Why hasn't Linus tried soldering on an IHS?

[Julian5]

23 minutes ago, M.Yurizaki said:

On an aside, now that I had some time to stew what I said, you could probably just do it yourself by:

• Delid the CPU
• Put enough solder on the IHS where the CPU die is to hopefully make good contact
• Bake it in the oven at just over the melting point of the solder (so like what, 200C-220C?)

... and then realize the solder probably just spread out all over the place and now you have shorted out components.

He could do another collab with louis rossman? He seemed to have the stuff necessary for this kinda job.

[Rockstar83]

Just now, M.Yurizaki said:

Intel doesn't use "catshit" TIM. Reportedly they use Dow Corning TC-1996. The only review I've found of it is at https://www.hardwaresecrets.com/thermal-compound-roundup-august-2011/5/  and it shows it's a pretty decent performer.

 

Also some of Dow Corning's other stuff is pretty impressive. The National Renewable Energy Lab did some testing on various thermal compounds (https://www.nrel.gov/docs/fy08osti/42972.pdf). Dow Corning's product (TC-5022) is a much better performer than Arctic Silver 5, while costing a lot less per unit mass.

 

The only thing missing is the long term viability of thermal compounds. Considering that people here are still using Core 2's (which appear to be the earliest thermal grease was used) without reporting any real thermal issues, that seems to breed confidence that the TIM Intel chose lasts well after the processor becomes irrelevant to the mainstream.

I don't know man.  My personal experience delidding is way different than what should have been if this is true.  

[Rockstar83]

1 minute ago, M.Yurizaki said:

Are we talking about the IHS application? It has been reported that yes, Intel spaced the IHS too far from the die and so the thickness of the TIM was just a wee bit too much.

You could be exactly right there.  Maybe they should try to con intel out of a sample so they can test that theory.

 

[flibberdipper]

4 minutes ago, M.Yurizaki said:

Are we talking about the IHS application? It has been reported that yes, Intel spaced the IHS too far from the die and so the thickness of the TIM was just a wee bit too much.

Yeeeeep. 

[Mira Yurizaki]

Maybe Intel or AMD should just give enthusiasts what they really want: bare dies.

 

You should be pro enough to torque the screws to spec

[porina]

No reference, but I vaguely recall apparently it isn't tin based solder, but indium that's used. Lead is long gone outside rare niches due to environmental concerns. Indium apparently melts at 156C, a lot easier than 232C of tin. Tin-lead as commonly used in general electronics before the ban could go as low as 183C. 

 

Also I vaguely recall reading somewhere if you get the process wrong, the different cooling rates of the solder and die could cause thermal stress and break it. That might be a further contributing factor. In a manufacturing process they could control the temperature profile. To do it by hand is much more risky.

[DildorTheDecent]

10 minutes ago, porina said:

No reference, but I vaguely recall apparently it isn't tin based solder, but indium that's used. Lead is long gone outside rare niches due to environmental concerns. Indium apparently melts at 156C, a lot easier than 232C of tin. Tin-lead as commonly used in general electronics before the ban could go as low as 183C. 

 

Also I vaguely recall reading somewhere if you get the process wrong, the different cooling rates of the solder and die could cause thermal stress and break it. That might be a further contributing factor. In a manufacturing process they could control the temperature profile. To do it by hand is much more risky.

https://overclocking.guide/the-truth-about-cpu-soldering/

 

Based De8auer and his handy "guide".

[porina]

15 minutes ago, DildorTheDecent said:

https://overclocking.guide/the-truth-about-cpu-soldering/

 

Based De8auer and his handy "guide".

Looks kinda familiar, might have been that I saw before. Indium isn't as expensive as mentioned in the guide, at least the lump I got off ebay wasn't. Used it in part to make brew liquid metal so I'm not going to run out any time soon...

[Canada EH]

2 hours ago, Rockstar83 said:

liquid metal probably creates almost identical

the $1000 per ml liquid metal isnt as good as solder or other traditional electrical connection methods.

 

 

Whats IHS?

[Crunchy Dragon]

-Thread cleaned-

 

Please keep comments respectful and on-topic.

[AskTJ]

Why would he? It is over complicated, and delidding is cheaper, and simpler.

[TetraSky]

They need specialized equipment for that and I'd much rather for Linus to write titles that actually tells us what a video is about instead of the current ones that have little to do with the products they are looking at.

Seriously they are showcasing an Asus ROG phone and they make zero mention of Asus, ROG or even Phone in the title... but they effing put Razer for some reason. What is wrong with the person writing the video titles.

[AskTJ]

On 06/06/2018 at 12:10 PM, TetraSky said:

They need specialized equipment for that and I'd much rather for Linus to write titles that actually tells us what a video is about instead of the current ones that have little to do with the products they are looking at.

Seriously they are showcasing an Asus ROG phone and they make zero mention of Asus, ROG or even Phone in the title... but they effing put Razer for some reason. What is wrong with the person writing the video titles.

I highly agree. 

[Sniperfox47]

8 hours ago, M.Yurizaki said:

In order to melt lead based solder you need about 190C. Add about 20C at the worst for lead-free solder.

 

If you don't do it right, and it's likely very hard to do it right, you're going to outright fry the CPU.

Sadly it's nowhere near that simple. You need to pretreat the die and IHS and then use a specialized indium based solder, rather than your typical Tin solder. And even then doing it at home you'd be *very* likely to destroy the CPU with the thermal expansion of the pretreat and solder. I'd suggest giving Debauer's guide as mentioned by @DildorTheDecent a read. It's a good read and gives an idea of just why they moved away from soldering them.

 

6 hours ago, DildorTheDecent said:

https://overclocking.guide/the-truth-about-cpu-soldering/

 

Based De8auer and his handy "guide".

 

[JoostinOnline]

6 hours ago, porina said:

Looks kinda familiar, might have been that I saw before.

Did you read this? The diagram we used is very similar, although his article isn't talking about the same thing and it's far better written.  My post was simply about how much difference getting rid of (or minimizing) the amount of adhesive makes.

[ARikozuM]

7 hours ago, M.Yurizaki said:

Maybe Intel or AMD should just give enthusiasts what they really want: bare dies.

 

You should be pro enough to torque the screws to spec

I've been thinking about this for a good while. Why not give the K-series chips with unattached IHS's? Intel can supply each box with gasket maker (RTV or whatever else might work) and a basic list of TIM's that they have approved. They can then have a page where they ask for community feedback, with data to back up claims, for other third-party pastes that Intel can then test first-hand for qualification. This would make their enthusiast K platform marginally better than the respective non-K CPU. Those CPU's have to be pasted and adhered with approved products and carry a paid-for (by the consumer) warranty for overclocking and stock performance for a longer time by requiring a picture in a specific format where the TIM, adhesive, and IHS are shown for Intel to prove against should a warranty claim be needed. 

 

Is it more work for enthusiasts? Yes. 

Is it something we need? Not really.

But it would show that they do care for small segments of their community and enthusiasts aren't exactly smaller than 20% of their sales. 

[JoostinOnline]

23 minutes ago, ARikozuM said:

I've been thinking about this for a good while. Why not give the K-series chips with unattached IHS's?

Because you're far more likely to break it, and they'll either have to give no warranty, or they'll constantly have to replace them.

[dalekphalm]

11 hours ago, TetraSky said:

They need specialized equipment for that and I'd much rather for Linus to write titles that actually tells us what a video is about instead of the current ones that have little to do with the products they are looking at.

Seriously they are showcasing an Asus ROG phone and they make zero mention of Asus, ROG or even Phone in the title... but they effing put Razer for some reason. What is wrong with the person writing the video titles.

I agree.

 

What I especially hate is when they review or showcase a product, and don't mention the product name or model number in the title.

 

It'll be like:
"Is this the BEST GAMING MONITOR Ever???"

I don't even have a problem with that kind of title, if it was like this instead:

"Is this the BEST GAMING MONITOR EVER??? ROG Swift PG247Q Review"

 

Just, throw me a bone. I don't want to waste my time watching a review of a product I'm not interested in.