4 Attachment(s)
New life for 2.2kw BLDC Spindle
So maybe the Mikini isn't quite dead...
If there's anyone reading this who owns (or owned) a Mikini, you'll know that the spindle motor was one of the fundamental problems with the mill. Numerous reports of the motor cutting out at higher speeds and pathetic torque at lower speeds. Various threads in this forum all suggested that the drive - regardless of version - was at fault.
Broken tools and terrible performance resulted from the retarded drive. Much unhappiness.
There have been many threads discussing sourcing a replacement drive, but I've not see anything followed to completion. My investigation years ago only resulted in a couple of >$800 drives from off-shore sources with non-English manuals. Most threads wound up with the owner pitching the motor & drive and switching to an induction motor + VFD, or some other solution (AC servo). All of this usually requires retrofitting a new controller as the Mikini spindle output signals are an odd-ball format.
I believe Swath (see threads here) has deciphered the 'code' and created an intermediate board to translate between Mikini and industry standard 0-10V (or PWM) signals. The rest of us either ditched the entire controller or pitched the mill. Because I was not interested in an $800 BLDC drive experiment or a $4k AC servo experiment, I went with a 2hp induction motor and VFD, along with a PMDX-126/107 and use UCCNC. There's nothing Mikini left on the control side.
This issue has been bugging me for years. There's no reason it should be hard to run the BLDC with a generic drive, but at the time I was rather electrically challenged. Still am, to be honest. Frankly, I was tired of what appeared to be a perfectly good motor sitting around doing nothing and taking up space. Not like I could Ebay it - who wants a motor with no way to drive it?
But a few things have come together over the past couple of months and I decided to have another go at the motor.
First, I discovered through an older thread in the Servo forum that servo drive/amplifier manufacturers used to make 'dumb' analog drives designed for brushless servos (some still do). Old news to some here, but that was a revelation. These drives were designed to use various feedback types (encoder/resolver/etc) and take a +/-10vdc input for torque or velocity control. They were designed to be 'generic' amplifiers - not like typical drive/motor combos from a specific manufacturer with proprietary signals. Maybe one of these could be purchased on the cheap? And some of them were rated for quite high amps & volts.
Older threads here indicate the Mikini motor is a 13A @320Vdc unit, and there are a few of the older drive which appeared to be able to supply enough juice on 240VAC single phase to get the motor doing its thing. As for the rest of the motor characteristics, Mikini refused to provide 'standard' data for the motor, perhaps in an attempt to discourage reverse engineering. Most of the data normally provided by motor manufacturers is simply missing (rotor inertia, rated speed, inductance, etc.)
But how to take the 120 degree Hall signals from the Mikini and send it to the amplifier? It's got to be commutated in some way. And could I do better than Halls? My internet education indicated that the Halls-only feedback is never very good for low-speed velocity control - one of the defining characteristics of the crippled Mikini spindle. No low speed torque - no drilling steel, hunting for RPM.
Second - about the same time I stumbled across commutation encoders. Again, forgive me if this is old news to readers - but it appeared I could get a generic differential output encoder with Hall outputs built-in. The problem there is how to align the Hall signals with the motor phases. As I mentioned I'm electrically challenged - motor phasing and 'back EMF" are black magic. I don't have a two-channel O-scope or the experience to fiddle with this stuff. I need Legos.
A bit more scrounging and internet digging and I came across the CUI Devices AMT-31 series of encoders. $40 got me a programmable encoder with A/B/Z plus 120 degree Hall signals. AND - the AMT software will auto-phase the Hall signals. Holy crapski.
Maybe a generic drive plus a generic encoder plus a generic Hall signal for commutation would get the motor working. And it wouldn't be just a dumb BLDC, but with the right setup the motor could be used in positioning mode - i.e. an actual servo. Maybe indexing or homing the spindle for a tool changer? Too much to ask?
So looking at some of the old brushless amplifiers - ebay and datasheets - and I see that they are all really, really dumb. They use potentiometers & dip switches to adjust the motor tuning & control loops. I have zero experience with this and don't want to go down a entirely new hobby rabbit hole of becoming a motor expert.
I'll get to the point here. After some discouragement, especially because the older drives are now commanding decent money on ebay, I eventually stumbled across the Copley Controls Xenus line of amplifiers. These are their newest line of drives - software programmable and one of the rare drives which can be driven from 240vac single phase with some serious amps. Expensive drives brand new, but not only can they take just about any feedback and control input possible - they also have an auto-tuning AND auto Halls phasing feature.
Seriously - plug some 3-phase random motor with a random commutation encoder installed by a clown and the drive will figure it all out. Could it be that simple?
It appears the answer is yes - very, very Yes.
I bought a cheapish surplus Copley XTL 230-36 drive, and AMT encoder, and enough connectors to get it all talking. Making the connections has been fussy (encoder cables, about 8 different connector types, etc.) but it all just works now.
The old Mikini BLDC is now spinning happily. The motor can accelerate wicked fast and hold speed within a couple RPM. It also hold position just like a 'real' AC servo and is incredibly quiet using sinusoidal commutation. It also sounds happy at 6kRPM.
An ill-advised but very satisfying 'leather glove on the pulley" test indicated it has stupid high torque at 50RPM. Yes - 50. No PWM squealing, no hunting - just gobs of ass. Baby got back.
This amplifier is better than Nutella and bananas.
So what's the next step? Get a bigger motor, of course, and see if this Franken-motor-encoder & drive combo works. I'm going to see if the drive will work with a much bigger real AC servo and then I can get rid of the 80lb cast iron induction motor.
What this means for anyone reading it: you now have options. I have no clue how to integrate this drive in to the OEM Mikini electronics, but maybe someone can figure it out. More importantly, if you are (or already have) retrofitting the Mikini you don't have to replace the spindle motor. My inexpert analysis indicates the motor isn't grossly overrated when driven by a proper amplifier. 2-3HP is probably a reasonable motor given the mill's stiffness limitations. And the OEM motor is shockingly light & compact compared to an AC induction motor.
Copley drives aren't terribly expensive on ebay, and there are other manufacturers of 'generic' servo drives with the critical single-phase 240vac input plus PWM or other easy control inputs.
Almost as important... anyone who purchased one of the old Sangmutan/HiPool/Automation Technologies 1.1kw/2.2kw BLDC spindle motor & drive combos and was shocked at how poorly they performed can now breathe some new life in to the motor.
Don't ask for youtube videos - I don't post anything there. But I'll try to attach a couple of phone clips here.
Good luck.
