Yes to the second question. For the first, I can only tell you that real world results seem to indicate much better performance than the regular non-metal stuff.
-
Save 15% on ALL SFF Network merch, until Dec 31st! Use code SFF2024 at checkout. Click here!
S4 MINI Classic (S4M-C)
- Thread starter Josh | NFC
- Start date
Yes to the second question. For the first, I can only tell you that real world results seem to indicate much better performance than the regular non-metal stuff.
It's also corrosive and doesn't coagulate. But it does work. There are just solid reasons manufactures don't use it....
Depending on the liquid metal used, 5+ times the heat conductiveness of the best silicone based TIMs.
Tim = thermalpaste?
They are normally made of silicone? I thought silicone was an insulator
They are normally made of silicone? I thought silicone was an insulator
I believe they use a silicone matrix but I could be mistaken.
There are both silicone and non-silicone based TIMs. Liquid metal is Gallium Indium Tin (GaInSn).
I
Ionrent
Guest
quick question is it ok to reverse the connection for the sfc riser? v2.3 cant fit decided to make whats best for the gpu to mobo pics as reference thanks
https://drive.google.com/open?id=1EwZc_Eg6w8NPm5o6IjrteWl43LK5a652
https://drive.google.com/open?id=1EwZc_Eg6w8NPm5o6IjrteWl43LK5a652
Yup, those are solid GPU temp test methods! Thanks @cesk I'm gonna use either 65% 70% or 75% fan speed for my tests. I'll also be using furmark in addition to Unigine Heaven's extreme preset.For testing GPU temperature, you'll want a good methodology to get any meaningful data. For example, I've found excellent repeatability by doing the following:
First establish a baseline. Launch Unigine Heaven (extreme preset) and find a taxing scene by cycling through them with the enter key, then pause it with the space bar. Note the scene you paused at and use this for all future testing. This ensures a consistent load. We don't want the GPU boosting all over the place so let the GPU warm up and use MSI Afterburner to lock the voltage and frequency to your target values (Ctrl+F, click the desired point, and then press L). Write these values down and use them for all future tests. Set the GPU fan speed to a fixed value (you may want to test several fixed RPMs, but do just one at a time). Now let the GPU temperature reach steady state and record it. If you have a decent temperature sensor, also record the ambient, but don't sweat it if you don't. For best results, use the delta value for comparisons.
Now you can start investigating a test variable. That can be different fans, different thermal paste or liquid metal, lower voltage, etc. Pick one and repeat the above procedure.
The reason I went through all of that is because there's wildly inaccurate information online about the benefits of liquid metal, especially on GPUs. In reality, it's the reduction of the gap between the CPU die and heatspreader that contributes to most of the temperature drop when delidding -- not the liquid metal. What does this have to do with GPUs? Well, the bare GPU die is already right up against the GPU heatsink. You'll only see a large temperature drop with liquid metal if the stock thermal paste had poor coverage. In my tests with the 1080 Ti Mini, liquid metal only reduced temperatures by 3-4°C (all other variables held constant). By comparison, undervolting reduced temperatures by 9-10°C (all other variables held constant). There is also a risk of shorting if the liquid metal is applied incorrectly so careful prep work and masking of small board components is recommended.
With all that said, I think you should still go for it.
Don't do this! Always follow the instructions on the Riser and the little riser key!
Last edited:
Replied with urgency so no time to look for it - here you are https://smallformfactor.net/forum/threads/nfc-systems-s4-mini.96/page-192#post-62960 scroll to see what ultimately caused the problem, the riser plugged backwards
yeah, so double check your riser KEY and the riser itself.damn.. that's a close call thanks guys appreciate the quick responses
so does it look like this to you guys? im usung gigabytes mini, I dnt remember looking like this
https://drive.google.com/open?id=1Ds3xgwiRAvQku5yZUGE-9Aota36US1pq
front
https://drive.google.com/open?id=1ucXDIh2KKD8v5t8xrxDCUeiig59qqLwC
https://drive.google.com/open?id=1Ds3xgwiRAvQku5yZUGE-9Aota36US1pq
front
https://drive.google.com/open?id=1ucXDIh2KKD8v5t8xrxDCUeiig59qqLwC
looks fine to me bro. Just put some black gaffers tape over the top of it and it'll be like those prickly pin parts are invisible.
yes! managed to make it work no burn here thanks brother!
Yes, this is true, you have to remove everything at least if you want access to the GPU's fan header.
She fits like a glove! Now I need to clean up the Jags and paint it a little to cover the silver and put a vent plate and she will be good as new!
