Centrifugal fan DIY ...

[MrCrankyface]

Not so much machining related but I want some mechanical ventilation in my garage/workshop so I can quickly change out the air if it's dusty or smoky.
Looked at commercial alternatives and the price quickly gets out of hand ...
But I figured I already have a bunch of motors and materials, so why not just make my own.

Designed and printed a fan, 90% on intuition "what might work well" and 10% inspiration from various websites.
The 1-phase induction motor is rated for 320W and draws around 200W with this setup.
The airflow "feels" good but I want way more.


I printed a second wheel with more blades and another scoop angle, which increased the air speed a tiny bit BUT made me realize I need to be more scientific if I want to find out what works best.


So I designed this "modular" wheel. It can take 4, 6, 8, 12 or 24 blades and adjusting the angle on any one blade will force all others to follow along.
The blades are slightly convex which in hindsight would've been better if they were flat as this might skew the results.


Furthermore I added a long snout to the blower to force it to work against more back pressure and make it easier to get a reading on the airflow/airspeed.
I considered adding a tube so I could check static pressure but been too lazy so all I'm measuring is power draw and windspeed out of the nozzle/snout.
Testing all configurations on this, the power draw only differed by a few watts, going between 210 and 216W so it doesn't seem to be a major factor at this size.


Graphing things out, this is what I have so far.
"1" step is roughly 45* so the blades can move almost -90 to +90 degrees from neutral.
In hindsight I should've made this more sensitive to have more steps between -30 and +30 and skip the rest of the range.
Generally any blade quantity seems to perform well at neutral angle and 2 out of 3 seems to perform even better at roughly 45* forwards ie pointing in the direction of rotation/travel.
The lonely dot is a separate "fixed" wheel I printed out with 60 blades, because this wheel does not have the adjustment feature it's almost 10mm thinner and I could thus move it in and out on the motor shaft to test how the gap between inlet side and fan wheel side affects output.
Whilst not graphed it did seem to have a major effect, with ~8mm gap the airspeed in the output was 10.1m/s and with ~1mm gap the airspeed was up to 11m/s, improvement of ~9% just by closing the gap by 7mm.


And here's the test with the 60 blade fan.

I'm printing a 100 blade fan to see if that increases the airflow even further or if I'm starting to hit some kind of peak.
If I manage to find some kind of sweetspot when it comes to blade quantity, it'd be interesting to test different lengths of blades to see if I can up the pressure and flow further.
Another factor I'd like to try is RPM variation. It's difficult to test with this motor but I have a 24V 1700rpm DC motor that with a modified fan wheel could be modulated much easier. Whilst I'm guessing the flow is fairly linear to RPM, I won't know for sure until I try it.
It'd also be interesting to compare if there's specific RPM's where the flow is greater related to the sound levels being produced.
Whilst it doesnt need to be silent, it'd be nice if it's not louder than necessary.

 

[mattthemuppet2]

what a fun project! You must have won a lot of high school science competitions

 

[GrifterGuru]

Daft question, but would an aerofoil (aircraft wing cross section shape) blade be any better?

Had you looked at the fan basket from the HVAC unit of a vehicle? they have those down to a science, so that may provide some clues as to blades Vs diameter.

I'm also wondering if using a double cone in the outlet pipe (as per the below diagram) might help with nozzle velocity and shifting more air?

 

[MrCrankyface]

what a fun project! You must have won a lot of high school science competitions

Unfortunately I was both a troublesome student and went to schools where creativity was not encouraged or appreciated.

Daft question, but would an aerofoil (aircraft wing cross section shape) blade be any better?

Had you looked at the fan basket from the HVAC unit of a vehicle? they have those down to a science, so that may provide some clues as to blades Vs diameter.

I'm also wondering if using a double cone in the outlet pipe (as per the below diagram) might help with nozzle velocity and shifting more air?

I've hastily read through some different studies that seem to indicate that aerofoils, especially when curved backwards, are the most efficient.
The biggest reasons for testing flat blades at the moment is to remove factors at play.
Once I've come to some conclusion regarding amount of blades, RPM and angle, aerofoils would be an interesting addition!
Granted I imagine that the optimal amount, rpm and angle will change when aerofoil geometry is applied, so it might be chasing my tail.

I have looked at a bunch of different radial fans to get inspiration but so far haven't seen any clear pattern.
For the same applications some fans have lots of small wings whilst others have have few large ones.
Generally my impression from reading about it, is that more and smaller wings are better for high airflow but you reduce the static pressure it can achieve.

I'm not sure if I understand how it's supposed to work but seems interesting, will have to read up on it.
Intuitively it feels like the narrowing cone would increase pressure and block flow, which results in lower pressure and worse flow in the opening cone, but I've never seen that concept before.

 

[den-den]

Very interesting project. My guess is that if the diameter and width are fixed, testing various options will find the most efficient for a given back-pressure. Diameter x rpm will have the most effect on pressure capability, width will have a great effect on volume.

 

[MrCrankyface]

Very interesting project. My guess is that if the diameter and width are fixed, testing various options will find the most efficient for a given back-pressure. Diameter x rpm will have the most effect on pressure capability, width will have a great effect on volume.

That brings up a very good point, altering the back-pressure.
It should/might be enough to close off the outlet by varying amounts to get some more data for some kind of efficiency curve.

I'm thinking of a cone shaped restriction so it doesn't cause more turbulence than necessary, that can be slid into the outlet.

 

[markba633csi]

Looks like you are aiming at a good compromise of air velocity and air volume- You can only get so much of both
with any given amount of motor power
Interesting experiment though

 

[benmychree]

Reinventing the wheel.

 

[homebrewed]

To reduce the number of experiments you may want to look into factorial design. It allows you to vary multiple elements at once & extracts dependencies for each one. A good example of its use can be found here. It's regarding a different kind of problem but does a good job of explaining the procedure and why it does what it does. There are many resources that can be found to help design and analyze the data you get from your experiments.

 

[Weldingrod1]

Fans are N^2 on RPM and outside diameter.

I did similar stuff on ducted fans, trying to develop one that was both useful and delivered real thrust. Good fun!

Sent from my SM-S911U using Tapatalk