Gadgetory

deleting the AMDR 322 an orgy wasn't as clean as using the pre-built tools for Intel CPUs but we have a separate video that'll show the deal in process to expose the APU die this new apu uses thermal paste rather than andy's usual solder which isn't likely a cost-saving measure for the low-end parts we ran stock thermal tests on our 2200 g using the included cooler and a 280 millimeter X 62 liquid cooler then deleted it applied thermal grizzly conductor not liquid metal and ran the tests again today we're looking at that thermal test data to determine what sort of Headroom we gain from the process of removing the stock thermal paste and replacing it with liquid metal just like we deal with Intel CPUs before we get into that this video is brought to you by the EVGA sleeved wire combs people often ask me what my hair care routine is and it's the same as any self-respecting enthusiasts cable management routine I use the EVGA sleeves wire combs to neatly organize each individual strand of hair just like how they can tidy up individual cable wires in your system the cable comb kit includes two 24 pin 2 8 pin 4 PCIe and a couple of peripheral sleeved wire combs learn more at the link in the description below so this was actually really cool it's pretty fun because the delayed process here normally we can use something like this this is a deal a dime 8 X from their Bower or one of these from rocket cool and they make the process pretty foolproof and overall safe today we basically put a rise in CPU or the APU into a vise clamped the IHS on one side and clamps the substrate on the other side and they just applied pressure until they broke free with a prescription of razors included yes this is where I keep them so we got it all working put liquid metal on there it is thermal grisly conduct or not and the improvement is noteworthy it's very similar to what we see with Intel before getting into the data some disclaimers here about safety don't do this at home unless you're willing to risk the APU it's fortunately pretty cheap part so not a big deal to replace it and also on being worth it people like to say this about Intel CPUs where they'll say like you have to deal with them no you don't you don't have to deal it Intel you don't have to deal with the AP use for them to perform within reasonable spec when you should deal with them is when you would like to have some actual thermal Headroom so that you can reduce the fan speed so that extra Headroom means that we can bring the fan speed down and operate at the same temperature as before with a quieter mode of operation alternatively it gives you more thermal Headroom for overclocking or of course it gives you room to use a smaller cooler if you don't want to buy a 280 you can get a 240 or an air cooler and you've got the Headroom there because you've deleted it and the thermal transfer between the die to the IHS is now going through a 73 watt per meter Kelvin liquid metal rather than four or whatever it is for the thermal paste so this is a process that from an enthusiast perspective is worth it but from a practical perspective will depend entirely on your needs because you can just throw a bigger cooler on it and get basically the same effect so we're not trying a serious overclock here for our overclocked numbers quotes there we have it had 3.8 gigahertz that's not a big jump and we've got it at one point four volts the point was not to push it as far as we could the point was to generate a very reliable set of thermal data that could be done from just pushing a lot of voltage through it and frankly instability on these ApS is such that it was just hard enough getting them to work at all with any kind of overclock on any board we found the a C's crosshair six hero to be basically the most reliable for a thermal test data and so we used that we had issues with the asrock pro 4 for a few reasons weight issues with the MSI B through 50 board see the timing video on that issues all around but we use the crosshair finally we're using delta T over ambient for the numbers here so that means it is a delta between ambient and the core temperature which is why the numbers will appear lower if you don't know that already and also t control is gone and ii got rid of t control which is the offset that would apply a 20 to 27 degree offset value to t die previously and call that t control so they've gotten rid of that and yeah that's pretty much all the disclaimers so let's get into the thermal numbers this first chart just establishes our base lines with the stock IHS we don't have any delayed numbers on here yet before deleting took place and using the crosshair six hero with full auto settings and a crack in X 62 our baseline for TDI is 39 degrees Celsius delta T over ambient this is with a 10 second high at about 39 degrees as well so very consistent data we operated at around 2.6 degrees celsius when idle manually configuring the CPU to run at 3.8 gigahertz and 1 point 4 volts give us a core temperature of 47 degrees over ambient which was about 25 degrees during all tests this increased our temperature over baseline by about 7 degrees idle is subsequently higher running at around 4 to 4 point 7 degrees the air cooler ran at about 56 to 57 degrees over ambient and that put TDI at around 81 degrees for the reported temperature with the 3.8 gigahertz and 1.4 volt configuration failing to complete the test the CPU measured at 69 degrees before thread failed at which point we considered the test to be incomplete we weren't able to achieve stability with this high of a voltage despite the frequency not being overclocked in any meaningful way subsequently we also disabled the cooling fan momentarily to determine what t.j.maxx is on these ap use or what temperature the thermal protections it would be enabled this turns out to be about 95 degrees Celsius T died at which point we start seeing throttling on the frequency side of things and you'll eventually hit a shutdown if your motherboard enables it if we add our ambient to about 25 degrees back to the T die of 69 in the event of the incomplete or DNF test we're at 94 degrees celsius which happens to align with TJ max moving to the deleted thermals chart the one that you're all interested in we're using conductor not here for the application but still using our standard test pace on the IHS which we manually applied and spread over the service with a graduated syringe to make sure it was the same amount of paste each time we immediately saw a 14 degree reduction in temperature between the X 62 thermal paste at 3.8 gigahertz and one point 4 volts and the X 62 with liquid metal using the same frequency and voltage the dee-lighted variants using conductor not operates at about thirty two point six degrees over ambiens versus 47 degrees over ambient for the default and the if they're all pasted that's a significant improvement in thermal headroom and quick notes here we allow these to run for thirty minutes at a time so you can check the data on that and the article linked below if you need more info on the overtime performance and note also that we removed the silicon adhesive from the IHS but we left it on the substrate as a guide for remounting the heat spreader we've tested this on Intel before and I found that you might gain another one or two degrees max 3 by removing all silicon adhesive and it's just generally not worth it so we left the adhesive on the base of the substrate for a guide and thinned it out then remove the rest from the IHS this is already a major improvement idle remains about the same as expected liquid temperature is also functionally equivalent and within error margins but for the auto test with liquid metal which operated at nearly an identical power load we observed a twenty eight point seven degree over ambient core temperature marking an eleven degree drop from the stock they're all pasted with our air coolers the auto configuration drew only a couple watts more than the throne paced test entirely within variance but still managed to perform at forty three point six degrees delta T over ambient comparatively this puts the stock air cooler at a reduction of roughly thirteen degrees versus the baseline stock Tim from AMD that feels weird to say Tim from AMD it's not just Intel anymore and he is now using thermal paste as well although only on the low end stop with the air cooler running at 3.8 gigahertz and one point four volts sort of an overclock the performance report was now fifty two point five degrees over anime and for the liquid metal version or 69 degrees for the stock paste that's a 17 degree difference and it was enough of a change that we were able to actually fully survive the 30 minute burnin test pass when using liquid metal despite failure with the stock APU thermal interface material as for power consumption this chart illustrates the power consumption for a deleted and a stock test pass which ensures that we're pulling the same amount of current at the EPS 12-volt rails for each test this is part of our means to validate that our test data is and you can find more of this discussion in the link in the description below for the article will also include a power chart table in the article for test validation purposes and on the note of power we basically keep a current clamp on the 8 pin power cables going into the motherboard so that be the 12 volt cables and use that to measure power consumption which we can then plot over time versus temperature to make sure it's drawing the same amount of power for each test so there's no inconsistency that gets through from prime or a blender or whatever the test software may be for each test overall ultimately this is somewhat of a risky endeavor again there's no pre-made tool you can't just buy something from rocket cool order power and get something that makes it easy to deal with rather safely there Bowers do you add eye mates for example only exists for Intel why would anyone make something for Andy because they're all soldered in the past so if you want to do this you're basically on your own it's not that difficult but there is certainly more risk of screwing up personally speaking I have not deleted something by using a vise before but it worked fine and logically it made sense you just protect the substrate put it on one side and then clamp the IHS now we did do a good bit of thinning out that he so first the razor but it's not too hard overall so ignoring the aspect of getting it done is it worth actually deleting these CPUs or the AP isn't putting liquid metal on it it's actually it depends so it comes down to if you have an issue with the noise of your system like fan noise you could reduce the fan speeds by deleting it that gives you another 10 degrees or so of headroom under the workloads we tested which would then allow you to run the fans at a lower rpm while maintaining the same temperature as previously that's one of the advantages that people don't really talk about if you have a thermal Headroom issue like we did with the stock cooler and the somewhat aggressive voltage and the thermal paste test run then going liquid metal could help with that too if you just have a weak cooler this is primarily something that allows you to reduce fan speeds use a smaller or lower powered cooling device and achieve the same effect as previously in terms of temperatures or just get some extra Headroom for overclock and if you really need those last couple degrees so just like with Intel CPUs the change is not necessary you don't need to deal with the APU or the Intel CPUs either of them to get decent performance out of the box it's just it's a nice thing to do if you already have objectives in mind where it would help so certainly pretty fun to test and the results were good now you wait 13 to 17 degrees at the low and the high end that's a pretty big difference so for that we used just liquid metal manually applied it and then manually spread their own pace on the top mounted the cold plate and that was it so as always you can go to the channel check out the build video and other things related to the AP use subscribe for more apu content coming up and go to Stuart our gamers Nexus dot net to pick up a shirt like this one backorder one of our mod mats or back us on patreon patreon comm slash gamers Nexus thank you for watching I'll see you all next time