Gadgetory

everyone so this is gonna be kind of a vlog style video I've had to take the camera and show you some stuff in the other room and you're still out which means I'm shooting this myself hence the vlog style so we got a lot of stuff in front of me here like nail polish has purpose for that liquid metal and D lid kit couple other things CPUs in various states of lidded nests and yeah it was it was a long night of trying to get the D letting stuff and liquid metal working properly on the CPUs I'm working on so we're gonna walk through some of that before getting to that this video is brought to you by synergy the software that lets you share a keyboard and mouse between multiple systems if you have limited desk space and multiple computers to command synergy removes the need for separate peripherals or a KVM and where it's as over the network software use our link below to get 50% off the home or pro version with ssl okay so first of all I have some footage of the process of deleting I used their Bauer's new kit I think it is called the D lid dime 8x I believe is the sky like X version and shot the process of deleting using this 2d lid is actually trivial it took me minutes to get it done I used this previously at copy text it's had some revisions this thing was not present at Computex and the whole process was easy we didn't lose any components it went well what didn't go well was my first attempt at applying liquid metal so I'm sure some of you have used liquid metal before we actually haven't used it on something like a CPU service before nothing like that so this was new to me and that meant making some mistakes on the way now fortunately my mistakes were not that big of a deal and that they were all correctable but maybe some of you can learn from this so first of all this thing this is the thermal grizzly stuff that they're Bauer kindly sent to me and it comes out pretty fast so I got way too much on the die basically and then the trouble was figuring out what to do with it so more or less the process became spreading it around the surface of the die taking some of that put the IHS where it should be and ending up with way too much having spillover and and then with the spillover I had a shorted capacitor so basically it spilled off of the dye and because this is a sky like X CPU its 7900 X were working with today there's a reason we're starting with that one because it is the cheapest CPU that I will be working with so a $1,000 risk was a substantially lower risk than some of the other stuff I'm looking at doing but basically we had some spillover skylake X has some capacitors and resistors and things on the upper substrate very close to the die which is a bit different from KB Lake and that means if it spills over which I had too much way too much liquid metal on there so it did it will short stuff now fortunately for me I only shorted a capacitor and so I was able to clean it up basically use some of these black q-tips that Devourer sent over and cleaned it up got it off the capacitor booted and it worked so there's definitely a scare there that's a big scare for a moment of like to the point where I pull dope I open up new egg I looked at 7,900 X I was like how much do I want to finish this test because I'd already done half the test in three days of work and I'm looking at it like man I'm not gonna make ROI if I just spend a thousand dollars but I really want to finish the test so I had to add into the cart after some some frustration some cleaning it up things like that to realize that one of the capacitors did come off of the substrate now fortunately it's just a capacitor so it's not a critical component everything boots all the scores are the same in Cinebench everything validates just fine and gaming and production workloads all the scores are identical to previously and the overclock is about the same so I think this may have just been some sort of something that contributed to voltage ripple suppression or or something like that I'm speaking with their Bower he and his team have run into that several times before where you lose a capacitor generally the CPU always still works afterwards so ideally you don't lose components but this one wasn't a big deal so we didn't kill anything critical that was good that was I mean basically you have to understand my process here so over the period of the night it was basically starting at like 12:00 a.m. or somewhere around there 11:00 p.m. I deleted the thing there was a it was pretty easy and obviously I was uncertain of how that would go but I mean this is it's really well built and it's it's pretty hard to screw up you just you tighten it down until you get one to two millimeters movement on the IHS once you get one to two millimeters movement referencing the RFID chip just to kind of get a point of reference of one the dials moves you stop you back it off and then pull the heatsink straight up don't don't like push it but pull it straight up and everything's good then you proceed with cleaning things and part of the cleaning process remove all the Tim and then the other part so der Bauer had suggested to me to keep it easy leave the silicone adhesive on the on the actual substrate rather than just removing it all and his idea there was to give me a guy the reference point of where the I adjust goes back when I'm ready and so I scraped down to just a small black layer for an outline but what I misunderstood him with was he specifically said leave it on the substrate he did not say leave it on my HS so I left telling IHS as well and the process here was a lot of things that could have been done better at once which is why we started with the cheap CPU so thing number one too much liquid metal spilled over shorted a capacitor is able to clean that up and fix it no big deal fortunately but it was very scary because you're looking at a thousand dollar purchase the next thing number two that could have been improved was once I cleaned up the liquid metal I put the I adjust back on this time with a lot less liquid metal but still ultimately too much and so it looked kind of like a pool of metal that was reflective rather than a thin film which is what it should have been and eventually I got towards that direction but basically this just it squirted out a lot more than it expected and that meant I didn't really I wasn't sure how much I needed it how to get rid of it so yeah so that was thing number two was it was basically after getting the IHS back on there the second time and getting it to boot this time without the smell of smoke without the like oh it's not posting what happened fear got it to boot but the temperatures were twenty Sol's he's higher so obviously something went wrong spokes at their bower spoke to VSG from thermal bench he's got a great website you should check it out thank you to both of them for the help and basically after talking with them finally Durbar pointed out like send me a picture so it did and I had too much silicone adhesive around the IHS still so went through removed it all reapplied it and booted and actually I'll take you over to the other room and show you the where we stand right now with the thermals you know I should start with his uh here's a graveyard of of paper towels from all the thermal paste and liquid metal I went through so much thermal paste alright so we have it's gonna be dark for a second sorry we have two of these and all of them is almost completely empty now and that's thanks to all of the skylight CP is being so damn big so we went through two damn fins well not two we went through almost this entire thermal compound container anyway I've got it all running now and it's finally working so this is my automated test that I scripted we are was that 400 seconds left out of like 1400 pretty far in running prime95 28.5 and it is in fact using basically all the CPU here's the cool bit so there's our current numbers ignore that max column that's from something else here out the left column so we're down in the 40s and 50s and I'll have the full numbers for you in a separate video so we're gonna do like a full thermal and power video we've promised this for a while now with sky like X now that we have all the parts it's time to execute and do that alright so we're back in the set room so there's a lot more to do and I am going to have to run a lot of tests on this deleted and then compare them to the original test results so so far everything validates the same scores outside of thermal limitation scenarios we are roughly very roughly ok let me just very roughly normally we do charts right you all know that what I showed you was a screen we don't normally do that but I'm excited about this process it took me all frickin night to get it right it should not have so I'm hoping that this helps some of you who want to do this in the future so you don't go through all the stupid things I went through like really just too much liquid metal it should be a really thin film and then it's all good to go smooth sailing from there so we will have charts I'm gonna do a whole separate video on it we might start with 7900 X and then move laterally from there and it's it's looking good so far so very roughly 15 to 20 Celsius decrease but don't go running to the hills at that number yet so we've got a lot more to do that was with a 1.1 5v ID 3.6 gigahertz locked on a 700 X neither of those settings are required but they restrict everything so that we get the same testing when it had Tim on it as we do with liquid metal you don't want the voltage to be auto control they'll jump all over the place so yeah it's looking good I'm pretty excited about it I mean seriously like this is a big move for us to finally get somewhere with the skylake power and thermal stuff we were waiting for the current clap got that in so we can finally take measurements at the rails we've been doing that for maybe a month or two now probably two months now and then waiting for one of these got this in and then we were waiting on really just the schedule to clear up from Vega this came in when Vega came in so it got buried but yeah pretty exciting this will be just in time for some other stuff so I guess that's it really um yeah thank you for watching I know it's informal I know there's not a lot of data in here but I'm not trying to present it as a lot of data it's just like here was my experience a very quick recap just for anyone who wants to do this in the final video I'll talk through the process more or maybe we'll do a separate video we might do a separate like do you a tutorial video it's trivial to use but it's also a very expensive mistake to make if you miss one of those trivial steps so we might do a video but either way just to recap this is easy you put the CPU in there clamp it down with this thing and slowly tighten to the point where you'll hear a pop and see very slight movement and just immediately stop back it out it's always easier to apply more force later than it is to to undo the application of excess force so just stop pull the CPU out you basically kind of like rock the IHS and then it should come straight up like don't push it you'll rip rip caps and resistors and things so that's easy liquid metal trying to use a small dot when I use this and just it way too much came out and I just wasn't really positive how much I needed on there so learn from that if there is a an issue with boot and oh actually another so another tip before I get to that VSG and they're bauer both suggested to me using either nail polish or basically like not a captain tape but you could use an electrical tape that's rated for like 105 Celsius or something the nail polish worked well so you basically put that coat that on the pastures and things surrounding the the dye resistors capacitors other SMD parts and then that will propel the it'll repel the liquid metal so they'll keep it out of there so I learned that a bit later after the first short and it prevented a second short so that's pretty cool trick to use very easy to do cost a few bucks to get one so that's that's most of it from there you end up with a thin film on both the dye and the IHS that's important don't forget the IHS because you will have worse thermals without it and if it shorts if if you can't boot don't panic stop trying to boot just turn the thing off remove the processor I know it sucks to redo if they don't pay is to remove it all and look under the hood if you've had spillage clean it up very carefully you can use q-tips you can use rubbing alcohol you can use nail polish remover just make sure it doesn't have too much acetone in it you can dilute it with water if you if you're uncomfortable and try again that should get you there you just want to be very careful when you're cleaning the thing not to rip off components like if you're trying to scrub with your nail or something else that's kind of harder and won't conform to services be careful because you can rip off tiny component if it's a capacitor you might very well be okay you might leave some voltage ripple suppression or might lose something but probably not performance if it's something else not a capacitor you might be in trouble so just be really careful but yeah that's it for this one thank you for watching kind of different but we'll have more for you in the near future with this benchmarking so patreon.com slash gamers Nexus tops not directly subscribe for more I will see you all next time you