*** Official Clevo P870DM/Sager NP9870-G Owner's Lounge - Phoenix has arisen! ***

Hey guys, so I have a tiny (possibly big) problem.

My Sager came in last week and it's been great. Cruised this thread and did the good ol' undervolt method and had 4.2x 4.1x 4.1x 4.1 running at -150 offset and ran a 5 minute test with peak temps at 74. But I'm an enthusiast, and since I derped and forgot to order the copper heatsinks, I watched some delid methods and wanna make use of the CLU I've got laying around that I bought from my now recently sold AW 18. Needless to say I wanna go a little bit faster while maintaining relatively cool temps: Hoping to get a high 70s or lower at 4.3GHz

So I attempted to do my first delid today, and was able to get the IHS off of the processor. I cleaned the bits, scrapped most of the adhesive off (very little is there), applied CLU lightly (as I've already done a CLU repaste with my previous AW 18), put everything back together and vwala! My computers lights turn on but nothign happens. No boot, no POST, nothing. Now there is something to consider:

I'm currently using the "floater method" (aka no adhesive replacement) so I, to the best of my ability, have held the IHS in place to the best of my ability, and it appeared pretty proper. Maybe an adhesive is required?

Will the computer fail to boot if any heatsinks aren't fully bolted down? Because the main GPU heatsink (I run an SLI 980m config) has to be removed before I can access the 6700k, and the manufacturers gave me 2 pretty stripped screws, which were a hassle to take out the first time and I'm fairly sure for the second they're just "mostly" bolted down (enough to put some pressure on the GPU but not fully bolted). I obviously took the 2 pretty good ones and installed them opposite of each other to maximize tension.

Anyways I'm typing this from my old Alienware 17xR3, any help would be most definitely appreciated! Thanks!

 

I do believe that they sound different.

I just was studying how the actual keys look like. From what I know there are different languages versions of the same keyboard, but I believe they are made with the same constructive design.

 

You know when you delid that you have to take extreme care to not even get close to touching the internals. Maybe the tip of the razor touched and left a very tiny scar on an electronic internal.

It should start even without any cooler on it, but probably would be toast in 5 seconds.

If you get no post no nothing, maybe the CPU was not inserted enough in it's socked. OR maybe something is not connected.

 

Probably a bit to aggressive undervolt for your chips. Reset bios with taking out the cmos battery. How to reset nvram https://biosmods.wordpress.com/2015/10/21/sky-x9/4/

 

Thanks for the quick reply. I was extremely careful with pushing through the adhesive, to the point where it took me almost an hour lol. Doesn't mean that I can't make mistakes, as I remember I could have possibly touched it once. However when I looked at the processor everything looked spotless. I'm uploading a picture I took but granted: This is with the CLU still on it.
Also, I found some leaked IC Diamond compound (very little) in front of some of the many tiny dots underneath the processor. Could this have obstructed the boot? I've since cleaned up the heatsink that had some leakage and properly cleaned all other aspects and will try to boot this again.

As for the undervolt, it was done through XTU, so shouldn't it just revert the settings if it's too low? I also did a quick 5 minute test and ran the computer for about 2 hours yesterday and it worked fine with that voltage. I do understand that voltage can be tricky and sometimes it works for a while then it doesn't.

After I put all the cleaned material back together I will try booting. If it still doesn't work I will try the CMOS battery reset and hope that fixes things. If not, I can't really think of anything else. I will probably upload a picture of the center DIE more zoomed and without the CLU to see if we can spot any damage (though I know it could be very tiny and not enough to see).

If worst comes to worse, I heard you can send the processor back to Intel, and if they determine that physical abuse wasn't an issue they will replace it. Is this true? If so I'll probably go this route if I cant fix this within the next 2-3 hours.

Thank you guys for your suggestions and comments thus far!

 

Wow, 1 hour is like too long lol. Took me 5 mins to do mine.

From the picture it isn't really clear, you might want to take a close up with a good camera or something with better lighting. Also are any of the pins on the socket bent by any chance?
I don't think, Intel honors warranty for delided CPU's. And since the the issue happened after de-lidding there is something that went wrong around that time.
Gotta be very careful when using a razor to delid, i got a pack of those really thin double sided shaving razers and covered one end with tape. They seemed way safer than the thicker ones and sliced through the glue easily without any effort. Just need to make sure the bade is never pointing downward towards the PCB since there are very thin traces under the layer.

My IHS is floating as well, so that isnt an issue, take a closer look at the contact points under the processor and the socket pins. When cleaning the pins, DO NOT USE A CLOTH.

 

Haha I mean I'm kind of exaggerating, it was more like 30 minutes-ish. Still a very long time; I was trying to be super careful (maybe no no avail) .

Well, just put everything back together and same result. Did a CMOS reset and same result as well. I could take a picture, but I think it's safe to assume I did something wrong and did puncture the PCB somewhere. As a matter of fact I have a suspected area that I can see on the picture I upped; I'll re-upload with the area in question circled. As for bent sockets I didn't see any when I inspected the belly of the beast.

I'll just same day ship a new one from Amazon (RIP $350 ;- so it gets here tomorrow. Might try an RMA, I'll look more into it.

Don't really regret it if I messed up though; $350 is definitely a setback, but If I never tried to begin with I would've never been able to mod the many laptops in the past for great gains (thanks you reading you guys on these threads for the last few years!).

EDIT: I'll probably inspect the belly where the pins are closer tomorrow. Working on this for hours has exhausted me -_-

 

Take a soft plastic card (credit card, costco card etc etc) and scrape away the remainder of the silicon, and then clean it with alcohol. Will be easier to check if you messed up or if it is something else. Did you by chance use the razer to scrape the glue off?
I guess if he processor isn't the issue, you can always return the new one, god bless amazon for that

Also those GPU bracket screws are a ***** to work with. They get stipped super easily.

 

It may look like the PCBs are somewhat damaged from what I can see on the picture(but my vision is not all too good ). A better picture can confirm it better. Regardless. $ 350 is not calamity.

 

Yeah, I used the razor to scrape the glue off. I did it very lightly as to not damage the PCB, but I got only bulky top layers doing this. I started using an old membership card after that (and then I realized... probably should've done that to begin with. But in the circled upped pic, its possible that trying to get inside is what did it. I know I went at the adhesive from an angle a few times and that might have done it.

Seriously though! I love Amazon for that, and I kind of feel bad because (this just came to me), I could simply use the Amazon processor and order a new 6700k from silicon lottery already delided. It is 50 bucks though so meh, after having done it now I know the mistakes I made and it feels like I could nail it doing it a second time. But at least this 50 would eliminate all risks and technically save me some labor (which if I minded doing I probably wouldn't have done this in the first place.

It's good to see I'm not the only one having issues with the GPU bracket screws lol. I mean I'm no glad we all have them, just glad we get to suffer together !

Also had a quick question: I didn't order the copper heatsinks (as I posted above) but I noticed that the underside of both GPU and CPU heatsinks (where the paste is) appears to be copper. The top layer leading to the fans however is painted dark black. Is this like a copper head and aluminum body? Or was I mistakenly given copper heatsinks?? I can grab some shots if requested. I mean, I've only taken this thing apart how many times today?

EDIT: and yeah apologies friends, I don't have a better camera than my phone I wish I could do better! (might invest in an actual camera now since this issue comes up often lol)

 

Yeah Silicon lottery saves you a lot of trouble. self delidding is more for people who know they got really lucky with the processor and want to push it even further, or really unlucky ones who want to get decent working temps and OC's.

The heatsinks are full copper, they are just painted black or did they account for heat transfer through black body radiation? We will never know.

Phone camera is good enough, just use better lighting and it should be enough.

 

Hope this was the last time such a thing happens to you . But this will only remain on your goof account. Many have done the same, so do not be too disappointed

 

@Papusan haha thanks buddy. This is my main account I just never ever post on it because I've never screwed up lol (and never bothered talking for some lame reason). I'm surprised it doesn't show join dates, because this accounts been here for years!

@bloodhawk So the heatsinks are cooper then? Which means when the new one comes in I can use CLU on it. I repasted this current one after the delid with IC Diamond, because I hadn't ordered the copper ones and despite it looking like they were copper (at least the head of the heatsink) I didn't want to risk it since it seems from what others are saying that using CLU on aluminium will act more as an insulator than a conductor of heat.

And reading more on silicon lottery's page: it seems like they guarantee giving you a processor that's at least decent? It says "56% tested at 4.7GHz or higher, and when selecting the processor to purchase it says "i7-6700k @ 4.7GHZ" which leads me to assume that they do gurantee recieving at least a decent running processor (which is very interesting because the pricing is very good). Unless theres more options that address my question at checkout, which I'll check out right now.

 

The copper heatsinks that are mentioned on a lot of websites are extra ones that a lot of vendors stick of the existing ones. The heatsinks that come with the system are full copper as far as i can tell.
One heads up about using CLU with the heatsink, it will leave the heatsink surface grey. So it will be very difficult to clean it back to the copper layer. I guessing a level of oxidation does this?

Also the SiliconLottery processors are tested using desktop boards and cooling which are way more robust than what we are working with. So do not expect the exact same results. But slightly less, its pretty easy to hit 4.5-4.6Ghz on most 6700k for 24/7 use.

 

@bloodhawk Yeah that's true about the grey, I've used CLU on my last two Alienware laptops and I now swear by it since the performance is amazing lol. You're probably right about the oxidation, it makes sense. Good to know that I had copper heatsinks and my instincts weren't off, I mean these temps were pretty good after all.

I'll probably order the delid + 4.7 GHz from silicon lottery, since the 4.7GHz is the same price as a normal processor. I'm eyeing the 4.8 GHz as an investment value, but its a whole $120 more and I'm already at a deficit from losing this processor . The decisions ahhh!

I'll probably order and use the Amazon one in the interim. I feel kind of bad doing it, but then again I give them soo much business that I kind of don't .

EDIT: @Papusan Oh my goodness I completely forgot about this video! Haha I saw it before I had done the CLU on my AW 18. Thanks for reminding me: It's alot worse than I described lol.

 

So I'm getting massive, headache-inducing screen flickering when trying to run things full-screen. It happens at 1080 or 4K, G-sync on or off. Has anybody else experienced this before? Using 362.03 with the Prema VBIOS.

 

Does this happen on an external display too?

 

hello all, just picked up a new Sager NP9870S. I overclocked the cpu to 4.4ghz and undervolted the core to -100mv. I also followed previous steps and increased the power throttle to 200 watts. ive noticed in intense games like the divison that my cpu will not hold 4.4ghz and will throttle back to the stock 4ghz speed.Temps are below 85C so I'm sure it is not thermal. Is this because I'm limited with the 330watt psu? Anything I can do to keep that 4.4hz clock? My specs are in my signature.

 

That's probably a bug with G-Sync and 4k. Plus the brightness is always maxed out , so it's really bad for the eyes.

 

Alrighty guys, so back with a 2nd day update: My new Amazon 6700k came in and so I did the following steps to install it: Cleaned CPU heatsink, applied IC Diamond, pressed and verified good contact and paste distribution, placed the processor in and booted (very simple stuff). No delid, no nothing, just cleaned and installed. I still have the same non booting, no POST problem. I guess I can try a CMOS reset, but is one really needed (I presume not) when you install a new CPU? This may mean that I may have in fact not messed up my old CPU, but that something else is wrong: maybe something mobo related? I'm not sure of the scale of issues that have the symptoms of not BOIS, no POST, no boot, just simple LED lights on and waiting endlessly.

I might call up the tech guys at XoticPC and see if they can help me out a bit!

UPDATE: After letting it run for 30-40 seconds, it booted notifying me of the CMOS reset I did yesterday and everything is working fine with the new CPU.

 

did you have any overclock in the BIOS before? one should always have the BIOS at stock before installing a new CPU

 

@pheonix Thanks for the quick reply, all overclocking and voltage control I do through XTU. It was not stock XTU settings however (so was still OCed and UVed). Though I should probably update my previous post, but I just turned it on again and after leaving it on for roughly 30-40 seconds it just randomly turned on, notifying me of the CMOS reset (which I had done yesterday). It then restarted and I was at my desktop with everything seemingly working fine.

Now a second question: Should I try reinstalling the "bad CPU" and attempt booting it a few times (but leaving each boot running for like around a minute)? Afterall, I admit that when I tried booting it yesterday I waited for around 20 seconds max (probably around 10 seconds on average). And with this current result, I may have not given it enough time to boot.

As always, any advice is appreciated! Thanks!

EDIT: to Alleviate anymore issues, I will be setting XTU to default settings.

Also, I thought software overclocking will reset factors to default settings (OC & UV) in the event a boot failure occurs vs. BOIS (like with XTU). Is this correct or incorrect?

 

@luciela > Even an expert will tell you that each time one lays hands on a machine it becomes exposed to risk, regardless who you are. Since the swap's still fresh in your mind, & you know the routine involved, I would test that other cpu today, but that is me, since I would reqwire an answer, kuriosity would kill my kat. A delidded part of unknown working status can be sold for a loss of course, to offset new cpu. The other option is to run the new one, hold back the 'spare', & should the new one need removed for some reason later, for any reason, then you can test it.

However, if you have a liberal return policy which allows u to keep it for some time, perhaps bench the new one 1st: if you have old data concerning old cpu, perhaps play the silicon lottery, see which was the better performer. When all these variables pile up & render a clearer picture, you'll be in better position to know what to do. Owing to the fact you had no joy the 1st time u dropped in cpu, a small chance exists you botched the delid install for same reason, who knows, I don't. That is why I would test. If delid works? I would do my silicon lottery comparison in order to decide keeping new one or not. This is a dilemma, do what u think is best. Sleep on it if need be.

 

Try installing the old one and check if it boots. Might as well as this point.

The system takes way longer to boot after a CMOS reset, so gotta keep that in mind.

 

@bloodhawk & @Cass-Olé

I think I'll go ahead and try it once more, the curiosity is too compelling right now lol. The only reason I don't is for the very point you mentioned: every time the system is open it's exposed to risk.

I'm thinking along the same lines @Cass-Olé , thanks for the info; much appreciated.

@bloodhawk I didn't know the CMOS reset takes much longer to boot. That's probably the reason why this one took a while. But I suppose there's always hope for the first one!

Thanks again guys. I'll take it apart right now and report back my findings.

 

Yes it will re-initialise the micro code on a new cpu being installed and take it longer to boot.

 

Alrighty so the old one didn't boot, put the new one back in and on it right now. Gonna try my shot at the silicon lottery with the new one, and figure out what I want to do with the old one later. I'll first do some of the XTU steps in this thread and get a decent 4.3 with UV. from there ill work my way up and see how it goes! Wish me luck !

EDIT: Oh, and I've got this thing propped up with two large fans for airflow. So I should be able to lower the temps a bit to help push the OC without going over the thermal limit (since I haven't applied my CLU yet).

 

Do you have Prema BIOS or stock ?

 

Currently running stock.

 

Ah gotcha, XTU is pretty decent, but you will be able to get much better results using ThrottleStop or a custom BIOS. Try reading up on that.

 

What is everyone throttlestop settings? I'm still getting throttle during intense games from 4.4ghz down to 4ghz. Temps are excellent and undervolted -100 Might be power lmited. Stock bios using XTU

 

@bloodhawk I remember ThrottleStop a while back, is it still that much more effective than XTU? I only learned limited information about it so I'll definitely need to look further.

Completed a 5 min stress test @ 4.5 GHz with (IMO) very aggressive voltage (-20mV +160 Turbo Boost power). Hit the thermal limit with 1% near the end with temps peaking very high (of course) @ 98c.

Going further will increase the temps, so I'm gonna try to trim voltage down and probably look into ThrottleStop from here on out.

 

Good time/place to post this: [SOLVED] Which silicon glue for my CPU? (Permatex 82180)

You're back in b'ness, so, there's always that. Multiple ways to test the 'bunk' 6700k elsewhere. My local PC Shop is 5minutes away; I could contact friend / co-worker with skt 1151 & hand deliver / mail it; open call in a pertinent forum, mail there / mail back; message your vendor for instructions / advice ...

 

Mr. Fox, can I add, one very impotent for user item, to this order:

4. Clevo "desktop laptop" 17-inch or larger panel must be with OLED IGZO technology from Sharp.
I know is dream, but nice dream, and we know - tomorrow beginning today…

 

Not sure as I haven't used throttlestop yet, but that information would be useful. We could probably find some good numbers using search for this thread.

 

Dang thats high, my 4.5Ghz is running with an offset of -75 (actual voltage will be different than yours) and this is with 4.6ghz with -60mV :
https://i.gyazo.com/4d53d255a423fbbf1531229192551d84.jpg
I personally barely touched XTU, My max power in BIOS is set to about 125W i think. (Which seems to be enough for my 24/7 4.5Ghz to 4.6Ghz.

 

Are your max power limits set properly? usually it throttles at those clock speeds when it isnt getting enough power. Its unlikely you are hitting a thermal throttle playing games.

 

I use a combination of XTU and TS for my 4.5 setup at -20mv. I find the only way to keep static 4.5 is with both running side by side. Then no issues.

 

Completed another 4.5 GHz 5 minute test, still on XTU. Currently lowered voltage to: -70mV & Turbo boost power = 130 W. Max temps were 86c, which is much better. I'll probbaly try a -75mV & 125W test. If both check out I'll start with a 4.6 GHz test.

 

Try doing the long stress test with TS if you really want to check better for stability and heat.

 

Backing off 200Mhz and going for a lower voltage can be better for 24/7, different for benching of course.

 

@Meaker@Sager Definitely agree. I'll probably run it standard at 4.3 or something relatively low, mainly benching to check how high the clocks will run. I'll bump the 24/7 usage MHz after a delid, CLU repaste, or both.

@Firebat246 Yeah I'll run a much more demanding test once I hit that "hard" limit. Mainly checking to see how high clocks will go. I'm still using a non-delid with IC diamond pasting. I managed to run the -75mV and 125W turbo max successfully with 4.5GHz. Max temps were 86 and I ran it again and was 89.

Quick question: I might return this one to Amazon and order a silicon lottery + delid 6700k if this Amazon one isn't pulling good stuff (but so far seems okay). Can I repaste with CLU between the CPU & heatsink with no discoloration to worry about? I want to return the processor in its best condition. I've cleaned the heatsink with it before (during my old delid) and with alcohol it appeared just fine. I don't mind a stain on my heatsink, as whatever I wind up rolling with I'll be using CLU anyways.

 

You will need the CLU cleaning kit to remove it properly, but even then the heatsink will have a grey color.

 

Cool, I've got the cleaning kit and I don't mind the color. Now for more specifics: are you talking about the center square (which is silver-ish)? or the copper on the outside? I always thought you couldn't let the CLU seep onto the orange portion, as that could cause a shortage.

EDIT: Just booted ThrottleStop 6.0 and it gives me a CPU not supported error, which I assume means I downloaded an out of date version or there's a quick fix out there. I'm about to leave for dinner, but I'll research this when I get back. If any of you don't mind giving me an answer on this though it would be most appreciative! Though I feel guilty I've already asked so many questions

 

Old build

I mean the copper portion that makes contact with the processor IHS

 

Ahh okay gotcha. Yeah that's perfectly fine. As long as there's no discoloration on the IHS that's all I'm worried about. When I get back home I'll probably repaste with the CLU, then get back to it

 

You will get a Grey stain with CLU and no way you can get rid of it. I have the special liquid cleaning kit from Coollaboratory and even with its metal cleaner liquid it won't get rid of the stain. So I bought me another heats ink and used IC Diamond. I ordered Thermal Grizzly Kyronaut which I will be using from now on thanks to @Ethrem for letting me know about it

 

u can get rid of pretty much most of it with 100% isopropanol

Sent from my Huawei Mate 8 NXT-AL10 using Tapatalk

 

The colour should not impact performance