[Fish4Fun]

I am curious if anyone has thought of using BLDC motors/controllers from the RC airplane hobby world for CNC spindle motors. There are some very healthy motors (up to 8.4kW!) in the 12V to 60V range with matching speed controllers (granted not exactly set up for Mach3 control). The more powerful motors (2kW to 8.4kW) are in roughly 3" diameter housings, and can easily be re-wound to increase RPM or Torque. The stock RPM Range is from 8k to 20k. The concept is pretty simple, the electronics a bit more complex, but the potential for driving a spindle seems pretty good. The RC airplane market is considerably larger than the CNC spindle market, and so mass production prices have less engineering baggage than narrow market products. The 8.4kW motor mentioned (6.4kW continuous) has a price tag of ~$270 w/o controller. Here is a link to one particular motor TowerHobbies.com | GPMG4805 Great Planes Rimfire 65cc Electric Motor it turns ~8.8k RPM @ ~ 55V. Obviously these motors would not be ideal for direct drive of a spindle, but with a belt they could cover the RPM range from 800 to 40k RPM with as few as two "step pulleys" (1:5 5:1). If used to drive a quality spindle the potential appears to be very good at a reasonable price. It is a bit hard for me to grasp "switching" 120A (I consider 120A to be in the "welding" range), but it appears there are existing controllers designed specifically for the task. It seems probable that for extended run times, water cooling might become important, but with the savings over CNC specific designs, this is not an insurmountable expense.

Anyway, I was just wondering if anyone else had looked at these motors wrt building a spindle.

Fish

[atwooddon]

Quite a bit of discussion about using the RC brushless motors to drive spindles in the latter part of this thread.

http://www.cnczone.com/forums/diy-cn...ndle_idea.html

There are also several videos on YouTube showing this approach.

Don

[tskguy]

I have been using a 1kw RC brushless motor for a while now and it works pretty darn well! My motor is 1200kv and I run it at 15volts for about 18000 max rpm with a 1:1 ratio, I also have other ratios to slow it down as needed but mostly I run at faster feeds to deal with the RPM. The biggest problem with the larger KW motors will be supplying it with 55v at 150amps!! That will be one hell of a power supply!! My 15v 80amp supply is pretty beefy and is much smaller than that monster would be. One thing I am currently working on is actually controlling the rpm through Mach3. I’m pretty close but have a few issues that I’m still trying to work out. One major issue will be getting mach3 to initialize the ESC that these motors use. I’m sure a macro in mach3 would work but I haven’t tried that yet either. Any way good luck!

[CarveOne]

I have been using a 1kw RC brushless motor for a while now and it works pretty darn well! My motor is 1200kv and I run it at 15volts for about 18000 max rpm with a 1:1 ratio, I also have other ratios to slow it down as needed but mostly I run at faster feeds to deal with the RPM. The biggest problem with the larger KW motors will be supplying it with 55v at 150amps!! That will be one hell of a power supply!! My 15v 80amp supply is pretty beefy and is much smaller than that monster would be. One thing I am currently working on is actually controlling the rpm through Mach3. I’m pretty close but have a few issues that I’m still trying to work out. One major issue will be getting mach3 to initialize the ESC that these motors use. I’m sure a macro in mach3 would work but I haven’t tried that yet either. Any way good luck!

Deleted reply. Dumb idea I had wouldn't work.

[toughtool]

CarveOne,
Now we won't know what not to try, and why. Isn't that like "There are no such thing as a dumb question". Joe

[CarveOne]

CarveOne,
Now we won't know what not to try, and why. Isn't that like "There are no such thing as a dumb question". Joe

If I told you what it was, and everyone started laughing at me for thinking of it, I might feel obligated to lay on the CNC table and V carve "DUMB A--" into my forehead. We can't have that, now can we? :nono: (Kidding of course.)

CarveOne

[toughtool]

Tskguy,
Making a power supply for one of those motors shouldn't be too difficult. I would use the core from an old 2KW variable transformer, like the Variac. By removing the wiper, then insulating and tapeing the wires where the carbon brush traveled; a secondary winding can be wound with little difficulty. That is what I did to get the 15 amp supply for my stepping motors, as well as an additional small 8 volt secondary winding for a 5 volt reg. The AWG wire size of #4 (.204" dia) should handle the 120 amps. While #4 is a lot bigger than the #10 stranded I used for my 24 volt supply. I did make one once (for a solid state KW Amature Radio transmitter) using solid #8. At about .5 volts per turn, doesn't take a lot of turns to make a 15 Volt supply. If the voltage sags a little under load, just add a couple more turns. I don't think these motors care about a highly regulated and filtered supply so a chassis mounted bridge diode assembly and a hefty capacitor should do the trick. Joe

[Fish4Fun]

atwooddon,

I had only read through about page 10 of the spindle thread when I made this post, like always I should have kept reading prior to posting, but sometimes I think I have thought of something new and just can't resist posting :-) Thank you.

toughtool,

I have NOT played with any of these motors, or any power supply designed to pump out more than 20A at any voltage (except welders), but it seems to me a welder might be a good place to start, lol. It's a shame so many of the old "buzz boxes" have been scrapped for their copper, but perhaps a cheap modern welder could be retrofitted to act as a power supply for these high current motors? Something like:

Northern Industrial Arc 200 230 Volt, 200 Amp Arc Welder, Model AC4180 | Arc Welders | Northern Tool + Equipment

@ $149 might be worth investigating. Most stick welders have an Open Voltage ~80V, but the Voltage drops quickly under load. A fairly rugged rectifier could be fashioned from industrial SCRs in anti-parallel and the output VA could be controlled by a combination of the welder's settings and phase firing of the SCRs. I would think for < $500 one might be able to achieve a fairly reliable 50-100A supply in the 20V to 50V range. Filtering the 60hz chopped output could require a large capacitor bank if output ripple is a consideration.

This is just mind-candy for me right now, I have not looked seriously at these motors or building a suitable supply for them. There are probably several threads that deal specifically with building/sourcing high power spindle supplies that I just haven't read yet, lol. So much to read, so little time....

Fish

[Fish4Fun]

RomanLini,

The Thread Title of 8.4kW was chosen for shock value, not out of any need for an 8.4kW spindle. I have been "watching" these motors evolve for a number of years, and they seem to have matured very nicely in the last few with prices dropping and performance/reliability increasing. The primary advantage I see to them is the very large, competitive consumer market they are aimed at, much like the Bosch Colt is aimed at a much larger market than CNC router spindles; this large market greatly increases the performance to cost ratio. I feel like the 1kW to 2kW BLDC motors have reached a place in the cost/performance curve that we as CNC hobbyist/professionals should be looking at them seriously for spindle motors. While power to weight ratio is NOT a primary concern for most spindle motors, less mass can have advantages.

Thank you for offering to take a serious look at them.

To everyone who has responded,

Thanks! The thread is going exactly the direction I had hoped it would!

Fish

[CarveOne]

Some of these motors also have water cooled versions for the R/C boat guys. I don't know much about them and they may or may not be useful as a CNC spindle. I have installed some Hacker C50 geared motors in an Osprey wind tunnel model we built during the Osprey crash investigations, so I know that something like that is available for lower speed higher torque needs.

CarveOne

[blighty]

found this one HK-3026-880KV Scorpion Brushless Motor For Protos

just to give you a rough idea on whats out there.

Specifications
Stator Diameter ............................ 30.0 mm (1.181 in)
Stator Thickness ........................... 26.0 mm (1.024 in)
No. of Stator Arms ................................................ 12
No. of Magnet Poles ............................................... 10
Motor Wind ............................................ 10 Turn Delta
Motor Wire ..................... 14-Strand 0.25mm (30 AWG)
Motor Kv ............................................. 880 RPM / Volt
No-Load Current (Io) ................... 2.30Amps @ 6 volts
Motor Resistance (Rm) ............................. 0.026 Ohms
Max Continuous Current ............................... 52 Amps
Max Continuous Power .............................. 1450 Watts
Weight ......................................... 199 Grams (7.02oz)
Outside Diameter .......................... 37.5 mm (1.476 in)
Shaft Diameter ............................. 5.00 mm (0.197 in)
Body Length ................................. 48.4 mm (1.906 in)
Overall Shaft Length ...................... 74.2 mm (2.921 in)

[tskguy]

Ive heard ood things about the Scorpion motors. Im a big fan of Neu motors
Welcome

They have a line of Heli motors that have internal fans to help cool them. They are pretty pricey though.

Eric

[groswald]

RomanLini,

I'm not sure what the source for your data is, but you may want to do some more research. The motor is this airplane is an e-Flite Power 160. Rated at 2700 Watts burst (37V at 75A), 2200 Watts continuous.



My own direct measurements confirm those numbers. Swinging a 22x10 prop and running on 10 fresh 5000MaH LiPo's (36.8V on the meter) it peaks out at just over 2700 Watts drawing 75 Amps. I'd say that is pretty accurate. I cannot say how much of that energy is lost as heat, but I can say performance is certainly equivalent to a two stroke 1.6c.i. engine putting out 3.7 rated hp (2759 Watts). I have to believe that efficiency is in the 80 to 85% range based solely on the motor's demonstrated ability to haul this very draggy, 22lb. air-frame around with authority.

Regards,

Randy

[Fish4Fun]

I do not have any first-hand knowledge of the ESCs used with these motors, but I have done some reading, and the reading appears to present some contradictions. The motors themselves have specifications implying RPM is directly related to input Voltage (KV = K rpm per Volt). This seems straight-forward enough; however, these motors are permanent magnet, 3 phase motors, which implies that the RPM is tied directly to the frequency of the drive signal (Output of the ESC) and number of poles, and should have little to do with the DC input voltage applied at the ESC input. This, of course, led me to more reading and some thinking. Finally a little bell went, "ding ding" in my head, and I think I understand how/why these seemingly conflicting viewpoints on the motors and their ESCs actually come together :-)

First thing is to throw out every preconception you might have from working with conventional industrial VFDs and 3 phase motors. A lot of information is in the name "Brushless DC Motor".

In a conventional Brushed DC motor the commutator alternately energizes some number of coils, and the amount of current that is drawn is dictated by the supply Voltage, the winding Inductance, the winding Resistance and the period of Time any particular coil is energized. For any given Voltage the physics of the motor dictates a "no-load RPM". From this base-line maintaining any particular RPM at any given load is handled by increasing or decreasing the Voltage, and the torque at any given RPM is limited only by the motor's ability to handle current and your speed controller's ability to supply current. Pretty straight-forward.

A conventional 3-phase VFD provides a 3-phase signal to a 3-phase motor. The motor rpm is directly linked to the frequency of the drive signal. The amount of current required is dictated by "rotor slip". Rotor Slip is a measurement of the phase angle difference between where the rotor "should be" vs where it actually is. Under no-load conditions there should be very negligible slip, as the load increases, more current is required to keep the rotor "where it is suppose to be", and the "slip angle" increases. The ability of a 3-phase motor to provide torque at any given RPM is dictated (much like the brushed DC motor) by the current capabilities of the motor and the Supply.

With the RC BLDC motors and ESCs things are a bit jumbled. The required RPM output margin for error is quite large, so the expense of a true VFD is not requisite. In the RC BLDC's design 2/3s of the coils are energized at any given time, the other third of the coils are used to provide timing information to the ESC. Each set of coils takes a turn being the "sensor". The ESC itself is a simple PWM current regulator, and it doesn't care how much time is spent supplying that current to a particular set of coils, only that at particular positions in the shaft rotation that it changes which coils are energized. In this it is much like the commutator in a conventional Brushed DC motor. We don't need to know anything about how the ESC achieves current limiting or shaft position sensing to take control of it; we simply need to add a shaft postion sensor to monitor RPM and then adjust the "throttle position" in the ESC to increase or decrease RPM/Torque. Our ability to regulate speed will then become a question of ESC response time. Unlike a conventional VFD, the ESC does not have a "target frequency", rather it simply "responds" based on the shaft sensing. That is, the Voltage/current can be increased w/o a corresponding increase in RPM.

While it would be ideal to design an ESC from the ground up for our purposes, this could greatly increase the cost. Experimenting with various ESCs and testing their response time is likely to prove the least expensive approach. Assuming a single shaft sensor and 10,000 RPM, this would imply a uController would get new RPM data at 10uS intervals. The ESC interface is most likely controlled by a standard RC pulse width controller, so it responds to pulse widths from .5mS to 2mS every `20mS. This means that up to 2000 RPM samples could be averaged between adjustments to the ESC, and that up to 50 speed adjustments could be made per second.

While 50 adjustments per second might prove adequate, a dedicated design has the potential to greatly improve the feedback loop. Perhaps some of the ESCs use FLASH based uControllers and could be re-programmed via JTAG. (Did some checking, here is an excellent link on controlling an ESC via I2C/TWI! : Converting TowerPro 25A type 2 ESC's from PWM to TWI/I2C control - RC Groups ). In a casual browse through the ESC link (1149 posts!), it appears that everything we might want to do to these ESCs has been done! It appears many of the "older" ESCs used an atmega8 uController, while many of the "newer" ESCs have moved to Silabs based processors (formerly Cygnus). Both processor families are well documented and easily re-flashed. The RC hobby folks are extremely weight conscious, so it might be reasonable to port the same basic design to a larger DIY PCB with heat sinks and larger FETs. I am certain any commercial attempt at selling "borrowed" designs would end badly, but DIY PCB's for the CNC crowd would likely not ruffle too many feathers. It all looks very promising.

Fish

[tskguy]

Fish4Fun,

That is very exciting stuff!! Great info, lets hope Roman can use it. I like the idea of a larger diy driven ESC for these motors. Sadly I wouldnt have a clue on were to start. I am no eletrical engineer!
I do have a crappy esc and motor we could use for testing if needed. Let me know.

Eric

[blighty]

Ive heard ood things about the Scorpion motors. Im a big fan of Neu motors

just put that up as a guide to whats out there, but yes Scorpions are expensive.

as Fish4fun says, the voltage for the motor is fixed and the ESC changes the frequency. only thing is i don't think there infinite i.e you can't get 1001, 1002, 1003rpm that's if a brushless esc is working like a brushed esc. i have made a few brushed esc for rc and they also work buy frequency, but it was in block form. 200rpm then jumps to 500rpm then 800rpm etc..... on the esc that we did this was down to the size of the memory chip, we didn't have enough room on it for bigger look up table. this "stepping" was more noticeable on the cheaper esc.

if this is still the case with brushless esc then you might get a problem when trying to get specific rpms. i have a load of brushless esc and motors sitting around i could do some rpm tests if that would help.

[tskguy]

Oops!! That was supposed to read..."Ive heard good things about the Scorpion motors. Im a big fan of Neu motors"

Eric

[PaulRowntree]

I have built controllers for BLDC motors from hard disk drives using an optical encoder wheel to indicate true position. This can control the frequency to any value. Commercial controllers can used current sensors on the non-powered leads to determine position, and to set the timing need for the frequency control.
Why would RC plane ESC's not do something similar?

[Al_The_Man]

In a conventional Brushed DC motor the commutator alternately energizes some number of coils, and the amount of current that is drawn is dictated by the supply Voltage, the winding Inductance, the winding Resistance and the period of Time any particular coil is energized.

A conventional 3-phase VFD provides a 3-phase signal to a 3-phase motor. The motor rpm is directly linked to the frequency of the drive signal. The amount of current required is dictated by "rotor slip". Rotor Slip is a measurement of the phase angle difference between where the rotor "should be" vs where it actually is. Under no-load conditions there should be very negligible slip, as the load increases, more current is required to keep the rotor "where it is suppose to be", and the "slip angle" increases.

Fish

I know I came to the game late but I read through all the posts and maybe can make some things clearer.
The brushed DC motor current relies on BEMF to limit the current, with no load, the BEMF voltage approaches the applied voltage and current is minimum, as the load increases, the rpm drops and hence the difference between BEMF and applied Voltage increases and in turn current.
A normal 3ph induction motor can never be synchronous, as it requires slip to induce a magnetic field into the rotor.
the lower the slip, the lower the current, apply a load, the rpm drops and the slip increases in turn as does the current.
The difference in construction between a AC P.M. motor and a BLDC is indistinguishable.
The difference is in the commutation, the BLDC is so named because it represents a DC brushed motor turned inside out, and only two of the 3 windings are energised at any given time.
The P.M. AC has three phases applied to the stator 120deg apart, and can be truly synchronous.
The amount of poles of a BLDC or P.M.AC indicates how many electrical revolutions or cycles there are per mechanical revolutions.
e.g. a 8 pole motor has 4 electrical revolutions per mechanical.
Al.

[JohnZ]

The difference in construction between a AC P.M. motor and a BLDC is indistinguishable.
The difference is in the commutation, the BLDC is so named because it represents a DC brushed motor turned inside out, and only two of the 3 windings are energised at any given time.
The P.M. AC has three phases applied to the stator 120deg apart, and can be truly synchronous.
The amount of poles of a BLDC or P.M.AC indicates how many electrical revolutions or cycles there are per mechanical revolutions.
e.g. a 8 pole motor has 4 electrical revolutions per mechanical.
Al.

I've got to ask a question I have been wondering about for a long time now.

Since a BLDC and P.M. AC motor are constructed in essentially the same manner, can you run a BLDC motor with a VFD designed for a three phase AC motor, and on the flip side, can you run a P.M. AC motor with an ESC designed for a BLDC?

Thanks - John Z