Getting away with a randomly placed radiator.

[yoadknux]

I would guess you had air in the pump. The coolers have a little bit of air in them and you have the pump as the highest point so the air will collect there. That would explain the noise and the poor results. I read a few stories of people having problems trying to mount an AIO on the bottom of the case with a G10/12.

I would say the optimal setup would be to have a case with room to put the CPU rad on top and the GPU rad on the back both as exhaust. Then have front/bottom intakes. Or have two 120mm rads and mount them both on the front.

I think you're right, because the pump itself when plugged "freely" makes less noise, it's still noisy but not as much. This narrows the list of compatible cases (for 240mm GPU rad) by a lot, unless I do front exhaust.

An interesting, general question now is, what should give better temps, a radiator mounted "freely" followed by exhaust fans (see first image on this page), or a radiator mounted in front of an exhaust port (standard mounting for radiators, see 4th image on this page for CPU rad as an example).

Radiator mounted freely -> extract heat more efficiently, fans run quieter because they deal with less pressure. Problem: may end up circulating hot air through the rad.
Radiator mounted on an exhaust port -> fans do more noise, air does not flow as freely through the radiator

 

[Dermen]

I think you're right, because the pump itself when plugged "freely" makes less noise, it's still noisy but not as much. This narrows the list of compatible cases (for 240mm GPU rad) by a lot, unless I do front exhaust.

An interesting, general question now is, what should give better temps, a radiator mounted "freely" followed by exhaust fans (see first image on this page), or a radiator mounted in front of an exhaust port (standard mounting for radiators, see 4th image on this page for CPU rad as an example).

Radiator mounted freely -> extract heat more efficiently, fans run quieter because they deal with less pressure. Problem: may end up circulating hot air through the rad.
Radiator mounted on an exhaust port -> fans do more noise, air does not flow as freely through the radiator

You are forgetting that freely mounted in this case has an extra exhaust fan. The radiator fan and it's airflow would make less noise, but that noise will just be shifted to the exhaust fan plus extra fan motor noise. I think it would make little difference other than being more cluttered looking and more complicated. It is kind of like a push/pull setup. It might help a tiny bit but I prefer to keep things simple over the few degrees difference it could possible make.


I'm kind of curious now how a 120mm would compare to a 240mm. I couldn't really find a direct comparison on a GPU.

I haven't shipped the H105 back yet and have to remove the GPU cooler to install the new fan when it arrives on Thurs. Maybe I'll keep the H105 until then and do a direct comparison of the H55 and H105.

 

[yoadknux]

You are forgetting that freely mounted in this case has an extra exhaust fan. The radiator fan and it's airflow would make less noise, but that noise will just be shifted to the exhaust fan plus extra fan motor noise. I think it would make little difference other than being more cluttered looking and more complicated. It is kind of like a push/pull setup. It might help a tiny bit but I prefer to keep things simple over the few degrees difference it could possible make.


I'm kind of curious now how a 120mm would compare to a 240mm. I couldn't really find a direct comparison on a GPU.

I haven't shipped the H105 back yet and have to remove the GPU cooler to install the new fan when it arrives on Thurs. Maybe I'll keep the H105 until then and do a direct comparison of the H55 and H105.

I would be interested in watching it, since in my case I was simply unable to do this comparison. If you're cracking your video card open anyway may as well just give it a try. Will your case allow for a 240 radiator around the GPU area though?

 

[Dermen]

I would be interested in watching it, since in my case I was simply unable to do this comparison. If you're cracking your video card open anyway may as well just give it a try. Will your case allow for a 240 radiator around the GPU area though?

Yeah my case is very shallow so it is no problem for the cooler's lines to make it to the front of the case where I can mount up to a 280mm rad. In some cases it is too shallow, it only has 330mm space for the gpu with a 25mm fan in the front. A radiator makes it less, and in the case of the H105 (38mm) would limit me to 292mm GPU or less. Plenty of the new AIOs are thinner and perform as well or better though.

 

[Evilsizer]

not sure what size case your ok with but even though im MATX. if i were to go water cooling, i would still get this case with a MATX motherboard LOL.

https://www.newegg.com/Product/Product.aspx?Item=N82E16811112584
man the possibilities of a completely silent build with cpu+gpu being cooled... i digress, i dont have money for such a setup LOL

 

[Dermen]

Well my shipment got delayed so I went ahead and ran the test between the H55 and H105 today. I didn't test the H55 with it's stock fan but tested it with a stock H105 fan. Tests were done with the radiators as front intake in a Phanteks P300. I did a single Heaven benchmark and recorded the max temp. Waited 10 mins between tests to allow GPU to cool down.

H55 with Phanteks F120MP
1500rpm....69C
1800rpm....68C

H55 with Corsair SP120L (stock H105 fan)
1500rpm....75C
1800rpm....72C
2600rpm....61C

H105 with Corsair SP120L (stock fans)
1500rpm....54C
1800rpm....53C
2600rpm....50C

Just out of curiosity I ran the test again with the H105 pump at 1500rpm (pump at full speed is 1800rpm). 1500rpm is what I consider quiet (and it stops buzzing) and it is the same rpm as the H55 runs at full speed.

H105 Pump@1500rpm with Corsair SP120L (stock fans)
1500rpm....55C
1800rpm....53C
2600rpm....50C

 

[Evilsizer]

clearly looks like the phanteks fans has higher static pressure then the corsair fans. if you arent i would switch to those, but then im not there to hear the difference between fans at different speeds.

 

[Dermen]

clearly looks like the phanteks fans has higher static pressure then the corsair fans. if you arent i would switch to those, but then im not there to hear the difference between fans at different speeds.

To me the Phanteks is noticeably quieter, most reviews of it agree.

The H55 results are a little odd to me. I have always ran it mounted in the back of the case as an exhaust and never had temps that high. I assumed having it as an intake would give me better temps. However, I returned my system back to it's normal setup and ran the test again.

H55 mounted in rear as exhaust with Phanteks F120MP. Two intake fans, Rosewill 120mm at 1500rpm and a Rosewill 140mm at 1200rpm.
1500rpm....63C
1800rpm....60C


When I ran the original tests I had two front intake fans (SP120L x2 or Sp120L + F120MP) and one exhaust fan (Rosewill 120mm PWM). Rosewill ran at 1500rpm all the time, and the SP120L ran at 1500rpm when I tested the H55. When I tested the H105 it ran at the stated speeds.

 

[Evilsizer]

with it mounted as an exhaust, is the fan mounted as a push or pull on the rad? if the intake on the front case has a lot more restriction on pulling air in then it makes since it would run hotter that way vs being at the back of the case exhausting the air out.

im really not sure how to explain it but its all in the setup and case. having to do with temps as in being mounted up front vs in the back.

 

[yoadknux]

Well my shipment got delayed so I went ahead and ran the test between the H55 and H105 today. I didn't test the H55 with it's stock fan but tested it with a stock H105 fan. Tests were done with the radiators as front intake in a Phanteks P300. I did a single Heaven benchmark and recorded the max temp. Waited 10 mins between tests to allow GPU to cool down.

H55 with Phanteks F120MP
1500rpm....69C
1800rpm....68C

H55 with Corsair SP120L (stock H105 fan)
1500rpm....75C
1800rpm....72C
2600rpm....61C

H105 with Corsair SP120L (stock fans)
1500rpm....54C
1800rpm....53C
2600rpm....50C

Just out of curiosity I ran the test again with the H105 pump at 1500rpm (pump at full speed is 1800rpm). 1500rpm is what I consider quiet (and it stops buzzing) and it is the same rpm as the H55 runs at full speed.

H105 Pump@1500rpm with Corsair SP120L (stock fans)
1500rpm....55C
1800rpm....53C
2600rpm....50C

The H105 seems to do a better job. I wonder if it works as well under sustained loads. Why radiator on front intake and not exhaust?

 

[Dermen]

with it mounted as an exhaust, is the fan mounted as a push or pull on the rad? if the intake on the front case has a lot more restriction on pulling air in then it makes since it would run hotter that way vs being at the back of the case exhausting the air out.

im really not sure how to explain it but its all in the setup and case. having to do with temps as in being mounted up front vs in the back.

It was pull as intake, push as exhaust. Most tests I've seem comparing push vs pull the difference is 0-1C. I thought about it more and it probably has something to do with having 2 intakes and 1 exhaust. The positive pressure in the case is helping push air through the exhaust and fighting against the intake.

The H105 seems to do a better job. I wonder if it works as well under sustained loads. Why radiator on front intake and not exhaust?

I was expecting about a 10C difference. What surprised me was how much better the Phanteks fan did. It would have been interesting if I had a 2nd one and could have tested the H105 with them. I didn't try the H105 as exhaust because the case and dust filters are designed for front intake and rear/top exhaust. Plus there are so many different things to test you could do this for weeks. I probably should have tested with the radiators in open air so the case design/airflow/pressure wouldn't have mattered.

 

[Evilsizer]

The H105 seems to do a better job. I wonder if it works as well under sustained loads. Why radiator on front intake and not exhaust?

i would hope the H105 is better considering it is a 2x120 vs the H55 being 1x120 rad.

 

[yoadknux]

A friend has offered me his Corsair Obsidian 750D Airflow edition for basically peanuts. I'm thinking about switching, because it should allow me to throw the H105 with proper mounts on my 1080ti. It will NOT fit top, so the radiator can only be placed at the bottom or at the front. Bottom is very tight fit with my PSU... so all that remains is front.

Now comes the question: Should I front intake or exhaust?

I was thinking about doing something like this:

FRONT: H105 2x120 or 4x120 (depends on fit) exhaust
Bottom: 2x120 intake
Top: H80i 2x120 exhaust
Rear: 1x120 intake

Kind of inverted airflow, what do you guys think? Should I change one of the fans before I build everything?

 

[Dermen]

I think the H105 won't reach the front of the case. If it is close and you have a couple old fans you can gut them and use them as spacers, you will need to grab some longer screws from a hardware store. People used to do this anyway to reduce the dead spot from the fan hub. Radiator|gutted fan|rad fan. It is hard to tell from just a picture if that will be enough, I guess you could do two layers of gutted fans if it isn't.

Then I would just test it. Try bottom/front intakes with top/rear exhausts, then swap the rear/front to intake/exhaust, then swap the bottom/top to exhaust/intake.

 

[maxfly]

the 750d af was designed for excellent flow using the front/bottom for intake and top back as exhaust. the 750d airflow is a sweet case.