i7-7800x Delidding Catastrophe? Resoldering micro resistors?

Bense

Limp Gawd
Joined
Aug 7, 2013
Messages
130
After wussing out on following through with the delidding of my i7-7800x (socket 2066), Saturday night I pulled out the delid tool that has sat in my desk drawer since August, and I gave it a go. I was nervous and very careful. I read several of the guides that I'd found online, and watched a handful of videos.

The Good: I did manage to do exactly what the guides / videos demonstrated, and I removed the IHS successfully.

The Bad: While I was using a q-tip with rubbing alcohol and removing the glue/residue, on one of the corners of the chip, I managed to knock off two TINY resistors and a TINY black chip. In hindsight, the guides/videos that I'd followed must have been for socket 115x, or some other cpu that doesn't have these. Or maybe they did say it, and I just overlooked it because I'm an idiot.

Be that as it may, as of right now I've got the two resistors and the black chip carefully folded up in a piece of paper, and placed within a ziplock bag.

My question is; do any of you have any suggestions, or know of any place that I might be able to send this cpu + tiny black chip + resistors to, that might be able to resolder them back on?


Picture I took right before I began the delidding. It shows the chip + two resistors in tact.
i7-7800x_delid01.jpg



The carnage
i7-7800x_delid02.jpg



Another view:
i7-7800x_delid03.jpg




Any help or advise is greatly appreciated.
 
With a magnifying glass, tweezers, a steady hand, and a fine enough soldering tip, you could likely put the two resistors back, but that chip is going to be a bear. This type of soldering is possible by hand but takes a great deal of skill.

One thing you'll want to be careful of if you do find someone to do the work for you - make sure you can identify without a doubt where each component goes. The resistors (as long as that's what they are) won't care about polarity but may be of different values, and you'll want to be sure they went back where they came from. The chip will need to go back in the same orientation too, being treacherously symmetrical.

This isn't something I would attempt yourself unless you're very comfortable (read- professional experience level) with a soldering iron. Just be sure that whoever you hire to do the job has the necessary info to put it back right.

Best of luck in your search. This kind of thing always sucks, sorry it happened to you! =(
 
As luck has it, my best friend messaged me today telling me about his new engineering job. Said that he spent his entire day soldering with his micro soldering station.

So that gives me hope that I will be able to fix this. Apparently, that black chip is an RFID chip that's not used on i7.

My next concern is that I may have knocked off one of the smds that is closer to the die. I will inspect tomorrow.
 
What type of delidder did you use?


$11 ebay tool. It was not the delidding tool that knocked it off. I noticed the chip before I began the process. I took extra care to ensure that the tool did not interfere with this chip. What got me was the q-tip / removing the adhesive material.
IMG_20190204_030203_438__1000x1000.jpg
 
Booya:
https://www.gamersnexus.net/hwrevie...nchmarks-game-streaming-vr-premiere?showall=1
About That Skylake X RFID Chip

Let's start by revisiting our earlier 7900X discussions. First, the RFID chip on the corner of the 7900X is actually disconnected – the CPU can't read from or write to the chip, and it stores no data. We asked Intel about this at a recent event, and the company told us that the RFID chip is present because the HEDT and server CPUs use the same package, and the RFID module is used for the enterprise division. The SMBUS actually isn't connected on the HEDT CPUs, meaning it's a dead chip that's only there because Intel is just making a lot of the same package for both enterprise and server. The enterprise tracking chip happens to be on both – but the SMBUS only connects on the server CPUs.

This could be something like RMA tagging or service life tags, among myriad other maintenance utilities that enterprise groups would make use of.


Okay, so next issue: It looks like I knocked loose a smd closer to the die. see bottom right of the die in the above pictures. Hopefully, I can get this resolved.

It's starting to seem like like the best option is to just delid the cpu, wipe off the die, and the die only. Then use one of these delid guards. It's not like laptop cpus have IHS.
IceMan-Cooler-Delid-Die-Guard-cpu-open-Cover-Protector-Intel-LGA-2066-CPU-X299-Support-7820.jpg
 
This guy is pretty good at soldering small stuff, if you still need the assist. He seems to generally do old console work, but couldn't hurt to ask.
 
I installed it back in my motherboard just to see if it would POST. It did not. Lol.

As luck has it, my best friend messaged me today telling me about his new engineering job. Said that he spent his entire day soldering with his micro soldering station.
I dropped it off at at my friend's place last night.


Here's a picture that shows the resistor near the die that I may have knocked off:
i7-7800x_delid04__1000x1000.jpg



A closer view:
i7-7800x_delid04_closeup__800x800.jpg



I'm nervous. :/
 
He was able to get the two caps back on, There is in fact a resistor missing. He said he measured all them up there, and that they were all 5 ohms.

Where do I even order these things?
 
I have delid faulted i7-7800X it has 2 missing resistors and I did measure other resistors and got several values from them 51Ohm, 49Ohm and 99Ohm not single 5Ohm one stated above, it would be really helpful if some one could measure them to me marked in pic below.
CPU_whole_marked_small.png
 

Attachments

  • upload_2019-2-7_12-14-32.png
    upload_2019-2-7_12-14-32.png
    5.6 MB · Views: 0
Thank you for the reply thuttu77.

About two hours after my friend told me they were 5ohm, I sent him the URL that capt_cope was kind enough to share, and asked which size I should order.
He responded: 'Already on it. They are actually 1uF capacitors'



Here are the two pics he sent me:
7975011242809998053__1000x1185.jpg

.
.
6602454453260799270__1000x1204.jpg


Also, he said that he does not have access to their equipment that he would need to reattach the RFID chip. I am less concerned about this chip, because I believe that the reason the CPU won't post is attributed to the capacitors that I carved off my $400 CPU with a machete.


If this thing actually winds up working, I'm going to crap my pants lol.
 
Bense Yeah those brown parts in your cpu are caps not resistors but I am looking values of 2 resistors on my CPU whicha are rounded with red in my pics.
 
Still waiting for parts to arrive from China. My friend said they may be in as soon as Friday.
 
  • Like
Reactions: N4CR
like this
Thanks for the update. I just de-lidded a 7980XE and your experience made me extra careful around those little devices, they are much smaller in person!
 
Thanks for the update. I just de-lidded a 7980XE and your experience made me extra careful around those little devices, they are much smaller in person!

Surely the insight that this thread has provided you in ensuring that your astronomically priced CPU is worth ample payment.

Let me know when you are ready for my paypal address ;)
 
Surely the insight that this thread has provided you in ensuring that your astronomically priced CPU is worth ample payment.

Let me know when you are ready for my paypal address ;)

Hah! Like I have two nickels to rub together after buying the thing! :)

But I know as a [H]ard Brot[H]er, the only payment you'd ask for would be a hearty "Thanks!" (y)
 
Sorry for the delayed response.
While my friend had my CPU, I began to wonder if the reason it hadn't posted before was because there wasn't sufficient 'clasping' pressure being exerted on the CPU in it's socket when I had inserted it without the IHS.
So, after I got it back, I installed it into the socket once again without the IHS, and noticed that the CPU was in fact 'loose' in the socket. So I put some of my cheap thermal paste onto the die, set the CPU back in the socket, placed the IHS on top of it, clasped the latch, cleared CMOS, and got the machine to POST.

All the moving around of my radiators + GPU + waterblock on my 11month old hose managed to create all kinds of leaks and other bs. So I tore down my coolant loop. Before I'm able to put everything back together, I've gotta order a few parts (O-rings, etc).

From what I'd read, using the 'liquid metal' thermal paste between the CPU die and the IHS requires cleaning/reapplying every 12 months or so. Which, would require further, routine removing/reinstalling of the IHS. -- Not only does all of that sound like a pain in the ass, but its also more scenarios I could damage the CPU. So I decided to go ahead and order the 'delid guard' that I posted a picture of above. If the delid guard works like it seems like it will, I might go ahead and order another one for the 2066 cpus that have the larger cpu die (for the higher core count 2066 cpus).

IMG_20190223_192823107_HDR__1362x1000.jpg
 
I've noticed that my Koolance CPU-370si waterblock needs to be cleaned. I haven't disassembled it yet, as I would like to get replacement o-rings from koolance. They do not have this waterblock listed in their O-ring Comparisons Table that they have on their website. I've tried contacting them through the contact form on their website, as well as messaging them on their facebook page, but I have yet to hear a response. Guess I'll have to call them.

On the other hand, I managed to damage the plastic or 'acetal' portion of my Heatkiller R9 290x Waterblock that I've been using for 4 years now (I replaced my R9 290 with an R9 390 last year). I must have stripped the threads when I was cranking down my 45 degree fitting last year. Fortunately, Heatkiller sells this replacement part on their website for a very fair price. I will order that, along with it's corresponding, replacement o-rings, as well as a replacement o-ring kit for R9 290x waterblock.

My R9 390 is a bit dated, and I'd like to upgrade it. But I don't mind spending ~$25 to replace parts on the waterblock that have been in use all day, everyday for over 3.5 years.

I'll keep you guys posted.
 
Glad to see it posted!

I only de-lidded one other CPU before the 7980XE, a 7700K that was a pretty toasty example. Anyway, it'll be 2 years in a month or so and it still runs same temps as when I re-lidded it - no plans to do anything to it. Used Coolaboratory Liquid Ultra (CLU) and just a tiny dab of black silicone sealant in the corners to hold the IHS on until I got it mounted. Same formula I used on the 7980XE and it's running very well.
 
This gives me anxiety. Ive had computer problems that were a PITA to fix, or a few that were never resolved. Ordering from China, stress, involving high-level soldering, time, frustration, googling. It’s like I watched a horror movie at 6 years old again
 
This gives me anxiety. Ive had computer problems that were a PITA to fix, or a few that were never resolved. Ordering from China, stress, involving high-level soldering, time, frustration, googling. It’s like I watched a horror movie at 6 years old again


What's the fun in completing a high performance build if I'm not able to infuse immense drama into it? :D
 
Seems like a good portion of the stories I read concerning the direct die mounting end up unhappily. Of course there's always going to be more posts from unhappy users than users who just install and have no issues - and one guy gave up on the Der8uer version and was gifted the Ice version you have pictured (given to him by a guy who had given up on it) and had success so far. To me, re-lidding or even re-de-lidding a processor isn't much of a hassle. I ran that hot 7700K (that I'd delidded and used CLU on a couple of years ago) for hours straight in the Boinc Sprint this weekend, 100% CPU load at max voltage and never saw 80C, and this was in a room with 3 other HEDT rigs running at the same time - it was pretty warm in that room. Breaking down the loop (that you'd have to do regardless) is way more of an obstacle than putting the lid/CLU back on IMHO. So save that lid, see how the direct die goes, but I would have never jumped into de-lidding if it meant re-applying the liquid metal annually.
 
Seems like a good portion of the stories I read concerning the direct die mounting end up unhappily. Of course there's always going to be more posts from unhappy users than users who just install and have no issues - and one guy gave up on the Der8uer version and was gifted the Ice version you have pictured (given to him by a guy who had given up on it) and had success so far. To me, re-lidding or even re-de-lidding a processor isn't much of a hassle. I ran that hot 7700K (that I'd delidded and used CLU on a couple of years ago) for hours straight in the Boinc Sprint this weekend, 100% CPU load at max voltage and never saw 80C, and this was in a room with 3 other HEDT rigs running at the same time - it was pretty warm in that room. Breaking down the loop (that you'd have to do regardless) is way more of an obstacle than putting the lid/CLU back on IMHO. So save that lid, see how the direct die goes, but I would have never jumped into de-lidding if it meant re-applying the liquid metal annually.

I ran my i7-5820K @ 4.4GHz with 1.35v vcore with Einstein@Home running on all cores. Never got above 65c on water. So far this x99 -> x299 jump has been very underwhelming.

What kind of problems were people encountering with one of those de-lid guards?
 
I ran my i7-5820K @ 4.4GHz with 1.35v vcore with Einstein@Home running on all cores. Never got above 65c on water.

Yeah, I had my 5960X @ 4.6 running on this sprint, too (one of the rigs in the same room). It ran a lot cooler than that 7700K @ 5.0, peaking in the mid 60's about like your 5820K. I used the 7700K as an example, its temps were unmanageable before the de-lid, but it hasn't gotten worse in 2 years of use with CLU under the lid.

So far this x99 -> x299 jump has been very underwhelming.

My X99 to X299 experience has been very good. My 5960X/ RVE never could get 32GB of ram to OC at all, 2666 is the best it ever did and I tried for weeks to crack that. This 7980XE / RVIEO did 64GB at 3600 right off the bat and OC'd to the same 4.6Ghz at 1.2v vs 1.3v for the 5960X

What kind of problems were people encountering with one of those de-lid guards?

Just inconsistent hot spots. Some cores would be running cool while the next ones would be at TJMAX, just a matter of getting everything lined up perfectly - not just kinda close, but micro perfect.
 
The thing with resistors is... they're kinda sensitive to resistence. And manual soldering tends to add extra (unwanted) resistence. Good luck with your friend. I hope he is as skilled as yo say he is.
 
CPU appears to work 'fine'. -- Or at least, how it worked before I delidded.

I have thoroughly cleaned my cooling system. I have tore down my CPU and GPU waterblocks, and I cleaned the microfins. I replaced several o-rings, and spent a considerable amount of time cleaning my cooling system. I have the delid guard installed right now, and I am using regular thermal paste that I have applied across the die as evenly as possible.

Using OCCT 4.5.1, running stock clocks, my CPU cores are sitting at the following temperatures.

Core #0 - 55c
Core #1 - 44c
Core #2 - 47c
Core #3 - 52c
Core #4 - 49c
Core #5 - 61c

Talk about a wide array of temperatures. Dullard, did warn me of this. But still. This is ridiculous. Meanwhile, my R9 390 is chillin at 42c.

I thought there was supposed to be a substantial drop in temperatures after getting rid of the horrible lid / crappy thermal paste intel used?
 
CPU appears to work 'fine'. -- Or at least, how it worked before I delidded.

I have thoroughly cleaned my cooling system. I have tore down my CPU and GPU waterblocks, and I cleaned the microfins. I replaced several o-rings, and spent a considerable amount of time cleaning my cooling system. I have the delid guard installed right now, and I am using regular thermal paste that I have applied across the die as evenly as possible.

Using OCCT 4.5.1, running stock clocks, my CPU cores are sitting at the following temperatures.

Core #0 - 55c
Core #1 - 44c
Core #2 - 47c
Core #3 - 52c
Core #4 - 49c
Core #5 - 61c

Talk about a wide array of temperatures. Dullard, did warn me of this. But still. This is ridiculous. Meanwhile, my R9 390 is chillin at 42c.

I thought there was supposed to be a substantial drop in temperatures after getting rid of the horrible lid / crappy thermal paste intel used?


These temps are not that horrible per se, for full load. The cores of my old i7 6700k (unlidded) all sit around 59c on air (with a fan-profile of 'silent'). But your delidded CPU should indeed be way lower, especially on water cooling.
 
CPU appears to work 'fine'. -- Or at least, how it worked before I delidded.

I have thoroughly cleaned my cooling system. I have tore down my CPU and GPU waterblocks, and I cleaned the microfins. I replaced several o-rings, and spent a considerable amount of time cleaning my cooling system. I have the delid guard installed right now, and I am using regular thermal paste that I have applied across the die as evenly as possible.

Using OCCT 4.5.1, running stock clocks, my CPU cores are sitting at the following temperatures.

Core #0 - 55c
Core #1 - 44c
Core #2 - 47c
Core #3 - 52c
Core #4 - 49c
Core #5 - 61c

Talk about a wide array of temperatures. Dullard, did warn me of this. But still. This is ridiculous. Meanwhile, my R9 390 is chillin at 42c.

I thought there was supposed to be a substantial drop in temperatures after getting rid of the horrible lid / crappy thermal paste intel used?
Maybe only getting good contact in the center of the die? Could lap the die a bit, or try a different block maybe?
 
Maybe only getting good contact in the center of the die? Could lap the die a bit, or try a different block maybe?

I'm not about to try lapping my CPU die. That's crazy talk. My CPU waterblock is very smooth and flat. I wetsanded it with fresh 5000 grit.
 
I'm not about to try lapping my CPU die. That's crazy talk. My CPU waterblock is very smooth and flat. I wetsanded it with fresh 5000 grit.
Lapping is crazy talk after you decided already lol? But yeah I still have 10 degree range with my 7820x after deliding. Temps only dropped around 10c from the deliding. Need a good lapping.
 
I'm not about to try lapping my CPU die. That's crazy talk. My CPU waterblock is very smooth and flat. I wetsanded it with fresh 5000 grit.
Right, but the top of your CPU die is not flat. You could possibly get better results with relidding or a concave CPU block, but the concave block could damage the die if it's too extreme. Or, you could lap the die, but you don't want to do that, so those are your other options (or just live with it, it's really not that bad).
 
I am not lapping the CPU die. I draw the line at that. I will admit that the thermal paste that I am using is not the best (arctic silver, etc). But with the same setup, my i7-5820k @ 4.4GHz with 1.35v, would not exceed 65c.
 
What if you used something like this to measure the clearance at different contact points?

https://www.plastigaugeusa.com/plas...l?Session_ID=b1620563838a9b46b6418fbefca04197


Seems to me like, indeed, his CPU isn't making proper contact. I doubt the OP is a n00b in the field, but yes, delidding (obviously) means you need to accomodate for 1 mm or so surplus room between CPU and heatsink. A good liquid metal compound (not applied too thin) should do the trick, IMHO. At least wortth looking into, as improper contact, as the result of delidding, seems a plausible cause. Lapping the die is something I probably wouldn't do either (and no need to, I think: delidding should normally cause temps to drop way enough 'as is'),
 
What if you used something like this to measure the clearance at different contact points?

https://www.plastigaugeusa.com/plas...l?Session_ID=b1620563838a9b46b6418fbefca04197

It's been years since I've bought plastigauges! Very clever idea.

Seems to me like, indeed, his CPU isn't making proper contact. I doubt the OP is a n00b in the field, but yes, delidding (obviously) means you need to accomodate for 1 mm or so surplus room between CPU and heatsink. A good liquid metal compound (not applied too thin) should do the trick, IMHO. At least wortth looking into, as improper contact, as the result of delidding, seems a plausible cause. Lapping the die is something I probably wouldn't do either (and no need to, I think: delidding should normally cause temps to drop way enough 'as is'),

I ordered 1.0g of Phobya liquid metal last summer. I wasn't ready to use it at the time, and I later thought it was lost during a move. So I went to order it again, and only the 0.5g tube was available, so I ordered that. -- When I pulled out my box of thermal cpu stuff, I found the 1.0g tube.

I've got far more of the liquid metal than I need, so it is not an issue of that perse. It's more so that I wanted to make sure the CPU would work before using the fancy stuff on it. I am, however concerned that some of it may 'leak' off of the die and get onto the substrate and mess things up. I plan on re-doing it and using the liquid metal, however, due to how involved it is to pull everything out, I will wait until I have other parts to install.
 
Back
Top