From an ex vacuum engineer - it isn't as simple as either 'max flow against no back pressure' or 'max pressure pull with no flow'. What matters is the inbetween, and whether your impeller performs well in series with the back pressure of the vacuum's filters. Vacuum cleaner companies use a metric 'airwatts' to characterise it - pressure times flow rate (in SI units) actually cancels out to watts, and you can compare your electrical power in to your air moving power out!
It’s like the saying that anyone can build a bridge with enough steel and concrete, but an engineer will build one that is efficient in use of materials, adhering to time, safety, and budget constraints.
You should take a look at the way cutting discs are attached to angle grinders. They are held on by a nut that's threaded opposite to the spinning direction. The centrifugal force keeps them tight, they will never come undone on their own.
that would be hard to implement here, then again he can fix a coupler to the shaft and drive a bolt into the coupler from the top of the impeller but that solution would be bulky.
@first-thoughtgiver-of-will2456 at the risk of sounding pedantic (and maybe antagonistic) it's actually the interplay of normal/frictional/inertial and drag forces. Centrifugal plays no real role in holding the nut on, and acceleration only plays a part at the start of the motion and in the form of vibrations.
To transfer the torque better and make it easily removable you could also use equaly spaced grub screws. or if you want something non removable you could grind a divit in the motor shaft and also some in the 3d printed hole and then glue it. That way the glue will not just keep the rotor from slipping but instead mechanicly lock it on the shaft
@@quelixfenzer5108 once you've gone to the trouble of grinding a axial slot in each, just go ahead and install a key: that's the traditional solution to transferring torque from a shaft
@@RossReedstromYou cannot key a shaft this small, that looks to be smaller than ∅5mm. Well, you can - it just wouldn't make very much sense, shafts this size are usually secured using "dog point" set screws, with according holes in the shaft, or with cone point set screws if the shaft must not be weakened further. Removing material for a keyhole from an already undersized shaft isn't a great idea. I'd bet a regular pair of set screws with flats would serve him right and it would all be achievable at home using a file and a tap.
@lazar2175 if he really wanted to he could use an edm machine to make the initial hole and some specialized tap for hardened material. I don't know much on the tap side, but an edm can cut through any hardness of metal.
The closed impeller design they used can make a massive difference for static pressure. On impeller to motor connection just get some metal(alluminium) hub for plane/car model motors and make the impeller connect to the hub instead of the motor shaft. This will introduce much higher surface and possibly hard connection to it.
you should try getting the impellers SLS or SLA printed. It should help with airflow by having a smoother surface, and the SLS one can be CF Nylon, which should be plenty strong and light for this application
For part three, maybe include a "turbo button" with a 30s timeout that runs the motor at ~95% of the maximum that the motor and controller can handle and run it at whatever level the lady deems acceptable all other times.
the actually vacuum cleaner is a turbo, but instead of using exhaust gases it has an electrical motor, whatever it sucks instead of going to the engine goes to the bag
I'd like to publicly acknowledge how much Scott cares for his girlfriend. Clearly, her ability to quickly and efficiently vacuum his apartment - and, therefore, *her* quality of life - is hugely important to him.
Be sure to check out the Tineco Pure One. We have several clients with handheld Dysons that we occasionally use if we're just adjusting something or didn't make a big enough mess to grab our Festool shop vac from the truck. One of these clients replaced theirs with the Tineco when their Dyson died and couldn't be repaired. The Tineco is way quieter and feels less like it's been overly value-engineered (just feels more solid in general).
I am an engineer, i designed and 3d printed impellers for the wash and drain pumps on a dishwasher. I designed them for 8% improvement. The logic was that the motors had at least a 10% safety margin, and thus, 8% would be safe. I used ABS. That was 7 years ago. I still hear the stronger Swoosh-swoosh sound when it's used... and i made test impellers, which were thicker layer height for testing fit and function.... never removed them.... 7 years, and the rough test fit impellers are still going..
I'm not a vacuum designing engineer like the commenter below but I used to engineer centrifugal liquid pumps so I know a little bit about this subject. The static pressure will go up with impeller diameter and/or RPM squared. Google "fan laws", it will be valuable information. The blade angle of attack won't affect static pressure much because there isn't any flow through the blades during this test (weird I know). One cool thing about these kinds of devices is one might think the shaft torque would go up if you restrict the flow, but actually it goes way down. Notice that you hear a vacuum cleaner motor speed up when you cover the hose. The load on the impeller comes from the flow rate of fluid that is being accelerated so more flow = more torque and power required. What happens to the motor power in your experiment when you go from open hose to the static pressure test? If you cover the outlet the flow goes down but the fluid density goes up. I think it will still consume less power (I designed liquid pumps with constant density so I have no actual experience with this). If you can design a "volute" on the outlet of your pump that gradually expands the cross sectional area, it might improve the performance although I don't think you have the room to do this in your housing. The purpose of the volute is it acts as a diffuser to have pressure recovery as the fluid decelerates.
A grub screw will help with keeping the impeller on. It will have to be factored into the design. As far as the impeller is concerned you could try carbon fibre filament, it may be stronger. The other option is creating a silicone mold of it and using a casting resin or making a clay mold,and using aluminium to cast the part.
30 year home repair contractor here, and I use the exact extractors in the video without any problem. You need to apply a lot of force on the drill, and go really slow- allowing the bit to bite.
Hello Scott! Do you consider printing the impeller with other materials, for example ABS? Because in general the density of ABS filament is about 10% lower than PLA and PETG. Maybe saving a bit weight on the impeller can help the efficiency. Thanks!
@@harrylenon9594Weight certainly DOES play a factor. Weight = LOAD. The more load, the harder the motor has to work. When the motor is capped at a set power limit.. it will reduce the overall RPMs with a heavier load. The load might also be amplified, as you move the vacuum around.. as the extra mass with create more vectoral forces.
@@johndough8115 no, if the load has the same friction with its surroundings, it will reach the same top speed. Acceleration will be higher with lower inertial load, but top speed will not change. At constant speed, weight plays no role in power draw.
@@harrylenon9594 There is a channel that makes Custom fan blade designs. Certain designs are heavier, use more blades...etc. They all spin at different RPMs, despite the exact same motor being used.
@@johndough8115 yes, different DESIGNS.... the design will determine the drag on the propeller while spinning, not the weight. If you did a test with two identical blades, only difference that one is heavier, you'd find the lighter blades spin up to max speed faster, but have the exact same top speed
I think you can take it a bit further. I'm very familiar with RC motors and we never use a press fit gear as they'd come right off the shaft like you're seeing. A simple fix would be to integrate a 48p RC pinion right into the impeller. You'd need to add some gcode to pause the print so you can insert the gear directly into the part and then keep printing to seal it in. This will give you teeth on the gear to engage the impeller directly and a small hole in the side of the impeller will give you access to the pinion's set screw. The set screw in the pinion will lock the whole assembly onto the motor.
GLAND SEALING SYSTEM. is a technique used in steam turbines to make the bushing "air tight" without having contact with the shaft. Try adding it to your design.
I would highly recommend screwing the impeler in place, either on the top or side of the shaft. I believe it is necessary for any fan system to function correctly.
I believe the cap on top of the fins, on the original, was to increase static pressure at lower rpm's .. also static pressure increases with multiple stages. People often emulate automotive turbos without taking into account the facts. They have very tight tolerances, and operate at extreme rpm's, some at over 100,000 rpm. At those speeds and pressures, the chamber itself becomes the cap on top of the fins. The cap on the fins will leverage the lower rpm window for your materials of choice, and those often used for mass produced products.
@@triaxelongd3337yes grub screws would be great for that, especially if he makes a lil groove in the shaft for it to bite in, or even better use an actual pinion RC gear for one of those motors and something in the impeller that can hold onto that
I wonder if a duplex impeller, with a single inlet and two outlets would balance the thrust forces against each other? There would have to be a new inlet scroll... The "two" impellers would be "attached" at the inlet end.
You should try a resin 3D printer for that. It is possible that bc FDM printers aren't known for printing air tight enclosures. It is possible that there are small holes in the enclosure that prevents it from producing pure vacuum. Resin solves this problem
Those holes are pretty insignificant in size compared to the input and output of the impeller tho, even more so considering more than 2 walls, the overlap of the material, even with small voids, compensates it (think of the Swiss cheese). With resin I would be concerned about the material strength itself, it's also not too great with creep, and even temperature. It would be like doing the part out of PLA but weaker still, with the benefit of it being extremely (visually) precise, sure there are good resins but they're quite expensive as well.
@@Kalvinjj depends on the type of resin. There r many different types of resin out there other than the standard form labs. There are metal based resin and high temperature resins out there also. But I think in terms of cost, buying new vacuum would be cheaper than trying to go all out in a high end resin printer.
Do NOT purchase a Dyson. They are pretty and have great suction, but that's about where their benefits end. They get clogged up super easily and the plastic is super brittle. Mine was under warranty and they said I can't get it fixed because they don't have the parts for it. WTF??
Great work. Vapor polishing would improve the smoothness of your prints and should provide some level of performance gain as the airflow over the blades should be smoother. You're probably losing a lot of efficiency to airflow past the clearances between the impeller and the housing. Perhaps SLA printing would be a better option, as it's higher resolution and it also produces smoother surfaces.
This is where people mess up when cooling PC's too. All fans are not built the same. Some are made for high static pressure and some are passive. Don't use a passive fan for your cpu or any situation where the fan needs to pull or push through something.
du hast noch ein problem mit dem Luftauslass - die luft, die der impeller ansaugt, muss auch wieder raus. und da sind die kleinen fensterchen quer zur strömungsrichtung der luft. geschickter wäre die fensterstege längs, tropenförmig zur strömungsrichtung zu machen und denen so viel querschnitt wie möglich zu geben. dabei ist zu beachten, dass nicht nur die fläche entscheidend ist, sondern auch der querschnitt je fenster. wenig große fenster sind besser als viele kleine, auch wenn der gesamtquerschnitt gleich ist. luft mag keine ecken, jeder kanaleinstrom und ausstrom sollte es der luft so leicht wie möglich machen und möglichst kein labyrinth oder ecken entgegenstellen.
Just an idea... use a small turbo. Match the current motor rpm to a compressor map of the turbo. Then use the comressor, then turbine and see how they both perform.
You might be able to alleviate the impeller pulling itself off by implementing some set screws. This of course though means redesigning the impeller and also adding some rotational weight.
You must work with the surface finish on the impeller and the housing. Surface roughness greatly affects flow attachment and its properties (Reynolds numbers, fluid dynamics, bla-bla-bla; you can easily find a lot of free articles). Long story short, surface smoothness is critical in such applications, and losses in the system scale exponentially with the speed of the blades.
Let me know if you're interested in more details -- but I've made a similar type of impeller (and propeller too) with 3D printing -- the best way to secure it (I found) was to use small brass inserts and insert them into the round bit in the middle (in-between the blades) to screw into the motor shaft. Since an iron won't fit in the small gap, I would screw in the head-less bolt into the insert, hold it with needle-nosed pliers, then heat it up with a blow torch - then pushing the insert into the plastic between the blades. I would then cut down one of those head-less bolts (forgot their name) so they wouldn't stick out and use an allen-key to tighten them on EITHER: 3 sides | OR on the top and bottom of the motor shaft for more even securing. And BOOM can run it as fast as you want and everything -- hope this helps :)
Use acetone to smooth the propeller, stick pieces together and make stronger joints. Just putting the 3d print in a acetone bath for a couple of minutes will not only smooth it but will increase the strength to a large extent.
Maybe make the impeller closed with a skirt to minimize boundary layer drag and skin friction between the impeller and the enclosure? Polish those impellers and minimize any layer visibility, every ridge creates extra drag, which creates trubulence. If you're doing Part 3 and want to try with an aerospace engineer, I'm down to take Bosch on 😁
Some BLDC motors just have a nut on the end that you can use to ensure the prop doesn't walk off the shaft. Some are even reverse threaded such that when the motor spins, the screw tightens down on the part. You could also go for a keyed shaft to reduce rotational slippage.
You just need a longer shaft so you can thread a nut or stopper on the end. That's how turbos keep their impellers on, or even some sort of C-clip would work.
If you want, I design and work on these for a large turbomachery company (the big blue ones), and can offer you some pointers, both with how the impeller is attached, and why you are having such issues with static pressure
If you decide to buy one get a Miele at least, they are fantastic vacuums! There is not only Dyson! Maybe you can also ask them for some infos about theirs since they're also from germany. (I think there vacuums are made in asia though, still great quality)
You could have used a set screw and a motor with a shaft that has a flat on it to counter the impeller slip.. glue never holds…. And the tolerances between impeller and compressor housing has to be very tight for that style to work.. is why most vacuums use that flat style..
You use an open impeller design, which causes recirculation flow around the impeller blades, lowering the efficiency. Try using a shrouded impeller and reduce the clearance around the impeller and the shroud.
Vapor polishing would improve the smoothness of your prints and should provide some level of performance gain as the airflow over the blades should be smoother.
Perhaps with some post processing the FDM prints(sanding, smoothing and sealing), You might be able to get a few percent higher efficiency there. I recall rctestflights videos where he mentions a significant performance penalty with propellers and impellers that are FDM printed without post processing work being done.
I wonder how much improvement you'd see from smoothing out the 3d prints. In theory all the ridges from the 3d printing create disturbances in the air which would reduce efficiency. There's a reason why propellers are as smooth as possible.
There is a reason why they are not. Actually, performance fans are rarely glossy smooth. The only example of consumer fan with glossy blades I can remember is an office fan, the big thing that blows and swings around, definitely not a performance hero. I don't know the exact explanation but I'd say near-surface air layer is unimportant, at least at relatively low speeds, so nobody cares. Edges and, in case of static pressure, gaps are far more important IMO.
U can etch a dimple on one side of the motor shaft and add a security screw to keep it in place so it doesn't fall off. Belt pulleys use it like that to keep them in place.
If you sand a rought patch on the motor shaft and then put the glue in the bottom of the impeller before pushing it on it will handle alot more speed before the glue fails
To prevent the impeller from running away from the shaft, maybe you could make a grove at the end and put a circlip there. Also, cut a flat surface on the side of the shaft to prevent the impeller from slipping on it. I woodworking, it's well known glue is always the weak part of an assembly: you need to think your assembly without glue and add it as reinforcement, not as the primary transmission.
@@Scotty-vs4lf you wouldn't do that, because the pin is going to tear into the plastic at startup dur to excessive torque. you really want to maximise contact surface and avoid pins. half moons would work.
@@RegisMichelLeclerc yeah i figured the inertia of the impeller was low enough that it would be fine but i have no experience so ill take your word for it
In my opinion You need tighter clearances to improve the seal. Unfortunately I’m not sure how realistic it is to do this using a 3D printed impeller. Perhaps a resin 3D printer could help u get a more precise print?
heya, if you ever look for a nice vacuum that doesn't cost as much as the dysons, you could try some of the xiaomi ones, they are actually surprisingly good
Impeller balance will play some role in negating performance at higher rpm’s. That being said I believe you’re on the right track, I figured you’d buy a metal turbo impeller and try that but I respect the commitment to designing your own!
The stock propeller seems to have a smaler inlet 🤔 this is probably better for creating pressure or vacuum. I bet you lern much by reading about turbos, comparing turbine maps and the physical demencion and drow some concussion between turbine well size on inlet and bottom in order off pressure vs airflow map
there should be a distinct eye in the impeller, and quite long blades for this. if you end up with a smaller eye, and more surface area on the blades then you should be able to have a lower static pressure, meaning more suction, therefore having a potentially better vacuum.
These two videos just reminds me again how hard and complicated fluid dynamic is, and paying the price for geniuses to do it right is well worth it for my dumb mind.
Have you considered finding a motor with a threaded shaft and using embedded nuts in the impeller to connect it? When spinning in the correct direction, it will self-tighten itself to the shaft.
You could have used original impeller and new motor that might be the best way to achieve efficiency. And for benchmarking you can use dc motor with new impeller as bldc tends to consume more power at higher rpm and provides very less torque.
the axis of the motor is smooth therefore it doesn't bite well into the plastic. also I see sometimes a bolt perpendicular to the axis, that prevents the axis from slipping from its position, translation AND rotation are blocked this way.
Out of curiosity, a question to our mechanical engineers: How much does the surface finish affect the outcome in such a system? The printing looked quite rough tbh.
Maybe look at an Angle Grinder...they've (Some at least), have a nut screwed on the wrong way. I think that in your case, you could even look into drilling a hole into the shaft then bolting your impeller to it. A bit drastic, I know, but that's how I bolt my large 20 Inch drone props to my drone and it's worked fine for upwards for 5 years now.
I got a Bosch Ultimate 7. lighter than most Dysons, sucks about as much, uses normal Bosch DIY battery's and has a nose so you can stick it in corners easier. oh and costs about a third compared to a Dyson.
Wow very interesting video! I have done a lot of airflow and suction measurements and most important for good cleaning power is the airflow from the floor nozzle. I have dedicated airflow/suction box for that. I can seal floor nozzle to the box and get very accurate readings.
one big factor that seems to impede your design is the use of a 3D printer: the end products its produces are far from smooth, and if you know anything about aerophysics youll know that that is quite significant. consider sanding down your prints or making them ever so thinner and then applying a layer of lacquer.
Idea: Why not texture the shaft a bit and print the impeller with a bit of an undersized hole and use a soldering iron to heat the shaft to melt the plastic onto it? Like a heatset thread insert.
Why not using a copper core of a wire strip to joing the impeler to the axle? The screws will hold tighter to the motor shaft and double as a key while being close enought to the core to balance it out easily.
Or at least a brass collar that holds two grub screws 180* apart from each other, the grub screws interface with both the impeller and the shaft of the motor to hopefully act as an anchor point to prevent the impeller flying off the shaft.
Id like to see multiple fan stages, im no genius, but it can help like used in jet engines. Also id recommend to go back to hobby edfs and demo the right diameter one, i think they can have some inspiration.
Ever look at the shaft ends of a turbocharger? They use lock-nuts on them for a reason, and they use impellers designed much like yours... only more radical in some cases.
Design a motor with a larger diameter shaft to be able to handlle a keyway or even better make it with a shaft that goes past the compressor wheel that is splined, past the wheel the shaft can be threaded for a locknut or have a deppresion for a washer and lock ring.
At the speeds you're seeing, running unfinished FDM prints is really holding you back. You need to sand all of your layer lines so they are smooth, and you need to hand clearance the rotor with the housing (likely by printing the housing a little undersize and sanding it to fit, then ramping power until it rubs, then clearancing some more. Steps need to be done in that order to ensure performance. Or just print it with a resin printer. You'd still need some balancing and clearancing, but, it's going to be way less work overall. From there, it's just about optimizing your rotor design, which is probably close to fine as-is.
If it keeps pushing itself off the shaft why not just an interference fit? Cool the shaft with something like dry ice, make the hole undersized so it barely goes in when cooled and as soon as it gets to room temp it will NOT come out. Fluid dynamics isn't exactly my normal field so I can't help much in the impeller design but I can suggest using a high-percentage infill to counteract the forces trying to pull it apart. At higher speeds you'll definitely notice the sizing differences between stationary and spinning impellers.
Hi! Why didn't you start from original form propeller to create a new, more big or with other angles? Other thing, original propeller with a non brushed motor? Simply... Happy to see you! Continue!
You should use propellers made from resin mix Siraya tech tanacious and normal resin 50% 50%, this will give a very strong blades yet have some flexibility and not brittle like normal resin. There is a video discussing this in details search for it, I can’t mention it otherwise the comment will be removed.
I guess one way to have the turbine stay on the shaft is if the sucking end of it was pointed towards the motor, that way it would only pull itself more securely on. That being said making something like that work from an airflow perspective and work within the packaging confines of the original vacuum would be extremely difficult.
have you tried to get rid of the lower half of the blades(blend them in the middle) and at the lower part area to create a cone like shape to redirect airflow.
Great video again! Always something new and interesting even for us, the "older engineers"! One question - which tool do you use for your mechanical designs - FreeCad or another?
From an ex vacuum engineer - it isn't as simple as either 'max flow against no back pressure' or 'max pressure pull with no flow'. What matters is the inbetween, and whether your impeller performs well in series with the back pressure of the vacuum's filters. Vacuum cleaner companies use a metric 'airwatts' to characterise it - pressure times flow rate (in SI units) actually cancels out to watts, and you can compare your electrical power in to your air moving power out!
Thanks for the feedback :-)
Amazing info! Your old job sounds cool as heck too btw.
@@DrGooseDuckman it sounds like it sucked... a lot.
@@Komeuppance Get out.
also called P/Q-Curves
It's almost like the Bosch engineers knew a thing or two about designing a vacuum cleaner 😆
Haha true.......next time I beat them ;-)
It's like the one time where bosch engineers designed something that sucked and it was a good thing
It’s like the saying that anyone can build a bridge with enough steel and concrete, but an engineer will build one that is efficient in use of materials, adhering to time, safety, and budget constraints.
And have a little more resources available than just 3D printer
Actually the only major difference is that Bosch uses Glass Fiber PA6 nylon. So the strength is WAY up. But you can easily 3d print it.
You should take a look at the way cutting discs are attached to angle grinders. They are held on by a nut that's threaded opposite to the spinning direction. The centrifugal force keeps them tight, they will never come undone on their own.
that would be hard to implement here, then again he can fix a coupler to the shaft and drive a bolt into the coupler from the top of the impeller but that solution would be bulky.
@s.sradon9782 he could try to thread it, or get one that's threaded
At the risk of being pedantic wouldn't that be the equal but opposite tangential acceleration?
@@first-thoughtgiver-of-will2456 You're asking that with the bold assumption that I know what that means...
@first-thoughtgiver-of-will2456 at the risk of sounding pedantic (and maybe antagonistic) it's actually the interplay of normal/frictional/inertial and drag forces. Centrifugal plays no real role in holding the nut on, and acceleration only plays a part at the start of the motion and in the form of vibrations.
I hope you do part 3 and keep leveling this up. Even if you buy a dyson, its fun to learn this stuff!
Please do a part 3 and beat the original
Put a groove on the end of the motor shaft - spin it up full speed and use a triangular file - and then a circlet will keep the impeller in place.
Oh boy. Love that idea. Will try it in the future
To transfer the torque better and make it easily removable you could also use equaly spaced grub screws. or if you want something non removable you could grind a divit in the motor shaft and also some in the 3d printed hole and then glue it. That way the glue will not just keep the rotor from slipping but instead mechanicly lock it on the shaft
@@quelixfenzer5108 once you've gone to the trouble of grinding a axial slot in each, just go ahead and install a key: that's the traditional solution to transferring torque from a shaft
@@RossReedstromYou cannot key a shaft this small, that looks to be smaller than ∅5mm.
Well, you can - it just wouldn't make very much sense, shafts this size are usually secured using "dog point" set screws, with according holes in the shaft, or with cone point set screws if the shaft must not be weakened further.
Removing material for a keyhole from an already undersized shaft isn't a great idea.
I'd bet a regular pair of set screws with flats would serve him right and it would all be achievable at home using a file and a tap.
@lazar2175 if he really wanted to he could use an edm machine to make the initial hole and some specialized tap for hardened material. I don't know much on the tap side, but an edm can cut through any hardness of metal.
I just want to say that also printing an impeller removal tool was a very nice touch & not something I expected to see. Great job!
All thanks to Marko ;-)
@@greatscottlab Nice! He was definitely thinking ahead there.
The closed impeller design they used can make a massive difference for static pressure.
On impeller to motor connection just get some metal(alluminium) hub for plane/car model motors and make the impeller connect to the hub instead of the motor shaft. This will introduce much higher surface and possibly hard connection to it.
3d printing is not suitable (not yet) for something that needs to be as precise and strong as an impeller.
Create a black hole
Maybe next time ;-)
Buying the famous vacuum brand will create a black hole in his wallet all by itself
Add a boost button to overclock the motor for short periods of time, if you get in trouble spots
Good idea
overclocking vacuum cleaner, sounds cool
I can hear his girlfriend complaining that her knickers got sucked into the vacuum when she pressed the boost button!! 😆
I never would have thought I'd watch a 3 part series about a vacuum cleaner. Lol
3 parts? Only at 2 parts so far. But who knows what comes next ;-)
@@greatscottlab I'll be waiting. 😉
@@greatscottlab this person is ready for a 3rd part, just like me :)
you should try getting the impellers SLS or SLA printed. It should help with airflow by having a smoother surface, and the SLS one can be CF Nylon, which should be plenty strong and light for this application
For part three, maybe include a "turbo button" with a 30s timeout that runs the motor at ~95% of the maximum that the motor and controller can handle and run it at whatever level the lady deems acceptable all other times.
Dyson already does that
@@fusseldieb Oh my god. Noo waaay.
Roger that, I own one. I was curious about how our friend here would implement a similar function.
the actually vacuum cleaner is a turbo, but instead of using exhaust gases it has an electrical motor, whatever it sucks instead of going to the engine goes to the bag
@@chikogotaThey said a _turbo button,_ not a _literal turbo._
I'd like to publicly acknowledge how much Scott cares for his girlfriend. Clearly, her ability to quickly and efficiently vacuum his apartment - and, therefore, *her* quality of life - is hugely important to him.
He for sure is a huge feminist and not old school
@@gamesnic????
@@gamesnic???? What?? Caring about women is not feminism....
@@nikkiofthevalley it's exaggerated to be a joke.
If he really cared for his GF he would do the vacuuming himself.
Be sure to check out the Tineco Pure One. We have several clients with handheld Dysons that we occasionally use if we're just adjusting something or didn't make a big enough mess to grab our Festool shop vac from the truck. One of these clients replaced theirs with the Tineco when their Dyson died and couldn't be repaired. The Tineco is way quieter and feels less like it's been overly value-engineered (just feels more solid in general).
I have one and like it a lot. Can vacuum with the kids in bed.
Filters clog up though, every none dyson bagless system is kinda crap.
I am an engineer, i designed and 3d printed impellers for the wash and drain pumps on a dishwasher. I designed them for 8% improvement. The logic was that the motors had at least a 10% safety margin, and thus, 8% would be safe. I used ABS. That was 7 years ago. I still hear the stronger Swoosh-swoosh sound when it's used... and i made test impellers, which were thicker layer height for testing fit and function.... never removed them.... 7 years, and the rough test fit impellers are still going..
I'm not a vacuum designing engineer like the commenter below but I used to engineer centrifugal liquid pumps so I know a little bit about this subject. The static pressure will go up with impeller diameter and/or RPM squared. Google "fan laws", it will be valuable information. The blade angle of attack won't affect static pressure much because there isn't any flow through the blades during this test (weird I know). One cool thing about these kinds of devices is one might think the shaft torque would go up if you restrict the flow, but actually it goes way down. Notice that you hear a vacuum cleaner motor speed up when you cover the hose. The load on the impeller comes from the flow rate of fluid that is being accelerated so more flow = more torque and power required. What happens to the motor power in your experiment when you go from open hose to the static pressure test? If you cover the outlet the flow goes down but the fluid density goes up. I think it will still consume less power (I designed liquid pumps with constant density so I have no actual experience with this). If you can design a "volute" on the outlet of your pump that gradually expands the cross sectional area, it might improve the performance although I don't think you have the room to do this in your housing. The purpose of the volute is it acts as a diffuser to have pressure recovery as the fluid decelerates.
A grub screw will help with keeping the impeller on. It will have to be factored into the design. As far as the impeller is concerned you could try carbon fibre filament, it may be stronger. The other option is creating a silicone mold of it and using a casting resin or making a clay mold,and using aluminium to cast the part.
Or simply buy the cold side wheel from a car turbo which is pretty much the same shape.
30 year home repair contractor here, and I use the exact extractors in the video without any problem. You need to apply a lot of force on the drill, and go really slow- allowing the bit to bite.
Was in Germany last week and met some Rohde and Schwartz engineers. Lovely guys
Awesome! :-)
Try to saturate the printed parts in super glue or clear nail polish it helps solidify the prints and can achieve a smoother finish. 👌
Thanks for the tip!
Or - vapor polish some ABS :)
Would be a costly surprise if it actually sucked up all the water and spilled to your lab bench.😅
Happened off screen with the commercial solution. Luckily it was not too much.
@@greatscottlab that must pack quiet a punch!
@@greatscottlab Actually, RC Test Flight reports it happened to him too 😀
Hello Scott! Do you consider printing the impeller with other materials, for example ABS?
Because in general the density of ABS filament is about 10% lower than PLA and PETG.
Maybe saving a bit weight on the impeller can help the efficiency. Thanks!
Weight makes no difference once it's already spinning. Also more ductile plastics will deform at high rpm and rub on the outer housing
@@harrylenon9594Weight certainly DOES play a factor. Weight = LOAD. The more load, the harder the motor has to work. When the motor is capped at a set power limit.. it will reduce the overall RPMs with a heavier load. The load might also be amplified, as you move the vacuum around.. as the extra mass with create more vectoral forces.
@@johndough8115 no, if the load has the same friction with its surroundings, it will reach the same top speed. Acceleration will be higher with lower inertial load, but top speed will not change. At constant speed, weight plays no role in power draw.
@@harrylenon9594 There is a channel that makes Custom fan blade designs. Certain designs are heavier, use more blades...etc. They all spin at different RPMs, despite the exact same motor being used.
@@johndough8115 yes, different DESIGNS.... the design will determine the drag on the propeller while spinning, not the weight. If you did a test with two identical blades, only difference that one is heavier, you'd find the lighter blades spin up to max speed faster, but have the exact same top speed
I think you can take it a bit further. I'm very familiar with RC motors and we never use a press fit gear as they'd come right off the shaft like you're seeing. A simple fix would be to integrate a 48p RC pinion right into the impeller. You'd need to add some gcode to pause the print so you can insert the gear directly into the part and then keep printing to seal it in. This will give you teeth on the gear to engage the impeller directly and a small hole in the side of the impeller will give you access to the pinion's set screw. The set screw in the pinion will lock the whole assembly onto the motor.
GLAND SEALING SYSTEM.
is a technique used in steam turbines to make the bushing "air tight" without having contact with the shaft.
Try adding it to your design.
I would highly recommend screwing the impeler in place, either on the top or side of the shaft. I believe it is necessary for any fan system to function correctly.
I believe the cap on top of the fins, on the original, was to increase static pressure at lower rpm's .. also static pressure increases with multiple stages.
People often emulate automotive turbos without taking into account the facts. They have very tight tolerances, and operate at extreme rpm's, some at over 100,000 rpm. At those speeds and pressures, the chamber itself becomes the cap on top of the fins. The cap on the fins will leverage the lower rpm window for your materials of choice, and those often used for mass produced products.
Use resin printing instead for strong and better quality prints. Also use some kind of locking system for impellers.
Strongly fixing the rotor to the shaft will solve a lot of future problems.
Yep. That was the biggest problem. I hope to reliably solve that in the future.
@@greatscottlabi would prrsonally use a small worm screw that goes trough the impeller and grabs the shaft.
@@triaxelongd3337yes grub screws would be great for that, especially if he makes a lil groove in the shaft for it to bite in, or even better use an actual pinion RC gear for one of those motors and something in the impeller that can hold onto that
@@greatscottlab I would make the shaft something like a screw, or add a groove and use a retaining clip.
I wonder if a duplex impeller, with a single inlet and two outlets would balance the thrust forces against each other? There would have to be a new inlet scroll...
The "two" impellers would be "attached" at the inlet end.
You should try a resin 3D printer for that. It is possible that bc FDM printers aren't known for printing air tight enclosures. It is possible that there are small holes in the enclosure that prevents it from producing pure vacuum. Resin solves this problem
Those holes are pretty insignificant in size compared to the input and output of the impeller tho, even more so considering more than 2 walls, the overlap of the material, even with small voids, compensates it (think of the Swiss cheese).
With resin I would be concerned about the material strength itself, it's also not too great with creep, and even temperature. It would be like doing the part out of PLA but weaker still, with the benefit of it being extremely (visually) precise, sure there are good resins but they're quite expensive as well.
The surface finish will also be way better
@@Kalvinjj depends on the type of resin. There r many different types of resin out there other than the standard form labs. There are metal based resin and high temperature resins out there also. But I think in terms of cost, buying new vacuum would be cheaper than trying to go all out in a high end resin printer.
Do NOT purchase a Dyson. They are pretty and have great suction, but that's about where their benefits end. They get clogged up super easily and the plastic is super brittle. Mine was under warranty and they said I can't get it fixed because they don't have the parts for it. WTF??
That....sucks. Thanks for the feedback :-)
Yeah, you try vacuuming up a bunch of cocoa cereal with a Dyson cordless and it'll choke. The hole that goes into the debris chamber is so tiny.
@@kylekirby6424 That's what she said?
@@plixplux 😂
Great work. Vapor polishing would improve the smoothness of your prints and should provide some level of performance gain as the airflow over the blades should be smoother. You're probably losing a lot of efficiency to airflow past the clearances between the impeller and the housing. Perhaps SLA printing would be a better option, as it's higher resolution and it also produces smoother surfaces.
Or dipping the blades in dilluted plaster & letting it dry? (or epoxy)
This is where people mess up when cooling PC's too. All fans are not built the same. Some are made for high static pressure and some are passive. Don't use a passive fan for your cpu or any situation where the fan needs to pull or push through something.
du hast noch ein problem mit dem Luftauslass - die luft, die der impeller ansaugt, muss auch wieder raus. und da sind die kleinen fensterchen quer zur strömungsrichtung der luft. geschickter wäre die fensterstege längs, tropenförmig zur strömungsrichtung zu machen und denen so viel querschnitt wie möglich zu geben. dabei ist zu beachten, dass nicht nur die fläche entscheidend ist, sondern auch der querschnitt je fenster. wenig große fenster sind besser als viele kleine, auch wenn der gesamtquerschnitt gleich ist. luft mag keine ecken, jeder kanaleinstrom und ausstrom sollte es der luft so leicht wie möglich machen und möglichst kein labyrinth oder ecken entgegenstellen.
Just an idea... use a small turbo. Match the current motor rpm to a compressor map of the turbo. Then use the comressor, then turbine and see how they both perform.
You might be able to alleviate the impeller pulling itself off by implementing some set screws. This of course though means redesigning the impeller and also adding some rotational weight.
You must work with the surface finish on the impeller and the housing. Surface roughness greatly affects flow attachment and its properties (Reynolds numbers, fluid dynamics, bla-bla-bla; you can easily find a lot of free articles). Long story short, surface smoothness is critical in such applications, and losses in the system scale exponentially with the speed of the blades.
I'd file a small flat section on the motor shaft and use a grub screw to secure the impeller in place.
Let me know if you're interested in more details -- but I've made a similar type of impeller (and propeller too) with 3D printing -- the best way to secure it (I found) was to use small brass inserts and insert them into the round bit in the middle (in-between the blades) to screw into the motor shaft. Since an iron won't fit in the small gap, I would screw in the head-less bolt into the insert, hold it with needle-nosed pliers, then heat it up with a blow torch - then pushing the insert into the plastic between the blades. I would then cut down one of those head-less bolts (forgot their name) so they wouldn't stick out and use an allen-key to tighten them on EITHER: 3 sides | OR on the top and bottom of the motor shaft for more even securing. And BOOM can run it as fast as you want and everything -- hope this helps :)
Use acetone to smooth the propeller, stick pieces together and make stronger joints.
Just putting the 3d print in a acetone bath for a couple of minutes will not only smooth it but will increase the strength to a large extent.
Maybe make the impeller closed with a skirt to minimize boundary layer drag and skin friction between the impeller and the enclosure? Polish those impellers and minimize any layer visibility, every ridge creates extra drag, which creates trubulence. If you're doing Part 3 and want to try with an aerospace engineer, I'm down to take Bosch on 😁
Some BLDC motors just have a nut on the end that you can use to ensure the prop doesn't walk off the shaft. Some are even reverse threaded such that when the motor spins, the screw tightens down on the part. You could also go for a keyed shaft to reduce rotational slippage.
Weird suggestion but what if you sand the impeller blades so that the air can slip better. Overall great vid!!
Drill and tap a small hole in the shaft of the motor and put a washer around a screw to hold the impeller on.
You just need a longer shaft so you can thread a nut or stopper on the end. That's how turbos keep their impellers on, or even some sort of C-clip would work.
If you want, I design and work on these for a large turbomachery company (the big blue ones), and can offer you some pointers, both with how the impeller is attached, and why you are having such issues with static pressure
Have you tried to use the original rotor on the different motor? For science
If you decide to buy one get a Miele at least, they are fantastic vacuums! There is not only Dyson! Maybe you can also ask them for some infos about theirs since they're also from germany. (I think there vacuums are made in asia though, still great quality)
You should consider making a closed impeller design. They are great for static pressure.
Just a heads up PCBWay does polymer SLS printing. Maybe they'd sponsor a V2/MK2 of the 3D printed impeller.
JLCPCB does that too ;-)
@@greatscottlab
Ya know, I might be wrong and it's actually JLpcb. I'm not sure now, I'm gonna ask the almighty google 🤣😂🤣
Edit: it's both : )
^
You could have used a set screw and a motor with a shaft that has a flat on it to counter the impeller slip.. glue never holds…. And the tolerances between impeller and compressor housing has to be very tight for that style to work.. is why most vacuums use that flat style..
You use an open impeller design, which causes recirculation flow around the impeller blades, lowering the efficiency. Try using a shrouded impeller and reduce the clearance around the impeller and the shroud.
Vapor polishing would improve the smoothness of your prints and should provide some level of performance gain as the airflow over the blades should be smoother.
I'd be so curious to see the work/math behind impeller and housing design. It is such a mystery to me, and the resources online aren't too great.
Perhaps with some post processing the FDM prints(sanding, smoothing and sealing), You might be able to get a few percent higher efficiency there. I recall rctestflights videos where he mentions a significant performance penalty with propellers and impellers that are FDM printed without post processing work being done.
I wonder how much improvement you'd see from smoothing out the 3d prints. In theory all the ridges from the 3d printing create disturbances in the air which would reduce efficiency. There's a reason why propellers are as smooth as possible.
There is a reason why they are not. Actually, performance fans are rarely glossy smooth. The only example of consumer fan with glossy blades I can remember is an office fan, the big thing that blows and swings around, definitely not a performance hero. I don't know the exact explanation but I'd say near-surface air layer is unimportant, at least at relatively low speeds, so nobody cares. Edges and, in case of static pressure, gaps are far more important IMO.
U can etch a dimple on one side of the motor shaft and add a security screw to keep it in place so it doesn't fall off. Belt pulleys use it like that to keep them in place.
If you sand a rought patch on the motor shaft and then put the glue in the bottom of the impeller before pushing it on it will handle alot more speed before the glue fails
To prevent the impeller from running away from the shaft, maybe you could make a grove at the end and put a circlip there. Also, cut a flat surface on the side of the shaft to prevent the impeller from slipping on it. I woodworking, it's well known glue is always the weak part of an assembly: you need to think your assembly without glue and add it as reinforcement, not as the primary transmission.
or drill a hole in the shaft and thread it so he can put a screw in perpendicular to the shaft
@@Scotty-vs4lf you wouldn't do that, because the pin is going to tear into the plastic at startup dur to excessive torque. you really want to maximise contact surface and avoid pins. half moons would work.
@@RegisMichelLeclerc yeah i figured the inertia of the impeller was low enough that it would be fine but i have no experience so ill take your word for it
In my opinion You need tighter clearances to improve the seal. Unfortunately I’m not sure how realistic it is to do this using a 3D printed impeller. Perhaps a resin 3D printer could help u get a more precise print?
Dyson is not the only obvious choice when getting a new portable vacuum. We have a Xiaomi G10. Very strong and reliable so far.
i wonder if a resin printer would give a smoother finish to the impeller blades and if that would maybe help it pull air more efficiently
Probably it will decrease the amount and strenght of the vortexes so i think yes. Maybe just a minimal amount
Nice effort, part 3 will entail installing a wall mount with 2 screws for your shiny new Dyson.
heya, if you ever look for a nice vacuum that doesn't cost as much as the dysons, you could try some of the xiaomi ones, they are actually surprisingly good
Thanks for the tip
Impeller balance will play some role in negating performance at higher rpm’s. That being said I believe you’re on the right track, I figured you’d buy a metal turbo impeller and try that but I respect the commitment to designing your own!
The stock propeller seems to have a smaler inlet 🤔 this is probably better for creating pressure or vacuum. I bet you lern much by reading about turbos, comparing turbine maps and the physical demencion and drow some concussion between turbine well size on inlet and bottom in order off pressure vs airflow map
there should be a distinct eye in the impeller, and quite long blades for this. if you end up with a smaller eye, and more surface area on the blades then you should be able to have a lower static pressure, meaning more suction, therefore having a potentially better vacuum.
These two videos just reminds me again how hard and complicated fluid dynamic is, and paying the price for geniuses to do it right is well worth it for my dumb mind.
Have you considered finding a motor with a threaded shaft and using embedded nuts in the impeller to connect it? When spinning in the correct direction, it will self-tighten itself to the shaft.
Its always a good day when you upload 😊😊😊😊
Happy to hear that!
You could have used original impeller and new motor that might be the best way to achieve efficiency.
And for benchmarking you can use dc motor with new impeller as bldc tends to consume more power at higher rpm and provides very less torque.
A pin through the rotation shaft interlocked with the impeller would solve your problems I guess.
the axis of the motor is smooth therefore it doesn't bite well into the plastic. also I see sometimes a bolt perpendicular to the axis, that prevents the axis from slipping from its position, translation AND rotation are blocked this way.
Out of curiosity, a question to our mechanical engineers: How much does the surface finish affect the outcome in such a system? The printing looked quite rough tbh.
not much i think. You're trying to generate a huge negative pressure, you don't need turbulent free airflow for that.
Maybe look at an Angle Grinder...they've (Some at least), have a nut screwed on the wrong way.
I think that in your case, you could even look into drilling a hole into the shaft then bolting your impeller to it. A bit drastic, I know, but that's how I bolt my large 20 Inch drone props to my drone and it's worked fine for upwards for 5 years now.
I got a Bosch Ultimate 7. lighter than most Dysons, sucks about as much, uses normal Bosch DIY battery's and has a nose so you can stick it in corners easier.
oh and costs about a third compared to a Dyson.
Wow very interesting video! I have done a lot of airflow and suction measurements and most important for good cleaning power is the airflow from the floor nozzle. I have dedicated airflow/suction box for that. I can seal floor nozzle to the box and get very accurate readings.
Wow Scott, watched your videos many years ago.. never ever did I thought to see the man behind the voice after all these years
one big factor that seems to impede your design is the use of a 3D printer: the end products its produces are far from smooth, and if you know anything about aerophysics youll know that that is quite significant. consider sanding down your prints or making them ever so thinner and then applying a layer of lacquer.
Looking forward to seeing part 3
I bet SLA prints work much better, also an air flow simulation could help to evaluate your designs.
Idea:
Why not texture the shaft a bit and print the impeller with a bit of an undersized hole and use a soldering iron to heat the shaft to melt the plastic onto it?
Like a heatset thread insert.
Why not using a copper core of a wire strip to joing the impeler to the axle? The screws will hold tighter to the motor shaft and double as a key while being close enought to the core to balance it out easily.
at this point, you need to concider metal insert in the plastic and press fit on the shaft of the motor. nice video keep it up :)
Or at least a brass collar that holds two grub screws 180* apart from each other, the grub screws interface with both the impeller and the shaft of the motor to hopefully act as an anchor point to prevent the impeller flying off the shaft.
One thing you could have done is reduce the intake diameter for the casing that houses the dirt to increase the suction pressure
Id like to see multiple fan stages, im no genius, but it can help like used in jet engines. Also id recommend to go back to hobby edfs and demo the right diameter one, i think they can have some inspiration.
10:35 in part 3, i think you will be going to buy a brand new vacuum cleaner with your girlfriend 😂😂😂😂😂 have fun 😁
Haha
Just a tip man.. maybe if you finish the surface of your 3d printed impeller you can decrease the friction of air on it, thus better performance.😊
Ever look at the shaft ends of a turbocharger? They use lock-nuts on them for a reason, and they use impellers designed much like yours... only more radical in some cases.
You should use a reverse thread screw to secure the turbine blade to the shaft with a lock washer to keep it from coming loose.
Design a motor with a larger diameter shaft to be able to handlle a keyway or even better make it with a shaft that goes past the compressor wheel that is splined, past the wheel the shaft can be threaded for a locknut or have a deppresion for a washer and lock ring.
At the speeds you're seeing, running unfinished FDM prints is really holding you back. You need to sand all of your layer lines so they are smooth, and you need to hand clearance the rotor with the housing (likely by printing the housing a little undersize and sanding it to fit, then ramping power until it rubs, then clearancing some more. Steps need to be done in that order to ensure performance.
Or just print it with a resin printer. You'd still need some balancing and clearancing, but, it's going to be way less work overall. From there, it's just about optimizing your rotor design, which is probably close to fine as-is.
You can print a bolt hole from the side of propeller, and bolt another filament bolt .That might hold it
If it keeps pushing itself off the shaft why not just an interference fit? Cool the shaft with something like dry ice, make the hole undersized so it barely goes in when cooled and as soon as it gets to room temp it will NOT come out. Fluid dynamics isn't exactly my normal field so I can't help much in the impeller design but I can suggest using a high-percentage infill to counteract the forces trying to pull it apart. At higher speeds you'll definitely notice the sizing differences between stationary and spinning impellers.
Hi!
Why didn't you start from original form propeller to create a new, more big or with other angles?
Other thing, original propeller with a non brushed motor? Simply...
Happy to see you! Continue!
You should use propellers made from resin mix Siraya tech tanacious and normal resin 50% 50%, this will give a very strong blades yet have some flexibility and not brittle like normal resin. There is a video discussing this in details search for it, I can’t mention it otherwise the comment will be removed.
I guess one way to have the turbine stay on the shaft is if the sucking end of it was pointed towards the motor, that way it would only pull itself more securely on. That being said making something like that work from an airflow perspective and work within the packaging confines of the original vacuum would be extremely difficult.
You could also always modify the top of the rotor shaft for mounting
This is the first video I’m watching of yours but I just wanted to say your intro was beautiful
have you tried to get rid of the lower half of the blades(blend them in the middle) and at the lower part area to create a cone like shape to redirect airflow.
Great video again! Always something new and interesting even for us, the "older engineers"!
One question - which tool do you use for your mechanical designs - FreeCad or another?
Probably fusion 360
He uses Fusion 360, eg ua-cam.com/video/0P4R91hsMf0/v-deo.html.