Satellite P300 Thermal Solution

[Moesly_Armlis]

Satellite Thermal Solution

This applies to my Toshiba Satellite Pro P300 laptop that has thermal issues.  The Thermal interface material became dried up; the laptop is over eight years old.  New thermal pads have been ordered but wanted to try this trick using copper (also getting impatient it has been over four weeks since placing the order also my kerbals informed me that they miss me ) .  The idea is to replace the thermal pads with thermal grease and copper wafers. 

The spacing between each chip and heatsink was measured using two sided tape.  Multiple layers were placed on top of the CPU.  When the tape adhered to both the heatsink and chip the thickness was measured using calipers.  The three measurements taken are 0.25 mm for the CPU, 0.60 mm for the GPU and 1.00 mm for the NB.

A copper pipe was cut and hammered flat and used to fabricate three wafers that were relatively uniform to the required thicknesses (0.21 to 0.24 mm, 0.56 to 0.61 mm and 0.96 to 1.02).  A dry fit was done first and without having any resistance to assembly.  Thermal grease was applied and each copper wafer was pressed and smeared into the chip to remove air gaps. 

The heatsink assembly was secured and now waiting for further assembly and to boot up the machine.

 

Album

 

 

 

[LordFerret]

Holy cow... you went wild with it. I would have waited for the right material. Hope it works.

I've had to deal with Toshiba laptops overheating ... two of them. Both of them, the problem was solved by disabling the keyboard illumination feature (keys light up in the dark). Don't know if that's the same problem you were having?

[Elthy]

Its important to get the surface of those copper panels as flat as possible. Not sure how you could improve it, though...

[Moesly_Armlis]

The issue was the laptop would shutdown after an hour the fan would ramp up and then the computer would shut down.  Unclogging the radiator was done first and the computer would boot and remain active for more than an hour but playing KSP would cause the same problems, the fan ramps up followed by shutdown.  After removing the heatsink I discovered dried thermal interface material between all the chips.

3 hours ago, Elthy said:

Its important to get the surface of those copper panels as flat as possible. Not sure how you could improve it, though...

I agree and the fabrication method was poor at best.  Although hammering has given the copper a smooth surface and the thermal grease should bridge the surfaces.   I will be doing the final assembly and firing up the computer in a few hours and I will post the results.

After this experiment I will have to raise my redsmithing skill by fabricating some daggers or just wait for the T777 Phase-Change Thermal Interface Pads.

[Flymetothemun]

15 hours ago, LordFerret said:

Holy cow... you went wild with it. I would have waited for the right material. Hope it works.

I've had to deal with Toshiba laptops overheating ... two of them. Both of them, the problem was solved by disabling the keyboard illumination feature (keys light up in the dark). Don't know if that's the same problem you were having?

My toshiba lappy's overheated a few times. It's always the exhaust, it gets clogged up with dust and detritus and then I have to open the sucker up, unscrew the metal top on the fan, and clean off the fan blades and the heat pipe radiator. But judging by how hot the thing gets sometimes the thermal compound and heat pipe contact is alright.

And Moesly, I've done stupid stuff like this. Before the right side cut out completely (dropped from 6 feet onto hard floor), I was keeping the right speaker of my headphones going by bending the wire to where it would sound and I superglued it to the side. But I hope you took a wire brush and/or sandpaper to your copper plates. And for goodness sakes, next time you do this don't use a little pair of wire cutters to cut your copper, Go down to your local hardware store, get a nice pair of channel lock dikes and a bottle of 3 in 1 oil. Use the dikes to cut your copper and keep all the metal surfaces oiled (you might as well stay stupid if you oil the rubber grips). Don't be stupid and try to cut bolts or heavy wire with it, if you take care of a pair of channellock dikes (or any channelock tool really) you'll give them to your grandkids. You could use cooking oil but I like the 3 in 1. Cooking oil's too heavy for my taste, I don't feel confident that the oil's everywhere in the cracks where I can't slather the oil on with my fingers. Whatever you do, make sure the oil's light, don't use motor oil or any slow moving oil. If you don't keep it oiled, it'll rust and your tool will be close to junk.

Edited December 6, 2015 by Flymetothemun

[Moesly_Armlis]

The laptop is back together and operating.  After 20 minutes of idle the heatsink fan RPM did not increase.  I loaded up and played a video and after 1/2 hour the fan RPM increased slightly.  The video is 1 1/2 hours and at the end of it the fan RPM has increased slightly.  The temperature around the heatsink is 31 degrees Celsius in a room temperature of 26 degrees Celsius.  The fan RPM has never peaked.  Now I will let it cool and then recycle.  Then it might be go time for a more intense burn and load up KSP and craft some insane part count vessel.

 

@Flymetothemun The initial cutting I used pipe cutter and then yellow handle aviation snips and for the 0.60 mm and 0.25 mm I used heavy scissors.  The red electrical snips I used to tap into the heatsink fan and powered it with a five volt USB charger.  I wanted to run tests on the heat pipes to ensure they did not go sour and were able to draw the heat.  I used my fingers and after running the fan I could feel them cooling.  Also used a heat gun and increased the temperature of the copper pad to 80 degrees Celsius and measured how quickly the temperature dropped.

 

The idea of sanding did cross my mind but the hammered copper was remarkably smooth.  There is copper foil available for multi layer PCB fabrication and is available in a variety of thickness and I might order some.  The copper foil compared to the PCTI pad does have the advantage that it would not dry up and much better thermal conductivity.  The only disadvantage is now it is susceptible to impacts.  The PCTI pads would offer some cushioning if the laptop is dropped.

I also am slightly worried about the different thermal co-efficient of expansion.  The copper was fabricated thinner than what was measured to allow for expansion.  When the thermals are maxed the expansion rates for the CPU wafers and copper are different and might fracture the wafers.

Tomorrow I will play some KSP using the laptop and will post the results.

 

I butchered the keyboard ribbon connector pressure retainer and will need to order another unless someone can suggest a fix.  I use a USB keyboard anyhow.

Edited December 7, 2015 by MoeslyArmlis

[LordFerret]

3 hours ago, MoeslyArmlis said:

keyboard ribbon connector pressure retainer

Most folks would likely jump to tell you 'superglue / crazyglue', but it tends not to work well with those plastics. My suggestion is JB-Weld (epoxy), that stuff is stronger than steel and bonds to almost anything. Use sparingly.

Flymetothemun, as I recall: The Toshiba laptops were prone to overheating (a certain model/series of them anyway). One of the contributors was the heat generated as I described (illuminated keyboard keys). There was another issue, but I forget now what it was. Anyway, Toshiba had issued a driver update to help alleviate it. If you do some searching on Toshiba's support site for 'yourModel' overheating, you'll likely find it.

[Moesly_Armlis]

11 hours ago, LordFerret said:

Most folks would likely jump to tell you 'superglue / crazyglue', but it tends not to work well with those plastics. My suggestion is JB-Weld (epoxy), that stuff is stronger than steel and bonds to almost anything. Use sparingly.

 

The JB-Weld is exactly my choice too.

 

The laptop is now running and have installed a monitoring tool called Open Hardware Monitor.  The temperature is around 60^o C when idle and after playing  four hours of video the temperature remained at  60^o C  and the fan RPM never increased speed.  When I started Steam to download KSP 1.0.5 the fan increased and increased some more. Heat was felt from the exhaust.  I thought all is good if the heat is being removed.  The fan speed kept going full speed and I thought this is just downloading and not even playing.  These were the signs of thermal overheat and I was anxious to see if it was going to fail.  The temperature was approaching 80^o C.   When the fan RPMs started to spool sown I felt a wave of relief.  The laptop went through a high demand thermal cycle and survived.

 

So onto playing KSP and for two hours I gave the Caveman challenge a go and monitored the temperatures.  The fan RPMs cycled and the temperature remained below 80^o C.  The fan RPMs are not maxed like before and now the RPM speed cycles I am a happy kerbal.  I will post an update again later in the week.

 

P.S.

I just realized that I am still using the laptop and have submitted photos to imgur while posting all this.  My laptop runs much better than my olde workstation.

Edited December 7, 2015 by MoeslyArmlis

[Guest]

19 hours ago, MoeslyArmlis said:

The laptop is back together and operating.  After 20 minutes of idle the heatsink fan RPM did not increase.  I loaded up and played a video and after 1/2 hour the fan RPM increased slightly.  The video is 1 1/2 hours and at the end of it the fan RPM has increased slightly.  The temperature around the heatsink is 31 degrees Celsius in a room temperature of 26 degrees Celsius.  The fan RPM has never peaked.  Now I will let it cool and then recycle.  Then it might be go time for a more intense burn and load up KSP and craft some insane part count vessel.

 

@Flymetothemun The initial cutting I used pipe cutter and then yellow handle aviation snips and for the 0.60 mm and 0.25 mm I used heavy scissors.  The red electrical snips I used to tap into the heatsink fan and powered it with a five volt USB charger.  I wanted to run tests on the heat pipes to ensure they did not go sour and were able to draw the heat.  I used my fingers and after running the fan I could feel them cooling.  Also used a heat gun and increased the temperature of the copper pad to 80 degrees Celsius and measured how quickly the temperature dropped.

 

The idea of sanding did cross my mind but the hammered copper was remarkably smooth.  There is copper foil available for multi layer PCB fabrication and is available in a variety of thickness and I might order some.  The copper foil compared to the PCTI pad does have the advantage that it would not dry up and much better thermal conductivity.  The only disadvantage is now it is susceptible to impacts.  The PCTI pads would offer some cushioning if the laptop is dropped.

I also am slightly worried about the different thermal co-efficient of expansion.  The copper was fabricated thinner than what was measured to allow for expansion.  When the thermals are maxed the expansion rates for the CPU wafers and copper are different and might fracture the wafers.

Tomorrow I will play some KSP using the laptop and will post the results.

 

I butchered the keyboard ribbon connector pressure retainer and will need to order another unless someone can suggest a fix.  I use a USB keyboard anyhow.

Are you up to soldering by hand?

[Moesly_Armlis]

17 hours ago, CliftonM said:

Are you up to soldering by hand?

Soldering the ribbon cable would prevent access to the main board.   The aluminium plate in the picture covers the mainboard and the keyboard fastens above it.  Currently I just plug in a USB keyboard so this is not a critical component.

 

The laptop is still running fine. When playing KSP I did not want to take any chances so the laptop was moved in front of the open window to make use of the chilly December air.

[Guest]

14 minutes ago, MoeslyArmlis said:

Soldering the ribbon cable would prevent access to the main board.   The aluminium plate in the picture covers the mainboard and the keyboard fastens above it.  Currently I just plug in a USB keyboard so this is not a critical component.

 

The laptop is still running fine. When playing KSP I did not want to take any chances so the laptop was moved in front of the open window to make use of the chilly December air.

Solder a header onto it.

Or you can add a new one of the exact same type.

[Moesly_Armlis]

Just now, CliftonM said:

Solder a header onto it.

Or you can add a new one of the exact same type.

Soldering is not something I want to do just yet.  Going to try to identify the part manufacture and order one.  Maybe swapping out the retainer bar might work.

[Guest]

2 hours ago, MoeslyArmlis said:

Soldering is not something I want to do just yet.  Going to try to identify the part manufacture and order one.  Maybe swapping out the retainer bar might work.

It probably would.

[Flymetothemun]

On 12/7/2015, 12:35:42, LordFerret said:

Most folks would likely jump to tell you 'superglue / crazyglue', but it tends not to work well with those plastics. My suggestion is JB-Weld (epoxy), that stuff is stronger than steel and bonds to almost anything. Use sparingly.

Flymetothemun, as I recall: The Toshiba laptops were prone to overheating (a certain model/series of them anyway). One of the contributors was the heat generated as I described (illuminated keyboard keys). There was another issue, but I forget now what it was. Anyway, Toshiba had issued a driver update to help alleviate it. If you do some searching on Toshiba's support site for 'yourModel' overheating, you'll likely find it.

My machine doesn't have backlit keys. Although I'll have to see what the new keyboard does; the old one finally gave up the ghost after 5 years and a new one's on order; I'm just using an old USB keyboard I had laying around, although if my machine doesn't have backlit then it probably doesn't have the capability in the ribbon connector for backlit. I dunno. And I am certain that it is the fan getting clogged; it was overheating just a couple weeks ago and I opened it up, opened up the fan, and there was a 1/4 inch carpet of built-up dust blocking the air from going through the radiator and over the heat pipe. I usually have to clean the thing out once every 6 months, but every time I do it runs as cool as it ran when it was new.

Edited December 9, 2015 by Flymetothemun

[Nuke]

i would have honed the copper to a mirror finish (at least the areas that serve as a thermal interface). keep using finer grit sand paper until shiny.

Edited December 12, 2015 by Nuke

[Moesly_Armlis]

On 12/12/2015, 4:31:53, Nuke said:

i would have honed the copper to a mirror finish (at least the areas that serve as a thermal interface). keep using finer grit sand paper until shiny.

IPounding lightly and repeatedly resulted in very smooth finish.  Here is a close up of a piece I flattened to 0.12 mm.

The results were in my opinion satisfactory and I think sanding it would have put in scratches. If I were to use copper again I would buy some copper foil that is used for PCB manufacture.  This was intended as a quick fix until the ordered PCM arrived.

 

Update:

The laptop is still going and it has had several high temperatures moments.  The fan RPM  was maxed (temperature measured from OHM was 88^o C)and the laptop cooled down and the fan RPM reduced.  I am still hesitant to push the limits.  The laptop sits in front of the window.  Made some ports for it because winter is coming.