Related
Warning: Attempt this modification at your own discretion. I am not responsible for any damages which may occur.
After attempting to OC my phone, I noticed hot spots when stressing my CPU at 1.4ghz. Looking at the anatomy of my phone (http://www.ifixit.com/Teardown/Samsung-Galaxy-S-I-9000-Teardown/7122/1), I noticed the structural mid-plate is a metallic alloy. This modification tries to cool the CPU and other heat generating ICs by using the phone's mid-plate and EMI shielding as a heatsink.
Test:
CPU stressed for 10 Min at 1.4ghz starting from a minimum cooled state. Application: “Stability Test”
Before Mod: 53oC (127.5oF) at 7 min 48 sec (I cut the test short as I did not want to overheat my phone)
After Mod: 43oC (109.5oF) at 10 min 10 sec
Procedure:
1. Teardown your phone as per the ifixit.com site and extract the mainboard.
2. Remove the simcard & MicroSD EMI Shield
3. Apply thermal compound to all the raised ICs. This allows the EMI shield to act as a small heatsink.
4. Reinstall EMI Shield and press down firmly to allow proper contact with the thermal compound
5. On the other side of the mainboard, liberally apply thermal compound onto the exposed ICs as per the pictures. Note: There is a relatively large gap between the surface of the ICs and the mid-plate.
6. Reinstall mainboard and press firmly against the mid-plate. If the thermal compound is making contact with the midplate then there should be some slight resistance when trying to lift the mainboard. If there is no resistance, then you will need to apply more thermal compound.
7. Install remaining components and close up your phone.
I used Arctic Silver 5 as I had some laying around. It is probably better to use a non metallic thermal compound, but I have yet to experience any problems with electric components.
My phone has been modded for about 2 weeks, running @ 1.5ghz and I have yet to experience any ill effects with operating system or with the OLED screen.
Though this mod is reversible, it may be messy cleaning up the thermal compound from your phone.
Happy Modding
interesting idea...
although it sounds like your just trying to squish as much pb&j between the plate and the ic's.
whitewindhtc said:
I used Arctic Silver 5 as I had some laying around. It is probably better to use a non metallic thermal compound, but I have yet to experience any problems with electric components.
Click to expand...
Click to collapse
indeed, i've never had trouble with AS5 but would probably go with a non metallic compound for this.
PS wow, that's a LOT of AS5... must have cost you $5 just in compound just to do that
Either way, conduction is almost always better than convection in terms of cooling a heat source. It might be a little messy, but the mod is quite effective.
The amount of thermal compound is quite a bit, but I wanted to ensure that I made contact with the 'heatsink'. There are many cheaper thermal compound alternatives than AS5, but I ran out of the generic white stuff.
very good idea.. thanks for sharing. do you think using thermal pad in this case would be better solution than thermal compound? or do think that thermal pad. won't work?
Sent from my Nexus S using xda premium
Although thermal pads would be cleaner, they are also more difficult to use in this application due to the variation in gap spacing between the component and heatsink.
Sent from my GT-I9000 using XDA
Wanted to fix the GPS antenna connector, ended doing this aswell, although just RAM and CPU. The amount of thermal grease is really enourmous for a proper contact. To anyone who will do this aswell, take your time, its not hard, just be patient
wow, great!
Btw, in which kernel you use OC 1.5ghz?
NDeaz said:
wow, great!
Btw, in which kernel you use OC 1.5ghz?
Click to expand...
Click to collapse
Devil kernel can do that
Sent from my GT-I9000 using xda app-developers app
1.5ghz it's 1400mhz*107%liveoc? or what?
P.S. interesting to see benchmark results on 1.5ghz
Thank you for the idea. Just done that with thermal pads since it's more clean.
jorgemiguel4 said:
Thank you for the idea. Just done that with thermal pads since it's more clean.
Click to expand...
Click to collapse
Where do you get that stuff? I couldnt find any seller in my country. Ebay/amazon?
Sent from my GT-I9000 using xda app-developers app
R4DK0 said:
Where do you get that stuff? I couldnt find any seller in my country. Ebay/amazon?
Sent from my GT-I9000 using xda app-developers app
Click to expand...
Click to collapse
Ebay
How are the temps with the termal pads? Do you notice any improvement?
Does the phone still heat up when oc and playing games?
Sent from my GT-I9000 using xda app-developers app
darkknigh_t said:
How are the temps with the termal pads? Do you notice any improvement?
Does the phone still heat up when oc and playing games?
Sent from my GT-I9000 using xda app-developers app
Click to expand...
Click to collapse
I havent't measured it with any app, but noticed that the heat is more spreaded over the phone after some time playing a game... That's a good thing, I guess.
Looks like a good idea. Getting the most out of this little device always helps.
Do you guys get stability over 1.2-1.3GHz? Achieved with "oc" and/or "live oc"? After applying this mod, this thermal compound, the phone definately heats up more evenly.
Sent from my GT-I9000 using xda app-developers app
Thanks for the idea.. i'll try this one..
better not damage my SGS . (i hope :angel
jorgemiguel4 said:
I havent't measured it with any app, but noticed that the heat is more spreaded over the phone after some time playing a game... That's a good thing, I guess.
Click to expand...
Click to collapse
How thick pad do you recommend? Would 0.5 mm be too much?
elektroda said:
How thick pad do you recommend? Would 0.5 mm be too much?
Click to expand...
Click to collapse
I have used 1mm thick on mine.
how this enormous OC affected battery life? will it last through the day?
UPDATE:
I wish I'd done some temperature benchmarking beforehand, but instead sourced the community to help out. Check out the various results on XDA... there are also results from two TF700 owners (in Europe) who were gracious enough to help me out.
http://forum.xda-developers.com/showthread.php?p=28051518
First, it seems clear that my mod really didn't mitigate the heat issue enough... so don't bother.
Second, comparing the numbers between other A700's and the TF700... well, I'll let you draw your own conclusion on that one. :-(
---
Hi all,
As many of you have experienced, the A700 can get very hot under certain conditions. Mine got extremely hot to the touch after just an hour of gaming (Heavy Gunner, NFL Flick QB, and Cut the Rope), and crashed 2 or 3 times.
Well I decided to do something about it. Reading through the Service Guide (thank you paugustin!!!) & A700 teardown photos from another site, I came to some conclusions. First, from the looks of things, the back panel has some kind of metal plate lining, and 3 "pads" that closed the gap between the metal plate and several surfaces of the mainboard. My hope was that these pads were not simply foam pads but were actual thermal pads. I guessed that a bit of thermal paste might help the efficiency of those thermal pads... and if I got really ambitious, I could replace the thermal pads with larger ones (and paste those too).
Service Guide:
http://forum.xda-developers.com/showthread.php?t=1716922
I followed the instructions in the Service Guide & opened up my A700 earlier tonight. I'm pleased to report that as long as you take your time, it's super easy to open up.
And lo and behold, my conclusions were accurate... the back plate is copper and the pads were thermal pads!
Back Plate:
https://www.dropbox.com/s/um556g6vtl7y99e/C360_2012-06-26-20-14-30.jpg
MainBoard:
https://www.dropbox.com/s/jka8gooiyqn9r2d/C360_2012-06-26-20-14-46.jpg
Closer Look @ the Thermal Pads:
https://www.dropbox.com/s/0j4e7wjfodwryk9/C360_2012-06-26-20-15-06.jpg
I took some CPU thermal paste I had lying around & dabbed some on:
https://www.dropbox.com/s/kvv99k3ghx73a00/C360_2012-06-26-20-17-24.jpg
Afterwards, I slapped everything back together, fired up NFL Flick QB and played for a half hour. I'm pleased to report that the A700 got warm but nowhere near as hot as it did in my prior gaming session. The "level of warmth" was what I'd consider mildly warm, comparable to my fiancee's iPad3, so definitely within reasonable tolerances.
I still intend to put the A700 through some more intense gaming tomorrow, but I wanted to write this post up and share it with everyone first. Hope folks find this useful and insightful!
Nice discovery, but yikes! Thermal grease is meant to go on in a translucently thin layer or it ends up having the opposite effect! If you put that much between a CPU and it's cooler, you'd kill the CPU! :S
superawesome!
Yet still i'd like to wait for more people to evaluate the hardwaremodding before i go in there myself.
Still, kudos to you. :good:
FloatingFatMan said:
Nice discovery, but yikes! Thermal grease is meant to go on in a translucently thin layer or it ends up having the opposite effect! If you put that much between a CPU and it's cooler, you'd kill the CPU! :S
Click to expand...
Click to collapse
Except in this case, the thermal grease isn't going up directly against the CPU... it's going up against a thermal pad & large metal plate that is covering the entire mainboard.
Additionally, unlike a heatsink & CPU which are fitted together tightly, the back panel of the A700 & thermal pads do not contact anywhere near as tight to the mainboard plate. I fitted the two halves together, then pulled them apart again, to see how much the paste really spread out. This way I could better gauge how tight the contact was & remove any excess that got squished out. However, not much paste really did, telling me that the compression between the two isn't that great. That is why I'm toying with the idea of replacing the thermal pads entirely.
Hey, good stuff. Keep us in the loop on this....
Thicker thermal pads might be in order, or a better paste job under the plate on the CPU/GPU...
Bigger pads to spread the heat out more since the back looks to be one copper heat sink..
http://www.frozencpu.com/cat/l2/g8/c487/list/p1/Thermal-Thermal_Pads_Tape.html
Beknatok said:
Hi all,
As many of you have experienced, the A700 can get very hot under certain conditions. Mine got extremely hot to the touch after just an hour of gaming (Heavy Gunner, NFL Flick QB, and Cut the Rope), and crashed 2 or 3 times.
Well I decided to do something about it. Reading through the Service Guide (thank you paugustin!!!) & A700 teardown photos from another site, I came to some conclusions. First, from the looks of things, the back panel has some kind of metal plate lining, and 3 "pads" that closed the gap between the metal plate and several surfaces of the mainboard. My hope was that these pads were not simply foam pads but were actual thermal pads. I guessed that a bit of thermal paste might help the efficiency of those thermal pads... and if I got really ambitious, I could replace the thermal pads with larger ones (and paste those too).
Service Guide:
http://forum.xda-developers.com/showthread.php?t=1716922
I followed the instructions in the Service Guide & opened up my A700 earlier tonight. I'm pleased to report that as long as you take your time, it's super easy to open up.
And lo and behold, my conclusions were accurate... the back plate is copper and the pads were thermal pads!
Back Plate:
https://www.dropbox.com/s/um556g6vtl7y99e/C360_2012-06-26-20-14-30.jpg
MainBoard:
https://www.dropbox.com/s/jka8gooiyqn9r2d/C360_2012-06-26-20-14-46.jpg
Closer Look @ the Thermal Pads:
https://www.dropbox.com/s/0j4e7wjfodwryk9/C360_2012-06-26-20-15-06.jpg
I took some CPU thermal paste I had lying around & dabbed some on:
https://www.dropbox.com/s/kvv99k3ghx73a00/C360_2012-06-26-20-17-24.jpg
Afterwards, I slapped everything back together, fired up NFL Flick QB and played for a half hour. I'm pleased to report that the A700 got warm but nowhere near as hot as it did in my prior gaming session. The "level of warmth" was what I'd consider mildly warm, comparable to my fiancee's iPad3, so definitely within reasonable tolerances.
I still intend to put the A700 through some more intense gaming tomorrow, but I wanted to write this post up and share it with everyone first. Hope folks find this useful and insightful!
Click to expand...
Click to collapse
Fantastic for the XDA community, but the average buyer will be like the cheesey pop group Power Station and will "Feel the heat".
rushless said:
Fantastic for the XDA community, but the average buyer will be like the cheesey pop group Power Station and will "Feel the heat".
Click to expand...
Click to collapse
Then be happy that you're amongst the enlightened!
CAUTION!!!!!
OK maybe that's a bit strong, but as an engineer I have to question the logic employed.
First of all, the heat generated by the CPU, battery, etc. must exit the device. ALL OF IT!
Temperature of the back plate is determined by two factors alone: thermal resistance (e.g. insulation, heat pipes, air flow, fins, etc.) and the amount of heat transfer. I'm serious!
It appears that you have done nothing to affect the thermal resistance between the back plate and the air (e.g. adding cooling fins or increasing air flow). Therefore if your back plate temperature has decreased (and the air temp stayed the same), then we would conclude that heat transfer between the back plate and the air has DECREASED.
If hea
If heat transfer from the plate to the air has decreased, then that means the heat transfer from the CPU etc. to the plate must have decreased as well!
My guess is that your thermal paste is producing a higher thermal resistance, as predicted by FFM! And it's deceptive because it's insulating the back plate from the CPU etc.
So where's the heat going? (I hear myself asking... myself.)
My guess is that it's dissipating through the tablet, getting distributed and absorbed and ultimately emitted more uniformly from the device. This would raise the internal temperature of the device, including the CPU! Hence the CAUTION!!!!!
Your idea could have some benefit, though, and here's how i see it playing out:
If you decrease the thermal resistance between the CPU and the back plate, then the CPU would be closer to the temperature of the back plate (i.e. lower). But the back plate temperature won't really change because you still have the same total heat coming through and the same thermal resistance with the air.
As for spreading the temperature across the back plate more, I'm not sure you can do it more effectively than the copper plate that's already there. Not without insulating the most direct heat transfer path!
yeoldeusrename said:
CAUTION!!!!!
OK maybe that's a bit strong, but as an engineer I have to question the logic employed.
First of all, the heat generated by the CPU, battery, etc. must exit the device. ALL OF IT!
Temperature of the back plate is determined by two factors alone: thermal resistance (e.g. insulation, heat pipes, air flow, fins, etc.) and the amount of heat transfer. I'm serious!
It appears that you have done nothing to affect the thermal resistance between the back plate and the air (e.g. adding cooling fins or increasing air flow). Therefore if your back plate temperature has decreased (and the air temp stayed the same), then we would conclude that heat transfer between the back plate and the air has DECREASED.
Click to expand...
Click to collapse
Right, I'm aware of this.
The exact problem I was attempting to mitigate, is the transfer of heat between the mainboard cover plate and the back panel copper plate. I was theorizing that the thermal pads being used were not fitted well, thus not functioning optimally as designed to begin with. As a result, not enough heat was being transfered to the back panel copper plate for proper dissipation.
So either heat is being transferred to the back copper plate more efficiently now and dissipating more effectively... or the opposite as you proposed.
However, if heat transfer has decreased, wouldn't heat buildup would still occur within the small area between the back panel & mainboard? While it wouldn't be absorbing the same amount of heat directly from the thermal pad, heat would still build up over time and that should still be noticable no matter what, right?
I'd also note that the BETTER solution would be to NOT combine thermal paste with the thermal pads, since as you point out, that can adversely affect thermal resistance.... but rather to replace the pads with better pads.
But I decided to be a guinea pig and try it out anyway.
Sorry I got distracted before posting my conclusion:
So my conclusion (prior to your latest comments) is that you can effectively lower the temperature of the CPU but not of the back plate. (unless you're insulating it from the CPU, which will spread the heat around and raise the temp of the internals!!!!).
I will consider your latest comments and repost!
The way I see it, the heat was not effectively transferred to the radiator. We have to remember that the case heat-up experienced by touch is not only the radiator, but the case itself heated by both the radiator and the air.
I'd assume that lack of crashes indicates smaller temperature of the CPU\GPU, as the heat is conducted more effectively by the paste to the radiator and dissipates evenly, instead of building up in one place and transferred to the case by air.
It's times like this, that highlights the one minor missing flaw of the A700... there's no bloody temperature sensors that we can poll!!!
*bashes head against desk*
So unfortunately, all findings are really subjective unless yeoldeusrename's conclusions are correct and mine are totally incorrect, and my A700 decides to fry itself.
Beknatok said:
It's times like this, that highlights the one minor missing flaw of the A700... there's no bloody temperature sensors that we can poll!!!
Click to expand...
Click to collapse
Well, i wouldn't be that sure. The A510 has them (at least separately for cpu and battery) and can be read by certain apps from the store. (E.g system tuner and battery monitor widget) so i would guess the A700 has them too.
Perhaps that helps
Sent from my A510 using xda app-developers app
mearoth said:
Well, i wouldn't be that sure. The A510 has them (at least separately for cpu and battery) and can be read by certain apps from the store. (E.g system tuner and battery monitor widget) so i would guess the A700 has them too.
Perhaps that helps
Sent from my A510 using xda app-developers app
Click to expand...
Click to collapse
The two apps I tried, didn't show anything for temperature, but I just grabbed system tuner & it does register cpu temp! Yay! Will explore the app further to figure out how to record history for accurate metrics. Thanks for the hint.
Sent from my A700 using Tapatalk 2
Beginning to wonder if the heat and touch issues are why the 700 is priced the exact same as the 510. Nothing else apparently makes sense.
Anyone who is tracking this thread & wouldn't mind helping gather some additional data, please check out this thread:
CPU Temp Benchmarking - Request for Help
http://forum.xda-developers.com/showthread.php?p=28009068
Thanks!!!
I think I'm going to look into creating a heat shield to place against the rear cover.
2 to 3 inches across and 4 down on the right side should do it.
This is a great tablet otherwise and don't want to give up on it just yet.
Besides the Asus sound system is inferior.
For those of you who have a Tronsmart Promethius media box and are having lockup issues, there is a fix!
Nothing an old South Bridge low profile heat sink and some MX4 thermal compound can't fix!
Take a look at the attached photos for some shots!
A few more shots!
Here are a few more shots of the applied heat sink.
(Edit I had to change the spring loaded pins to shorter ones because I could not get the Prometheus to go back together due to the old pegs being too tall.
Good post
SovereignKnight said:
Here are a few more shots of the applied heat sink.
(Edit I had to change the spring loaded pins to shorter ones because I could not get the Prometheus to go back together due to the old pegs being too tall.
Click to expand...
Click to collapse
Great little mod, I did this to my box and it's improved the stability no end.
Thanks for posting.
All my old PC motherboards laying around just got harvested
Several others also reporting success at: http://www.armtvtech.com/armtvtechforum/viewtopic.php?f=133&t=1340
I also used a scavenged heatsink from my parts pile.
Apparently the mounting pattern for the holes is 60mm and the max height is 8mm.
As usual a heatsink for any of these android sticks/boxes is a great mod!
deadhp1 said:
Several others also reporting success at: http://www.armtvtech.com/armtvtechforum/viewtopic.php?f=133&t=1340
I also used a scavenged heatsink from my parts pile.
Apparently the mounting pattern for the holes is 60mm and the max height is 8mm.
As usual a heatsink for any of these android sticks/boxes is a great mod!
Click to expand...
Click to collapse
I've taken my mod one step further! I've removed the small heat sink, cut a hole in the roof, used a much taller heat sink and added a fan and soldered it in.
Now I'm running at 1.3 GHz and stable as a rock! Antu scores 7500., and a Quadrent of 3300. I've attached the hardware photos.
SovereignKnight said:
I've taken my mod one step further! I've removed the small heat sink, cut a hole in the roof, used a much taller heat sink and added a fan and soldered it in.
Now I'm running at 1.3 GHz and stable as a rock! Antu scores 7500., and a Quadrent of 3300. I've attached the hardware photos.
Click to expand...
Click to collapse
It really looks like a spaceship now. Or maybe some sort of futuristic hovercraft.
Nice and clean looking!
deadhp1 said:
It really looks like a spaceship now. Or maybe some sort of futuristic hovercraft.
Nice and clean looking!
Click to expand...
Click to collapse
Thanks. I tried to make it as clean as possible but cutting a hole in the top of tough plastic was a challenge with the tools I had to work with lol.
The reason why I changed the cooling to active is I thought about it. The heat sink will cover the processor but also the dram chips and the flash chip. The dram and flash chip will now share the heat of the CPU. I don't think dram chips and flash chips are designed to function at temps as hot as this CPU gets. Even if clocked at 1.2GHz all day long. I'd think this is bad in the long run without cooling down everything. This new mod gives me peace of mind. The CPU gets WAY to damn hot even with the extended heat sink passively. With a fan everything runs super cool. I think the Chinese designers of this product really did a half baked job on the design and really didn't think this trough. That seems typical now days.
I changed the thermal paste in my H910 and OMG it makes a crazy huge difference and you can immediately notice a difference. The LG V20 clearly throttles because of the heat.
Before I changed my thermal paste I did some benchmarks using the Performance I/O. Here are the results:
BEFORE
Antutu: Run 1: 123000 Run 2: 119000.
Geekbench 4 Pro: Run 1: 1656/3918 Run 2: 1657/3997
AFTER changing my thermal paste to Arctic MX-4 here is what I got:
It seems like Geekbench score doesn't change much. I suspect that it doesn't push the device hard enough to make it throttle.
Antutu on the other hand... 152000pts on the first run and 143000pts on the second run! HOLY CRAP!
EDIT: Here are some proof of the difference it makes: https://www.reddit.com/r/lgv20/comments/8ffn4b/lets_collect_facts_on_cpu_throttling_report_in
Anymore freeze/ lag or slowdown?
popwar said:
Anymore freeze/ lag or slowdown?
Click to expand...
Click to collapse
There is still some occasional slowdowns but I think it's because of poorly designed apps like Facebook. Overall it does feel snappier.
It also feels less hot when doing more intensive stuff and multitasking. For the trouble I say it was worth it.
It was pretty easy to do. It's almost like changing thermal paste in a laptop. It took about 20 minutes to do and that was because I was being careful.
Here is another Antutu that I did when the CPU was at it's coldest.
XblackdemonX said:
There is still some occasional slowdowns but I think it's because of poorly designed apps like Facebook. Overall it does feel snappier.
It also feels less hot when doing more intensive stuff and multitasking. For the trouble I say it was worth it.
It was pretty easy to do. It's almost like changing thermal paste in a laptop. It took about 20 minutes to do and that was because I was being careful.
Click to expand...
Click to collapse
is it possible for you to write a guide? i'm curious to try on my H910
adityajiwap said:
is it possible for you to write a guide? i'm curious to try on my H910
Click to expand...
Click to collapse
Check out this video. Just replace the pink thermal paste which is under the motherboard.
XblackdemonX said:
Check out this video. Just replace the pink thermal paste which is under the motherboard.
https://www.youtube.com/watch?v=https://www.youtube.com/watch?v=KNgSDmkn96c
Click to expand...
Click to collapse
Dont you know if there is another guide on this? The video its down and i cant find anything related to that
XsaucerX said:
Dont you know if there is another guide on this? The video its down and i cant find anything related to that
Click to expand...
Click to collapse
https://www.youtube.com/watch?v=KNgSDmkn96c
XblackdemonX said:
Click to expand...
Click to collapse
The one in the back of the motherboard with the copper looking thing?
XsaucerX said:
The one in the back of the motherboard with the copper looking thing?
Click to expand...
Click to collapse
Yep. You replace the pink paste on that copper heatsink.
How's the battery usage after this?
Sent from my XT1635-02 using Tapatalk
Wow, tried this yesterday with good results. I still haven't noticed big slowdowns like I did before, and the temperatures don't even get close to the original. Smooth as hell :highfive:. I put thermal paste on a bigger surface than the original stuff was; I completely covered the copper piece and a bit around it with thermal paste.
Edit 2021: The device does run fairly hot after all, slowdowns is less of an issue but still applicable. A bit of an underclock helps a lot, like for example the Lighthouse rom does.
XblackdemonX said:
I changed the thermal paste in my H910 and OMG it makes a crazy huge difference and you can immediately notice a difference. The LG V20 clearly throttles because of the heat.
Before I changed my thermal paste I did some benchmarks using the Performance I/O. Here are the results:
BEFORE
Antutu: Run 1: 123000 Run 2: 119000.
Geekbench 4 Pro: Run 1: 1656/3918 Run 2: 1657/3997
AFTER changing my thermal paste to Arctic MX-4 here is what I got:
It seems like Geekbench score doesn't change much. I suspect that it doesn't push the device hard enough to make it throttle.
Antutu on the other hand... 152000pts on the first run and 143000pts on the second run! HOLY CRAP!
EDIT: Here are some proof of the difference it makes: https://www.reddit.com/r/lgv20/comments/8ffn4b/lets_collect_facts_on_cpu_throttling_report_in
Click to expand...
Click to collapse
What paste did you use?
x86cpu said:
What paste did you use?
Click to expand...
Click to collapse
They used Arctic MX-4
Migush said:
He used Arctic MX-4
Click to expand...
Click to collapse
Wow... after reading this reply, I see it in the 1st post...opps... Thank you.
Just picked up a V20 on Swappa. This is the first mod I plan on doing. Quick question though, is that a copper looking thermal pad or an actual copper shim?
If that's a thermal pad, I've always been told not to combine a pad and paste. Use either or, but not both together since it is a thermally conductive bridge. Wouldn't it be better to remove that pad entirely, then apply an actual copper shim with paste on both sides?
[/COLOR]
MathewCNichols said:
Just picked up a V20 on Swappa. This is the first mod I plan on doing. Quick question though, is that a copper looking thermal pad or an actual copper shim?
If that's a thermal pad, I've always been told not to combine a pad and paste. Use either or, but not both together since it is a thermally conductive bridge. Wouldn't it be better to remove that pad entirely, then apply an actual copper shim with paste on both sides?
Click to expand...
Click to collapse
That's an actual copper shim, the thermal compound its the pink thing over(and below touching the plate touching the actual chip) you can just use thermal paste over it and also replace the thermal paste that goes on top of the chip(just as bad as the one outside of the plate) inside the metal plate(it has the name of the soc on top)
How do you get behind the heat spreader plate that's connected to the PCB, to see the 820 die?
I'm betting it's soldered or permanently clipped to the motherboard.
MathewCNichols said:
How do you get behind the heat spreader plate that's connected to the PCB, to see the 820 die?
I'm betting it's soldered or permanently clipped to the motherboard.
Click to expand...
Click to collapse
I watche a teardown of the xiaomi redmi note 5 and the guy just poped it out with a nail, its just being holded by some little tabs i believe.
This is the video: go to 6:45 and you'll see what he did(in the redmi video he used his nail to pop it out, in this one he just used some twezers)
Thanks @XsaucerX. You've been a great help!
I didn't see anyone mention pulling off the heatspreader of the mobo on XDA or Reddit. This has got to help. Most laptops I've pulled apart had access to apply paste directly on the CPU die.
The factory heatsing had thermal paste.
I got a new one from seller dince i had overheat problems at summer and i used thermal paste as well.
But the px6 is not flat. It has holes and such.
So i am wondering.
Maybe a thermal pad would be better than thermal paste? If it is what thickness?
Paste has always been better with uneven surfaces.
I can understand it exactly the opposite way. Even surfaces go for paste, non even with pad.
GnFgr said:
I can understand it exactly the opposite way. Even surfaces go for paste, non even with pad.
Click to expand...
Click to collapse
You may be thinking of EPOXY thermal compound, which in my opinion is garbage.
On the other hand any silicone based grease is far better at filling gaps than pads are. Most factories tend to use pads though because it's cleaner and there is less maintenance to perform on the factory assembly machines.
Good read here:
Thermal Pads vs Thermal Paste: The Best Choice for Mounting Your Heatsinks
I used to work as a construction assistant, where my primary responsibility was to level properties prior to the heavy concrete trucks laying down a level foundation. It was a tedious and taxing job, but something that is quite crucial for a building in order for it to withstand years of use and...
resources.altium.com
I never said about epoxy based. Never.
The px6 is not only uneven but has a metal shield on top. The metal shield acts as an interference shield ( like a faraday cage maybe?) And has holes. So the surface in not only not even it is full of holes.
I am not asking about even and not, and you stand correct, in surfaces that are either even or almost paste in my mind is the way to go.
But when the gap is bigger that x what is better?
So today i cleaned the paste with isopropyl alcohol and used a good pad from gelid i think. It is rated as 15W/mK and is 0.5mm thickness.
With external temp of 5 degrees Celsius the cpu was at 78 to 80 degrees after an hour. That is with thermal paste.
With the thermal pad the temperature after an hour is 65 degrees.
The usage is exactly the same at the one hour usage.
Maybe the metallic shield has no contact with the cpu and the pad manages through the holes to have at least some contact and achieve better heat transfer?
Well... some kind of cooling is better than none, that is for sure, but heat sinks are supposed to make full and direct contact with the device and if it can't then I suggest you have mechanical issues to correct rather than thermodynamic ones.
Damn.
I am not asking what is better as a general concept. I am asking what is better for a specific use that there is a shield above the cpu and there is no direct contact.
Anyway mate forget it.
GnFgr said:
Damn.
I am not asking what is better as a general concept. I am asking what is better for a specific use that there is a shield above the cpu and there is no direct contact.
Anyway mate forget it.
Click to expand...
Click to collapse
I doubt you can find many to answer a question on mounting a heat sink in such an improper fashion. Not withstanding, you seem to be convinced that the pad is the right way to go, so I'm not sure why you are asking the question in the first place.
I asked because i wanted to learn what is better. Haven't test both when i asked and thought asking in a forum might find someone that had data on it.
Sorry if i overwhelmed you.
Anyway just forget this thread to the void.
Keep on