Fan comparison data

R

raidriar

Member
NSC and others have alleged that the 1.75A nidec fan is the best fan that can be installed in the phat PS3s. Others have contended that the 2.90A NMB MAT (19 blade) fan is better. I tested two CECHA01 systems, APS226 PSU, liquid metal under the IHS and AS5 between IHS and heatsink. I have swapped the fans between the systems to verify the results are consistent, and I found them to be within a degree margin of error when swapped between consoles, so I'm confident with my data

Results:
- both fans are able to cool the systems
- the 15 blade 1.75A fan is quieter at duty cycle between 20-35%.
- the 19 blade 2.90A NMB fan is quiet between 20-30% duty cycle but quickly gets louder in the higher range.

the real kicker: testing the same game, the 1.75A nidec fan had to run at 35-40% to keep temps under 68C, while the 2.90A NMB fan ran between 23-30% to stay under 68C

conclusion: the 19 blade NMB 2.90A fan moves more air and does not have to work as hard to keep the system cool. I attribute this to the huge motor it has, and less so to the design or number of the fan blades. Under working load they sound compatible to each other, but the 19 blade fan brings temps down faster when load is removed. However the big difference is the 19 blade fan has more power in reserve it can use to move more air since the 15 blade Nidec 1.75A fan is being taxed more so.

so I can conclude between these two fans the 19 blade NMB fan is definitely better. I would like to test more fans, like 2.25A NMB fan and 2.65A nidec fan, 2.0A delta fan, but I'm having a hard time sourcing them for cheap :D
 
The results of that kind of test you did depends a lot of several factors, is needed to consider if the console starts from ambient temperature or if it was preheated, also is needed to do exactly the same while "ingame" in the same game, etc...


The accurate way to do this in a laboratory would be to connect the fan sealed to a "tube"... and at the other side of the tube some instrument to meassure the air flow (im wondering itf it can be made with another fan connected to a multimeter)
Then start spinning the fan at diffferent PWM speeds with some external controller (not the PS3)
And meassure noise levels too with the same instrument located at the same distance (it could be an smartphone with an app to meassure decibels)
 
sandungas said:
The results of that kind of test you did depends a lot of several factors, is needed to consider if the console starts from ambient temperature or if it was preheated, also is needed to do exactly the same while "ingame" in the same game, etc...


The accurate way to do this in a laboratory would be to connect the fan sealed to a "tube"... and at the other side of the tube some instrument to meassure the air flow (im wondering itf it can be made with another fan connected to a multimeter)
Then start spinning the fan at diffferent PWM speeds with some external controller (not the PS3)
And meassure noise levels too with the same instrument located at the same distance (it could be an smartphone with an app to meassure decibels)
Click to expand...
Well I don't have a lab so empirical data is all I can gather. Both consoles were "cold" and loaded into motorstorm, paused on the same track with the camera turning. Ambient room temp was the same as they were operating together at the same time, about 2 meters apart from each other. The decibel meter app I have in my phone is not very accurate. In a different test I played the same level over and over again for 30 minutes. It gets to a point where the system temperature level off and saturate the thermal solution and that's where we can draw some conclusions about fan and heat removal. The 19 blade fan stabilized at 27% after 30 minutes, the 15 blade 1.75A fan at 35%.
 
Last edited:
The quality of the decibel meter phone app depends of 2 things
If the value given is not much realistic (lets say you meassure a PC fan well documented and the result varies a lot from the specifications)... thats bad, but we dont care because we are going to use the same phone/app in all the tests so incase there is a deviation is going to affect all the tests... so is fair :)
The most important is the sensitivity of the phone microphone, if the phone have a good mic it should expose the differences of noise levels in between the different fans very well

The problem of trying to imagine the air flow based on the ingame temperaures is that we are doing it indirectly, probably there is a big delay and is very hard to keep the CELL/RSX workloads constant
When talking about temperature tests i use to suggest to do a basic test in idle, starting from ambient and waiting more than 30 minutes, because this way you are sure everything is stable (to the point that you should not see any notable change in the temperatures or fan speeds)
 
sandungas said:
The quality of the decibel meter phone app depends of 2 things
If the value given is not much realistic (lets say you meassure a PC fan well documented and the result varies a lot from the specifications)... thats bad, but we dont care because we are going to use the same phone/app in all the tests so incase there is a deviation is going to affect all the tests... so is fair :)
The most important is the sensitivity of the phone microphone, if the phone have a good mic it should expose the differences of noise levels in between the different fans very well

The problem of trying to imagine the air flow based on the ingame temperaures is that we are doing it indirectly, probably there is a big delay and is very hard to keep the CELL/RSX workloads constant
When talking about temperature tests i use to suggest to do a basic test in idle, starting from ambient and waiting more than 30 minutes, because this way you are sure everything is stable (to the point that you should not see any notable change in the temperatures or fan speeds)
Click to expand...
I can do XMB test as well if you think that would be better, but I'm still confident in my finding that the 2.90A fan is the best one for all fat PS3s especially for CECHA/B/C/E

in addition I've noticed in both my 2.65A and 1.75A nidec fans they have no ventilation holes on the bottom of the fan. The NMB fan assemblies have holes for ventilation which helps cool the motor, a good design choice
 
Never minds, i just mentioned it incase someone wants to do the test i mentioned, with the fan connected to a tube, and meassuring air flow at the other tip of the tube (but without PS3)
That experiment would clarify this mistery for once :D
 
sandungas said:
Never minds, i just mentioned it incase someone wants to do the test i mentioned, with the fan connected to a tube, and meassuring air flow at the other tip of the tube (but without PS3)
That experiment would clarify this mistery for once :D
Click to expand...
Which fan do you use?
 
sandungas said:
Never minds, i just mentioned it incase someone wants to do the test i mentioned, with the fan connected to a tube, and meassuring air flow at the other tip of the tube (but without PS3)
That experiment would clarify this mistery for once :D
Click to expand...
Update I did the test in tandem on XMB idle in the friend section (blank) across 3 CECHA01: two of them have NMB MAT 2.90A vs the 1.75A Nidec fan. They have been idling for more than 1 hour. results:
2 x NMB MAT idling at 25% fan speed under 68C limit, 42.5 and 43.1 DB average (probably due to differences in bearing wear can attribute to the difference in DB)
1 x Nidec 1.75A idling at 35% under 68C limit, 40.3 DB average

All three have liquid metal under the IHS cell/rsx, AS5 between IHS and heatsink, all three using APS226 83% efficiency PSU.

If somebody can provide 2.65A nidec and 2.25A NMBMAT fan for test that would be appreciated. Perhaps @squeept has some he can sell me
 
Turns out my DECHA00A has a Nidec 2.65A fan. AS5 above and below IHS and APS 226 power supply. It stabilizes at 27% duty cycle @ approximately 47dB after 1 hour in XMB. Now I just need the NMB 2.25A fan for a final comparison but it looks to me that the 2.90A fan is the best. The Nidec 2.25A fan seems to be rare and it's interesting that it is the same part number as the 19 blade 2.90A fan according to ps3devwiki. I wonder if it is a revision/efficiency improvement over the 19 blade 2.90A fan or a cost cutting measure.
 
If anyone is ever interested, Here is a small lines.qrc mod that could be used to simulate a consistent load, So people who want to compare their fan speeds and temps do not have to get the same game and hope they are playing it the same etc. This is not a large load, but its more of a load than idling on a normal XMB.

So for example testers could set their fan to manual speed 40% while using this wave, and then see what temperature their console stabilizes at after an hour or something. I don't know if it's a viable idea, but anyway it's an option if anyone wants it.

Obviously room temp needs to be considered too, and I do not know if this is putting more pressure on the CPU or the RSX.


 

Attachments

Last edited:
squeept said:
@raidriar I'll set aside some rebuilt fans for you for a bit, but the 19 blade that everyone loves is $30.
Click to expand...
I have 5 of the NMB MAT 2.90A fans, I just happened to ruin one of them trying to relube the bearings and shaft. :(

I did end up finding a Nidec 2.65A in my DECHA00A, but the 2.25A NMB MAT fan is proving to be a challenge. From photos on PS3 dev wiki it doesn't appear to be very efficient. The diameter of the motor housing is smaller in the 2.25A version, but there is no increase in blade length compared to the 2.90A version, so there is wasted space. I'm not convinced it's going to be a good option, but I would like to try and see anyway. The Delta 2.0A fan cable is too short to fit in OG phat boys, as is the later 1.75A Nidec from H01/P01 systems, but I suspect it is the same build as the 1.75A Nidec from G01 systems


You are always welcome to contribute data!!! You probably have more systems than I do!
 
raidriar said:
I have 5 of the NMB MAT 2.90A fans, I just happened to ruin one of them trying to relube the bearings and shaft. :(

I did end up finding a Nidec 2.65A in my DECHA00A, but the 2.25A NMB MAT fan is proving to be a challenge. From photos on PS3 dev wiki it doesn't appear to be very efficient. The diameter of the motor housing is smaller in the 2.25A version, but there is no increase in blade length compared to the 2.90A version, so there is wasted space. I'm not convinced it's going to be a good option, but I would like to try and see anyway. The Delta 2.0A fan cable is too short to fit in OG phat boys, as is the later 1.75A Nidec from H01/P01 systems, but I suspect it is the same build as the 1.75A Nidec from G01 systems


You are always welcome to contribute data!!! You probably have more systems than I do!
Click to expand...
If you're going to continue testing I might suggest placing a thermometer in the same spot next to your consoles when taking the measurements. Even if the ambient room temp goes up 5C from when you tested another model you can simply adjust the other result by 5C to get a roughly equivalent result.

DeViL303 said:
If anyone is ever interested, Here is a small lines.qrc mod that could be used to simulate a consistent load, So people who want to compare their fan speeds and temps do not have to get the same game and hope they are playing it the same etc. This is not a large load, but its more of a load than idling on a normal XMB.

So for example testers could set their fan to manual speed 40% while using this wave, and then see what temperature their console stabilizes at after an hour or something. I don't know if it's a viable idea, but anyway it's an option if anyone wants it.

Obviously room temp needs to be considered too, and I do not know if this is putting more pressure on the CPU or the RSX.


Click to expand...
I like to use a solid black png as a background as I've noticed that it does change the power draw from the wall vs the normal XMB.
 
IMO on my cechg the 15 blade was quieter, but it is true that the 19 blade can stay at lower rpms with the same cooling (just sounds way louder tho)
 
MItch07 said:
IMO on my cechg the 15 blade was quieter, but it is true that the 19 blade can stay at lower rpms with the same cooling (just sounds way louder tho)
Click to expand...
Check your fan and see if you have Nidec 1.75A or NMB 2.25A, I am curious which version you have.
 
MItch07 said:
@raidriar what is the difference between the 1.75 and 2.25amp fans? Does one run cooler?

Sent from my SM-S102DL using Tapatalk
Click to expand...
I don't have a 2.25A fan to test but I imagine it will perform worse than the 2.90A fan as the design has shorter fan blades. On the lower amp fans definitely nidec is the better design
 
You must log in or register to reply here.

Similar threads

Back
Top