How to apply ducted fan theory to real world fans 

Hi, I really enjoyed your video. I'm a mexican aeronautical engineer and I have made some simulations on this phenomena. I can tell you that, in principle, the effect of a duct on propeller efficiency can grow up to 30% in static thrust. However, the effect is faded away when the system is travelling with respect to air, mainly because the drag of the duct. On the other hand, the calculation that you are doing with Newton's 2nd law is called in aeronautics the 'Rankine Froude approximation'. For a better approximation one needs to use the 'blade element theory'. In that case, you will need the lift and drag curves of the airfoil (or airfoils) used along the blades. Indeed, the 'lost power' in your calculation is due to drag on the airfoils: Rankine Froude approximation cannot take that into account. I offer my help if you desire to learn about blade element theory. Congratulations for the great work!

Testing a static fan, you probably want to think about putting a "bell-mouth" inlet in front of the fan. When the fan is stationary the capture area (the area at which the amount of air flowing through at the flight speed of the fan) goes to infinity (pesky divide by zero thing). Of course you can't put an inlet on it with infinite area, but you can put an inlet on it that has a shape that follows pretty close to what the streamlines relatively close to the inlet. This type of area smoothly accelerates the air as it moves from the stationary room air (very very low velocity at very big area) to the inlet of the fan (higher velocity and smaller area) You can actually see the streamlines if you put a little smoke from a small source in the air in front of the fan. A great way to do that is with an incense stick (as long as the air isn't moving so fast that the smoke is stretched out so much you can't see it. Smoke is the best way to visualize the flow field since the smoke particles follow the curvature of the streamlines and allows you to see the entire flow field by introducing smoke just along the outer edges of the area you want to visualize the flow field. Another way is to put a short (L/D of maybe 10) tuft of yarn on the end of a very thin stick (so as to disturb the airflow as little as possible) and then move the stick at a range of axial distances from the fan and radial distances from the centerline. When the air is moving relatively quickly the tuft of yarn will follow the streamline pretty closely. Fluff up the yarn to decrease the density and increase the drag so that it follows the streamlines at lower velocities. At some lower airspeed the weight of the yarn is such that the drag from the moving air isn't enough to keep the yarn in line with the streamline. Either with smoke or yard on a stick, mount a camera at a stationary location perpendicular to the plane of the fan and in front of the fan by 1 fan diameter. Then move the incense stick along the perimeter or the tuft of yard in a raster scan across a 2-D slice starting at the centerline about 3 fan radii in front and then move upward to about 2-3 fan radii above the centerline. To visualize the flow field grab individual frames and superimpose them and you should end up with a visualization of the entire flow field in front of the fan inlet. If you make an inlet that follows the shape of the streamlines going into the tip of the fan you should end up with a bell-mouth inlet. Here is a picture of a bell-mouth along with sizing information. forum.ih8mud.com/threads/1hd-t-intake-manifold-modifications.651365/page-4 I hope this was helpful.

We’re still here and lots of us, haha! Your humility and candor are refreshing as you give us a peek into what researchers long ago had to go through to develop the fantastically efficient ducted fans which power every airliner with massive bypass engines; most of the air doesn’t even go through the combustion chamber. It’ll be super to see what principles lead to powerful fans with more than four blades.

After another month of poking around the internet, this is still the best resource I have found for calculating the pitch along a swept blade. Keep it up.

Wow! Not only is your content really interesting, but it also has really intelligent comments posted from some obviously well-tenured individuals. Big respect to you and your commenters!

Somehow, in an extremely sleep deprived state, ended falling down the 3D printed EDF RU-vid rabbit hole and landed here. This was AWESOME! I made it to the end and subscribed, but when I went to check your other videos, noticed this was the last one you made. :( I get it, though. This project must have taken an utterly absurd amount of time and money to put together, which might have burnt you out early. Or... you know, maybe life happened and you haven't had the time or motivation. Still, I subbed and will look forward to a new video if we ever get one! :) Don't know why, but fan design fascinates me so much. Especially 3D printed ones.

Excellent video and theory description! Experiments and searching the truth are amazing. Thanks to you,I finally figured out the twist of propeller’s blade! Simple. Easy. 100% correct. Awesome job,man!

Great job. Not only made it all the way to the end but I watched it three times!!!!

I realize I'm late to this design party, but I have a intense interest in ducted fans, and have 3D printed quite a few of my own up to 200 mm diameter, and I got some very useful information from this video. .so my question to you is, could you publish your formulas in a spreadsheet format.? Is that would be extremely helpful to us backyard designers. Thanks and keep up the good work

Nothing wrong with metric, but the statement that all these calculations need to be done in metric units or that the formulas only work in metric are flat out false. The relationship "f = m x a" works for all units of force, mass, and acceleration. Mixing measurement systems adds a touch more work, but is also fine. Physics works the same, regardless of the arbitrary naming and magnitudes used in measuring it.

As a fromer duct fan designer, I can only tell you one thing : Welcome to the rabbit hole :D d'ont worry there is no bottom to this one, you'll just be falling from now on. Also congrats, these models and the way you make hyphothesis is really nice and a really "hand on engineer" approach I like it :D ofc you're making mistake even though some make me grit my teeth, mistakes are fun as long as you're trying to find them or checking your math with experiments. And last thing ^^', Is you're motor ok? 400W of power loss... this thing might have been burning in there?

If you want a very easy way to check your RPM measurement you can simply install a free app like decibels, or any other that gives you the instantaneous sound spectrum of the fan noise emissions. Indeed, a fan emits discrete tones at the blade passing frequency (BPF) and harmonics. The BPF is defined as BPF = B Omega/60, where B is the number of the blades, Omega is the rotational speed expressed in revolutions per minute (rpm). So, you can look at the dominant tone and derive the rotational speed Omega with very high precision from the formula above. This works if the blades are (almost) equally spaced.

this is such a fantastic and charmingly made video and i really enjoyed it! i'm now back a few months later, and see that you haven't posted again since, so i really hope you're doing ok and tat you might find the time eventually. wishing you all the best, and thank you for the interesting experiments!

This video is so well made! Awesome stuff dude!

Still here. I’ve been looking for this video for a while now. Thanks for making it.

I'm grateful that I stumbled upon this. What a lucky day...

What a great opening sequence! Yes! I already love you - a nerd, an entertainer, a comedian, a cinematographer, well done!!!

I like this video. I was thinking about fun home projects. I thought more propellers were better. After this video, not only is the number of propellers very important to take into account but getting the right pitch. Good stuff.

Great stuff, had some similar ideas pop into my head and nice to see someone else mathing over some of the concepts. I see a lot of RC plane EDF's are optimized for high power/high rpm but thought a larger, lower rpm, high efficiency EDF's could be appealing for a glider type RC plane or even a solar plane (rctestflight style). Next episode soon please :) I know at least the wikipedia page mentions having very small clearance between the blades and duct can be a large factor for efficiency. Also i imagined multi-stage/multi prop ducted fan may have some more thrust gains too potentially? Thinking like, running on the same motor axis a larger high volume intake fan at the front with a second internal blade optimized for high static pressure a bit like a turbofan engine? Or even just another of the same blade internally may extract slightly more efficiency from the same input power. The next step would be to 3d print or redo the duct frame in a lightweight material and see what kind of thrust/weight ratio you can get from the whole thing.

I loved this. Watched it twice over the last year. So funny and well explained. Thank you!!

Dude, this is an awesome video!! Well done, I have a feeling you’ll have a few more subscribers soon. Would be interesting to measure air pressure inside the duct, pressure sensors are relatively cheap. Most iPhones have one now

this is a great video. making mistakes, finding and breaking them down is the best way to learn something - and the best way to teach something. I do hope you continue this project. It'd be fun hearing about the problems in more detail (Especially after reading some of the informative comments on this video)

Also be very interesting to compare your real world with a CFD program like sim scale which has a free public version

these vids are amazingly edited, hopefully this channel will soon become huge, great experiment

I just found this randomly and it's great. I subscribed and immediately looked for the follow up.

Great explanation! Its cool to see how far this basic physics will get you, even with something as complicated as aerodynamics

you could do a variable size exit cone and empirically check for best speed / thrust, also the increased speed in a reduced duct is a venturi effect (speed increased, pressure decreased) and this is used in the ICE engines throttle bodies etc.

Feel like I stumbled upon a future million sub channel before it took off. We need more projects! This is great!

I am one who watched the whole thing - and enjjoyed it. I know that the running time of the video is 19:16. Very interesting and a good example of the pursuit of the science. As for interest to me, it maybe is something abpout my background. I have had a lifelong interest in aviation, but a one point I was medically disqualified from flying. My father was an aerospace engineer, and his father spent 30 years in the RAF. In my military services, I spent three years with a USMC fighter squadron, and six years as an aircraft maintainer in a USAF heavy airlifter wing. I also worked for a while with a major airline. And I retired from a career in measurement science. So yes, this is VERY interesting to me! Very well presented. Thank you.

"If the ball park is big.. it will get you in the ballpark" love this.

WOW. I love you actually get to "calculate" the parameters, even if it is at "napkin" precision. Thank U greatly.

It's not just the Coanda effect that makes an inlet flare a Good Thing. A sharp-edge inlet forces separation at the inlet. Good picture of the resulting flow is that of a Borda mouthpiece, which essentially reduces flow area to below that of the actual passage. Superb treatment of the instrumentation that you created! Skepticism combined with keen observation produces very useful results.

I made it to the end, thanks for that. FYI, I'm watching because I have a very limited knowledge and my full home HVAC system sucks. So, I'm trying to make it better, thinking I'm going to put in an inline duct booster. I basically can't imagine living in a wind tunnel with temperature hot spots though, so I ain't going to be happy till the mailman has trouble putting paper in my mailbox though. Thanks for fueling my crazy adventure. Subscribed

Good video, I may have watched this to get ideas for fans for the Fan Showndown. Definitly learned me some napkin aerodynamics.

Nice , came to this vid having missed the previous ones, so was a bit behind the facts for a bit, but you showed how the math and physics worked quite well, good illustrations of the various calculations. I don’t remember all the physics I got but it made sense, well done.

That was really cool! On next iteration try to keep clear the distinction between power and energy and their symantics. Great work! I look forward to the next one!

Love stuff like this! Wish RU-vid had more of this content. Nice work! By the way, in order to get the motor RPM up did you consider gearing reduction? In aviation, especially turboprops, extensive gearing is used to reduce the RPM of the turbine to the propeller. Curious how that would have impacted your experiments (notwithstanding the added friction and heat losses from the gearing)

Awesome video! The explanation of the design methodology is great!

7:00 Another way to express the second law is momentum change per time. Also, the velocity change is of the air, not the fan. Obviously the air speeds up when entering the fan, however not as much as airspeed increases

This will go a long way to winning your doctorate! Great job.....

Gave you a thumbs up for figuring out something as minuscule as the hz rate of the electricity in your lights 🤣 genius

Remember that the diffuser at the inlet increases the mass flow because it slows down the air raising its pressure and density, the greater pressure difference between the inlet(before the fan) and outlet(after the fan) adds the mechanical displacement of air from the fan. In other words, with a diffuser, air begins to pile up before it interacts with the fan.

Hey. First of all two things: Congrats & thanks. I adore napkin math because it shows up that "vital few & trivial many". I read in the comments about drag and air density inside the cone and many other things, but for measuring coanda effect maybe you can add some measurements of the velocity of the back flow around the duct at a certain distance. Then you could average the speed over the "influence ring" around the duct and calculate the change in momentum (vectorial mv). That wouldn't be too hard or too complicated and could give you the theoretical deviation / validation of your measurements without the duct. You can also check the results against the Bernoulli given impulse due to differences in static pressure.

I have been sitting here for 13 months waiting for a follow-up video. I know it's not "all about me" but come on, man. I should probably check my postbox. Maybe get something to eat....

I'm just a home handyman. This explains why duct-taping a common floor fan onto the front of a honeywell air cooler when the original fan stops working does not work out to be very effective. I have tried it and quickly realized there must be a lot more science and technology in the design of fans and fan motors than I thought. Nothing works as well as the original parts. I am having trouble sourcing the right replacement parts in Australia though. Thanks for the video, I watched it all and I found it very informative.

I liked how you tried different number of blades. I wonder what the correlation is between the number of blades vs rpm vs power vs efficiency. Since you have all those fans made up why not? I learned a lot from this video, great information, thanks

Don't let the tips hit the air flow . Have the tip recess into a recess. This will greatly reduce air blow back around edge of tip..

Fantastic video. You do a great job teaching, really helps me on my project, thank you.

I have to say, thanks for working through the power formulas. This let me 'solve' something I'd been curious about in a moderately different domain- a diver propulsion vehicle (DPV), depending on the model, is just a glorified ducted fan in water. For one, I have a DPV, and based on possibly wrong specs, I know it probably caps out at 300W, with a 25cm duct diameter. Thanks to math, I can now say, at most, the water is going through the DPV at ~4.0MPH (or 23 gallons / sec), which should be a ceiling on max speed. Then for two, I'm wanting to try building my own, using a similar ducted fan design but using an an edge driven axial flux motor prop concept thing. But I've had no idea what kind of specs to aim for other then the 'let's try it and see' method. Now I have a relation between flow velocity, power, and prop diameter. (granted, applying all of this in water probably adds a whole host of other issues, but if it's within a factor of 2 or 3 (parking lot of the ball park), I think that's fine) P = area*density*(m/s)^3 ... or ... P = (volume/s)*density*(m/s)^2 ; these need to be posted somewhere. I have no idea what to call them because it's an incredibly specific niche, but they need to be easier to find. (or I need to get good and realize that this was always a simple derivation of the base units of kg⋅m2⋅s−3 ) Srsly, everytime I would go looking for a power formula for pumps/fans/etc, they all focus on power in the context of a known delta-P. Which does make sense to me- in those systems you're adding energy to the system by moving mass 'up' a pressure gradient- simple case of energy/time. I'm still scratching my head on the interpretation of this formulae / the situation of no pressure change, I think it's down to "we've imparted velocity on a previously stationary mass of fluid". Which also kind of makes sense. It's a different way of getting to the energy change. All this to say, thanks for the video, it was one of the more useful/informative make/testing videos I've seen on the platform!

Put together well thanks for sharing.

You, sir definitely deserve my subscribe, absolutely awesome

That sharp edge of the duct when not using the rounded inlet geometry might just cause a flow separation in the front part of the duct leading to vortices forming in this low-pressure zone. Hence, the mass flow is reduced. When using the round inlet geometry you avoid those vortices, enabling a higher mass flow and exit velocity. Exactly as you measured!

December 2022 checking in. Looking forward to an installment about diffusors and exit velocity.

Hi, all video was very interesting for me, because I develop electric pusher for my paraglider. So, I am impressed by your results reach 2kg force from 350W !

Wow amazing video learnt a lot .

Thank you. Like the delivery.

The decreased outlet velocity without a curved inlet lip probably comes from a flow seperation just behind the sharp edge. The seperation effectively decreases the inlet stream tube area so the outer parts of the blades are not moving air, this is called blockage. As the nozzle area stays constant, the decreased mass flow leads to lower velocities after the nozzle.

In your calculations you dont include drag it seems so that is where the additional power goes. the propeller is like a wing that rotates so there is a drag there and the acceletaed air in your tube has friction with the inside of the tube. in both cases the power used to fight the drag grows like the cube of the speed. by running one setup at different RPM you should be able to derive the drag "constant" ( of course an oversimplification but a good approximation ). my 2 cents

Hi, very interesting video! Love how you tackle the physical problem from a very experimental point of view. In one of my project I use a commercial EDF to generate a kind of vacuum cleaner to apply a tuned suction force on a plane below the EDF, playing around with the RPM. I would like to estimate this "vacuum suction force" as a function of the rotation speed. From your opinion, would you say that such suction force is equivalent to thrust? Or is it something different? Thanks a lot in advance, Jules

How did you choose the height of the blades? About the angle of the blade, the air also moves outwards due to centripetal forces, should’t the outside of the blades push more air? Should the blades spiral forwards? Should the blades have notches for air separation on the leaving edge? Are the blades just tilted rectangles or is there some wing shape to them? Etc… This video is answering questions that are not addressed on youtube, I am really looking forward for more explanations.

Great video! Didn’t even realize it was 20min hahaha. Are you a aero / related field engineer? Can’t wait for the more concise explanation vid!👍

from mistakes is where we learn and make progress

From my theoretical research: The intake cone should be an ellipse with a 3 to 1 ratio. Same for the nose cone. Nose cone should be stationary so the air is not centrafuged outwards. The stators holding it will help straighten airflow in gusty/windy conditions, adding efficiency. There's a study out there on this. For a static fan, trying to pump air: the exhaust cone should increase in diameter with an angle of around 7 degrees vs axis. This is called Static Pressure Recovery. If the fan (and vehicle is moving through the air; static pressure recovery doesn't work and your current exhaust duct thinking is correct. Same study. Where a wing/blade meets the body/hub the corners should be radiused to reduce turbulence and increase efficiency. Questions: I see your stators are curved in a way that should better straighten the twisting airflow. Does this work? Is there any maths behind your design? What is the gap between the blade tips and duct? The smaller; the less tip vortex; the better.

I loved this. I might have missed a point, but am baffled that the measurements fitted the 2 or 3 blade fan. Obviously, experience (ie commercial designs) tells us that more blades are "better". In fact, as I read it, bigger fans are more *efficient* because they accommodate more blades. I'm looking for clues! Another thing (where I'm probably way out) is that compressing air (unlike water) leads to more complex calculations involving specific heats. I wonder if that is part of the disparity.

would you happen to have these files available for download id like to build some fans for my shop

Please keep up, i love your scientific approach!

Killer video. I'll have to think about this one.

Very simple and informative. I stayed to the end of your Video. Good one 👍

I know is a long time ago since you made the video:) It would have been very interesting to see the trust of the motor and propellers without the duct --to see the theoretical vs measured gain

Fantastic video and I'm hoping to gain some additional knowledge here. I have been playing around with bldc ducted fan as a form of forced induction on combustion engines and I am far from an engineer but understand enough to get in trouble. My question is how to calculate/estimate how much if any, pressure generated. I figured in some format thrust would convert or be part of pressure however I have yet to find a direction. I would like to find some baseline to begin testing "plumbing " configurations as well as utilizing the ducted fan in a compound turbo configuration feeding a "dormant " standard turbo. Again, awesome work!

I had to design a duct for the lift fan on a hovercraft I'm building. In the end I had a straight duct with no bell mouth as I found out that there is not much air velocity as the skirt holds most of it in. So the fan is really just pressurising the air and adding bits as air is spilt past the skirt.

Wyman : It is a pleasure to hear your explanation...sounds well thought out ! Best Wishes towards success .

Please continue to improve this design! I would love to make something like this one day for mounting on my bike

power measurement is the power input. the calculated speedx Force is power output. Both them can be used for calculating efficiency

Really excellent work.

Good video! hope to see more content. It's inspired me to put my all ready printed variety of fans/blades/pitches I printed onto my test bench finally!

Hey, where is the follow up to this video? I really enjoyed this one and in the end u promised more, was hoping to see more content like this.

Great video my friend. Now you need a friend with an ultralight to test fly this “electric jet” EDF… I love👊🏻

Very nice and good explained video. You got me hooked immediately. Let's raise your subscribers to infinity!!!

Yaa.. i still here😂 i love your videos😊

That was excellent. Good, straightforward application of Newtonian physics to a low speed ducted fan! You might want to look up momentum theory and try to analyse it that way as well, for comparison?

What do you think would be the maximum speed of the fan before the blade came off.

I learned something today! Thanks

curved outlets/inlets have quite a lot less resistance to flow. you can look this up in engineering tables for flow

Awesome build dude! Great explanation and transition to the various topics. I really liked how you went in deeper than pretty much everyone else

theres some basics to consider. first off, the less blades, the more efficient. secondly, the larger the diameter of the blades, the more efficient. now, when you use too many blades and a small diameter things get messy. The main reason to use lots of blades is to stuff as much work into a small area as possible- at the expense of efficiency. its a space constraint problem. The next is blade pitch. at static conditions there's a sweet spot and if you go too fast things get messy too. So consider plotting the results in terms of thrust v rpm and NOT maximum thrust and maximum RPM. Obviously when flying theres also a sweet spot but thats another story. And then there's the intake lip. You just cant test an EDF without one. And you cant just use any old shape either. For rules of thumb, power is proportional to diameter and RPM and speed is proportional to RPM.

I enjoyed this! You earned my sub. Thank you for this.

1) Duct air mass isn’t equal to ambient air mass. If not corrected the calculation can be larger than actual. 2) It is also known that air speed approaching the inlet accelerated to X m/s and further accelerated to 2*X m/s right after the blade trailing edge.

With the two bladed model, create ridges on the low pressure area, concentric to the shaft, about 3 mm deep and 5 mm wide. The cannelures/ridges should touch, with sharp edges.

I am trying to understand jet drives in performance jet boats well enough to make improvements to jet drives intentionally . I could not have paid more attention to your video from start to end without going into attention debt. Particularly interesting were the bits where you found that the volume taken by the fan in the duct did ot count against the dynamic airflow volume very much. Also the effects of intake bell mouth shapes and presumably sizes. Even though jet drives and ducted fans part ways in in most ways at the far ends of their impellors, the inlet side of both share a lot of similarities and room for improvement on their inlet sides

Great work and interesting results. Thank you!

I made it to the end but I wasn't subscribed... But I am now!

I reccomend you checking out 'velocity stacks'. This is why the tourus thing worked so well.

I made it all the way. Very interesting. I watched a video the another night where Ross does a critique of the Raptor aircraft's proposed ducted fan. Pretty dismal.

Love how he said "YES" LOL

Great video.

Definitely gonna print some donuts for my CPU fan instead of my shroud to see if I get the same effect.

Most important inlet effect is the venturi effect. Industrial Ventilation Conference.

Great Video! I made it all the way. Subscribed at the end, BTW.

You deserve much more views. Subbed and shared