Why You Can't Measure Beneficial Effect OF Silver Greases

[Mikewarrior2]

TonyOU.

You not only posted this information once on your website, but then you posted it again after cautery made his thread. And you come to this bbs constantly saying how you got results with radio shack grease coming out on top.

Well, since you obviously belive you are indeed reading cpu diode temp, let's examine this scientifically.

Your cpu, at 600mhz @ 2.3 volts, being a .25 micron chip would put out roughly 40watts of heat. YOu were using an Alpha P3125 heatsink. YOu got 31C with Radio SHack Grease. That would ABSOLUTELY REQUIRE an ambient temp of 15C, at the very highest, and probably much lower because it is almost impossible to achieve a perfect connection between cpu and heatsink. That would mean you had an ambient temperature between 51-59F. What a crock. THe week before you published these "cpu temp measurements"(adn the week before) were ridiculously hot in Southern california. 100S + in F. And stop saying you were using some "special" testing case when you were running these tests. I have seen that case. You cooled your chip with CONVENTIONAL Cooling. Yeah, with FANS. THere is no way for FANS to drop the ambient case temperature below ambient temp. So, unless you were running your AC at 55F or lower, there is no way you would have gotten those results.

Were you mistakenly reading the mb temp? I find your constant posts supporting people who say RadioShack grease is better than silver grease appalling. I find the fact that you left your "tests" up on your website despite being told that you could have done something wrong is also appalling.

You did post a link to Cautery's thread, but I never saw anything saying that your scores were suspect. YOu never questioned your own results.


Mike

 

[tonyou]

Mikewarrior2, let me take my second and final post on my website regarding this thermal grease issue:


<< As you can see from the picture, I've got some numbers with these thermal interface material/compound/paste/grease after two rounds of testing.... The only problem is that my results are highly questionable (Radio Shack stuff beating the Artic Silver?) so I won't release them yet until I am sure they are valid. Sorry for the delay... >>



Obviously I never released anything on my website after this, which shows my test are &quot;invalid.&quot; If you had misunderstood my statement above then I am sorry for my bad English.

You are absolutely right about the CPU readings I got are scientifically impossible, Clay Autery said the same thing. But the point I am trying to make is that with thermal diode, all I am looking for is consistency/repeatibility not accuracy. johncar's very first post and a couple following posts talked about this,

<< You could try reading internal thermal diodes on Intel chips, as they are on the cpu side of interface. Diodes are not &quot;accurate&quot;, but they are &quot;repeatable&quot;...ie, tell same lies every day.
John C. >>



I am sorry also if you find my other posts in threads where people question Artic Silver's superiority appalling. I was only trying to help and point out what my mistakes were and perhaps they can figure out what's wrong with their tests.

the ignorant one,
Tony

 

[Mikewarrior2]

Then your scores are consistently wrong.

While internal diodes are not accurate, they do vary say +- 3degrees C, there is still no way that you could have gotten the temps you got. You were not measuring Internal diode. YOu were measuring temp of a thermistor. Bottom Line.



Mike

 

WOW this thread moves fast =D.
Yea If you want to get that exact about the measurement, then there would be a slight discrepancy.
becasue of the thickness of the chip and the silicon, etc between the grease and the core.
John, invent a small sensor that will safely fit between the HSF and the chip core.
That would be a nice little toy to really allow clockers to get their jollys.
But then again, how much can .001 degree of heat make on a processor?

And what about thermal paste?
How effective is that?
I need to toss some HSF's on my chipset and geforce.

 

[johncar]

Sammy Son,
Re inventing a sensor that fits btween hs and chip...can't be done without seriously compromising the interface. But both Intel and AMD
show thermocouple, TC, thru hole in hs, touching top center of chip.

Believe Intel says to glue/epoxy bead to chip...disagree, lump of epoxy would interfere with cooling, cause large hole to be drilled, and might even add add'l temp drop from case top thru epoxy to TC bead...not to mention threading TC thru the hole which should be as small as possible.

Amd says to glue TC bead in the hs hole with &quot;slight protrusion&quot;...
sure...so you make poor hs to chip contact.

We've done variation of AMD...but polish off the &quot;slight protrusion&quot; so bead is flush with hs surface. Use bare wire TC with .005&quot; wires and thin teflon insulation...omega.com...let's us install thru only very small .031&quot; hole for minimum effect on cooling. Of course this also does not measure case top/die top...need to add temp drop across arctic silver interface...max temp drop = max Watts power dissipation times C/W of Artic silver adjusted for case top/die top surface area.
That's the best we know how to do...but now can &quot;directly compare&quot; that adjusted measurement to mfr's Tcase top max/Tdie top max.
John C.

BTW...we can't test for effectiveness of interfaces cause sensor is on wrong side of interface...(many previous posts)...but don't care...we &quot;know&quot; silver grease/arctic silver are the best interface, from mfr's published conductivity stats. More important to us is knowing we're not exceeding mfr's safe operating temp limit with directly comparative measurement and only uncertainty being actual power dissipation...+- few deg C, can live with that.

 

[tonyou]

Mikewarrior2, may I remind you that I don't have any thermister on my system? The Soyo 6BA+IV I used had thermal diode monitoring!

Look, I appreciate the fact that you actually post your opinion on what I did with my tests and how &quot;wrong&quot; I was in this thread. I would have never known that my messages would mislead you or possibly others in this forum. I know what I said here has less weight than you simply because I am a dealer, my opinions are mostly likely biased in my favor (although not in this case, since I am considering selling Artic Silver too). Continue with your good work here at the forum, but please think twice before coming up with a conclusion and calling people names (ignorant, appalling,... etc). It hurts, especially when my intents were portrayed as devious acts and this coming from a person whom I really enjoy reading informative posts from.

By the way, I've got the chance to try out those Enermax PS that you recommended and they are great. We will definitely be carrying them.

Tony

 

[eia430]

Johncar you said

&quot;You've used a simple and sometimes valid statistical experimental test analysis in your thinking, but have ignored or were not aware of all the subtleties that occur in a multi-path heat transfer process, and which conequently invalidate that &quot;one variable&quot; change approach.&quot;

One variable that I was refering to and you understood was Silver grease. You just stated above that because of multi-path heat transfer process, that invalidates accounting for temp loss/gain due to the change in a single variable. So if your above statment is correct, you inferred the need to take into account (measure) all the variables (secondary pathways) Because measuring just &quot;one variable&quot; is invalidated by multi-path heat transfer process. These are your own words.
I can't beleive that I have to sit here and force feed you what you posted yourself. And since you insist on taking personal swipes at me I will do the same. When I see that I've made a mistake I owe up to it and say so. But I guess the perfect powerplant builder Johncar can't possibly make a mistake. Even when faced with direct quotes of your own words you are in denial of even the mere possibility that you contradicted yourself. Do yourself a big favor and read over your own past posts before you tell me that you did not contradict every point that I've made that the &quot;full&quot; benefit of silver grease can be quatified by measuring in 1 location. Then we settle the difference in accuracy between sensing temp on one side of the die case vs the other side.

It seems that there is a need to restate both sides of the debate now. You and Nevin manage to just jump between scattered points and personal attacks and have succesfuly muddled the debate itself. Let's restate with clarity your side and how it's impossible to measure the &quot;full&quot; benefits of silver grease and my side that it is not impossible.

 

[Mikewarrior2]

I don't see where JohnCar said it was impossible, rather, I did see where he states why there are cases where people see NO improvement with silver grease.


Mike

 

[eia430]

Sammyson I am just amazed at how well you manage to understand my points. Then you state them so clearly.

&quot;But then again, how much can .001 degree of heat make on a processor?&quot;

I wonder how long it will take Johncar to understand that the vast majority of us do not care about acuracy past reasonable. My personal limit is 1 degree F of resolution. I could care less about laboratory precision when it comes to making a decision on weather to buy this grease or that.

Give me your opinion on this one... what do you think would be the temp difference of the die case where the die chip itself is mounted on (non flip chip) between the side of the case where the die is fused to and the other direct other side of that sheet metal? Both having common heat pathways. The insulator directly below the chip being still air. Johncar in his illustrious wisdom seems to have this delusional idea that the heat difference will be to such a degree that it would not show &quot;full&quot; benefit of silver grease. Now my question to you is this, in your opinion do you think the difference will be to a degree that people will care? Or is the illustrious Johncar just being stubborn and is just refusing to concede the possibility of him being infallible?

 

[Mikewarrior2]

WHAT? Have you not seen any posts or reviews where SILVER GREASE shows no improvement over Silicone Grease? JohnCar made this post for those people/reviews where some person has said that silicon stuff either outperforms silver based greases or teh difference is nil.

For example, there was a thread about this a mere week ago. Some person claimed his Duron ran 3C hotter with AS than it did with Radio Shack stuff. And he isn't the only person on this bbs ot believe this.

JohnCar took the time to make his thread to help THESE people try to understand why they got the results they did. If you have nothing to gain from this, then don't post anything.


Mike

 

[eia430]

Mr. Warrior, it would help greatly if you addressed your post to a particular person. In the post above, which post/person are you addressing?

 

[Mikewarrior2]

It was made to you.. You are obviously unaware of the number of people not only on BBS's but that publish reviews on their websites where people do say that artic silver sucks, or it has no performance advantage compared to radio shack grease.

JohnCar made his post for these people. He never said that silver grease was ineffective, but he was offering a explanation to these people who get results where silver grease isn't as good as Radio Shack grease or only provides a 1 degree difference.


Mike

 

[eia430]

Mr Warrior To address your second to last post above. Where did Johncar say it was impossible? Read the whole post above yours, it explains it. To summerize, John car attributed the impossiblity of accounting for the &quot;full&quot; benefits because of the invaladation that multi-path heat transfers does on &quot;single&quot; variable change approache that he thinks I used. These are his words not mine. So to &quot;fully&quot; measure the benefits you would have to measure the delta t of all the heat pathways to account for their loss/gain effect on the single change. The quotes that I've used in the post above yours is found on page 1 of this forum thread, Johncar's posts middle to bottom of first page.