Weird experience with Mini Ninja and Athlon 4850e
Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee
Weird experience with Mini Ninja and Athlon 4850e
I've seen posts noting that certain CPUs can be run fanless with the Mini Ninja heatsink, so wanted to report my experience, and see if it was consistent with that of others.
Just finished a new build using the Gigabyte MATX 780g chipset board, a Athlon 4850e, and Mini Ninja. All parts were new.
I first assembled and tested the system outside of the case (Antec 3480). Because I had previously tested the Minja on another build (same board, but with 5000+ BE CPU), I let it run in a relatively cold room without a fan on the HS. With the 5000+ BE, the Minja felt cool to the touch for a very long time. With the 4850e, I was shocked that the heatsink was very warm (even on upper fins) after just a few minutes. CPU temp in the BIOS read 55C. I was worried that I had improperly seated the HS, but (because I'm stupid!) chose to attach the fan before shutting down the system. After attaching the Scythe 80mm fan to the Minja, the temperature dropped rapidly, to 30-31C, within minutes, and the HS fans were cool to the touch.
Now running in the case, with only additional (non-PS) fan being a Nexus 120mm in the back of the case, and temps are still in the low 30s. I've never reseated the HS, and I'm pretty comfortable that I did that right. I don't have much experience with socket AM2 HSFs, but have built about a dozen systems and so generally comfortable with PCs. Applied a "pea" of AS5 to the center of the CPU before mounting.
Just finished a new build using the Gigabyte MATX 780g chipset board, a Athlon 4850e, and Mini Ninja. All parts were new.
I first assembled and tested the system outside of the case (Antec 3480). Because I had previously tested the Minja on another build (same board, but with 5000+ BE CPU), I let it run in a relatively cold room without a fan on the HS. With the 5000+ BE, the Minja felt cool to the touch for a very long time. With the 4850e, I was shocked that the heatsink was very warm (even on upper fins) after just a few minutes. CPU temp in the BIOS read 55C. I was worried that I had improperly seated the HS, but (because I'm stupid!) chose to attach the fan before shutting down the system. After attaching the Scythe 80mm fan to the Minja, the temperature dropped rapidly, to 30-31C, within minutes, and the HS fans were cool to the touch.
Now running in the case, with only additional (non-PS) fan being a Nexus 120mm in the back of the case, and temps are still in the low 30s. I've never reseated the HS, and I'm pretty comfortable that I did that right. I don't have much experience with socket AM2 HSFs, but have built about a dozen systems and so generally comfortable with PCs. Applied a "pea" of AS5 to the center of the CPU before mounting.
-
- Posts: 2049
- Joined: Thu Dec 15, 2005 11:06 am
- Location: Klamath Falls, OR
Re: Weird experience with Mini Ninja and Athlon 4850e
If the heat was getting to the upper fins quickly, then you definitely had the HS seated properly. Perhaps you could tell us how you did it? (There's a ton of posts here at SPCR from folk who did not seat it properly!)Piblokto wrote:With the 4850e, I was shocked that the heatsink was very warm (even on upper fins) after just a few minutes. I've never reseated the HS, and I'm pretty comfortable that I did that right.
Of course you're right! Still, the speed and warmth concerned me, especially since I did not have the same experience with the 5000+ BE.If the heat was getting to the upper fins quickly, then you definitely had the HS seated properly.
I see in the responses to some of those posts that folks think the trick might be to press down on both sides (levers) of the mounting mechanism at the same time, instead of first one side than the other. I can see how mounting the HS would be a pain inside of a case, especially if the case is on the small-side. Since I had such a case, I mounted the HS with the motherboard on a workbench, one side first and then using a flat-edge screwdriver to push the other side down. Took more force than I initially thought would be needed, but nothing excessive.Perhaps you could tell us how you did it? (There's a ton of posts here at SPCR from folk who did not seat it properly!)
-
- Posts: 2198
- Joined: Thu Feb 10, 2005 11:20 am
- Location: TN, USA
Your CPU temps do seem strange. My fleet of AMDs with stock heatsinks have never run over 40 during operation according to BIOS or the temp on the MicroFly case. They are all undervolted though. I am listening to music and surfing at 33.3C on a 3600 dual core now.
I bought a Q6600 on impulse for $179 and just put it together tonight with a mini ninja on an Asus P5E-VM HDMI, default 1.26 volts. I don't have an OS installed, but sitting here running for an hour in the temp screen on the BIOS it started at 28C and eventually topped out at 30C. I do have the scythe 80mm fan running though. When I shut down, the mini ninja was cool to the touch from top to bottom on 2 sides. I am currently running in a Antec Lanboy case with a 120MM right behind the mini and the 120MM Corsair PSU fan above it. I thought this pig would run hot, kind of surprised. The north bridge heatsink was a little warm however.
I checked the heatsink and the CPU spreader and they both looked flat enough, so I did my usual spreadding a thin layer of AS5 across on the CPU. I never liked their "pea" or "line" way of installing AS5, old school and stubborn I guess. DOH.
I bought a Q6600 on impulse for $179 and just put it together tonight with a mini ninja on an Asus P5E-VM HDMI, default 1.26 volts. I don't have an OS installed, but sitting here running for an hour in the temp screen on the BIOS it started at 28C and eventually topped out at 30C. I do have the scythe 80mm fan running though. When I shut down, the mini ninja was cool to the touch from top to bottom on 2 sides. I am currently running in a Antec Lanboy case with a 120MM right behind the mini and the 120MM Corsair PSU fan above it. I thought this pig would run hot, kind of surprised. The north bridge heatsink was a little warm however.
I checked the heatsink and the CPU spreader and they both looked flat enough, so I did my usual spreadding a thin layer of AS5 across on the CPU. I never liked their "pea" or "line" way of installing AS5, old school and stubborn I guess. DOH.
-
- *Lifetime Patron*
- Posts: 207
- Joined: Wed Aug 08, 2007 1:50 am
- Location: Xhystos
Minja Installation
I agree with the last poster.
Spreading a layer before installing is my preferred method. I've now done 3 Minjas and use a small spatula to coat the top surface.
Only problem has been that this makes the surface very slick. If the fingers slip when trying to fix the second latch, then the Minja can skate sideways and fall off.
Then it's time to restart, clean the die and try again.
I've found more temp variation due to the quality of the thermal compound. Arctic Silver is my best choice, but the pack comes without a spatula. I reused one of my DIY tools for the job - plastic not metal of course,
Spreading a layer before installing is my preferred method. I've now done 3 Minjas and use a small spatula to coat the top surface.
Only problem has been that this makes the surface very slick. If the fingers slip when trying to fix the second latch, then the Minja can skate sideways and fall off.
Then it's time to restart, clean the die and try again.
I've found more temp variation due to the quality of the thermal compound. Arctic Silver is my best choice, but the pack comes without a spatula. I reused one of my DIY tools for the job - plastic not metal of course,
Re: Minja Installation
If you ask me, goop is goop. My Scythe Mine works perfectly fine even with the stock silicone-based gel. All those TIM hypes, are well ... hypes: http://www.dansdata.com/goop.htmvictorhortalives wrote:I've found more temp variation due to the quality of the thermal compound. Arctic Silver is my best choice, but the pack comes without a spatula. I reused one of my DIY tools for the job - plastic not metal of course,
This test pretty much says it all, the choice of method and brand makes a big difference. Spreading the TIM results in higher temps (about 5°C under load), as the test shows for all the tested TIM's.
But I have no idea why a 5000+ BE would run much cooler than a 4850e, it doesn't make sense.
But I have no idea why a 5000+ BE would run much cooler than a 4850e, it doesn't make sense.
My "spatual" is an old credit card cut a little wider than the CPU. I run a good bead of AS5 along the top edge and less in the middle. I hold the credit card at an angle and press while spreading the AS5. It usually takes 2 passes and I have a perfect, even, thin coat. I do the AMD CPUs on the motherboard, but did the Intel chip by holding it with the protector on the bottom. It was clean and neat.
I have a Gigabyte 690G mATX board and the chipset cooler was HOT, couldn't hold my finger on it. I took it off and found the factory thermal crap looked literally like baked pink dirt. I used my finger nail and scraped it off the chipset. It came off in crusty layered chunks. I used razor blade to scrap it off the heatsink, cleaned them both, applied AS5, reassembled. It made a world of difference and I can hold my finger on it now without frying.
If someone wants to put toothpaste on their CPU, fine with me. I'll spend the $10-15 for a large tube of AS5 and have enough to do 5-10 PCs my way.
I have a Gigabyte 690G mATX board and the chipset cooler was HOT, couldn't hold my finger on it. I took it off and found the factory thermal crap looked literally like baked pink dirt. I used my finger nail and scraped it off the chipset. It came off in crusty layered chunks. I used razor blade to scrap it off the heatsink, cleaned them both, applied AS5, reassembled. It made a world of difference and I can hold my finger on it now without frying.
If someone wants to put toothpaste on their CPU, fine with me. I'll spend the $10-15 for a large tube of AS5 and have enough to do 5-10 PCs my way.
-
- *Lifetime Patron*
- Posts: 207
- Joined: Wed Aug 08, 2007 1:50 am
- Location: Xhystos
Spatula
Good idea. I'll remember that for next time. Better than a recycled DIY tool.
EDIT
Just had a look at the article referred to above. Wasn't aware there was special "cleaner". I just use ether bought from the phamacist in 250ml bottles. What do others use ?
EDIT #2
Well, I looked at the Arctic website and with a clear case of RTFM after the fact, I'll use the blob technique from now on !
EDIT
Just had a look at the article referred to above. Wasn't aware there was special "cleaner". I just use ether bought from the phamacist in 250ml bottles. What do others use ?
EDIT #2
Well, I looked at the Arctic website and with a clear case of RTFM after the fact, I'll use the blob technique from now on !
Odd. If the heat was conducted better to the heatsink, then I would expect the heatsink to be warmer (the underlying thermal source would be cooler, of course).najames wrote:the chipset cooler was HOT, couldn't hold my finger on it. I took it off and found the factory thermal crap looked literally like baked pink dirt. I used my finger nail and scraped it off the chipset. It came off in crusty layered chunks. I used razor blade to scrap it off the heatsink, cleaned them both, applied AS5, reassembled. It made a world of difference and I can hold my finger on it now without frying.