Removing redundant "slave" fan in Clevo laptops.

After a couple days of stress testing, I've concluded that my Clevo P870TM laptop would benefit from having the 2nd GPU fan removed from the system. I have a single GTX 1070 in my system and under full load, the fans bring my GPU core temp down to around 67°C. While the temps are great the noise emitted from the 3 system fans isn't.

I've noted that the noise of the system is GREATLY improved when removing the slave GPU fan from the system. However, when the redundant fan is removed, the motherboard freaks out and shuts the system down after a series of beeps. I believe this is known as a VR panic, and it's caused by the motherboard sensing a fan failure.

My proposed workaround is to construct a " fake fan RPM sensor" which simulates the "tach" signal from a fan. This should fool the motherboard into thinking a fan is present and spinning even when there is none.

The following schematics originate from the user " mohonri" on this thread. The updated diagram originates from user " ACalcutt" on this thread.
I've never built a sensor from raw electric components like this but it seems fairly simple. Any advice or thoughts if this would work as intended (before I go on a circuitry shopping spree haha ).

 

You could just split off the sensor cable from the first fan and feed it into both.

 

Hold on WHAT?!
What kind of BS is this?! Not even my P375SM-A is that poorly made, if there is no 2nd GPU the fan will not work, end of story. Whe the hell is your 2nd fan still running?!

 

The fans (particularly the 2nd GPU fan) are used to draw air over the internals so say you are running a hot set of M.2 drives they wont melt in summer. So it still gets used.

Do try and stop a second and question why rather than jumping to your conclusions.

 

Okay, I see what you're saying. I can see when the fan is removed there is a hole designed to pull air from the components under the keyboard and push it out the rear of the laptop.

Spoiler: Fan Profile

Spoiler: Fan removed

I have to ask, is it necessary for the 2nd GPU fan to ramp up in speed along with the 1st GPU fan. Why is it matching the 1st GPU fan's speed?...

 

It was designed for the really high end setups, hence that. However if you don't have M.2s in every slot then you should be fine tricking it.

 

I appreciate your advice. Your idea on splitting the sensor cable sounds much more simple and elegant. I'm purchasing a pack of 3 pin 1.25 mm JST connectors off amazon and I'll simply split the tach signal off to the slave.

Here's the connector in case others were wondering: https://www.amazon.com/Connectors-1-25MM-Female-Connector-Cables/dp/B073WPQLVX

 

Let us know how you get on

 

Make sure you test it thoroughly. Strongly recommend a graphing app that shows how the temps vary over time. I use OpenHardwareMonitor, but there are several others out there.

To each his own, but I would rather tweak the fan profile to make things quieter. I'm from India where it gets really hot in summer (room temperatures of 35C are common). That said, I use headphones while gaming, so fan noise doesn't bother me.

 

what not just get a bigger heatsink that uses both GPU fans? that way you can be like me have fan mostly stay on silent simply because the heatsink capacity is huge enough to passively dissipating heat.

 

Temps/noise on the default profile will likely not drop that much, as the difference between ambient and core temp so small it means diminishing returns.

 

I'd rather tweak the fan profile too, however, the fan profile seems quite limited in customization options.
This represents the next best option in my mind. Rembmer we're not talking about a small decrease in noise emissions, the laptop was dramatically quieter without the 2nd GPU fan.

I'm planning to use the following tools to stress the system:
CPU ID HWinfo - For monitoring temps and component usage
AIDA64 - Stress test CPU + FPU + cache + system memory
CrystalDiskmark - Stress test 970 Evo 500 GB
Unigine Superposition (4k optimized game mode) - Stress test single GTX 1070

Edit: I'm thinking of adding an EKWB heatsink to the M.2 drive I have under my keyboard to help with thermals. I understand generally the way to go is putting a thermal pad and let the fan do the rest. (adding more metal isn't always a good thing). But in this case(no pun intended), there won't be much airflow and it might work best that way. I'll also move my RAM out of that compartment so the only heat source will be that one M.2 drive.[/S]

 

What drive is it? Have you checked if it's getting hot?

 

I'm concerned that removing the 'extra' fan is making things DRAMATICALLY quieter because the air from the other fan now has an easy path, instead of being forced to go through the fins of the heatsink. (Note: I have a different model - in my case the second GPU fan is NOT redundant.)

Keep your eye on this thread:
http://forum.notebookreview.com/thr...-by-obsidian-pc.801464/page-148#post-10761630

The OBSIDIAN-PC guys are developing a Control Center replacement that will also have fan control. Still in early alpha, but very promising.

 

It is a Samsung 970 EVO 500GB
I've run several concurrent passes of crystal disk mark to stress test the drive. Max temps were 67°, idles around 50°

Thanks anytimer, I looked into that tool it looks amazing

Honestly, between the obsidian tool and RLECViewer, I may be abandoning my fan removal project.

Using RLECViewer I was able to set the VGA2 Fan speed to a low static value (like 20%). This means it's nearly silent but still allows some airflow which is really what I wanted.

I'm a bit nervous to keep using it though. Prema warned of the dangers associated with using the tool...

I may end up waiting for the Obsidian tool as it sounds much safer for long-term use.

 

I would not worry about that SSD, worst case use a pad on the controller or behind it to the PCB. Remember, don't cool the NAND.

 

So it's okay to put thermal pads on the controller and DRAM modules but not the two at the far left, correct?... Why is that?

 

NAND wears less and keeps data for longer if it runs warmer as the electron tunneling causes less damage to the less rigid structure of the silicon.

 

actually there is question regards to that. keeping it at high temp will keep nand cell alive, at higher temp what they suggested was like several hundred Celsius so thats not happening in laptop.

under our usage however, high temp is always an issue. how many GPU died due to high temp of poor vrm/memory solder materials. not saying samsung chips got poor solder, i'd rather have everything low than high.

 

https://www.anandtech.com/show/9248/the-truth-about-ssd-data-retention

The NAND is not getting as hot as an active chip like that. Certainly not in the solder danger area.

 

This is a snippet from the AnandTech website Meaker@Sager referenced.

Wow, my mind is blown that heat is actually beneficial in regards to the longevity of a device But correct me if I'm wrong, they were saying in the end, the higher endurance likely won't matter (for most users) on these new SSDs because they'll likely be long obsolete and replaced by the time they die.

 

that is not talking about keeping it several hundred degree celcius to re-nurture the life of nand flash.

https://www.tomshardware.com/news/flash-nand-dead-heat-heals,19491.html

keeping it high temp will causes other issue isnt something we should recommand in order to keep SSD alive. SSD usually dies not because of cell wear, but of controller, poor soldering, bad soldering material for flash/controller, or cracks in the circuitry pcb. When SSD dies of endurance because of wear level, it becomes read only and its still retain the data due to how firmware is written.

high temp will cause most of the described problem if a SSD is made poorly. as for retention of data, the SSDs nowadays with very advance firmware help towards it. TLC SSDs were to said only last 500 P/E cycle but nowadays they last a lot longer. no need to keep SSD temp high, best to keep it low and it'll likely last years, low temp also means no throttling in performance.

 

You keep the controller cool, you let the NAND stay warm. Hence pad the controller, leave the NAND alone.

That's why the heatspreader in the samsung SSDs copper label spread heat from the controller to the NAND. The NAND is no going to temperature throttle (it does not really produce much heat of its own) and the temperature range is not high enough to threaten the packaging/soldering.

I know that seems wrong to some people but it comes down to physics and an understanding of the power levels involved in what chip.

Keeping NAND as cool as possible is like keeping a CPU at 80C all the time regardless of load, it increases the wear speed on it and increases the chance of premature cell death.

 

the likeness of cell dying before any of the other factor i mentioned is wayyyy less likely. have had plenty of SSDs die, non of them are cell dying. you'll need to keep the temp low for the other factors to be safe, cell dying is the last thing we worry about. and yes, keeping it high temp allows only flash to last longer and kills everything else.