Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
2020 Air Heatsink Modification Thread
- Thread starter ilikewhey
- Start date
- Sort by reaction score
Oh trust me I know.
I have guarantee for sustained performance beyond the base clock with my MBP because it's cooling is adequate.
Lol, no one is saying MBP cooling is great, it's just adequate for the CPU which is as it should be.
[automerge]1584951662[/automerge]
Point is my MBP can use all of my CPU's performance. The 2020 MBA can't.
[automerge]1584952041[/automerge]
Of course it's not a "race" CPU, that's just how the OP of that wrote it. I'm not comparing to to higher wattage CPUs. I'm saying the i5 has much more to give than it cooling allows compared to the MBP. How do you not understand that?
Last edited:
Yeah that’s totally true but thats still normal in such small devices. But the cooling isnt that bad, it could be better but its still a portable notebook. Nobody would use it as a gaming machine. Even the MBP sucks at gaming, just get a decent tower pc for gaming and heavy load.
I'm not trying to be patronising, but even after multiple explanations some still don't seem to get it.
Yes, we can talk about thermal gradients if you wish, and I touched on it earlier when I mentioned that the heatpipe in the MPB and earlier MBA is "cooled" by warm air that has already been sucked from the inside of the case and blown out, whereas the 2018-20 MBA draws fresh air in over its heatsink.
So all things being equal, thermal gradient is another area where the heatpipe design quoted in this thread may be at a disadvantage.
This is really the crux of the issue: how do you even know what the MBA's "ultimate" performance is? Do we need to super-cool the CPU with refrigerants to achieve it? Do we need to approach 0 Kelvin to extract every last bit of performance out of it?raynamnh said:
I think it is silly to speculate what the CPU's "real" performance is when there are so many other factors to consider in addition to CPU temperature.
The 2020 MBA CPU's performance is what it is. Go and stand in a freezer if you want it to clock faster for longer. Otherwise just accept that everything is a compromise and that the 2020 MBA design is nowhere near as bad as some people are trying to make out.
[automerge]1584953008[/automerge]
In motor racing there is a saying: You have a choice of speed, reliability and low cost – the catch is, you can only have two of those things.
This thread needs a cooling system🥵
btw I still ordered the i5 Air because I mostly use for web browsing and productivity.The heat sink issue doesn’t bother me that much.
btw I still ordered the i5 Air because I mostly use for web browsing and productivity.The heat sink issue doesn’t bother me that much.
Depends on whether the power/clock configuration is not already optimised for the hardware as it is.
If this was AMD we'd all be talking about whether it can be under-volted, but this is intel, and a brand new architecture on a new 10nm process (as opposed to yet another 14nm+++ skylake refresh). What i'm saying is that right now, this processor in this machine is uncharted territory.
Maybe tweaking the power delivery down may allow it to clock higher for longer without spiking heat and clock and hitting max temperature quicker then clocking back.
Maybe there are GPU vs. CPU power trade-offs that can be made depending on what the user is doing. Maybe the thing is down-clocking the CPU in some tests in favour of GPU for things that are GPU accelerated? After all, the integrated GPU in this is relatively strong, even vs. the current 13" Macbook Pro. Maybe the "omg it down clocks cpu to 2ghz" is irrelevant for what it is doing at that point?
But i suspect probably not much to do in the SMC. Until more of them get into people's hands who actually know what they're doing (vs. apple's favourite youtube-celebrity people) - and they have time to tweak, we won't know for sure.
By ultimate I mean being able to sustain close to max boost speed. The i5 goes up to 3.5GHz. Thermally, however the MBA allows it only to sustain at 2.0GHz. That's the problem. By comparison, my single fan heat piped 2016 MBP with a 15W 2.0GHz i5 sustains 2.9GHz @ just over 20W.
In other words the MBA with i5 can only thermally sustain 57% of its max boost speed, while my MBP can sustain 94% of its max boost speed.
Yes of course you could cool it with liquid nitrogen but you know what I mean, come on!
Look at this youtube review. At 3:49 the cinebench test starts. A minute or two into the bench test it's at 100 deg C, pulling 11W at ~1.7GHz while fans are only spinning at 50% (about 4000rpm). He then (at 5:33) puts the fans to 100% (8000rpm) and the temp does not drop! Granted, he only follows temps for a few seconds, but when I do the same with my MBP, temps drop immediately. Cooling is clearly very restrictive to the i5, do you honestly not see that as an issue?
Nobody ever said performance was bad. We're saying it's a shame the thing thermal throttles so much otherwise it would be much more flexible machine, could do some video editing, light gaming etc.
Lol, I'm a car guy myself. Yeah it's one of those generic sayings you could say about almost anything in life... Performace, reliability, price, pick any two.
All a little amusing; Apple has always been OK with leaving performance on the table for the sake of the aesthetic and quieter operation, nothing's changed in 2020. I do object to that premise with the 15" & 16" MBP as these are purportedly designed for heavy professional workloads. Every Intel based Mac I've owned and used has always ran hot, nor has that ever been overly detrimental in my own experience. Likes of the Air and 13" MBP is far less of an issue given the intended usage.
The Y series CPU's are deliberately power limited, I own a W10 UMPC 2in1 with an i7 Y series CPU that's rated to 4.1GHz and it reacts very much similar to the new MacBook Air's and it has an active cooling system. The CPU will always be railed in by the power limits, even with an active cooling solution, similar CPU temps are aggressive. Being a W10 device I could likely unlock any power limits set be the manufacturer, equally thinking it's going to sustain anything close to 4GHz is completely unrealistic.
Some notebooks can sustain full Turbo Boost, this Asus can hold full boost across all cores, however that is more the exception than the rule with portables as this requires a very substantial power & cooling solutions. TBH if you feel that your Mac is running a little too hot simply raise up the rear of the notebook half an inch or so, you might be surprised...
Apple may also optimise the firmware in time, I recently resurrected a 2011 15" MBP and it now 9 years on performs far better (10.12). New the MBP could barely hold 2.5GHz at 103C under full load today it holds 3.1GHz across all cores, with lower temps and remains to be completely stock. These models are notorious for running hot and burning out dGPU's. TBH I wouldn't be surprised if Apple does, update the Air's firmware once the new 13" MBP is released with 10th Gen Intel as Apple likes it's product lines to be neatly segregated.
Simple match the notebook to your usage, if you need more performance than the MBA offers opt for a MBP. Like as not, Apple will release an update to the rest of the range with the new keyboard, given the Butterfly Keyboard debacle...
Q-6
Last edited:
not a thermodynamics expert but from the years of building custom servers and gaming pcs, a heatpipe with fins with a fan pushing air right in front of it is generally more effective as a cooling solution than the setup being presented in the 2018 and up. it offers a much higher static pressure for the cooling fins.
one example i can provide is thermaltake has a case that has a compartmentalized interior layout, meaning they separated the interior into different chamber, one chamber for cpu, one for gpu, and another for everything else. same concept as you stated earlier. although there were sufficient airflow in that chamber, there was a lack of static pressure to effectively wick away the heat from the fins on the cpu heatsink. we wanted to only use high airflow case fans and one exhuast fan, but ultimately had to install a purposeful cpu fan on the cpu heatsink to increase static pressure.
Irrespective, the performance gains of the 2020 MBA vs the old model are indeed real. I understand that I am not getting an MBP, but it is a better MBA than the 2019.
Where do you think the heat being ducted along the copper pipe goes? The heat is transfered to the atmosphere. What is our atmosphere made up of? Air.
[automerge]1584983009[/automerge]
Max boost speed is not meant to be continuous boost speed. It is meant for short-term bursts.
And so that I understand you correctly, you're saying your MacBook Pro can sustain a 45% increase in base clock speed (2.0GHz>2.9GHz ), but you think the MacBook Air cooling system is to blame because it can only sustain an 82% increase in base clock speed (1.1GHZ>2.0GHz)?
Kinda put things into perspective, doesn't it?
The rest, I'm afraid is semantics. You may believe what you wish, but in my opinion there is nothing wrong with the heat-ducting system on the 2018-20 MBA for its intended market and subsequent build.
Last edited:
The difference is that in a proper cooling design the heat is dissipated into the air at the fin portion of the heatpipe where there's more surface area and right at the aperture of the blower fan and chassis exit. The current MBA design generates heat more quickly than it can get rid of relying on air as the transfer medium and thus throttles.
There is a formula for heat exchange and you are correct that a proper heatsink with forced air over it can be an efficient system, but it relies on the surface area of the sink, the number of air molecules over that surface in a given time, and the temperature gradient between the two. It is also not the only efficient way to cool a CPU in a confined space. The radiators on the end of the MBP/early MBA heatpipes are not large and sit in front of a fan blowing warm air from the case, and in no way resemble the cooling systems of a server. I don't know how you can compare the two.
[automerge]1584984806[/automerge]
Air is always the transfer medium. It comes down to how much air you can move over a given surface area. A heatpipe with a small radiator in front of a fan blowing warm air is not necessarily exchanging more heat than a fan pulling cool, fresh air through a larger heatsink. The heat exchange formula does not care how convection is achieved, only the rate of energy transfer.
Last edited:
Better doesn't mean safer. It still can die or shut down because of thermal throttling and overheating.
Again, not true. My MBP sustains 2.9-3.0 GHz and it's max boost speed is 3.1 GHz.
I'm well aware the MBA can sustain double its base clock. But again that's not my point. Seriously, what do you not understand. Read over what I said previously. It's about the fact that the i5 MBA goes up to 3.5GHz yet can only sustain 2GHz. It's a waste of performance. It's not hard to understand my point.... The cooling is restrictive.
Except for my single fan heat piped MBP cools my CPU pulling 20W+. This MBA has to throttle down to pulling 10W. That's the problem. As you say, wattage goes to heat. And tell me why my MBP can dissipate 10W more heat? Heat pipe. End of story.
Ehmm I think It can’t sustain 2.0GHz, on the Appleinsider stress test it’s stuck at 1.6GHz after a minute of boost.
I'm glad that we have so many world-class engineers on here, there may even be enough to create a competitor to Apple.
Seriously, Apple designed a general-use laptop for the 80% that isn't designed for extreme performance constantly. They have the MBP for that and expect buyers to know if they need to move upmarket.
There is no perfect one-size-fits-all laptop. If you want MacOS and thin and light for general use, you get the MBA and it will do everything you need at a great price. If you want MacOS and need performance above anything, you get the MBP and it will do everything you need at a higher price.
Seriously, Apple designed a general-use laptop for the 80% that isn't designed for extreme performance constantly. They have the MBP for that and expect buyers to know if they need to move upmarket.
There is no perfect one-size-fits-all laptop. If you want MacOS and thin and light for general use, you get the MBA and it will do everything you need at a great price. If you want MacOS and need performance above anything, you get the MBP and it will do everything you need at a higher price.
I can't speak to Apple's intentions, but their advertising says:
A lot of people are going to be expecting 2x the performance of the old 1.6GHZ i5. Based on the thermal throttling observed in these videos, I think it's unlikely they're going to get that under sustained load.
Why can't people understand that if Apple wanted the $1000 MacBook Air to do the same applications as effectively as the $1300 MacBook Pro does, and not buy the $1300 MacBook Pro to do those things, they would have designed the MBA with a better cooling system.
Come on guys, this is simply product differentiation. It isn't an engineering or cost problem. I've specified and designed high tech products for over 40 years. It isn't a problem at all as far as Apple is concerned. It's about maximizing profits.
Come on guys, this is simply product differentiation. It isn't an engineering or cost problem. I've specified and designed high tech products for over 40 years. It isn't a problem at all as far as Apple is concerned. It's about maximizing profits.
that i can agree with, apple cannibalizing the air so folks have a reason to get the pro.
The problem is people looking at clocks rather than what the machine can actually do.
This machine will likely be faster real world than your 2016 Macbook Pro at the lower clock speed due to superior GPU and higher IPC of the new processor.
In a server or desktop environment you are not extremely space limited. In a server you are nowhere near as budget constrained as a bottom of the range "affordable" notebook.
A heat-pipe does nothing other than transfer heat. In a space constrained notebook that is space you could otherwise use for heat-sink area. Or battery. Or other components.
If you have sufficient space, a directly mounted air cooler (maybe itself containing heat pipes) is more efficient than a heat pipe plus radiator, assuming you have the same airflow and surface area for your heat exchanger.
The ONLY reason to run a heat pipe or other method of heat transport (liquid, etc.) is to enable you to run more surface area (or get better airflow across the surface you do have) somewhere else away from the heat source for cooling. Yes, there are air coolers with inbuilt heat pipes, but that is to enable larger surface area of fins on the cooler and more efficient heat transfer through the LARGE heat sink.
Now... what you're (perhaps?) suggesting, a "heat pipe with fins" is different to what will work effectively in this macbook.
There's no space to put fins on a heat pipe in that space. Given that fact - You'd just be running a heat-pipe to transport the heat... to where? All a heat pipe would do in that box is reduce the ability to shift air in that chassis.
Where are you going to fit a LARGER heat exchanger than that heat sink inside that device, that you can transfer heat to with your heat pipe?
[automerge]1585008472[/automerge]
Your MBP has half the size integrated GPU or less.
The big question is this: which one gets the job done faster?
Last edited:
what do you think a directly mounted air cooler is made off, heat pipe(s)
i'm not suggesting fitting a noctua d15 in there, just take a look at a 2015 mba that has been used for reference by previous users.
by using a heat pipe like the 2015 mba, you transfer the heat directly infront of the fan exhaust to get direct static pressure from the fan. therefore the fan would be much more effective at wicking away the heat than the heatsink being placed couple inches away from the fan's blowing direction.
you guys are really making this way more complicated than it is.