547,443 active members*
1,724 visitors online*
Register for free
Login
IndustryArena Forum > Mechanical Engineering > Linear and Rotary Motion > Custom Spindle (Custom Motor!), Very Gradual Progress Planned
Page 1 of 2 12
Results 1 to 20 of 32
  1. #1

    Custom Spindle (Custom Motor!), Very Gradual Progress Planned

    Hi all,

    Been lurking around for about a month in the forum and decided in this last week to start jumping in. I'm a Mechanical Engineer / Electrical Engineer with a focus on rotating machine design. Recently I went from a 90% desk 10% lab/shop position to a 80% shop 20% desk position involving a lot of fast paced prototyping in agricultural robotics and fruit handling. The result is that I've reconnected with just how enjoyable it is to cut material and end up with what you need to use at the end. Don't get me wrong, the 3D printer farm is also pretty nifty, but I'm all about smallest functional form (optimized) and 3D printers rarely cut it for final versions with that goal in mind.

    I've designed brushless and servo motors, frameless kits and fully housed custom solutions, as well as a couple of large (120-300kW) axial gap traction motors for EVs. Before motors I worked on custom continuously variable transmissions (belt and pulley type), as well as up-fitting class 3-8 utility vehicles and busses to plug-in hybrids before that was much of a thing.


    I have an end machine design goal, but before I jump into that, I figured it would be great to cut my teeth on individual subsystems to get a better understanding of the specifics of CNC machine design. I have access to a manual and CNC lathe (linuxCNC), and a manual Bridgeport knee mill and a Matsuura VMC with a ridiculously extensive chain type tool changer. I also have access to a solid 36"x48" work area router, a plasma cutter, and various other shop tools. My employer actively encourages side projects (on our own time) and will front a small bit of cost in tooling and materials. I do also get paid well enough to front most of the cost of materials and parts myself.


    My first project target here is to leverage my motor design background into a custom spindle. I'm looking into some novel active vibration management methods as well, but those will be beyond the scope of this particular build thread. I also may keep those proprietary (unless they don't work of course!).

    My intent is not to make a full ATC spindle, but simply the drive motor part. I want to walk through the design process for both an induction motor and a permanent magnet motor, fully housed into a spindle motor. Any references for standard interfaces would be very helpful. Things such as common spindle holder size, connectors, etc.

    For myself personally I'm looking at smallest viable product for cutting steels up to gear steels. Why? Well, again, I come from rotating machines and that means splines, case or through hardened surfaces, etc, often in <1" diameters and with <0.001" tolerances referenced to all datums. Yeesh. Yes, servo industry man, customers be like that. Also keep in mind that in a high performance motor, you might have a 0.010" nominal airgap, which you don't want to vary more than +/-25% from tolerances and also should never achieve contact under maximum loading conditions, with a factor of safety, while your working diameter is as large as 8". With stackups? You get the idea. Especially since unless you want to grind them, machining laminated stators is a risky proposition and they typically achieve between 0.002" and 0.005" profiles on the OD and ID, which already just ate your entire airgap for breakfast, worst case. YEAH.

    Anyways, let's break down the separate projects and please give me feedback on the target performance I would need:

    High Speed Machining Spindle (Minimum Specs for gear steels, such as 4342, and stainless such as 416)

    RPM >=?rpm
    Power >=?Watts (I see 2.2kW as a common rating, but how much do we actually NEED, at the tool?)
    Assembly dimensions (for interface with the machine)
    Connectors?

    Induction Spindle: Simplicity and old school reliability
    - Indexing? (Absolute Encoder/Resolver) or Non-Indexing?

    PM Spindle: A servo and a gentlemen (unless you're metal)
    - Indexing. DUH.

    Coolant options?

    I should mention I specialize in cooling motors a bit more aggressively than any spindle I've seen on the market... and I won't be violating any patents (because I'm on them )

  2. #2
    Community Moderator Jim Dawson's Avatar
    Join Date
    Dec 2013
    Posts
    5014

    Re: Custom Spindle (Custom Motor!), Very Gradual Progress Planned

    Quote Originally Posted by StrawberryBoi View Post
    Hi all,

    Anyways, let's break down the separate projects and please give me feedback on the target performance I would need:

    High Speed Machining Spindle (Minimum Specs for gear steels, such as 4342, and stainless such as 416)
    For cutting 4xxx and 4xx series steels, with modern cutters you are able to run in the 200 to 800 SFM range. Cutter diameter and other cutting parameters are pretty much dictated by your machine, mostly the mass of the machine. (for the purposes of this discussion I'll use pounds as the imperial unit of mass, and yes I know this is not correct)

    Spindle power and machine mass are closely related, you need a machine frame that is capable of resisting what ever cutting forces you apply. You could eventually accomplish the same part using Dremel tool as a spindle as you could with the Matsuura VMC, but using the Dremel would take a lot longer because of less HP. On the other hand the machine with the Dremel spindle could be a lot less massive. Using modern cutters, spindle power and machine rigidity are the limiting factors, the cutters will take just about any amount of power you can provide to them.

    The standard BP knee mill with a 2J2 head is about 2300 lbs., 2HP, and will reasonably cut with a 1/2'' dia endmill in steel, has an R-8 spindle. About 4200 RPM max. My BP clone is a larger frame, but the same 2J2 style head, 3500 lbs., 3 HP, with a BT30 spindle, same RPM range as a BP. Will reasonably cut with a 1/2'' end mill in steel. Not really much gain there, but I can push it a bit harder than a standard BP because of the larger frame. Spindle power is not the limiting factor, machine rigidity is. I rarely run at greater than 70% spindle load.

    Then we move up to my Haas TM-2P. This is a VMC, bed mill design, about 5500 lbs., 7.5 HP (servo?) spindle, 6000 RPM max, and wish we had the 8000 RPM spindle option. It will generally run parts about 4X as fast as my BP clone. A much more rigid design and will comfortably handle 5/8'' endmills in steel. #40 spindle. We generally run it at 70 to 80% spindle load. This is still a lightweight machine.

    RPM >=?rpm
    Maybe 10,000 RPM max

    Power >=?Watts (I see 2.2kW as a common rating, but how much do we actually NEED, at the tool?)
    The answer is: It depends. The machine needs to be massive enough to resist the cutting forces. More mass and rigidity allows higher HP. Higher HP allows higher Material Removal Rates (MRR).

    Assembly dimensions (for interface with the machine)
    There is really no standard. Buy the spindle and match the mounting.

    Connectors?
    I assume you mean tool holders. R-8 is standard for BP type machines. #30 is OK up to about 4HP, #40 is standard for 5 to 10 HP, and #50 above that. These holders come in several flavors; CAT, NMTB, and a few others.

    Induction Spindle: Simplicity and old school reliability
    - Indexing? (Absolute Encoder/Resolver) or Non-Indexing?
    You normally need indexing for use with many ATC systems to align the spindle dogs that engage with the tool holder.

    PM Spindle: A servo and a gentlemen (unless you're metal)
    - Indexing. DUH.
    Nice to have. I just replaced the 5.5kW induction motor with a 7.5kW servo on my lathe so I could have spindle indexing capability.

    Coolant options?
    High pressure flood coolant is normally the best for most operations, but requires a full enclosure of the work area. Coolant does three things, cools the tool, lubricates the tool, and maybe most importantly blasts the chips out of the cutting area so you are not recutting the chips. For lighter duty applications, spray mist coolant works, and is a lot less messy.

    I'll try to be more specific as you have further questions later on in this process, but for now this is a general overview.
    Jim Dawson
    Sandy, Oregon, USA

  3. #3
    An incredibly excellent response! Thank you, I'll have to respond to any specific elements properly on a PC tomorrow or Monday.

    What of high speed spindles in the 22krpm range? These kinds of speeds or higher are totally achievable in low side load machining, as bearings are the truly limiting factor and not the motor components. Mechatron spindles for example are running at these speeds. I've designed some larger motors turning at >35krpm and smaller motors that could with air bearings go >100krpm and in the 2kW range power-wise. I'm not intending to tackle anything like an air bearing, though I suppose for chips and giggles I could make a magnetic bearing version for micro-machining at extreme rpm.


    After thinking over the last few days I'm planning on starting with the induction motor, both because it's something I've only made in FEA (unlike the PM motors) and I want to build one and also because if I make an axial gap induction motor it is modularly stackable and easily produced in the machine tools I can access. For example I can cut the complete rotor cage on a mill or plasma cutter even (if balance is not a concern). I can work with Somalloy or a Chinese alternative SMC prototyping pick, which machines with chipped edges like a cast iron but makes functional test parts. Production of SMC parts is actually inexpensive in volume too.

    For a test motor I would hand wind, in volume I would make a simple 2 axis coil winder or buy and upgrade a cheap Chinese one. Generally a good design approaches 1-5 turns per coil, so I may end up bending square copper rod with a Kapton wrap. Not that I expect to go to volume with this for fun project, but fail to plan and you plan to fail.

  4. #4
    Community Moderator Jim Dawson's Avatar
    Join Date
    Dec 2013
    Posts
    5014

    Re: Custom Spindle (Custom Motor!), Very Gradual Progress Planned

    Certainly high speed spindles have their place. You only have to use cutters small enough to keep the surface speeds within the recommended cutter operating range for the material, and feed fast enough so the cutter maintains a chip load of 0.001 to 0.007'' or so depending on the geometry of the cutter. Then maintain a step over, DOC, and cutting strategy that is compatible with the cutter. Failure to push a cutter hard enough is a way to kill it quickly, proper chip load and engagement is really the key, and modern adaptive tool paths do a fine job of insuring that the cutter is properly engaged.

    I'll be interested to see what you come up with for a motor. I have never heard of anyone starting with the spindle motor first, then designing the machine around that. An interesting concept.
    Jim Dawson
    Sandy, Oregon, USA

  5. #5
    Registered
    Join Date
    Nov 2013
    Posts
    1342

    Re: Custom Spindle (Custom Motor!), Very Gradual Progress Planned

    Hi,
    Jim Dawsons reply is masterful, the spindle power and machine rigidity need to be matched. I would add that spindle rpm need to be matched also.

    I have an 800W 24000 rpm Mechatron spindle and I use it daily for hours at a time, its been bloody marvelous BUT it's hopeless in steel.
    I use it for isolation routing of PCBs and making small instrument parts in aluminum, brass and plastics. You can cut steel, but ever so slowly.
    Its just so easy to stall the spindle with only 0.3Nm torque.

    High speed spindles, even up into the low kW range have low torque and are unsuitable for ferrous milling except with very small diameter tools.
    For example a 2.2kW 24000 rpm spindle has a rated torque of about 1Nm. That's really only enough for a 3mm tool in mild steel.

    I would suggest that for steel you want about 5-6Nm for tools up to 10-12mm diameter. A 10mm carbide tool will have a surface speed of 100m/min
    at 3500 rpm, and 100m/min is good for steel whereas 60m/min-80m/min is more appropriate for stainless. The highspeed spindles will baulk at going at such slow
    rpm. Were you to go for a 10kW spindle then you are talking about decent torque levels and could probably run them slow enough to run some decent sized
    tools, but such large spindles are not really hobby stuff, and will certainly require 3 phase power.

    You can achieve decent metal removal rates in steels, including hard steels with high speed/low torque spindle with CBN tools and very carefully planned constant engagement
    toolpaths. CBN is out of my price range....and Fusion is not the best choice for HSM toolpaths.

    When I built my mini-mill seven years ago out of cast iron and steel, I fitted it with the 800W Mechatron spindle thinking that I would be able to cut steel.....man...was I disappointed!
    The spindle has however, with that exception proved to be extremely useful. What I did was make a second spindle using a second hand 1.8kW 3500rpm 6Nm (cont) Allen Bradley AC servo.
    I used a cylindrical Rego-Fix ER25 toolholder and P4 angular contact bearings. It fits in the same bracket as the 800W Mechatron so I can swap the over in a few minutes.
    Given that its a genuine servo I can have it free-running in velocity mode or have it index and even coordinated with other axes, and of course being a PM motor its very compact
    and power dense. It gave me the steel capacity that I wanted.

    That is when I encountered the problem that Jim mentioned, now I had a spindle that had the torque necessary to make my machine flex. What had been
    adequate, in fact better than adequate, with a low torque spindle starts to look inadequte with a high(er) torque spindle.

    I have in recent weeks commissioned my new build mill. It has own design cast iron axis beds of 115kg each, a twin 'L' shaped frame in 32mm med tensile steel, 32mm C5
    ballscrews and THK HSR linear rails and 750W Delta servos all around. All-in-all a major step up in rigidity and size over my mini-mill. My Allen Bradley spindle is now
    really coming into it own. Its been a long haul getting to this point. I've had to learn heaps to get here and overall I would say I've enjoyed it.

    May I suggest that you decide early on what sort of machining capacity you really want or need. If for instance you want/need steel capacity the you'll want a low speed
    high torque spindle whereas aluminum, brass and plastics capability suggest high speed low torque.

    As I'm sure you are aware you CAN have high speed and high torque, but in your professional capacity you must also know that its an expensive combination....
    do you really want that as a hobby project?

    As I stated earlier my Allen Bradley spindle is position capable and so can do rigid tapping etc....but guess what...I hardly ever use it!! In my opinion while encoder
    feedback with or without position control is nice but not really required, 99% of the time you just want a powerful, stable free-running motor. What is most important is the the
    power/totque/speed of the spindle match the material and tools and most importantly the rigidity of the machine. Personal experience tells me that securing the most
    rigid machine possible at a given weight/cost is FAR harder to achieve than a spindle is.

    Craig

  6. #6
    Member
    Join Date
    Jul 2018
    Posts
    3248

    Re: Custom Spindle (Custom Motor!), Very Gradual Progress Planned

    HI Strawb - All prior comments very applicable. You have the machine resources to build a great machine. Usually a machine is designed to fit a purpose. If you try to design a general purpose machine then everything is compromised and you end up with the std machines that vendors sell and your build cost will be similar to theirs (as yours is bespoke vs theirs are serial production) unless your labour and materials are free. I suggest you find an exemplar commercial machine that does what you have in mind then you have a benchmark to learn from and work to. Machine rigidity vs cost will be your greatest hurdle then match that to the spindle or vice versa... Everything is connected. Keep at it we all will follow with interest the forum has great depth in this area.... Peter

  7. #7
    You've all given me some excellent things to consider and chew on.

    I should say, my rather pipe-dream machine is something that can cut in 5 axis and is a mill-turn, with spindle hand off and auto-feed for stock of at least 2" diameter. I'd like to take a crack at a swiss-type with am optional bushing.

    I'm looking at then a main and sub spindle and a live tool head, turning turret probably, and an ATC on the live tool spindle.

    So SO many parts in bespoke servos and transmissions require tight turned tolerances ID and OD and also have parallel, perpendicular, and even off angle milled features for oil flow to bearings, coolant or pneumatic features, hydraulic features... splines, relief features, o-rings and often half or full dovetail, etc.

    Many of them require 3+ setups due to double ended features, turning+milling+broaching. Length is a major thing, the length of parts will often be >10x the shaft diameter so turning the primary shaft features is ideal.

    I also want to be able to mill larger pieces up to 12"x12"x?". Ultimately the pipe dream is a small all-in-one (slow throughput) lights out production machine. Given depth of cut limits, I really can get away with 1/8" or even smaller milling tools if I have a 5th axis. Internal splines are broached.

    The one caveat would be polygon shaft couplings, which would be up to 1.25" depth internal features that are rare so may just be outside the capacity of the machine and that's ok.

    For this reason I am leaning towards high speed spindles specs and a total machine that can operate off of a 50A 220/230 plug. I happen to have one of these for my EVSE, but 40A would be good too since it's common for dryers. I would most likely be installing this for years at work though which has full three phase, but taking it home and bricking it due to no three phase...

    This puts the machine power budget as <7kW on a standard 40A 220V breaker. (Derated for true continuous limit).

    I wouldn't need to be turning and milling at the same time, so I can allocate low wattage on the lathe spindles during milling operations and none on the tool spindle during turning. All three spindles must index.

    There is really no reason though that I can't use a Mechatron spindle as a starting point for interface dimensions and tolerances. I can also look at copying their connector (electrical connectors) setup. I come from automotive on the wiring side of my experience, which has very different styles of connectors. All these gland style connectors are totally new to me.

  8. #8
    Registered
    Join Date
    Oct 2011
    Posts
    33

    Re: Custom Spindle (Custom Motor!), Very Gradual Progress Planned

    RPM >=?rpm
    Power >=?Watts (I see 2.2kW as a common rating, but how much do we actually NEED, at the tool?)



    You can use feed and speed calculator.

  9. #9
    Community Moderator Jim Dawson's Avatar
    Join Date
    Dec 2013
    Posts
    5014

    Re: Custom Spindle (Custom Motor!), Very Gradual Progress Planned

    Oh, what a fun project.

    First let me describe my machine, then I'll make a recommendation that will meet your goals.

    I have a Hardinge Conquest 42 with live tooling, 1989 vintage. 1 5/8 bar capacity, 10 tool turret, 5 of those live tool stations. No Y axis and no sub-spindle (I wish we had both). Originally had spindle indexing capability to 1 degree. Spindle motor 5.5kW continuous, 7.5kW intermittent, 5,000 RPM rated. Machine weight: 10,200 lbs. Originally had Fanuc OT controls.

    Once the machine was on the floor and wired up, I spent about a week trying to get the antique Fanuc controls to do what I wanted, that wasn't happening. I have a very low tolerance for antique controls. So with the exception of the spindle motor, I carefully removed everything on the machine that said Fanuc on it and tossed it. Replaced the four Fanuc servos with 1.8kW DMM servos (X, Z, Live Tool, and Turret). I ran the spindle motor with a VFD and lost any indexing capability, but this worked fine up until a few weeks ago. We did a parts redesign that requires full C axis capability. So remove the Fanuc spindle motor and replace it with a Delta 7.5kW servo system, we lost 2000 RPM (3000 max vs.5000 max) on the spindle, but we never ran it at greater than 2000 RPM for our work anyway. Now we have full C axis capability, and not one Fanuc labeled part on the machine.

    What about the software? I just used a Galil controller and wrote my own CNC software. When adding the C axis code to the software, I also added tool stations in the code to allow gang tools in the non-live slots, so now we can have up to 20 tools if not using live tools. This machine is running on a 240V, 50A single phase circuit, through a 15hp rotary phase converter that I built. Normally the line does not see more than 20A (this is another story) The machine runs lights out, and holds +/- 0.0005'' or better all day long.


    So to meet your goals I would concider a Hardinge Conquest T42 (not to be confused with the 42, they are different machines) with live tools, sub-spindle and Y axis. This gives you a solid platform to work from, and all of the hard work is already done. You can concentrate on the motors and controls and have a known solid base to work from.

    This one has live tooling and a sub spindle, but I can't see if it has a Y axis. I think all of the T42's have a 2'' bar capacity, but not sure about that.
    https://www.ebay.com/itm/26492789145...EAAOSwO-JfpbSn
    Jim Dawson
    Sandy, Oregon, USA

  10. #10
    Hardings Conquest T42, I'll keep an eye out for something like that. I don't have the space to host that kind of a project unless we have a need at work for them, but I could host a small build, likely an EG base.

  11. #11
    Community Moderator Jim Dawson's Avatar
    Join Date
    Dec 2013
    Posts
    5014

    Re: Custom Spindle (Custom Motor!), Very Gradual Progress Planned

    Of course you need one at work.


    But if that is not in the cards, then building your own on a smaller scale would make a great test bed and proof of concept

    Another option for a mill-turn machine is to use a 4th axis on a mill as the lathe spindle. I've seen a guy on YouTube with a kit that he builds that does turning on a mill using something like a 4th axis. Uses the mill spindle as the tool holder for the turning tools and just locks the mill spindle to operate with the lathe function. There are all kinds of configurations that I can think of that would use anything from a small benchtop mill to a big VMC as the base machine. Then it's just a matter of getting creative to make it all work together and produce the desired result.
    Jim Dawson
    Sandy, Oregon, USA

  12. #12
    Quote Originally Posted by Jim Dawson View Post
    Of course you need one at work.


    But if that is not in the cards, then building your own on a smaller scale would make a great test bed and proof of concept

    Another option for a mill-turn machine is to use a 4th axis on a mill as the lathe spindle. I've seen a guy on YouTube with a kit that he builds that does turning on a mill using something like a 4th axis. Uses the mill spindle as the tool holder for the turning tools and just locks the mill spindle to operate with the lathe function. There are all kinds of configurations that I can think of that would use anything from a small benchtop mill to a big VMC as the base machine. Then it's just a matter of getting creative to make it all work together and produce the desired result.
    I have to admit those little Chinese bench top mills are a tempting test bed...

  13. #13

    Re: Custom Spindle (Custom Motor!), Very Gradual Progress Planned

    Alright, so my plan for external dimensions is to just use the Mechatron HFP-8022-50-ER20 dimensions. The length I'll shoot for shorter with a higher torque. In the end I may just use the Mechatron ATC spindle, but I want to take a shot at improving the torque and power density of the unit.

    Rather than targeting 2.2kW though, I'll target ~3Nm at up to 20krpm and a top speed of >35krpm, or around 6.3kW at peak power. Yes, this is close to the machine budget, and yes this will be a continuous target, but with low side load forces I'm expecting it will fall within the budget in actual usage. Additionally there is little expectation it will run actually up to full power or anywhere near, as at that point you are just about to overpower the spindle.


    Liquid cooled of course. I'll be designing direct or near-direct to wire cooling as well as rotor cooling.


    Actual progress (other than spec selection, which is of course almost 90% of the engineering battle):

    1) Trying out a new, low cost to free motor simulation package MotorAnalysis-PM (paid version is MotorXP-PM and MotorXP-IM for induction motors). Due to the induction design being behind a pay-wall, I'll start with the Permanent Magnet motor for sizing the motor and determining feasibility. It will be the highest performance density anyways. My backup is to use FEMM and write up a full script for it, which also works, but is a lot more effort on my part.
    2) Verified that I can hit the tolerances on the in-house equipment that I'll need. The engine lathe spindle has never been aligned since installation, however the Hardinge turret lathe holds fine tolerances excellently, though it can't do threading if needed. The CNC lathe... is a Frankestein, it should hold very tight tolerances without issue but I have to learn to use it and the linuxCNC.


    Gradual progress, don't forget!
    Attached Thumbnails Attached Thumbnails Mechatron_HFP-8022-50-ER20_Dims.png  

  14. #14

    Re: Custom Spindle (Custom Motor!), Very Gradual Progress Planned

    MotorAnalysis works fine. EXCELLENT correlation with the results from more expensive software.

    As for using it for the design... I'm not going to hand plunk at optimizing this, I'm going to script this sucker. I have yet to have time to investigate scripting, however I know the paid version allows for at least some scripting. If I use MATLAB I can apparently fully script it, but I have to get a quote on the minimum level of MATLAB now. Would prefer a python or Lua or other option.

    In other news, I'll likely be doing some motor designs on contract that have very similar performance requirements, so I'll be utilizing the paid version of the software for that and I can piggy-back the effort.

  15. #15
    Community Moderator Jim Dawson's Avatar
    Join Date
    Dec 2013
    Posts
    5014

    Re: Custom Spindle (Custom Motor!), Very Gradual Progress Planned

    So when do we get to see a new motor design?

    On another note, I'm negotiating on that Hardinge Conquest T42 I linked to above. Like I need another project, but it will solve a manufacturing bottleneck in our process so worth it I guess.
    Jim Dawson
    Sandy, Oregon, USA

  16. #16
    Quote Originally Posted by Jim Dawson View Post
    So when do we get to see a new motor design?

    On another note, I'm negotiating on that Hardinge Conquest T42 I linked to above. Like I need another project, but it will solve a manufacturing bottleneck in our process so worth it I guess.
    LOL, well I guess for the initial part of the build documentation I can document the FEA design process. I'll tell you right off, motor design is way more boring that CAD models of hypothetical EG mills, of that you can be sure!

    Maybe as soon as next week I can start the process. It's mostly going to be 1-2 weeks getting a feel for MotorXP/MotorAnalysis and then the actual design part takes 1-3 weeks, unless I can script for optimization. Scripting it might take 1-4 weeks alone, then the runs start after that, but I can get a better design in about 3 days vs 1-3 weeks manually.

  17. #17
    I should add, after the FEA design comes the CAD, bearing selection, cooling, case, etc... which may also feed back into the FEA when inevitably I have more space or less space than originally anticipated.

    I guess I could do it the way a customer would and design the casing first with bearings and then inform the motor design. Hmmmmmmmm yes, let's do that!

    Got to start bringing my personal workstation laptop to work and plugging away during lunch!

  18. #18

    Red face Re: Custom Spindle (Custom Motor!), Very Gradual Progress Planned

    Time for an update!

    I am moving forward with the motor design contract job, so now I'm utilizing a 30 day trial of the paid version MotorXP-PM (Permanent Magnet). I also played a bit around with MotorAnalysis-IM (Induction Motor) and I am quite pleased with it too, though induction motors are definitely a new design experience for me. It's entirely one thing to understand them from an engineering perspective and to understand the design methodology and how it's supported by a particular software.

    Alright, so onward and spinward with a PM!

    I'll explain a lot of these later, but you can see the process on MotorXP-PM is actually wicked fast if you know where to start. I didn't end up with this geometry at the end of an hour of messing around, due to high frequency at high speed and the lower energy magnets (N30) default are actually more expensive than high energy N52s due to magnet lines just running N48-N52 magnets most commonly.

    The motor I ended up with at the end of playing is actually an ~1Nm continuous (theoretically, no thermal analysis done yet) motor with a 22krpm knee. This works out to be about 2.2kW.

    Pretty happy with this software.
    Attached Thumbnails Attached Thumbnails MotorXP-PM 9ampLoad.jpg   MotorXP-PM 20TurnResults25krpm.jpg   MotorXP-PM AckDumDum Mistake 001.jpg   MotorXP-PM BEMF YES.jpg  

    MotorXP-PM Blank Example.jpg   MotorXP-PM Cogging.jpg   MotorXP-PM CoggingSpectrum.jpg   MotorXP-PM Design Studio Default is HUGE.jpg  


  19. #19

    Re: Custom Spindle (Custom Motor!), Very Gradual Progress Planned

    Here is where I ended up, with a 12/10 motor at a 60mm length.
    Attached Thumbnails Attached Thumbnails MotorXP-PM EndDay1_Run4Shorty.jpg   MotorXP-PM EndDay1_Run4ShortySpecs.jpg  

  20. #20

    Re: Custom Spindle (Custom Motor!), Very Gradual Progress Planned

    Spent another hour working on the magnetic design. Noticed a glaring issue that was brought up in the flood of responses to my spec posting

    The spindle primary frequency for the 12/10 motor at the speed was >2kHz. While this would be totally fine for a quality inverter running 20kHz+ FETs, for an IGBT this would be terrible for drive losses if even supported by the drive supplier.

    I've played around with 6/4 (coggy without skew), 12/2, 21/2, and 9/8 configurations. Basically I'm looking for something with low natural cogging and high winding factor. Winding factor is the fractional measure of how perfectly the rotor poles line up with the magnetic poles of the stator phase. My favored 24/20 and 12/10 configurations have very high winding factors and thus very little of the copper is wasted in doing nothing for the motor torque. A 6/4 is something like 0.866 which is a good factor, typically acceptable, however 6/4 motors have the magnets line up perfectly with the stator teeth and so cogging is HUGE, which requires a skew (twisting of the stator or rotor) which is to reduce the alignment but also reduces the winding factor for the motor and...

    Anyways, the 9/8 and 21/2 configurations are GREAT for cogging and performance density (having high winding factor and very low cogging). The 9/8 is a fantastic arrangement for a high torque motor, it has excellent harmonic content, low cogging, and is a concentrated winding. That is, you wind around one tooth until done for that tooth, then go to the next tooth for the phase and wind around that until done... which means very little end turn (wasted copper and space). The 21/2 is a fractional pitch motor, which does mean it has so-called continuous windings. That is, the wire must jump multiple teeth as you are winding it around for each coil in the phase and the various phase windings all overlap and have to be wound simultaneously. This too has great harmonics and low cogging.

    I'm splitting the design into two motors, a high torque spindle and a high speed spindle, with the goal of limiting primary frequency to <800Hz at full speed operation. I'm going to actually target a peak torque at 500-600Hz, with field weakening at constant power until the 800Hz.

    Now comes the last decision I've yet to make, which is what kind of drive this will be using. If I use a 220Vac input AC drive, then I'll be able to connect it up to the direct power, which is GREAT. The downside is these drives are larger and thus I can't go the fun route of integrating it into the motor housing directly. The direct integration options are almost all 48VDC drives, which means I would need an active rectifier with PFC to convert from 220Vac to 48VDC somewhere in the system. Good thing there is I could run 48VDC servos with integrated controllers everywhere! That would be great for electrical noise. That would be terrible for wire size and the active rectifier would need to support the entire load meaning it's also going to be big. Additionally I'm not readily finding good options in this even though I know they exist.

    I'm resisting the urge to pull TI reference boards and make a new controller with full PFC active rectifier into a 160VDC bus. That would allow more voltage flexibility on the AC input size and downsize wires. I would though, have to use HV wiring safety protocols (orange wire, double insulated with shielding, interlocked connectors) *EDIT: If I used this as a power supply for the entire system, not just a single drive*. I mean, I know that stuff inside and out from working in the EV industry, but it's not on industrial stuff for a reason and that's unnecessary complexity.

    This puts me back to the following restrictions:

    - 500-600Hz knee point
    - 155VDC or 310VDC at the knee.
    - Runs from a VFD or Servo Drive
    -- Encoder built in to the motor for closed loop vector control and indexing
    - 6.3kW peak power (up to this point I've been toying with the 2.2kW level to match the benchmark from Mechatron)

Page 1 of 2 12

Similar Threads

  1. Mini Mill X2 Spindle Head + VFD + Chinese Spindle= custom!
    By diyengineer in forum Uncategorised MetalWorking Machines
    Replies: 9
    Last Post: 12-26-2016, 10:04 PM
  2. Custom Motor
    By gicos in forum Metal Working Tooling
    Replies: 2
    Last Post: 12-19-2015, 04:39 PM
  3. WTB Custom ATC Spindle Motor
    By Feeler_Gauge in forum Want To Buy...Need help!
    Replies: 0
    Last Post: 04-05-2015, 01:41 AM
  4. Custom bending a custom extrusion
    By brokenrinker in forum Bending, Forging, Extrusion...
    Replies: 10
    Last Post: 12-15-2007, 03:28 PM
  5. RFQ Custom Motor Endcaps
    By JustinM in forum Employment Opportunity
    Replies: 8
    Last Post: 07-20-2006, 07:03 PM

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •