Improving MacBook Pro Pre-Unibody Airflow/Cooling

[vddrnnr]

Hi all,

I'm trying to "boost" cooling performance in my late 2006 MBP 17" (MBP2,1).
Besides the usual re-paste/cleanup has anyone tried other ways of improving cooling
by replacing fans for example or other methods.
Please share your thoughts/experiences.

Best regards,
voidRunner

 

[Amethyst1]

Remove the keyboard/trackpad part of the case. Turn it over. Remove the black stuff that blocks the air vents.

 

[B S Magnet]

Remove the keyboard/trackpad part of the case. Turn it over. Remove the black stuff that blocks the air vents.

A thorough cleanout of vents, as @Amethyst1 described, along with an overhaul of your fans, by fully cleaning away all the fine dust accumulated on the blades. Follow this with a careful opening of the fan covers (you’ll need to take out the fans from the rest of the case) to pull the fan wheel itself from the spindle housing (where the copper coils live), cleaning the both spindle and the housing (with a fine cloth and a tiny enough object, like a paper clip, around which the cloth-covered clip can reach inside the spindle housing), and then carefully adding a drop of either mineral oil or a lubricant designed specifically for lubricating power tools (like drills), before reassembling the fan assemblies. (Other general lubricants can break down due to prolonged heat and ageing, and those can gum up things in the long run. Stick with a lubricant designed for high speed and moderate heat use.)

The above overhaul of fans is what I did with my A1138 and A1139 fans when they began to slow down. Now that I have the proper lubricant and a sense of how the Sunon fans in Macs are designed, this is my go-to for future overhauls.

Finally, after all the cleaning out and lubricating is done, grab Macs Fan Control for the OS version you’re using and set the fan speeds manually as needed. This ought to give you an improvement over your current situation.

 

[chabig]

The fan blades may also have accumulated dust on them. Clean that off.

 

[B S Magnet]

The fan blades may also have accumulated dust on them. Clean that off.

Yah, that goes hand in hand with the thorough cleaning steps I described. The dust which accumulates on those blades is extremely fine and cake-like, probably due to any oil particulates bonded with the fine dust particles over the years.

 

[vddrnnr]

Hi all,

Yes that's what I've done apart from disassembling the fans to make sure
all the dirt is removed.
One of the things I also noticed is that when I use a blowdryer with no heat is that
pushing air through the fan lowers temps considerably even ( off course ) to the point where
the fan stops because temps go into the 30s.
When I noticed this although my fans rev up all the way to 6000rpms when temps go up it
seems they don't push enough air to cool the cpu.
I don't know if a different design with more blades for example like the unibody MacBooks have and without the
top would allow for better airflow.

Right know using macfanscontrol I have the fans spinning at 2500rpm when the system is mostly idle
and cpu temps on the core around 45 celsius.

I'll keep you posted with new findings.

ps. I forgot to mention that my fans are from Delta Electronics

Best regards,
voidRunner

 

[B S Magnet]

Hi all,

Yes that's what I've done apart from disassembling the fans to make sure
all the dirt is removed.
One of the things I also noticed is that when I use a blowdryer with no heat is that
pushing air through the fan lowers temps considerably even ( off course ) to the point where
the fan stops because temps go into the 30s.
When I noticed this although my fans rev up all the way to 6000rpms when temps go up it
seems they don't push enough air to cool the cpu.
I don't know if a different design with more blades for example like the unibody MacBooks have and without the
top would allow for better airflow.

Right know using macfanscontrol I have the fans spinning at 2500rpm when the system is mostly idle
and cpu temps on the core around 45 celsius.

I'll keep you posted with new findings.

ps. I forgot to mention that my fans are from Delta Electronics

Best regards,
voidRunner

Digging through my (blurry) repair pics on my A1139 (first pic) and A1138 (second pic), I realize they use fans supplied by SEI, not Sunon.

Even so, comparing them next to the repair pics of the A1261 (third pic), which do use Sunon, the general construction appears to be of similar brushless, “maglev” construction, and the steps to refurbish them ought to be similar to what I described.

I imagine Delta’s fans are made to similar specifications, as the technical needs required of them by Apple would be more or less identical.

 

[Raging Dufus]

A thorough cleanout of vents, as @Amethyst1 described, along with an overhaul of your fans, by fully cleaning away all the fine dust accumulated on the blades. Follow this with a careful opening of the fan covers (you’ll need to take out the fans from the rest of the case) to pull the fan wheel itself from the spindle housing (where the copper coils live), cleaning the both spindle and the housing (with a fine cloth and a tiny enough object, like a paper clip, around which the cloth-covered clip can reach inside the spindle housing), and then carefully adding a drop of either mineral oil or a lubricant designed specifically for lubricating power tools (like drills), before reassembling the fan assemblies. (Other general lubricants can break down due to prolonged heat and ageing, and those can gum up things in the long run. Stick with a lubricant designed for high speed and moderate heat use.)

The above overhaul of fans is what I did with my A1138 and A1139 fans when they began to slow down. Now that I have the proper lubricant and a sense of how the Sunon fans in Macs are designed, this is my go-to for future overhauls.

Finally, after all the cleaning out and lubricating is done, grab Macs Fan Control for the OS version you’re using and set the fan speeds manually as needed. This ought to give you an improvement over your current situation.

Since you've recently done an overhaul of similar cooling systems in your PowerBooks, perhaps you'd be best suited to answer this query, but answers from anyone are certainly welcome. I'm curious as to what "quiet operation" really means in the context of pre-unibody MBP fans.

I know that the system spins the fans constantly, at a low level, by design. But in putting together an A1261 from parts (and I remember this also from Sunon fans in my formerly-owned PowerBooks), I've noticed a very slight whine coming from them. The best way I can describe it is a constant, distant mosquito. It's easily enough tuned out unless I'm using the MBP in a really silent place. Over time I could forget entirely that it's there. So my question is this: is it just me and my sensitive hearing, or is this an indication of impending failure? When you describe "quiet operation" of an MBP or PB with these fans, does that include a barely audible whine?

When the fans ramp up, they don't make any unusual noise, and no whine is discernable in the "whoosh" - if that makes sense. I don't want to go to the trouble of taking these fans out and apart to service them if: (a) they don't really need it, because the whine is normal; and/or (b) it won't make any difference.

So, how about it? Do your fans whine?

 

[B S Magnet]

Since you've recently done an overhaul of similar cooling systems in your PowerBooks, perhaps you'd be best suited to answer this query, but answers from anyone are certainly welcome. I'm curious as to what "quiet operation" really means in the context of pre-unibody MBP fans.

I know that the system spins the fans constantly, at a low level, by design. But in putting together an A1261 from parts (and I remember this also from Sunon fans in my formerly-owned PowerBooks), I've noticed a very slight whine coming from them. The best way I can describe it is a constant, distant mosquito. It's easily enough tuned out unless I'm using the MBP in a really silent place. Over time I could forget entirely that it's there. So my question is this: is it just me and my sensitive hearing, or is this an indication of impending failure? When you describe "quiet operation" of an MBP or PB with these fans, does that include a barely audible whine?

When the fans ramp up, they don't make any unusual noise, and no whine is discernable in the "whoosh" - if that makes sense. I don't want to go to the trouble of taking these fans out and apart to service them if: (a) they don't really need it, because the whine is normal; and/or (b) it won't make any difference.

So, how about it? Do your fans whine?

My fans on the PBs and the A1261 don’t whine or make a high-pitched buzz/hum/whirr like a mosquito might, no.

What I mean by “quiet operation” is the fans aren’t suddenly making an obnoxious drone or buzz which is often heard from dying fans clogged with a lot of gunk (and with old-style brush-motor fans on their death bed).

Rather, “quiet operation” here means when the fans spin up to their maximum, the only sound I can hear is that of air being moved — what I tend to think of as a “fan whisper”. At their minimum (and even at half-max), this whisper is either inaudible for the most part or just barely evident in a room with not much else happening.

I guess another way to put it is: if you brought the laptop to a public library, and the fan(s) are running at maximum, will anyone else notice? If so, as in distraction-to-annoyance, then no, this isn’t what I meant as “quiet operation”.

As for my A1278, whose fan construction, tech, and design are similar to those in the A1286 and A1297 (all are by Sunon), I’ve owned said laptop since 2011 and have had to replace its fan twice due to rising buzz/whirr — like, so loud to the point where it could be heard from several rooms over. Like, infuriatingly loud.

I replaced those fans long before figuring out how to bring new life to the ones in my PBs). I don’t yet know if the internal construction of the unibody-era fans are more or less similar or fundamentally different from the aluminium PBs/MBPs (aside from knowing they’re all brushless/maglev motors). As for the Sunon fans in my A1261, I think they’re the originals, and generally, the only one I tend hear at maximum (which is fairly often) is the one on the left (which is, no surprise, next to where the GPU lives).

 

[Raging Dufus]

My fans on the PBs and the A1261 don’t whine or make a high-pitched buzz/hum/whirr like a mosquito might, no.

What I mean by “quiet operation” is the fans aren’t suddenly making an obnoxious drone or buzz which is often heard from dying fans clogged with a lot of gunk (and with old-style brush-motor fans on their death bed).

Rather, “quiet operation” here means when the fans spin up to their maximum, the only sound I can hear is that of air being moved — what I tend to think of as a “fan whisper”. At their minimum (and even at half-max), this whisper is either inaudible for the most part or just barely evident in a room with not much else happening.

I guess another way to put it is: if you brought the laptop to a public library, and the fan(s) are running at maximum, will anyone else notice? If so, as in distraction-to-annoyance, then no, this isn’t what I meant as “quiet operation”.

As for my A1278, whose fan construction, tech, and design are similar to those in the A1286 and A1297 (all are by Sunon), I’ve owned said laptop since 2011 and have had to replace its fan twice due to rising buzz/whirr — like, so loud to the point where it could be heard from several rooms over. Like, infuriatingly loud.

I replaced those fans long before figuring out how to bring new life to the ones in my PBs). I don’t yet know if the internal construction of the unibody-era fans are more or less similar or fundamentally different from the aluminium PBs/MBPs (aside from knowing they’re all brushless/maglev motors). As for the Sunon fans in my A1261, I think they’re the originals, and generally, the only one I tend hear at maximum (which is fairly often) is the one on the left (which is, no surprise, next to where the GPU lives).

Thanks. Perhaps it really is just me then, or perhaps it's worthwhile to take them apart and try to kill the "mosquito." You mentioned a buzz, and I suppose what I'm hearing could be described that way as well. But it's really very low, not obnoxious.

 

[B S Magnet]

Thanks. Perhaps it really is just me then, or perhaps it's worthwhile to take them apart and try to kill the "mosquito." You mentioned a buzz, and I suppose what I'm hearing could be described that way as well. But it's really very low, not obnoxious.

Well, if it doesn’t bother you and doesn’t impact the cooling of your laptop, then why not leave it be for now and, pun not intended, keep an ear out for changes?

 

[Raging Dufus]

Well, if it doesn’t bother you and doesn’t impact the cooling of your laptop, then why not leave it be for now and, pun not intended, keep an ear out for changes?

Because I happen to have it open, or at least I haven't replaced the screws yet, because I'm still involved in testing/replacing certain hardware components. Before the pandemic borked the supply chain, I used to be able to tackle a project like this in an afternoon. But alas, waiting on parts these days makes it stretch to weeks. So, might as well take care of it while I'm under the hood

 

[vddrnnr]

Hi all,

After a long break I've started toying with this again.
To measure improvements I want to compare my temps with anyone with a similar model.
Can you guys help?

Right now I'm testing with an early 2008 17 inch model.

PS. Right now from my testing I've noticed that GPU temps rise when power is connected.
It seems the GPU changes the clocks if using power and so more heat.
Has anyone noticed this? Also in regular use I don't notice any performance benefit from this.
I'm not doing any heavy graphics usage just having the power connected is enough.

Thanx in advance.

Best regards,
voidRunner

 

[B S Magnet]

Hi all,

After a long break I've started toying with this again.
To measure improvements I want to compare my temps with anyone with a similar model.

Some good news!

Since you opened this discussion several months ago, I did some passive cooling improvement work on my A1261 (the same 17-inch Early 2008 you mention below), and the the results have been substantial (for my primary goal of keeping my legs from getting burnt!).

There are several parts I posted on this mod.

If you’re thinking of improving cooling, it’s worth reading through all four posts, in order, to help you decide whether this kind of project might be worthwhile for your needs.

Part 1 :: Part 2 :: Part 3 :: Part 4

Right now I'm testing with an early 2008 17 inch model.

View attachment 2132984

PS. Right now from my testing I've noticed that GPU temps rise when power is connected.

This is my observation, too. More specifically, the CPU and CPU temps drop when running from battery only — suggesting that battery-only usage send the system into a kind of lower-power mode. I have not tried running benchmarks on battery alone, so I don’t have the performance numbers to back that up.

It seems the GPU changes the clocks if using power and so more heat.

Possibly. I’m unsure whether it’s a clock speed change or a simple load average change which kicks up heat production under load.

When running from the mains (whilst having a working battery installed), I gathered that the temps spike because there is more demand on the GPU — rather than the GPU itself changing clock speeds.

 

[vddrnnr]

Hi B S,

I read your posts and will look into thermal padding .
But one thing strange is that I seem to already have much lower temps than you.
What iStat version are you using?
My values running with arctic fox open and doing light surfing are

Best regards,
voidRunner

 

[B S Magnet]

Hi B S,

I read your posts and will look into thermal padding .
But one thing strange is that I seem to already have much lower temps than you.

Two things:

I’m running the 2.6GHz (T9500) iteration of the CPU on the A1261; the temps you saw in my screen caps was whilst running from both the mains and the battery working together. The other thing is factoring in the room’s ambient temperature: as memory serves, the room ambient temperature was in the 25–26°C range when I made most of those screencaps. When I come home (and just before letting the heater warm the room), I do observe notably lower overall temperature readings.

What iStat version are you using?

I use iStat 4.22 on 10.6.8. You appear to be using iStat 3.19 (what I use on my 10.5.8 and SL-PPC PowerBook G4s). The sensor names, for whatever reason, differ between different versions of iStat Menus and also the sensor names listed in Macs Fan Control. I’m not entirely sure why that is.

My values running with arctic fox open and doing light surfing are

I’d be interested to know what your ambient room temperature is whilst running these applications, along with verifying whether this is with you using just the mains, just the battery, or both power sources together. Cheers.

 

[vddrnnr]

Hi B S,

I'm also running with battery and mains.
The room temps I think are in the 20 to 22 range.
I'm also using Lion and my iStat version is 2.0 but temps seem to be in-line with the ones
from Macs Fan Control which I'm also using to better control the fans.

PS. I think in SL I have a little higher temps also maybe 2,3 degrees difference but iStat is the
same version.

PS2. If I remove mains GPU temps drop between 3 and 5 degrees on both GPU diode and GPU Proximity

Best regards,
voidRunner

 

[B S Magnet]

Hi B S,

I'm also running with battery and mains.
The room temps I think are in the 20 to 22 range.
I'm also using Lion and my iStat version is 2.0 but temps seem to be in-line with the ones
from Macs Fan Control which I'm also using to better control the fans.

PS. I think in SL I have a little higher temps also maybe 2,3 degrees difference but iStat is the
same version.

OK. Gotcha.

PS2. If I remove mains GPU temps drop between 3 and 5 degrees on both GPU diode and GPU Proximity

View attachment 2133152

So I went ahead and did a couple of things since your reply: I let the system idle (and quit the three, high-resource applications I had been using, in order to establish a basic baseline which could be replicated) — first, on battery-only, then on both mains and battery.

The current ambient room temperature where the MBP is located is about 23–23.5°C (I have an analogue glass bulb thermometer, which is not medical grade, but is fish tank grade).

On battery only, after about ten minutes:

On mains and battery, after about five minutes following the previous:

The CPU on battery only was 43°C; with battery and mains, it’s rebounded to 46°C.
The GPU on battery only was 54°C; with battery and mains, it’s rebounded to 59°C.
The left (GPU side) fan was running at 2844rpm; with battery and mains, it’s at 3681rpm. The right (CPU side) fan is unchanged.

The range I’m seeing is somewhat higher than the 41°C CPU temperatures you’re seeing in Lion, but factoring the ~2–3°C higher temps you might see whilst running 10.6.8, with another 1°C thrown in there (to account for variations between the T9300 you’re running and the T9500 I’m running, as well as possible variations in that 20–22ºC room temp your system is running, and the 23°C mine is running), the sensor temperatures seem to be fairly pretty close to one another.

Hopefully these data help somewhat!


EDIT to add: I thought I’d add a bit about the passive thermal cooling method I’m using and when it makes the most significant difference. At cooler room temperatures, when not much is happening with the MBP, the system isn’t really taking much advantage of that passive thermal cooling/heat dissipation mod than when the system is under high loads and/or in a warm room. (After all, the two fans always run at a minimum of 2000rpm no matter how cool the system gets, so there’s always some active thermal cooling underway.) But by implementing the passive thermal cooling mod (atop the cooling implemented by Apple when they designed the system), I was more interested to have overall heat dissipation under high load to be spread across more of the case (namely, to limit the isolated hot spots on the lower case) than to lower temperatures at the heat-generation sources (i.e., the GPU, CPU, and memory controller diodes). It just so happens that while the passive thermal dissipation mod is moving more of that heat to a broader surface area of the aluminium case, it’s also lowering temperatures at the heat-generation sources and helping the fans to do just a little bit less work in active thermal cooling.

 

[vddrnnr]

Hi all,

Just to give you a quick update on my developments I've been testing ways to "improve"
airflow from the fans to the heatpipe / cpu / gpu by adding a cap like this to the fan.

I've used hard stickers which look like very thin plastic and the ones I have in
both macbook use all the space available inside the hole at the bottom of the fan housing
and are cut in a more rounded shape.

PS. I've bought the sticker at a Leroy Merlin shop in Portugal. I think in the UK you can find these at
Maxmat for example or order it online from ebay.

I'll post a picture with idle temps after some more testing.

Best regards,
voidRunner

 

[vddrnnr]

Hi all,

After a lot of testing I've ended with the folowing desing based on these two fans
one from a macbook pro unibody and another from a lenovo thinkpad t440p

The end result as an evolution of my previous post is something like this.

Both fans now have two "rings" which help fans keep a lower rotation while keeping
the temps a few degrees lower. In reality the temps keep a similar value but the fans are
able to keep those temps with a lower speed ( I'm using Macs Fan Control to better
adjust fan speed based on the temp of CPU and GPU sensors).

I think this helps because if you are using just the fans the air that is pushed "gets lost"
because as any "gas" you move with a blade it does not stick to the blades it'§s also
pushed up and down so the effect of the fan gets dispersed somewhat.

PS. If you do some cleaning on the fans, and remove the blades from the "motor",
be careful to place them back on the correct side ( left or right ) when reassembling.
Fans rotate in different ways on each side so the blades need to be "right" to
push the air correctly.

While doing these teste I've also noticed a new side effect from having the charger connected. If you are "charging" some components get hotter. One of those components is the memory controller. In my case
I noticed that if you are charging from a low battery percentage ( below 80% normaly ) the
temps went up all the way to above 50 degrees celsius. When the battery goes above 80% while charging
temps are around 42 - 45 degrees and when charging finishes they go down to high 30's .

One other thing I noticed while doing all these tests I've also found that Apple's "heatpipe" cooling system is very inneficient where contact between the chips and the "cooler" is concerned.
I think Apple also must have though this because all the unibody models have the "cooler"
directly and firmly attached to the logic board. So keeping all the screws tight firmly ( and having all
of them in place ) plays a significant role to keep the system cooler.

One other thing I did to help this is to put some "blocks" of the same "hard sticker" "glued" to
the underside of the keyboard to add more presure on the logig board. If you look at the
two screws near the memory slots they also do that but it's the only place it happens and because
the other screws that keep the logicboard in place are few in the spots where the CPU and GPU
are located it helped a lot to put this additional pressure.
I think that in some cases where the GPUs failed in these machines this additional "contact" pressure
might have helped avoid so many failures ( IMHO no real evidence of this but I'm using a baked
logic board and it seems to help ).

I also added another "tweak" that seems to help which is to put another ring
on the base of each of the fans "casing" to narrow the lower hole and so also helping
to move more air to the "cooler".

So in the end I have a cooler macbook pro.

Post your thoughts.

Best regards,
voidRunner

 

[vddrnnr]

new post above sorry

 

[B S Magnet]

Hi all,

After a lot of testing I've ended with the folowing desing based on these two fans
one from a macbook pro unibody and another from a lenovo thinkpad t440p

View attachment 2146311 View attachment 2146312

The end result as an evolution of my previous post is something like this.

View attachment 2146313

Both fans now have two "rings" which help fans keep a lower rotation while keeping
the temps a few degrees lower. In reality the temps keep a similar value but the fans are
able to keep those temps with a lower speed ( I'm using Macs Fan Control to better
adjust fan speed based on the temp of CPU and GPU sensors).

I think this helps because if you are using just the fans the air that is pushed "gets lost"
because as any "gas" you move with a blade it does not stick to the blades it'§s also
pushed up and down so the effect of the fan gets dispersed somewhat.

PS. If you do some cleaning on the fans, and remove the blades from the "motor",
be careful to place them back on the correct side ( left or right ) when reassembling.
Fans rotate in different ways on each side so the blades need to be "right" to
push the air correctly.

While doing these teste I've also noticed a new side effect from having the charger connected. If you are "charging" some components get hotter. One of those components is the memory controller. In my case
I noticed that if you are charging from a low battery percentage ( below 80% normaly ) the
temps went up all the way to above 50 degrees celsius. When the battery goes above 80% while charging
temps are around 42 - 45 degrees and when charging finishes they go down to high 30's .

One other thing I noticed while doing all these tests I've also found that Apple's "heatpipe" cooling system is very inneficient where contact between the chips and the "cooler" is concerned.
I think Apple also must have though this because all the unibody models have the "cooler"
directly and firmly attached to the logic board. So keeping all the screws tight firmly ( and having all
of them in place ) plays a significant role to keep the system cooler.

One other thing I did to help this is to put some "blocks" of the same "hard sticker" "glued" to
the underside of the keyboard to add more presure on the logig board. If you look at the
two screws near the memory slots they also do that but it's the only place it happens and because
the other screws that keep the logicboard in place are few in the spots where the CPU and GPU
are located it helped a lot to put this additional pressure.
I think that in some cases where the GPUs failed in these machines this additional "contact" pressure
might have helped avoid so many failures ( IMHO no real evidence of this but I'm using a baked
logic board and it seems to help ).

View attachment 2146346 View attachment 2146347

I also added another "tweak" that seems to help which is to put another ring
on the base of each of the fans "casing" to narrow the lower hole and so also helping
to move more air to the "cooler".

So in the end I have a cooler macbook pro.

Post your thoughts.

Best regards,
voidRunner

Interesting approaches re: the fans.

I’m going to have to think about the benefits of this.

Even with the adhesive rings you added, there is no contact being made between the rings and any other part of the fan assembly? Low ide is still ~2000rpm, etc.?

 

[vddrnnr]

Hi B S,

So regarding fan speed the lowest is still +- 2000 and it happens on cold startup
and idle on battery.
Posting some pics of idle on power and battery.

Best regards,
voidRunner

 

[B S Magnet]

Hi B S,

So regarding fan speed the lowest is still +- 2000 and it happens on cold startup
and idle on battery.
Posting some pics of idle on power and battery.

Best regards,
voidRunner

View attachment 2146430 View attachment 2146431

Thanks for the screen caps!

Out of curiosity, are you using ethernet or wifi on this MBP?

 

[vddrnnr]

Hi B S,

I'm using wifi and now that you mention it if the airport card is
being pushed ( large transfer for example ) it's temps also rise considerably.

In the end the conclusion i take from all of this is that in a laptop
heat is a real problem and Apple has done a very good job at trying
to eliminate / reduce it and it was/is an ongoing process.
If you look at the fans from a white macbook for example ( the last versions circa 2009 ) the only
fan in the system got bigger, had more blades and the "ring" was also a lot widder.

The lenovo fan is also a goot example as it is way smaller in size but the number of
blades and the "ring" is also huge in proportion to its size.

Besides fan tweaking I've also started to look at other software that may help keep
the temps down.
I found 2 applications that can also help with this: PStateMenu and PStateChanger.
These are apps used in "hacks" in the early days to achieve working powemanagement
for "non-compliant" Intel/AMD CPUs.
Both of them allow you to change the way CPU speed ramps up/down with CPU usage
and so try to keep CPU speed as low as possible when doing "lighter tasks" like for instance
writing this post ;-) ( if the CPU did what it wanted the speed would be max :-( ) this
way I can just run at the lowest speed possible without problems and so keep temps down.

PS. this should also help when running on battery

PS2. How hot does your CPU get for example loading gmail or some other CPU intensive site ( max temp )?

Best regards,
voidRunner