I was a service technician now a sales representative for a very large screw compressor manufacturer. To get high pressures usually have a two-stage compressor. The first one goes to 35 pounds the second stage of bringing it up much higher the temperatures of an oil less two-stage compressor at each stage is somewhere in the range of 360 to 420° and the rotors are cooling from the air that’s moving across them
That would be aweome, a two stage setup would be awesome. I don’t know that the air got that hot, I saw some of the dry compressors had intercoolers between the stages.
Great project! I'd recommend thrust bearings on the low pressure end and placing your timing gears (and motor) on the high pressure end. You can also offset the axial bearings according to the tolerances and play to get a better seal at full operating speed (knowing the forces and directions of force while in operation allows to minimise clearances by designing for the dynamic state rather than the static). Nikola Stosic has really done fantastic work to modernising screw compressor design and I'm very grateful for his work. I'm working on a long term project that incorporates variable geometry screw compressors, hoping to be able to CNC machine the main components for the prototype, but I'm not sure I'm up to the mathematics required!
Thanks! I love idea of a variable geometry screw compressor. That makes my head hurt just thinking about how you accomplish this. That's great to know you can slide the the rotors relative to each other. There's so little information out there about how these things are setup. I think in the future, when I add a thrust bearing I'll also make the thrust adjustable with some set screws. Filing the spacers by hand was a nightmare.
Bravo! This is very similar to the oil pumps we use for fuel injector oil pumps on our big boilers. Main reason is they can handle #4 oil which is dirty and they are very reliable. If you have a 2 meter drop on the intake pipe, you can get almost a perfect vacuum and if used for that, you can extract all gas exhausting into a warm tank. Try it! I have wanted to do this in plastic using water with mineral oil/sulfur which is basically cutting fluid for machining- but it would be easier to watch someone like you try it first😊
Impressive result, really! Definitely add some thrust bearings if you revisit this. I only just added them to my own hardware vitamin collection and I wish I'd done so sooner -- they're not expensive and are an easy way to add really solid axial constraint versus solutions that only rely on standard bearings.
@@IndeterminateDesign I know the feeling! That's where things stop being "not expensive". Kind of like when you want to refill your favorite fastener sizes without buying whole assortments with sizes you never use :P
Amen Indeterminate Design. I got a huge smile on my face for you and your accomplishments. You got it figured out! Great Job. I had been wondering if the main rotor had to be larger than the gate. The previous discharge will work now that the rotation has changed. Eaton has some additional slots adjacent to the opening that my guess if for sound....Pressure is not the performance metric you want, it's flowrate and air temp differential actually mass flowrate not CFM for application I am interested in. Got to have flow to keep the SC cool. My gut tells me plastics have too much expansion for the clearances needed even w/ the air flow to last any substantial time but look up the coefficients of expansion for various material and you can make the assessment. I believe the rotor exhaust end face to case clearance is a big contributor to performance any why thrust is so important. Tapered roller bearing and timing gear mod to include oil bath definitely needed in next model. I believe the main can not exceed 360* twist. If it does the air would be trapped in the gate at the sealing lines made by the two contacts by the main. The blow hole will have more of a negative affect on flow including at the suction port. Increasing temps could also. I hope you try again. It's easy for me to say since I'm not doing the work; but, if it were me I'd fix the damage, cut open the top of the case enough to get a good look at the rotors, put clear acrylic or hardware cloth over that and any other opening, timing gears, belt or pinch point for personnel protection (I like you humor), slow the motor down enough to see the rotors mesh (30 RPM maybe slower) then put it on my office desk...Demonstrations at lunchtime only. LOL I think you know what I mean. Really good job explaining a very complex and multifaceted project design. Best I've seen. No reason for me to do any more studying now.
I’m glad you liked it. I think the resin rotors would have been better for expansion but it’s hard to say. It’s really a mix and match game to with a bunch of copies of rotors to get the right fit . I think I could have an interference fit with resin rotors and just run them in. It would make a really cool demonstrator for your desk. I want to make one with no housing as a piece of kinetic art.
I agree with your assessment however that’s not very “fun”. Print it out of high temp filaments and go right for the full send. This is TH-cam not LogicalTube.😂
@@IndeterminateDesign Thermal Expansion is always consistent, you would need to calculate the formulas to get the dimension but all of that stuff is sorted out these days. Fail fast is the most efficient way of egineering. FarmCraft101 has a really nice video called "3D Printing Metal At Home?" It is incredibly useful and fun to watch.
Great stuff! I know the 3DP dimension fudging dance all to well. It's annoying because if you ever change to a better manufacture method, then you have to un-fudge them.
Good work on this. I think your original rotors may still be usable because screw vacuum pumps are a thing. They work well, but they are a HUGE pain when a contaminate gets in there. Someone below recommended thrust bearings. A thrust bearing+ball bearing would work well, but you can also get angular contact bearings which does both (but not as well as the 2 separate bearings).
I can only imagine. The mass of all that steel spinning at high speeds would be scary. What kind of rpm did you have to run? The SLA resin rotors were actually pretty heavy. I think it would be like 2500J if they let go, or a small shotgun blast 😂.
Used to work in a compressor/tool sake & repairs shop. (Untill stress triggered undlying Fibromyalgia) This is awesome work. I Am noticing that grinding noise. I would gues its from the screws making contact with rough print surfaces. Have you condidered marking an alignment point, coating the screws in a lapping compound, match them up, seassemble, let then gribd/match/self surface themselves in? At relatively low/rock tumbler speed, for a considerable time. That should let then polish/grind themselves out. This Would also tell you how Well they screws mesh.
The grinding noise is just the timing gears mainly, and possibly the poor bearings. The rotors themselves don't actually touch, they have about 0.2mm or .008 inches clearance between them.
You can enhance the frictional properties of 3D printed parts made from resin by filling the resin with a small amount of PTFE. At least if you print yourself.
given your research into the tooth profile, I have often wondered why something more common like the involute gear profile such as used in hydraulic gear pumps is not used?
These profiles are optimized to compress the air along the length of the screw and to pull in as much air as possible at the entry point. They're basically a directional gear because they can only spin one way. Also being a cycloidal gear the gate rotor spins faster which forces more air into the system. I have no idea how Lysholm originally came up with this kind of compressor though without the help of computers.
Man imma try printing it on my SLA machines in abslike and see what happens :D. Who knows, maybe I can get it working well enough to cast the parts and build one to attach to my car. If I succeed at all I'll try to get a video and of course I'll credit you with the design. (to be clear just for fun, i dont do anything commercial)
Yes, I've looked at this a bit. Technically in the US you can't even use them for personal use, but it would be very unusual for a lawsuit to occur. But putting a patented design on TH-cam or providing that design to others starts will cause more trouble than it's worth.
They're 3d printed rotors... Print the gears directly on the ends of the rotors, even inside the case maybe. Herringbone 3d printed gears work great, and you're already oiling it. You can cut the part count in half while also having gears at both ends (or other options)? Just spitballing.
I think that would be possible. The gears are a different diameter than the rotors so that might block some air flow. If I cut the timing gears the other direction like I should have, you essentially end up with a herringbone gear with the timing gears one way and the rotors the other. It’s strange because the rotors never touch, but the compressed air between them still exerts that force.
Rolling bearings if not specially designed for high speed usually fail at speeds higher than 10-15k rpm due to the dynamic loads. For high speed and high load applications the use of sliding axial and radial bearings should be considered. I don't use very often screw pumps at work but I haven't seen an industrial one to work at rpm higher than 3000 - 4000. But they are also with a larger diameter of the screws. I am really astonished by the amount of efforts you have invested in this project. Thanks for the video and the explanations.
Thanks. It's kind of a silly project, but the sound the compressor makes is all worth it. I think doubling up on the radial bearing I had in there would have taken the thrust loads but I'm finding some 8mm thrust bearings for $2 a piece but they're don't have an RPM rating.
Have you applied this to multi zone compressor screw s like would be useful in filament extrusion? The cad design of these things does not seem trivial (or within my capabilities, at least 😉)
Rather than directly driving the timing gears, you should use stepped gearing to get up to speed. Yes, it will introduce efficiency losses, but it will also increase the overall durability of the drivetrain and enable you to better address waste heat. Both linear geartrains and epicyclic ones have different pros and cons here. Another idea that might increase the life expectancy of the drive train is to use a couple of timing belts to isolate both the timing gears and the motor shaft from wherever you mount the geartrain suggested above. On a slight tangent from the topic of epicyclic gears... Do you think a cycloidal drive gearset with a strong helix to it might work well as a screw compressor, or would there be too much loss from the eccentric motion? What about using twin-screw style mating tooth profiles on an epicyclic gearset with the carrier also providing inserts between the planet elements as a constraint path for the pressure medium?
Also, I just remembered an idea that I had some time back about designing the screws with ports and internal paths for water cooling to account for the thermal vulnerabilities of thermoplastics... but that gets a lot more complicated to design.
That's a good idea. I really did think about using a set of GT2 belts as it would be much quieter and more reliable, but always the added complexity of how to tension those belts. I don't know if cycloidal drive set would work for compressing air if it was twisted. My gut says no because you don't have a virtual space that the air compresses along (Often called the "blow hole" for screw compressors). I did see a patent for a tapered single screw compressor with a stationary housing. Very similar to a cycloidal gear in it's movement.
I was really interested in trying water cooling. The initial parts I designed for MJF 3d printing had water jackets in the housing. It would take a lot more design and testing to get a design that worked though. I need to get a big industrial motor and try with some larger rotors and see what I can do.
@IndeterminateDesign I remember seeing that too. They had a whole, very informative, marketing site and everything. But they were only selling it to industrial clients and only by quote. You know what they say: If you have to ask the price, ~~you can't afford it~~ it's a scam.
@@IndeterminateDesign One concern with water cooling the screws internally is that almost the only place to put the ports is concentric with the drive shaft. Although, I suppose if you print the shaft as part of the screw continuously...
I wonder... If you want a bigger & better idea to persue next; The geometry Should work even with Oversized or uneven screw diameters.. (oooh, they're multistart screws..riiight; sorry, just found this content) Anyway, uneven screws; why? Because these pass unholy amounts of volume. So if you make a screw hollow, w a one wsy bypass Through it.. can you improve the efficiency via the air in motion creating suction effect? A venturi suction effect? If it wo4ks, does it work w double oversize screws?
That would be interesting, I don't have the math for that setup. I do have a 3:6 rotor profile meant for a supercharger, I think that would move a lot of air.
I had a real screw compressor like this brought into my shop to have the endplates touched up a bit. What I noticed was this 5:6 ratio between the rotors, BUT, the timing gears had exactly the same number of teeth, as though the rotors turned at 1:1. I couldn't get my head around this. I told my father about this, and remember that he remarked "That's impossible". So I invited him for a looksee himself. It was as i described it. What's the explanation?
That would be strange. I don’t understand how that would work because at some point in the rotation you would have to cover a partial lobe. It’s hard to say without seeing the rotors themselves. I don’t know if theres a special rotor profile that does this or what.
Could a double system help with thrust loads and balance? Like, two mirrored screws on each axle, that move air from each end, towards the middle? Having the timing gear in the middle, and bearings, might also help improve things.
I think I've actually seen a 2 stage screw compressor setup like this. It's definitely possible, I'm not sure on all the math needed to know how to balance it out for sure. I may pickup some thrust bearings for both ends since they're so cheap and can take loads far higher than anything this little guy can generate.
@@IndeterminateDesign Thanks for the answer! I was assuming mirrored screws on each axle would inherently cancel out each others axial parts of the thrust loads, reducing problems, known and possibly unknown ones. That is if they are identical except mirrored, and same stage. Also supporting something like an axle in the middle, as well as at the ends just seems so much better to me than just at the ends. But then again, I don't know, the mathematics is beyond me, that's probably why I like design features with inherent advantages, without actually knowing if they're worth it, or even if the net effect is positive. Anyhow, great videos! I find screw compressors intriguing, and you have increased my understanding of how difficult they are to design and produce.
That would definitely work, and I've seen commercial compressors with that setup. The key is having a tank/separator, so we have enough room for the air and water. I didn't try that with this version because I had dead headed the pressure gauge.
Yeah, I agree. I want to rebuild this compressor and at least see how much air it's flowing. What's the best way to measure the CFM for something so small?
@@IndeterminateDesign got any friends that do HVAC? There’s collapsible screens with a meter, that are used to measure air flow. Put the blower in a box, exhausting out the side, with the HVAC flow screen on top of the open box.
I've not seen the calculations for that, or what is a theoretical peak efficiency for a given screw compressor. Most of the flow details around the compressors I've seen were all determined using CFD. There are formulas for approximating the displacement of a compressor based on the diameter, length, twist angle, and number of lobes, but those are dependent on the fill factor for the rotor profile being used.
On my previous compressor I ran automotive buffing compound to polish the rotors and it worked well, except when the rotors expanded due to heat. This version has timing gears so this compressor's rotors don't actually come in contact with each other or the housing. Ideally you'd measure the thermal expansion of each of the parts and set it up to run very tight clearances when up to temperature, and then inject some abrasive compound.
I know that sound, you might get more with several microphones spaced out some feet away and mixed back together. Its not the same watching it with out the high pitch scream.
Yeah, I need to get a better microphone than just my iPhone’s built in mic 😂. I’m not even sure what the frequency range is for that noise. It’s definitely higher pitch than a supercharger on a car, I’m guessing because of the speeds and size.
I didn't measure the flow. In hindsight I should have done the flow testing before I just went straight to maxing it out to see how much pressure it made. It does move quite a bit of air, so it would be a cool parts cooler. Inefficient for sure but cool
Yeah. I’d love to get some of that IGUS bearing filament but it’s $$$. I may try Teflon tubing as the seals, but the compressor has to be larger for this.
Yes, the gear thrusts loads will cancel out. It's difficult to discern how much thrust load is being added by the air being compressed between the screws of the compressor (Obviously not very much in this model). I think either way if I can add some thrust bearings the compressor will be able to take loads well in excess of anything it can generate.
That wasn’t the invoice, just an example. The rotors were $15-20 each, and the housing was around $35 I think. But there are cheaper options, depending on the resin.
@@IndeterminateDesign Thank you for the update. There seem to be a number of similar companies springing up in China these days. Are they all around the same price and level of quality?
Coding algorithms has been my lifelong jam, my bread and butter. If someone wants to pay me I can code up a little app to generate profiles for compressorz.
Pistons would be easier. I may try one of those someday. I was always fascinated by screw compressors so I wanted to learn more about them by building one.
How high of temps have you been able to hold with that? The rotors were PETG and the intake/exhaust were nylon. I think these compressors should stay under 90C or so.
Yes, I think they’re super interesting and I know they’ve made working 3d printed versions. I’m not sure I can design them in Fusion 360 with that taper and I don’t have a profile for that style rotor.
Man, dont use nylon. To keep small tolerances you need high rigidity. PETG & Nylon has good chance to rub on the walls from centrifugal forces at high RPM. Since PLA is out of question (comlressing gas generates too much heat), then the next rigid material coming to mind is PC. ...or maybe ABS. After all it copes with heat better in short term (not melting) and long term (not creeping due to heat). And ABS can be vapor smoothed by cheap and easy acetone trick to get surface finish you are looking for.
I have some carbon fiber PC filament, my main concern with this shape would be warping. I think a glass filled nylon could do it as well. I read about some small screw compressors for the medical industry using nylon rotors.
A for moving air, the only ones I know of are vane style (teflon vanes) or diaphragm style compressors. I’m not sure what would be the quietest honestly.
There are peristaltic pumps. Often called dosing pumps as they're mainly used in medicine. They're basically a flexible hose bent in a U-shape around a disc with rollers on it. The rollers squeeze the tube and push the fluid in one direction. Quiet and low pressure is fairly easy. You just run your pump at low RPMs. The challenge is if you want low pressure, low noise, and high flow. But this also depends on what you consider low/high pressure and low/high flow.
I was a service technician now a sales representative for a very large screw compressor manufacturer. To get high pressures usually have a two-stage compressor. The first one goes to 35 pounds the second stage of bringing it up much higher the temperatures of an oil less two-stage compressor at each stage is somewhere in the range of 360 to 420° and the rotors are cooling from the air that’s moving across them
That would be aweome, a two stage setup would be awesome. I don’t know that the air got that hot, I saw some of the dry compressors had intercoolers between the stages.
Great project! I'd recommend thrust bearings on the low pressure end and placing your timing gears (and motor) on the high pressure end. You can also offset the axial bearings according to the tolerances and play to get a better seal at full operating speed (knowing the forces and directions of force while in operation allows to minimise clearances by designing for the dynamic state rather than the static).
Nikola Stosic has really done fantastic work to modernising screw compressor design and I'm very grateful for his work.
I'm working on a long term project that incorporates variable geometry screw compressors, hoping to be able to CNC machine the main components for the prototype, but I'm not sure I'm up to the mathematics required!
Thanks! I love idea of a variable geometry screw compressor. That makes my head hurt just thinking about how you accomplish this.
That's great to know you can slide the the rotors relative to each other. There's so little information out there about how these things are setup. I think in the future, when I add a thrust bearing I'll also make the thrust adjustable with some set screws. Filing the spacers by hand was a nightmare.
Bravo!
This is very similar to the oil pumps we use for fuel injector oil pumps on our big boilers. Main reason is they can handle #4 oil which is dirty and they are very reliable.
If you have a 2 meter drop on the intake pipe, you can get almost a perfect vacuum and if used for that, you can extract all gas exhausting into a warm tank. Try it!
I have wanted to do this in plastic using water with mineral oil/sulfur which is basically cutting fluid for machining-
but it would be easier to watch someone like you try it first😊
That's interesting, I may see what kind of vacuum I can pull. I think it would do quite well, especially if I spent more time tweaking the clearances.
Impressive result, really! Definitely add some thrust bearings if you revisit this. I only just added them to my own hardware vitamin collection and I wish I'd done so sooner -- they're not expensive and are an easy way to add really solid axial constraint versus solutions that only rely on standard bearings.
Thanks. I really need to order some. I always get distracted when I look at all the different bearings. I want to have them all 😂
@@IndeterminateDesign I know the feeling! That's where things stop being "not expensive". Kind of like when you want to refill your favorite fastener sizes without buying whole assortments with sizes you never use :P
Amen Indeterminate Design. I got a huge smile on my face for you and your accomplishments. You got it figured out! Great Job. I had been wondering if the main rotor had to be larger than the gate. The previous discharge will work now that the rotation has changed. Eaton has some additional slots adjacent to the opening that my guess if for sound....Pressure is not the performance metric you want, it's flowrate and air temp differential actually mass flowrate not CFM for application I am interested in. Got to have flow to keep the SC cool. My gut tells me plastics have too much expansion for the clearances needed even w/ the air flow to last any substantial time but look up the coefficients of expansion for various material and you can make the assessment. I believe the rotor exhaust end face to case clearance is a big contributor to performance any why thrust is so important. Tapered roller bearing and timing gear mod to include oil bath definitely needed in next model. I believe the main can not exceed 360* twist. If it does the air would be trapped in the gate at the sealing lines made by the two contacts by the main. The blow hole will have more of a negative affect on flow including at the suction port. Increasing temps could also. I hope you try again. It's easy for me to say since I'm not doing the work; but, if it were me I'd fix the damage, cut open the top of the case enough to get a good look at the rotors, put clear acrylic or hardware cloth over that and any other opening, timing gears, belt or pinch point for personnel protection (I like you humor), slow the motor down enough to see the rotors mesh (30 RPM maybe slower) then put it on my office desk...Demonstrations at lunchtime only. LOL I think you know what I mean. Really good job explaining a very complex and multifaceted project design. Best I've seen. No reason for me to do any more studying now.
I’m glad you liked it. I think the resin rotors would have been better for expansion but it’s hard to say. It’s really a mix and match game to with a bunch of copies of rotors to get the right fit . I think I could have an interference fit with resin rotors and just run them in.
It would make a really cool demonstrator for your desk. I want to make one with no housing as a piece of kinetic art.
I agree with your assessment however that’s not very “fun”. Print it out of high temp filaments and go right for the full send. This is TH-cam not LogicalTube.😂
Yeah, honestly the sound alone is worth it. I want to make one for my shop just to hear that compressor turn on.
You could make a mold out of 3dprint and cast metal with plaster, just make sure to make vent-holes to prevent trapping air.
@@Splarkszter I'd love to try something like that, not sure how accurate I could the parts with the shrinkage as the metal cools.
@@IndeterminateDesign Thermal Expansion is always consistent, you would need to calculate the formulas to get the dimension but all of that stuff is sorted out these days.
Fail fast is the most efficient way of egineering.
FarmCraft101 has a really nice video called "3D Printing Metal At Home?"
It is incredibly useful and fun to watch.
Great stuff! I know the 3DP dimension fudging dance all to well. It's annoying because if you ever change to a better manufacture method, then you have to un-fudge them.
Yeah. I hate that. Luckily this time I remembered to parametrize all my clearances so I didn’t have to go searching.
Good work on this. I think your original rotors may still be usable because screw vacuum pumps are a thing. They work well, but they are a HUGE pain when a contaminate gets in there.
Someone below recommended thrust bearings. A thrust bearing+ball bearing would work well, but you can also get angular contact bearings which does both (but not as well as the 2 separate bearings).
I hadn't thought about trying to pull a vacuum with one of these. A vacuum pump would be super useful to have actually.
Great job, I used to test the real thing. It could get exciting from time to time
I can only imagine. The mass of all that steel spinning at high speeds would be scary. What kind of rpm did you have to run?
The SLA resin rotors were actually pretty heavy. I think it would be like 2500J if they let go, or a small shotgun blast 😂.
Used to work in a compressor/tool sake & repairs shop. (Untill stress triggered undlying Fibromyalgia)
This is awesome work.
I Am noticing that grinding noise. I would gues its from the screws making contact with rough print surfaces.
Have you condidered marking an alignment point, coating the screws in a lapping compound, match them up, seassemble, let then gribd/match/self surface themselves in? At relatively low/rock tumbler speed, for a considerable time.
That should let then polish/grind themselves out.
This Would also tell you how Well they screws mesh.
The grinding noise is just the timing gears mainly, and possibly the poor bearings. The rotors themselves don't actually touch, they have about 0.2mm or .008 inches clearance between them.
Did you try pumping water? The higher viscosity might help compensate for the gaps.
I haven’t tried that. I bet it would spray water a long distance for sure.
You can enhance the frictional properties of 3D printed parts made from resin by filling the resin with a small amount of PTFE. At least if you print yourself.
I’ll look into that. I’ve been avoiding SLA printing for so long because of the mess. I need a dedicated space
given your research into the tooth profile, I have often wondered why something more common like the involute gear profile such as used in hydraulic gear pumps is not used?
These profiles are optimized to compress the air along the length of the screw and to pull in as much air as possible at the entry point. They're basically a directional gear because they can only spin one way. Also being a cycloidal gear the gate rotor spins faster which forces more air into the system. I have no idea how Lysholm originally came up with this kind of compressor though without the help of computers.
Can't wait to see your attempt at a Lontra Blade Compressor.
I’ll check those out. I always like new and weird kinds of compressors
Excellent work!
Thanks!
Man imma try printing it on my SLA machines in abslike and see what happens :D. Who knows, maybe I can get it working well enough to cast the parts and build one to attach to my car. If I succeed at all I'll try to get a video and of course I'll credit you with the design. (to be clear just for fun, i dont do anything commercial)
Sounds like fun, definitely be careful.
Well done
You can utilize patented designs for your personal use. You just can't sell them or use them commercially.
Yes, I've looked at this a bit. Technically in the US you can't even use them for personal use, but it would be very unusual for a lawsuit to occur. But putting a patented design on TH-cam or providing that design to others starts will cause more trouble than it's worth.
They're 3d printed rotors... Print the gears directly on the ends of the rotors, even inside the case maybe. Herringbone 3d printed gears work great, and you're already oiling it. You can cut the part count in half while also having gears at both ends (or other options)? Just spitballing.
I think that would be possible. The gears are a different diameter than the rotors so that might block some air flow. If I cut the timing gears the other direction like I should have, you essentially end up with a herringbone gear with the timing gears one way and the rotors the other. It’s strange because the rotors never touch, but the compressed air between them still exerts that force.
Rolling bearings if not specially designed for high speed usually fail at speeds higher than 10-15k rpm due to the dynamic loads. For high speed and high load applications the use of sliding axial and radial bearings should be considered.
I don't use very often screw pumps at work but I haven't seen an industrial one to work at rpm higher than 3000 - 4000. But they are also with a larger diameter of the screws.
I am really astonished by the amount of efforts you have invested in this project. Thanks for the video and the explanations.
Thanks. It's kind of a silly project, but the sound the compressor makes is all worth it. I think doubling up on the radial bearing I had in there would have taken the thrust loads but I'm finding some 8mm thrust bearings for $2 a piece but they're don't have an RPM rating.
Have you applied this to multi zone compressor screw s like would be useful in filament extrusion? The cad design of these things does not seem trivial (or within my capabilities, at least 😉)
Rather than directly driving the timing gears, you should use stepped gearing to get up to speed. Yes, it will introduce efficiency losses, but it will also increase the overall durability of the drivetrain and enable you to better address waste heat. Both linear geartrains and epicyclic ones have different pros and cons here.
Another idea that might increase the life expectancy of the drive train is to use a couple of timing belts to isolate both the timing gears and the motor shaft from wherever you mount the geartrain suggested above.
On a slight tangent from the topic of epicyclic gears... Do you think a cycloidal drive gearset with a strong helix to it might work well as a screw compressor, or would there be too much loss from the eccentric motion? What about using twin-screw style mating tooth profiles on an epicyclic gearset with the carrier also providing inserts between the planet elements as a constraint path for the pressure medium?
Also, I just remembered an idea that I had some time back about designing the screws with ports and internal paths for water cooling to account for the thermal vulnerabilities of thermoplastics... but that gets a lot more complicated to design.
That's a good idea. I really did think about using a set of GT2 belts as it would be much quieter and more reliable, but always the added complexity of how to tension those belts.
I don't know if cycloidal drive set would work for compressing air if it was twisted. My gut says no because you don't have a virtual space that the air compresses along (Often called the "blow hole" for screw compressors). I did see a patent for a tapered single screw compressor with a stationary housing. Very similar to a cycloidal gear in it's movement.
I was really interested in trying water cooling. The initial parts I designed for MJF 3d printing had water jackets in the housing. It would take a lot more design and testing to get a design that worked though. I need to get a big industrial motor and try with some larger rotors and see what I can do.
@IndeterminateDesign I remember seeing that too. They had a whole, very informative, marketing site and everything. But they were only selling it to industrial clients and only by quote.
You know what they say: If you have to ask the price, ~~you can't afford it~~ it's a scam.
@@IndeterminateDesign One concern with water cooling the screws internally is that almost the only place to put the ports is concentric with the drive shaft. Although, I suppose if you print the shaft as part of the screw continuously...
Put some Duraliner kind of a rubber layer on those rotors, and see what happens.
I wonder...
If you want a bigger & better idea to persue next;
The geometry Should work even with Oversized or uneven screw diameters.. (oooh, they're multistart screws..riiight; sorry, just found this content)
Anyway, uneven screws; why? Because these pass unholy amounts of volume.
So if you make a screw hollow, w a one wsy bypass Through it.. can you improve the efficiency via the air in motion creating suction effect? A venturi suction effect?
If it wo4ks, does it work w double oversize screws?
That would be interesting, I don't have the math for that setup. I do have a 3:6 rotor profile meant for a supercharger, I think that would move a lot of air.
I had a real screw compressor like this brought into my shop to have the endplates touched up a bit. What I noticed was this 5:6 ratio between the rotors, BUT, the timing gears had exactly the same number of teeth, as though the rotors turned at 1:1. I couldn't get my head around this. I told my father about this, and remember that he remarked "That's impossible". So I invited him for a looksee himself. It was as i described it. What's the explanation?
That would be strange. I don’t understand how that would work because at some point in the rotation you would have to cover a partial lobe. It’s hard to say without seeing the rotors themselves. I don’t know if theres a special rotor profile that does this or what.
Could a double system help with thrust loads and balance? Like, two mirrored screws on each axle, that move air from each end, towards the middle? Having the timing gear in the middle, and bearings, might also help improve things.
I think I've actually seen a 2 stage screw compressor setup like this. It's definitely possible, I'm not sure on all the math needed to know how to balance it out for sure. I may pickup some thrust bearings for both ends since they're so cheap and can take loads far higher than anything this little guy can generate.
@@IndeterminateDesign Thanks for the answer! I was assuming mirrored screws on each axle would inherently cancel out each others axial parts of the thrust loads, reducing problems, known and possibly unknown ones. That is if they are identical except mirrored, and same stage. Also supporting something like an axle in the middle, as well as at the ends just seems so much better to me than just at the ends. But then again, I don't know, the mathematics is beyond me, that's probably why I like design features with inherent advantages, without actually knowing if they're worth it, or even if the net effect is positive.
Anyhow, great videos! I find screw compressors intriguing, and you have increased my understanding of how difficult they are to design and produce.
What if you pump water mixed with air and seperate the two after compression? The water will seal the gaps while allowing more air through!
That would definitely work, and I've seen commercial compressors with that setup. The key is having a tank/separator, so we have enough room for the air and water. I didn't try that with this version because I had dead headed the pressure gauge.
Respect!
As to measuring the performance, I think measuring the volume of air, (CFM or lbs/min) would be a better scale of performance.
Yeah, I agree. I want to rebuild this compressor and at least see how much air it's flowing. What's the best way to measure the CFM for something so small?
@@IndeterminateDesign got any friends that do HVAC? There’s collapsible screens with a meter, that are used to measure air flow. Put the blower in a box, exhausting out the side, with the HVAC flow screen on top of the open box.
@@IndeterminateDesign After I posted, I searched eBay for “HVAC air flow meter”, and up popped several Anemometers,
@@IndeterminateDesign anemometer and a cardboard box tunnel is the Matthias Wandel dirtbag method. Flow = Velocity*Area
@@toolscientist Thanks, I look into that setup.
can you share how to calculate the suction mass flow rate theoricatcally? In case if we need to calculate the volumetric efficiency
I've not seen the calculations for that, or what is a theoretical peak efficiency for a given screw compressor. Most of the flow details around the compressors I've seen were all determined using CFD. There are formulas for approximating the displacement of a compressor based on the diameter, length, twist angle, and number of lobes, but those are dependent on the fill factor for the rotor profile being used.
spray on some soft self abbrasive coating to make better seal.
On my previous compressor I ran automotive buffing compound to polish the rotors and it worked well, except when the rotors expanded due to heat. This version has timing gears so this compressor's rotors don't actually come in contact with each other or the housing. Ideally you'd measure the thermal expansion of each of the parts and set it up to run very tight clearances when up to temperature, and then inject some abrasive compound.
@@IndeterminateDesign I mean a sacrificive coating , so it can fit automatically to tolerance
I know that sound, you might get more with several microphones spaced out some feet away and mixed back together. Its not the same watching it with out the high pitch scream.
Yeah, I need to get a better microphone than just my iPhone’s built in mic 😂.
I’m not even sure what the frequency range is for that noise. It’s definitely higher pitch than a supercharger on a car, I’m guessing because of the speeds and size.
did I miss the flow measurement? could be fun for as a compressor for part cooling on a 3d printer
I didn't measure the flow. In hindsight I should have done the flow testing before I just went straight to maxing it out to see how much pressure it made. It does move quite a bit of air, so it would be a cool parts cooler. Inefficient for sure but cool
man i was hoping youd get them printed in teflon
Yeah. I’d love to get some of that IGUS bearing filament but it’s $$$. I may try Teflon tubing as the seals, but the compressor has to be larger for this.
Mate, use a herringbone gear for the timing gears and the left and right axial thrusts will cancel out…
Yes, the gear thrusts loads will cancel out. It's difficult to discern how much thrust load is being added by the air being compressed between the screws of the compressor (Obviously not very much in this model). I think either way if I can add some thrust bearings the compressor will be able to take loads well in excess of anything it can generate.
you can't make it short ?
Please could you confirm the prices involved. The numbers on the invoice were a bit difficult to read in mu old age. Thank you.
That wasn’t the invoice, just an example. The rotors were $15-20 each, and the housing was around $35 I think. But there are cheaper options, depending on the resin.
@@IndeterminateDesign Thank you for the update.
There seem to be a number of similar companies springing up in China these days. Are they all around the same price and level of quality?
Coding algorithms has been my lifelong jam, my bread and butter. If someone wants to pay me I can code up a little app to generate profiles for compressorz.
why not with pistons like a driven 4 stroke engine ?
Pistons would be easier. I may try one of those someday. I was always fascinated by screw compressors so I wanted to learn more about them by building one.
If you want a good high temp 3d printing material, I have had great success with pet cf
How high of temps have you been able to hold with that? The rotors were PETG and the intake/exhaust were nylon. I think these compressors should stay under 90C or so.
Have you heard of conical screw compressors?
Yes, I think they’re super interesting and I know they’ve made working 3d printed versions.
I’m not sure I can design them in Fusion 360 with that taper and I don’t have a profile for that style rotor.
@@IndeterminateDesignI'd love to print one and have it as a desktop model, but yeah they seem hard to model
Man, dont use nylon.
To keep small tolerances you need high rigidity. PETG & Nylon has good chance to rub on the walls from centrifugal forces at high RPM.
Since PLA is out of question (comlressing gas generates too much heat), then the next rigid material coming to mind is PC.
...or maybe ABS.
After all it copes with heat better in short term (not melting) and long term (not creeping due to heat).
And ABS can be vapor smoothed by cheap and easy acetone trick to get surface finish you are looking for.
I have some carbon fiber PC filament, my main concern with this shape would be warping. I think a glass filled nylon could do it as well. I read about some small screw compressors for the medical industry using nylon rotors.
Are there examples of "soft" rotor compressors? Think peristaltic pump. The objective is a quiet low pressure compressor...
A for moving air, the only ones I know of are vane style (teflon vanes) or diaphragm style compressors. I’m not sure what would be the quietest honestly.
There are peristaltic pumps. Often called dosing pumps as they're mainly used in medicine. They're basically a flexible hose bent in a U-shape around a disc with rollers on it. The rollers squeeze the tube and push the fluid in one direction.
Quiet and low pressure is fairly easy. You just run your pump at low RPMs. The challenge is if you want low pressure, low noise, and high flow. But this also depends on what you consider low/high pressure and low/high flow.
Super charger for a very small car
Mouse car
You need to lube your timing gears. And don't use PLA for anything but decorations.
Everything was PETG or nylon. The only thing I used PLA for was the motor mount.
@@IndeterminateDesign still need lube
So much for the 500$ parts from JLC 😂
Way way less than that, under $100. SLA is very cheap. 3D printed metal like I had selected in there is not.
Saturn vehicles were no longer made.
Try a star rotor.
Those look really interesting. I hadn't seen them ever used for compressing gases before. I'll look into it.
promo sm 🌟