BeepBeep2
Temps were fine, it's just the quick temp swing (even if you are within operating limits) you get without the mass of the IHS that is the issue. These chips were designed to run WITH the IHS.Quote:
Originally Posted by smpsys
I apologize, 1.475v...
Quote:
1.475 is max i would bother with most chips for 24/7. TBH if I had the time and ripped my rig apart and put the amd system in thats been waiting for quite some time, I would be happy with 3.6 24/7, then again my 24/7 pc is mission critical, Quiet and cool.
The point at which is takes excessive voltage for minimal gains for 24/7 use is the point that you will degrade varies chip to chip but 1.475 is preety much safe for all chips.
Best way to find out is to remove the buffer (IHS) run on cool water at a max prime stable speed.
Let water get room temps then boot. It will instapop the chip think that happened at 1.5v IIRC but not 1.475 with quite a few test candidates.
If you can instapop, then it's fairly safe to assume you can degrade with the buffer in place.
Chil will die if you damage its silicon,otherwise no.You will notice 2-3 degrees,but it isn't worth the risk.In my opinion.
LOL, thats funny. It's all about contact, thats all.Quote:
Originally Posted by BeepBeep2
They can run without IHS, but its not easy to get perfect contact.
Quick temp swing is also BS, a chip won't die from that.
i ran a x2 5000+ without IHS for a few months on water, the one thing i did notice is that it put a nice dent in my waterblock that took hours of lapping to smooth it out.
i dont get the whole buffer thing though. the contact the IHS has should be no better than what we can get with a waterblock, and i can imagine someone building waterblocks around exposed chips for an incredible temp drop (5C is huge when you look at how low temps already are with a good loop)
Agree with Ownage,Quote:
Originally Posted by ownage
I had few AMD 90nm, 65nm dual core chips without IHS that had waterblocks on them with out any problems. I ran a Brisbane G1 4000+ at 3200Mhz at 1.6V that time. I believe the CPU was pulling over 200Watts by itself:shocked: Again never had any problems. As soon as I found out it was pulling so much power I upgraded to Athlon II 250 that did 3800Mhz.
I believe Athlon II dual core was the last 45nm CPU that did not have a soldered IHS.
I still don't understand why its OK to for laptop CPUs not to use IHS on QuadCore PII or any other AMD CPU:confused:
This is the latest AMD APU without the IHS: http://www.alibaba.com/product-free/...showimage.html
It looks nice and flat to me:)
Agreed, I ran plenty without an IHS at some pretty high temperatures for extended periods (months) at a time. If a heat spreader can stop temperature swing, a heatsink will too. Temperature swing killing naked chips doesn't gel well with me.
That being said, would I say that it's worth the effort? No, unless you have a terrible IHS. I used to do it because back then I wanted every 10MHz I could get out of my system for 24/7 use.
While I think that's what chew* meant I'd still like a little clarification myself...I cant see what else he could have meant.
Removing IHS kills voltage tolerance unless you solder your block to CPU. The IHS is a buffer.
Thanks for assuming i'm an idiot who has no clue about contact ;)
The chip died at 1.575 IIRC or an equivalent to that had chip been a retail. The spike in temps at boot is what killed it.
It was a HL ES chip with the sole intent to test exactly what it tested.
The newer dies are getting fragile, deneb could be chipped, thuban would just crack. I imagine BD will be even more fragile.
Also might note while the laptop chips appear flat the denebs we delidded were anything but. We had to lap the die on a few of them.
You guys are clearly misguided if you think that paste is equal to soldered.
Would you twist 2 wires together and call it good contact?
So I was right to an extent?Quote:
Originally Posted by chew*
Just like how you can have an LN2 pot with all the surface area in the world, but without mass in the base it's worth nothing? The contact and thermal transfer from die to IHS through solder is much better than paste to block/base would be and that's what kills the chip if I understand you right.
What i'm saying is If AMD thought they could get away with paste under IHS with all there current chips, Don't you think they would use it?Quote:
Originally Posted by BeepBeep2
Indium is not cheap.
chew*Quote:
Originally Posted by chew*
I guess I will be the smart ass to ask you a question;)
So you saying the only alternative to IHS is to solder the heat sink directly to the core using Indium since the thermal paste can't do the job?
.
yes that is what he is saying..
what cooling did you use on your 4000+
only asking because i had one that did 3100 @ 1.55 stable with the ihs on it.
Quote:
Originally Posted by smpsys
1. If you could solder the waterblock to the core, do you think any paste on it's best day would outperform it?
2. The answer is no they won't sell FX chips without IHS, and if it anything like thuban removing IHS = death.
Thuban.
http://chew.ln2cooling.com/qdig-file...%20shot%20.jpg
i dont remember the IHS being soldered down, when i removed my 5000+ IHS it was a little cutting along the edges and it falling right off.
The newer high end stuff has been soldered for a while now. Like AM2+ and up. Prior I was on a long vacation so I would know 754/939 which we all knew were paste.Quote:
Originally Posted by Manicdan
i know its soldered around the edges, but i had no idea if the center was still the same way. i also do not know the thermal specifics to solder and how different it might be from paste that is designed to fit in every little tiny pit and have some of the best thermal transfer possible. im willing to bet the solder they use i cant pick up at a local hardware store either, so its probably really tough to find out any real info on that part
while its obvious theres lots of dumb reasons to use an IHS-less chip, i just dont quite believe the issue with heat spikes since i would have expected to see it happen much more often than just once. also if its a chip that you could remove the IHS, then its also a chip that wouldnt have been soldered down, which then goes against the idea that it needed to be soldered to give it that thermal buffer. im not trying to fight about this, im just not coming to the same conclusion that you did.
I can only go by the results I did myself. Temps were a tad lower, the transition from load to peak temps was instantaneous way more significant than with IHS on. So take a realistic hot summer day with a dialed in significant OC in my case 4200 prime 95 stable........and boot a system "cold" where on that hot day is rather warm ( radiatior and watercooling won't help instantly ), then add to that the voltage being run for 4.2 then combine all that with XP load......which is a mother trucker, anyone that benches on ln2 knows XP boot brings the pain..........
Moral of the story in a very common everyday practice ( not everyone has a golden spoon with central air ) of booting a system the chip can and will pop.
Remember for AMD to qualify anything it has to pass the worst case scenario.....or it never happens.
The solder they use is Indium btw.
Believe me I discussed this in length with AMD when AM2+ launched.......and we ran quite a few controlled experiments.
Here's some fallout from them (experiments), that piece of foil on tray is indium.
Attachment 118979
very risky for me
chew,Quote:
Originally Posted by chew*
Can you expaint me again why its ok for laptops not to use IHS? I don't get it. Is it because they are not being overcloked? I wounder if there is a way to OC laptop CPU just for experiments and see its behavior.
Also, it looks like IHS was removed by force in the last picture. Did you try cooling the a CPU that was clean(without IHS) originally? Not sure if you were able to get your hands on CPU without IHS originally. I'm afraid melting indium again may cause the core to behave the way you described above.
Thanks for answering all my questions:)
Sure, so OEM's have a guidline to follow as far as pressure goes, not only that but they certainly are not installing a lot of weight on a core. The other reason AMD does IHSless for OEM laptops is clearance, they gain albeit a small amount but enough to matter. Laptops cpu's tend to also be on the lower end of the power consuming spectrum so heat is not as much of an issue.Quote:
Originally Posted by smpsys
The other reason AMD will do it is laptops are sealed and usually not serviced by endusers, in the department of "user error" AMD does not need to worry about endusers doing stupid things ( aka true copper bolted down with XXX torque cracking core than wanting rma )
I imagine you could pull a laptop Llano cpu for more experimentation but they really don't scale any better at say laptop oced temps to a good waterblock so they are a bad example of proving ocing headroom.
They just don't scale well.
Thats a bad pic and that cpu has been tossed around alot. Let me get a better shot.
http://chew.ln2cooling.com/misc/IMG_3937.jpg
http://chew.ln2cooling.com/misc/IMG_3939.jpg
As far as a cpu with a clean die, AMD never sent nor had them, least not anything on the enthusiast end of the scale that would meet our criteria for testing.
As far as the thuban picture, chip was already dead so we didn't need to be carefull, even with being carefull though it's near imossible to remove IHS without hitting one of those SMD's.
Only way it could be done is in machine shop like we did for a few chips. The IHS was milled then the IHS is heated with a fine torch, and the center part of IHS falls off core.
I wonder two things...have Chew always right?
When will Chew dont have time for this?(didn´t he make a thread i will leave or something?)
Temp swing?! why?, bad contact?, solder make ihs have contact on all sides around the core also?, Laptop dont have temp swing trouble?!(no ihs), easy to crush core?(have been from Slot-A time K500-K1000), SMD will get hurt if you try to remove ihs?(Socket A hade same trouble maybe not like Am2+ but).
The Goal is to be fastest....or trying to get faster without any risk no FUN! , why not buy 2-5 cheap Phenom II and try it out....dont listen to Chew always.....sometimes he may have wrong to!.
Sry bad english...
:rolleyes: :shakes:
There is good advice and bad advice.
Yours is to have him buy 5 of a product he doesn't need or seem to have interest in and delid them and this is relevant to IHSless or delidded BD how?
I would do that but I wanted the CPU to be without IHS already from manufacturer for better results.Quote:
Originally Posted by Lastviking
I might try removing it on some cheap CPU some day but not now.
.
LastViking's advice is the worst i had seen in a while :) All you can gain if you remove the IHS and it works after,is 2-3 degrees lower temperature,which is pointless including the risks and voiding the warranty.If you would gain 200 mhz for example it would mean something,but since you gain nothing...
If you were really curious as to wheather this would help or not, you'd probably be ALOT safer just lapping the IHS down...
Pretty sure Chew* demonstrated the risks of trying to remove the IHS on modern CPU's. :rolleyes:
Just for kicks, when the time is right I will have shots/pictures of this.
One has had IHS machined off already, one will be removed by hand.
At this time just based on component location I would say hand removal is near impossible but we shall try anyway just because i'm curious.
are you thinking more ln2 without the IHS again?Quote:
Originally Posted by chew*