johnrobholmes
10 MW
Its just hanging during the analysis. No error message. I'll post up the file when I get back home.
Using femm
- Thread starter Miles
- Start date
Biff
100 W
Quite often it will hang and run out of memory if you have a very small area that is trying to be meshed, such as an area that is created between an arc and line that is tangent to that arc. It usually gives a coordinate point at which it ran out of memory, you zoom to that area and see if there is a small problem.
I think others have suggested it, but it is good to re-state. Make your models simple, use chamfers or corners instead of small radii, it will make the simulation much smaller and will have about the same accuracy. once you get to a final type of design you can start meshing finer, and adding in the radii and stuff.
-ryan
I think others have suggested it, but it is good to re-state. Make your models simple, use chamfers or corners instead of small radii, it will make the simulation much smaller and will have about the same accuracy. once you get to a final type of design you can start meshing finer, and adding in the radii and stuff.
-ryan
johnrobholmes
10 MW
I'll go back in and get rid of some curves then. I just reran a version without curved internal corners and it did just fine.
johnrobholmes
10 MW
Anybody have a 9c stator modeled up? I just found my old burned up Bafang stator that I'm going to model.
Kingfish
100 MW
This is the latest iteration of the model Biff had sent me. Based upon the Rust & Rework project where I had direct access to the physical motor, I have corrected the model to reflect (IMO) reality.
Just on the side here for a moment: When I was working with this model and considering types of possible upgrades it occurred to me exactly why the magnets are the height that they are and why the hub motor is precisely the diameter that it is, and the answer is so obvious that it blew me away.
Enjoy, KF
Just on the side here for a moment: When I was working with this model and considering types of possible upgrades it occurred to me exactly why the magnets are the height that they are and why the hub motor is precisely the diameter that it is, and the answer is so obvious that it blew me away.
- The manufacturers are using 8-inch iron pipe (I suggest it's probably 1018) for the backiron. Go measure the ID and OD of the backiron and compare/contrast that to Standard Wall Steel Pipe.
Enjoy, KF
johnrobholmes
10 MW
Many thanks!
All the motors I have been building from scratch for the small scale world are made around the same thing. Cheap back iron. DOM tubing.
All the motors I have been building from scratch for the small scale world are made around the same thing. Cheap back iron. DOM tubing.
Man this is going to take me a while to get.
When I try to draw a stator tooth and copy rotation it makes it on an angle and starts at about 90 deg from the top... Im tying to find videos on the basic steps to draw a stator but most videos skip that part.....
When I try to draw a stator tooth and copy rotation it makes it on an angle and starts at about 90 deg from the top... Im tying to find videos on the basic steps to draw a stator but most videos skip that part.....
nieles
10 kW
you can draw the tooth geometry in a CAD program (like autocad,solidworks,inventor) and import a dxf file in femm
That's what I would do...nieles said:
http://www.3ds.com/products/draftsight/free-cad-software/
http://www.plm.automation.siemens.com/en_gb/products/velocity/solidedge/free2d/
https://www.plm.automation.siemens.com/en_us/academic/resources/solid-edge/student-download.cfm?
Ok guys so... I made it pretty far with google sketchup and I exported a dxf file. I even estimated torque with 10 amps on the phase wires... But I have some questions.
So the first one is when I want to estimate torque do I just use 2 phases with the phase amps planed and leave the third one with 0 amps applied to simulate WYE winding?
So the first one is when I want to estimate torque do I just use 2 phases with the phase amps planed and leave the third one with 0 amps applied to simulate WYE winding?
Biff
100 W
Answer 4) I don't know of a way to change the number of turns at one time. Sometimes I go in a text editor and edit the nodes, its not that difficult when there aren't very many coils, and it is very easy to just calculate what effect changing the number of coils will have in your post processing spreadsheet, so I haven't tried anything else. You might be able to select all the coils you want in one group (2 groups for each phase, one with + turns and one with - turns) and assign them all to the same group, then select that group and change them.
Answer 3) you can tell when things are saturated when flux starts travelling through the air rather than through the iron. But operating near saturation will have a negative impact on efficiency without much performance gain. A good rule of thumb is to operate a rotor at 1.8T and a stator max should be 1.7T If you are going for ultra efficiency you would go for a rotor at 1.6T and stator at 1.3T or so.
Answer 1 and 2) You put the current you want at that instant on each phase. If you want to simulate the torque produced at one instant when 100A RMS is being delivered, you need to know what phase angle you want to simulate (call it phi), then you would set phase A as 100 * sin(phi) Phase B = 100 * sin(phi + 120degrees) Phase C = 100* sin(phi - 120degrees)
Answer 3) you can tell when things are saturated when flux starts travelling through the air rather than through the iron. But operating near saturation will have a negative impact on efficiency without much performance gain. A good rule of thumb is to operate a rotor at 1.8T and a stator max should be 1.7T If you are going for ultra efficiency you would go for a rotor at 1.6T and stator at 1.3T or so.
Answer 1 and 2) You put the current you want at that instant on each phase. If you want to simulate the torque produced at one instant when 100A RMS is being delivered, you need to know what phase angle you want to simulate (call it phi), then you would set phase A as 100 * sin(phi) Phase B = 100 * sin(phi + 120degrees) Phase C = 100* sin(phi - 120degrees)
Biff
100 W
Arlo1 said:
Step 1) think of the AC current as a DC current for 1/12000 of a second
Step 2) figure out the position of the BEMF of one of the phases during that 1/12000 of a second, and the current being in phase with that BEMF. The other 2 phases are then 120 degrees out of phase of that so you can use trig to figure out what those currents are, then simulate that step in time. then rotate the rotor a small amount, freeze time and again, think of everythign as DC again at that new position.
Look at this video, the phases are the coloured waves. Pause the video and visualize that instant in time as DC. Maybe that will help.
http://youtu.be/V6ZWYFTELQA
-ryan
Hm, well these programs probably simply calculates stacking factor like the material is less dense. We saw that when you put a 100x higher stacking factor in Emetor. The flux everywhere around the iron will not be calculated correct if the flux would increase through the iron with smaller fill factor. Anyway, it's one of the factors which makes these 2d estimates not fully accurate.
I'm hoping Emetor is using the flux through the iron sheets instead of the average, when computing iron losses.
I'm hoping Emetor is using the flux through the iron sheets instead of the average, when computing iron losses.
OK, good. I just ran two sims to compare different stacking factors, which confirms that, i.e. the one with lower stacking factor had higher flux density. So everything works as expected in Emetor.
I think though, that the results in FEMM are not useless. You simple need to divide with the stacking factor, and you should get the same results as in Emetor. If you are using the the coloured scale to compare the results, you should use something like 2T * stacking factor as max. Then the one with less stacking factor should contain more pink.
The more I think about it though, the more I tend to agree with you, that it's a bug.
I think though, that the results in FEMM are not useless. You simple need to divide with the stacking factor, and you should get the same results as in Emetor. If you are using the the coloured scale to compare the results, you should use something like 2T * stacking factor as max. Then the one with less stacking factor should contain more pink.
The more I think about it though, the more I tend to agree with you, that it's a bug.
emetor
100 µW
That's how the stacking factor is considered in Emetor: https://www.emetor.com/blog/post/modeling-stack-laminations-finite-element-method/
Not sure how the stacking factor is handled in FEMM though.
Not sure how the stacking factor is handled in FEMM though.
