There's a lot of information out there regarding electric fans (efan, e-fan) and almost none of it agrees, which is exhausting, so I decided to set up a "fan lab" and do some testing myself. It's not perfect, but it's semi-controlled with methodology and equipment.

The Test Rig

There's a pressure drop when a fan is in the car, bolted up to the radiator etc., which lowers the actual CFM moving through the system (and probably raises current, too), compared with the same running in free air; therefore, I got a junk radiator and a junk A/C condenser from a friend. The rad is out of a 1985 60-Series Toyota Land Cruiser, and it's a good 3" thick. I clamped the condenser to the rad, then stood the assembly up in a vice, and clamped each fan to this assembly for testing. This should give results closer to actual performance in a car.



I used a 12V battery charger on 50A jump start mode for power. I connected one fan to my truck, just to confirm that running alternator voltage increases current draw and CFM by 10%, and adjusted the figures below accordingly. I used a cheap anemometer to measure airspeed, and a cheap RV power gauge to measure current. I used various clamps, jumper cables, pins, etc. to make connections. I measured low speed where the wiring made it easy; I skipped it if it was unclear, because I will ultimately not use low speed anyway, so I don't care.

CFM was determined, in every case, by crudely averaging airspeed (ft/min) from multiple points directly in front of each fan, then averaging the two fans (where applicable), and multiplying by the SqFt of the shroud. This is how the fans actually work – they pull across the entire shrouded area. There are high and low points all over, and they all average out.


1994 Ford Taurus SHO 3.2L
This fan consists of one 11" and one 14" fan in a 24” X 15” shroud that measures 3-9/16” thick. This fan made 3,575 CFM at 24.9A. The 14” fan was surprisingly weak – weaker than the 11”. This is also a fan that’s completely out of production, even for replacement motors, so it’s basically just another data point, and not a real option.




2007 Ford Focus
This fan consists of two 11” fans in a 23” X 14.3” shroud that measures 4” thick. This fan made 4,083 CFM at 25.0A.




2002 Mercury Cougar
This fan consists of two 11” fans in a 24” X 15.5” shroud that measures 3-Ύ” thick. On high, this fan made 3,552 CFM at 22.6A; at low speed, it made 2,558 CFM at 15.6A. This is supposedly the same as the Ford Contour fan that’s popular with Mustangs.




1996 Ford Thunderbird
This fan consists of one 18” fan in a 22” X 19” shroud that measures 5 9/16” thick. On high, this fan made 4,630 CFM at 20.2A; at low speed, it made 3,483 CFM at 7.0A. The tested fan was actually a VDO replacement fan – not factory Ford.



Here's a table of these results:




When I figure out how to run the MkVIII fan in my F-150 at full speed without intentionally overheating the engine, I'll post its CFM.

 

Its good to know the VDO is 100% comparable to stock. I will be going with this fan for my set up.

You Will Sell me the Tbird fan cheap....

 

i will sell you the tbird fan cheap....

 

Awesome job CO, this will be my winter project, I will be using the Tbird fan, next gotta decide on a controller,

 

I recommend the Flex-A-Lite Variable Speed Controller model 31163/107012. I have it on my F-150 and it works great.

 

VDO made the factory motors for sure.

 

I pulled apart the Focus and Contour fans. The motors and blades interchange, except the clocking is just different enough to pose a problem. I pulled the fan off of one of the Contour motors and found that it's a Bosch. Anybody know how to read a Bosch p/n? There's three rows of text on this fan motor, making it pretty tough to do a search for it.

 

1996 Lincoln Mk VIII / F-150 Retrofit
This fan consists of one 18” fan retrofitted into the factory 2001 F-150 32” X 18” shroud (thickness not measured). This single-speed fan made 6,000 CFM at 37.4A. It was much more difficult to make a crude average of the airflow, due to limited space and the difficult-to read screen on my anemometer, so a two grains of salt must accompany these results. I suspect it would measure higher on my test rig, just due to my inability to measure the points I used on the others. The Flex-A-Lite controller has provisions for manual override on and off switches; I used the override on to force 100% speed.

Here's the revised table:



And here's a new comparison table:



If X/Y is >100%, X = (cell value-100%) more powerful than Y.