Quick proof-of concept to see if you can 3D print a working single stage turbo compressor. Lot's of lessons learned and I hope to revisit this again. #turbo #turbocharger #3dprint #centrifugalcompressor #compressor
I'm amazed at how your channel is as small (relitively speaking i guess, still an unimagineable ammount of people (2.56 THOUSAND in a room?) ) as it is, but "Oh ya know, just designed and FDM 3D Printed kickass F-1 componet models" "Oh ya know, just researched and designed a Centrifugal Turbo Compressor then FDM Printed it" AMAZING stuff, keep up the awesome work !
I was surprised it was balanced decently. The runout was terrible too, but it’s so lightweight. Definitely want to try a piston style pump at some point here. There are actually so many types of compressors. There’s a couple of other styles that may have a chance at working.
Centrifugal compressors are fantastic for volumetric flow, but unfortunately not for pressure, in fact compressor surge is a major issue with these machines, if there is insuffiencient flow and high pressure, the flow can reverse going through the vanes and destroying the impeller. They're commonly used in industry for low pressure / high volume air applications, or on LNG tankers to send boil off gas to the ships boiler. Its a cool approach to arrange the stages axially, all of the multi stage machines I've seen simply take the output from scroll the volute and feed it into the inlet of another compressor stage
Yeah, I did get a massive amount of volume. I'm working on a V2 which has been decently promising. I think for 3d printing, the multistage volute to inlet configuration you mentioned will take a lot less material and time design.
You should look up regenerative blowers or side channel blowers. Neat way to get somewhat higher boost at lower flow than with centrifugal blowers. Also, invest in some colored water and clear tubing to make a simple U-tube manometer. You were probably making 5-10 inH20. Which is quite good for something with a tip speed not measured by Mach number.
If you want 5-10 psi and decent flow rates with a 3D printed pump, I'd suggest that you look at positive displacement diaphragm pumps. Just need a few sheets of rubber for the valves and diaphragm. (though thin printed reeds and bellows might work well too)
Fusion has its benefits like just being able to quickly design organic shapes, but solidworks, if you know what your doing, is a lot more precise, and can way more easily design production parts
I agree. I can work so fast in Fusion that it would be hard and expensive for me to switch to something like Solidworks. I've learned a lot more about how to do this properly, and I'm slowly working on a 2nd version that corrects some of my mistakes in this design.
Amazing designes, I work in factory automation and I happen to have delt with instrument air axial compressors. The ones we have are 4 stage compressors, first stage is the biggest stage sucking huge amounts of air then goes to interstage rad fan style cooler and then to the second stage, the process is the same throught to the 4th stage. Now the second is smaller than first and 3d is smaller 4th is the smallest. The ratio between the drive gear and the dreven gears, one was called high speed and the other is the low speed, was so high like 1:100 for low gear 1:150 for the high gear or even more " not exaggerating". I have actually seen inside of them during overhaul and noticed and linked what i have studied about axial compressors. In order to comperss you need rotor and stator blades. Rotor will move the air at very high speed, now this is were your designs stopped and in this stage it is a blower not a compressor. The secret is stator blades are essential to creat an effect in axial compressors called defusion. Air is comming at high speed by the rotor, stator will change air direction and reduce its velocity, now you are creating pressure and then the max amount of pressure will be limited to multiple factors of efficiency of the designe and clearences between the housing and rotor leakage. As for yor multistage design I have noticed that all stages are the same size which indicate no pressure increase throught the stages + the missing stator part in each stage. Hope i helped and i really look forward to you next videos. 😊
Thanks! Ideally I would create multiple compressor sizes but definitely more time consuming. I had thought about the stator vanes in the diffuser, but based on what read a vaneless diffuser has a broader functional range and that a vaned diffuser while more efficient has to be designed properly. Is that true? I don’t have a good way to CFD the compressor during design, but if I just copied the stator off another compressor would it work? It’s hard to find this information without buying expensive textbooks.
That's awesome! You must have had a massive motor. I know that Nissan test nylon impellers at one point so it's definitely a possibility for a long term upgrade.
I was actually looking for an open source "inline centrifugal fan" CAD (mainly for 3d printing it, and using as a low cost componet for a dessicant-humidifier AC) and by accident (was subscribed for some other video) you made exactly what i need essentially !
Cool turbo To smooth your surfaces: I´ve seen on another channel a guy, who puts his parts in a box and soaks the printed parts in acetone vapor (not in the liquid itself)
Well, I used a 540 brushed motor which was a. It too small really. Obviously there are a lot of motor choices, so it comes down to how much power you need, and how many RPM you’re willing to spin.
I didn’t do any stress test on this. You can use plastic materials for the analysis but the challenge of FDM 3D printing is that materials are much weaker when loaded against the layer lines. So the only way currently to test them is real world testing.
This is all great, but I have been looking through all that information, likely the same papers you've read, in an effort to get the Euler turbine equations narrowed down to generate geometry, but haven't been able to do so. Did you strike gold on a paper that you can point us to, or rather, can you make a video going through the design equations for a centrifugal pump?
I have to caveat this, but I’m not an engineer or expert on this. In my case I actually watched several demo videos of Ansys Blademaster and I actually took the detailed angles from their setup. This is also a very simple older compressor wheel design that allowed me to just use a few points and interpolate with splines. Not the most accurate, but it should exceed the FDM printer tolerances. This paper has the details on axial compressor design, which I used to lay out my blades. www.seitzman.gatech.edu/classes/ae4803/compressor_angles.pdf. Because my blades don’t twist at their root, I could lay out a blade essentially like an axial compressor with the angles from the ansys demos. I’m not sure if this is right or not. I’m working on a version 2 that I have an actual compressor wheel that I’m measuring and trying to model. I can tell you after playing with this, a dedicated CAD suite for designing turbo machinery is a must for designing a modern compressor wheel.
@@IndeterminateDesign First off, thanks for the reply. Second off, it's a shock to hear you say you're not an engineer on this; the only thing I can assume from that is that you're not specifically an aerospace engineer who specializes in turbomachinery, since my brain can't accept that you're not an engineer in any capacity short of that. If you actually aren't an engineer, I hate to break it to you but you are. :) I checked out the paper, it definitely helps provide vane geometry and overall velocity vectors, but what I was hoping for was a set of design or sizing equations that may provide equations to determine the following, which could be manipulated based off design constraint knowns or unknowns: - Vane count - Slight input on Vane Curvatures (fwd, bkwd, twist) - Staggered Vanes - Inlet diameter, outlet diameter, volute housing diameter ratio (if it's not linear growth) - Expected CFM - Expected Pressure/Velocity ratios - Expected efficiency And of course all of that would be based on the following requirements: - Gas properties (Reynolds, Z (compressibility factor), density, temperature) - RPM - Power - Pressure/Head I found so many papers on the design of it, all of which had the above variables, but didn't have the above outputs. I did find some that drove what type of fan to choose based off the application, etc. Regarding the V2 design you're working on, and the modeling, I devised pretty simple ways to model the geometry, but that's in SolidWorks, though I can't imagine Fusion won't have similar features. Essentially in SolidWorks, I: 1) Create an oversized disk that is the bottom of the centrifugal blade. 2) Create a plane that is the height of the blades away. 3) Create a "loft" geometry that essentially is two rectangular profiles clocked at a small angle from each other separated by the height that are lofted to each other on a curved, 2D profile. 4) Loft two circular profiles in the center to create the "body" of the blades. 5) Create a circular pattern of the loft dependent on the number of blades desired, ensuring that the rectangles overlap in the center to create jagged/overlapping geometry that i merge. 6) Cut the center out to be a constant diameter bore, cut bearing journals in, cut the disk against the blades to get the outlet diameter. 7) Chamfer the edges of the blades. 8) Fillet the geometry for machining and stress concentration concerns. I'm sure this is all TMI but figured I'd tell you so you'd have it if you wanted it. Thanks again for the reply man!
I'm an automotive performance builder and an electronical industrial control technician with a number of engineering skills acquired through years of experience. I've been experimenting with brushless motors and centrifugal superchargers for a few years now and have recently considered forced induction for a small block V8 consisting of four pairs of mirror image centrifugal bldc motor-driven direct-drive compressors each on its own stack to an individual-throttle-body (ITB) EFI setup. T-3s like they use on Isuzu 2.2 diesels would make a good model and since there's no turbine they don't even have to be metallic. Most small blocks ran 750 cfm carbs N/A so distributed a couple of hundred cfm each would probably be adequate. At least 15 to 20 psi static would be a great target. 40,000 is a pretty safe impeller speed and small ones would have little inertia. Just something to think about. There was a Dutch co. that did bigger T-4 versions with 48v brushless motors (Torqamp) but they're already history. Apparently covid blew it for them; but they were way pricy anyway.
That’s awesome. I hadn’t thought about using a T3 wheel, I actually have a couple lying around. I’m working on a v2 of this design that I’d love to take up to around 50k rpm. A brushless motor setup would be awesome. Given enough money and time, I’d love to recreate a split turbo like Mercedes uses on their F1 engine with a giant brushless motor to eliminate turbo lag.
@@IndeterminateDesign I'm wondering if I can use some of those high speed bldcs like they use for RC modeling to drive 3D printed multi-stage axial compressors to use on eight stacks for a couple of rows of individual 3D printed throttle bodies for my 1sr gen Chevy small block.
@@IndeterminateDesign That is a good point; but as long as you're sure you're not going to grenade the impeller I don't see it as any more risk than torqing your own bottom-end. I don't think 40K rpm on an OEM impeller is much risk at all especially on something small. Bench testing will find a defect if one exists. I'm thinking axial now though. That's what jets use; but I've not seen a single static dead-head pressure test done with one to date. All they're interested in is thrust which I'm sure is relatable to cfm and pressure somehow; but I've yet to learn to calculate it. I bet 8 of those bad-ass 50 mm ones with the high-speed bldcs would make some significant boost. I was even thinking of using a real fueled mini jet engine to push turbo turbines to see what kind of rpm and boost could be made that way. That radical enough for ya? LOL
That's an interesting application I hadn't really thought about. I'm slowly working on a version 2, I may have to try measuring the temperature differential of the air coming in and going out.
It may take me a bit, but I'll get this posted to my github. I wouldn't call it a complete success. I've started a second version because I've discovered several mistakes in that design. The motor is just a 540 brushed RC motor, 27T.
The types of projects you do, you should invest in angular bearings as they can do both thrust and radial loads. You need to mount them opposite orientation to lock thrust in both axial directions and keep them from falling apart. Another design flaw is that you need to calculate the intake area to a smaller exhaust ratio not the wrong way around as you have here. A ratio of 1.5:1 per stage for 3d printing would be stellar. Maybe a little less at first like 1.4:1 Don't give up, you'll get there. Keep up the great content.
That’s a great point. I did order some angular contact bearings. Aliexpress is great for that. I feel like I need a much larger motor and compressor really. The bigger the compressor the bigger the tolerances can be.
@IndeterminateDesign Lilly bearing is a good company to deal with and they have great prices. Done several orders with them over the years. Know what your project needs and don't over engineer the bearings for prototypes. There is no need for extra expense until you are engineering for lifespan longevity.
@IndeterminateDesign For higher compression use high speed RC motors 20000Kv or more and make sure to lighten the design as you move out on the radius of your impellers. Also flaten out the impeller design and add more stages. Because you are 3D printing you should use an enclosed impellerblade rather than open like you are currently doing. This will let you increase clearance around the case to allow for stretching at high RPM without losing compression efficiency. Hope this helps you with some ideas.
Maybe. There are a few axial compressor models that people have made. There are a couple of weird rotary vane compressors I'd like to try at some point too.
I think I've lost most of those links. I'd honestly recommend checking out the book: www.amazon.com/Radial-Flow-Turbocompressors-Analysis-Applications/dp/1108416675 For the second version of this compressor I just ended up copying an existing turbo compressor wheel and it was much better performing. I ultimately found it's really difficult to design a turbo compressor without CFD so you can tailor it to the application.
Honestly I found a RU-vid demo of Ansys Blademaster and I looked up each field from their config screen, like Beta angle. That led me to some grad papers with different designs that people were researching. I don’t know much about Solidworks, but unless there is an add-in package I imagine you would have to design this the same way I did with multiple sketches and surfacing.
You might wanna see If your design is a good candidate for a skirt compress for a (RC) Hovercraft. You could skip the coiling stage. So it would not be much heavier that a standard EDF unit but it might be more efficient. Anyways, nice demonstrator!
@@IndeterminateDesign I wasn't sure either. What would you use as a metric, Electric power needed for a given airflow at a Given static pressure and Aperture?
I'm not sure, I think you're right that it would be pressure and aperture would determine how fast the fan has to spin and the reynold's number. The centrifugal compressor will definitely create more pressure, but a single stage axial fan will move more air. A multi-stage axial probably wins for a full size hovercraft. It's hard to find information about this kind of thing on the internet. I'm sure it's buried in an expensive textbook somewhere.
It’s been a long time since I designed this and I don’t have the papers available. I do know for a leaf blower they usually use just straight vertical blades. This blower creates a lot of pressure but the flow isn’t very high.
Axial flow turbocharger...it is possible. A Jet engine is nothing more than a GIANT axial flow turbocharger that can run on its own exhaust energy. Take a look inside a high bypass turbofan engine. The multiple stages of compressors can create some serious HP.
choked flow is absolutely not the same as reverse flow 08:05 in terms of a compressor characteristic, these two states are on the opposite ends of the characteristic curve
If you are working on a new and improved design, can you throw the design files for this up on github? I'd like to use a well thought out design as a learning tool for myself and my kids and slap it on different rc toys and and teach them through playing and experimenting
github.com/indeterminatedesign/CentrifugalCompressorV1 Make sure you are wearing safety glasses when operating this. Print the wheel at the lowest layer height you can and at 30-35mm/s.
Yeah, it moves a lot of air for such a low power motor. I may need to pickup a cheap brushless motor and really crank this up. I also took apart a broken turbo yesterday and modeled it up. I found several mistakes in my design and little tricks the the big turbo manufacturers use. I didn't have a diffuser in the compressor housing, which could account for the lack of pressure and lots of turbulent flow.
Yeah, check out my 2 stage compressor where I used an adjustable restrictor to measure the peak pressure. ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-9YTMmGZ1d0g.html
That would be an awesome project. I’m not sure how it would work, the dust extractors I’ve seen have a simple vertical turbine blades, but they may just do that because of the cost. github.com/indeterminatedesign/CentrifugalCompressorV1
@@IndeterminateDesign you actually solved this problem for me as well with your multi stage design. I’m going to design a shroud around the impeller and hook that directly up to my ducting.
Definitely for FDM 3D printing on sub $300 machine. There are some cheap resin printers out now that are capable of 0.1mm accuracy. I just don’t know if I want to deal with the mess of resin.
you are using the wrong pressure tester! use a manometer from a medical tonometer! it is very sensitive and shows up to 0.4 bar. I hope I can see what you do in the next video as well! Thank you for your attention! if the arrow twitches, then just add a piece of sponge for washing dishes inside the pressure gauge hose
Yes, it was the wrong pressure tester and deadheading the compressor into a tank doesn't work for this type of compressor. Check out my Turbo Compressor Version 2 video, it's a much better setup with better results. ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-9YTMmGZ1d0g.html
1:55 I THINK I HAVE THE SOLUTION. Your blade is not efficient at all at the base. The mouth or top of the blade is fine and sucks in air like it should but the base of your blade is at the wrong angle. You have something that looks like the straight spokes of a wagon wheel looking from the top down. You need to angle them backwards more like the look of a hurricane. Your current setup pushes air tangent to the direction of the wheel and it can’t get out of its own way fast enough for more air to come in efficiently. Basically the fan on top is doing all the work. The idea as I’m sure you’re aware is to throw air to the outside of the housing. So by curving the base it will help push the air outward more and create better pressure. All the other dimensions are fine, you basically just need to grab the bottom corner of the blade and move it counter clockwise a little. The idea of a turbo is at its core an impeller with a fan attached on top. I think this is why you aren’t making any boost. Another idea you can try is something I saw on a rumba fan. You have trouble with the clearance because it’s 3d printed and it will never be good enough because that’s just how 3d printing is. The way rumba got around this is putting a cone over the blades and attached to them. This makes the compressor wheel a closed system and your tolerances between the wheel and the housing can be utter garbage and still work perfectly. I would recommend looking up a picture of their vacuum fan to understand what I’m talking about better. But I would try that above all else. Hope this helps and good luck. Can’t wait to see more.
hey bro, would ou like to share the file of it, beacuse i need to print something like this but i cant do in blender 😢that's why i am here :D ( if u want to share i would advice you to share it in parts, beacuse printer propably will connect the parts of compressor and it will not work)
I have the later version 2 on my GitHub. It is a single .3MF file but slicers can handle separating the parts automatically. github.com/indeterminatedesign/TurboCompressorV2
LOL A tenth tolerance isn't possible with FDM. Even high end CNC machines have difficulty achieving that. With shrink and deformation of plastic, it is highly unlikely. A .0010 tolerance on FDM would be more achievable - difficult, but more realistic. And at that point, one would need to consider surface finish as at that point you would only be measuring the high points of the surface with any traditional measuring method.
Even if you got that rotor within those clearances while stationary, runout from the bearings and expansion from the centrifugal spin force would totally muck them up.
Good observation. That was a huge issue with v1. If you check out the v2 compressor I added a large vaneless diffuser and actually made a little boost.
@@IndeterminateDesign also both v1 and v2 have an agressive blade angle of attack at the intake. And both seem to have a rather large intake which increases flow (and power needed to run them) but doesn't increase pressure. The pressure rise is determined by the radius of the impeller
1:35 You right.... its not easy job "Design a compressor wheel" 😑 And on some CAD software it becomes even more difficult. SolidWork = too expensive. AutoCAD = expensive. SolidEdge = poor performance. Fusion 360 = its a pooph cad 💩