DIY CPAP Style BLDC

[ntwallace]

In addition to exploring off the shelf 12VDC pump motors, it seems worth recreating a CPAP setup with 3D printed/off the shelf parts. Mark Rehorst's printable blower (specifically v2) + a cheap BLDC looks like a good option.

[necro-nemesis]

I see. So based on a working design the open area below the impeller was already incorporated in the design. I'm thinking the only enhancement to the rotor may then be to leave the blades stout but more closely contour them to the volute casing profile. I know it axial flow turbines this case clearance is critical to performance but this is a centrifugal air pump it probably trades off flow for pressure and may be more reliant on centrifugal force to than creating a tight tolerance to the case. If you think there's merit in recontouring the blades to the case profile I can give that a shot and leave the rest as it stands with the motor adaptation. As was stated earlier the lower area would ultimately exit out mind you there may be some explanation wrt to turbulent airflow which would likely have been modeled and tested for mass production of such a pump.

What are all the laterally placed holes in the case for? A specific mounting application?

[necro-nemesis]

FWIW I prefer your design as it meets a critical requirement in terms of time to reproduce. The other design does not even with me trimming the design down.

[hobodrifterdavid]

@necro-nemesis I see you're really fast with CAD. Would you be interested in helping me develop concepts for the overal packing for 'open-ventilator' (https://github.com/hobodrifterdavid/open-ventilator/blob/master/README.md)? I made a channel on slack to discuss. I'll just take a bit of downtime first. https://join.slack.com/t/open-ventilator/shared_invite/zt-cwh6py1x-C5tO4JewMUf2tRN7rwWB~g

[necro-nemesis]

I was. I'm troubleshooting my computer atm. I'll let you know when I'm up again. Software issue. Looks to be Windows update related. Way prefer Debian but my CAD is on this piece of crap.

[necro-nemesis]

Ok. fixed my issue. Something with roaming data corruption it appeared. I'm back up and running. Joined Slack. With mrehorst's input on my impeller concern I believe the course of action I would like to take is the follow:

-Complete the RC motor mount redesign using the motor I have in hand using the stout impeller of the initial design.

-Examine where I can trim the design to reduce material and print time. (possibly post prototype)

-Revise where necessary to print without support material.

-Commence building of prototype to collect data on operation.

-Source a suitable anemometer to measure flow.

*If collected data provides satisfactory results.

-Model various suitable common RC motors.

-Generate versions of mount to accommodate a range of motors.

[mrehorst]

I took apart another CPAP machine made by ResMed. The blower is quite small and power input for the machine was 12V @ 2.5A, so the motor would probably be no more than 18-20W. The impeller design is quite different from the first one that I modeled and printed. The motor is an out-runner that is exposed to the moving air inside the housing. I've been messing around for an hour trying to model the case - it's a snail-like spiral that I can't figure out how to model with any accuracy. The impeller diameter is 52 mm and the inside diameter of the housing is 64.8 mm, so there's lots of room between the impeller and the case. The input port (the hole in the top cover) is 17.4 mm. The exit port is 15.5 mm ID, but it's flared a bit- I think the diameter of the exit at the inner surface of the housing is about 9.5 mm, but the shape is complex.

Here's a CAD model of the impeller and motor: https://a360.co/33BGX5u

Pictures here: https://drive.google.com/open?id=1hQaCXwzunUPwm1Hkt-IuBgTkKlh1zWyR

[necro-nemesis]

Can you take pic of the case convert it to an SVG, clean up the SVG sufficiently then import that. Extrude the 2d SVG to use to give you some proportions to model around? You may have to do that in two dimensions of X, Y or Z axis to get the proportions right. I assume the snail like spiral is going to require some lofting between two separate circles of different diameters around an center axis. Possible more than two circles depending on how linearly the dimensions change over the length.

[mrehorst]

You can generate a coil in Fusion360, but the problem is matching the shape of the blower housing. It may not be all that critical anyway, since we don't have original specs on the blower. Approximating the housing without the coil will still work. I can generate that in a few minutes.

The impeller design doesn't really lend itself to 3D printing- I think you'd have to print it in two pieces, one mostly flat with recesses to receive the blades, and the other flat with blades, and cement them together. Not sure I'd trust that to hold together spinning at 30k rpm.

Car painting and spray tanning use something called HVLP turbine pumps to move lots of air. It might be better to think in terms of hooking 10 people up to one blower than to make a separate blower for each person...

[hobodrifterdavid]

Hi, we started a dicussion on packaging on open-ventilator.slack.com , 'design-packaging' channel.

,

[necro-nemesis]

@mrehorst while you work out the other pump internals I am focusing on parts that can be reused for mounting an RC motor which can be incorporated into whatever design. I have this collet arrangement in mind given the relatively high rpm of the stator. I will just print the collet to ensure it will print as it's rather finely detailed. I may have to make it more reliant on the flexibility of the material and reduce the level of detail to make it printable.

[necro-nemesis]

@mrehorst Mark how are you making out with reverse engineering the other pump? I would like to prototype one to test brushless motor performance. I am nearing having the required changes to the initial one you modeled. Which of the two in hand are you currently favoring?

[hobodrifterdavid]

Hi. I'm bouncing around between channels here, I can't contribute too much appart from some thoughts and ideas.

There are many eager volunteers on the helpfulengineering slack group (and elsewhere) who might be able to contribute usefuly (I saw one person who works with CFD), so, maybe you can move this discussion to new Github repo ('open-cpap-blower' or something), or, you can use the slack channel here for discussion: https://join.slack.com/t/open-ventilator/shared_invite/zt-cwh6py1x-C5tO4JewMUf2tRN7rwWB~g ('design-motor' channel). I can then direct anyone useful to that repo/channel. Maybe even get lucky with a turbomachinary guy.

Also, I really encourage you guys to put all interesting findings into a google doc (I think you have one) or a GitHub repo. The information in that document may save the time of people who come after you, and will allow others to contribute more effectively to the project. I know Mark (mrehorst) has a really nice blog. It's a bit annoying to write stuff down, but worth it.

[necro-nemesis]

@hobodrifterdavid I'm in there. I can't post design files to comments on github. There is a 3d design repo on the main repo that could be forked and PR'd but until we start to get more people interested in revising designs using the PR process then it's easier to simply post to Slack. I prefer Github in this regard for version control when it comes to collaboration but we aren't there yet with the number of people I see working the physical pump. Not sure the main repo is the best place to have design review carried out. Someone needs to take charge of a separate repo. Possibly mrehorst and I should be reviewing and approval authority for commits? When it starts to get out of hand I agree use a separate repo and get version control in hand using the PR process with a main and dev branch.

[mrehorst]

I have finished the modeling of the resmed blower as accurately as I can- that snail curve was a real tricky thing to do! https://a360.co/33BGX5u

This design does not really lend itself to FDM printing due to the complexity of the structures.

[necro-nemesis]

Great Mark I'll have a look at it. I think the FDM aspect may be a show stopper but I will see if I can arrive at a way and whether or not it merits doing it.

[necro-nemesis]

I think it could be reproduced if it's sectioned and the sections are slotted together for containment of the pressure but definitely more complex to assemble and longer to produce. Typical issue when considering the FDM over injection molding. I'm not sure injection molding has been ruled out for production but these are considerations. Going on the basis of it being available open source and replicated I am leaning towards the first pump. If the idea is to spool up production on pumps commercially then I see not need to present a 3d printable design and the focus goes to what can be produced in a mold. I see merit from the open source perspective to pursue a printable design as the original concern over access to commercially available pumps. On that basis recommend testing the theory of operation at the moment for the aforementioned reasons I'd go with the first design utilizing readily available brushless DC motors i.e. RC.

[nlambuca]

necro-nemesis where are those files? I am trying to test it also instead of bag system. thx

[necro-nemesis]

necro-nemesis where are those files? I am trying to test it also instead of bag system. thx

https://github.com/necro-nemesis/open-ventilator-centrifugal-pump/tree/dev

[nlambuca]

After doing some additional research I came across a design that was built and tested using a brushless RC motor. I suggest moving down this avenue. The drawing files and test footage are available.

Where?

[necro-nemesis]

You need to be more specific about what you want. The files as far as they are along are there. I'm not following this topic intently anymore to go back and see what you want in the context of any discussion. Probably this one I'm assuming. https://github.com/necro-nemesis/open-ventilator-centrifugal-pump/tree/dev/2205%20alternative. There's a whole thread on Slack with an open invite above if you want to read yourself in on where this went.

[nlambuca]

I was talking about the round clear design above. I found your design already. Looking good, but wanted to check that round one. Big Thanks.

[gaurav12294]

Hello everyone, I had built a working prototype with a commercial inflator pump which had a brushed DC motor; but it gave a maximum of 15 cm H2O. I want to improve it by keeping the same housing of the pump and replacing the motor by a better brushless or brushed DC motor.

Can you guys help me with the specifications required for the motor to get pressure of 40 cm H2O. I did not find a concrete motor which was selected in this thread. Any links will be very helpful.