[Windscreen]

I have a bed mill that I'm doing a CNC conversion on. The mill came with a 3 hp spindle motor and a built in drive. The drive is a proprietary unit, and the manufacturer won't provide any details or help to interface with it. Also, the PID tuning on it sucks (an understatement); it is extremely lazy and severely undershoots when loading down the spindle as you come into a heavy cut. So, I'd like to replace the drive with something my motion control board (Kflop) can interface with and control. Ideally, this would be an enabler for rigid tapping. I need to keep the motor, if at all possible, as I'm way over budget, already.

Here's the spindle motor
http://www.cnczone.com/forums/attach...d=286282&stc=1

I've been unsuccessful in tracking down any specs on this thing, but I'm certain it is of Chinese origin. I'm sure I'll also have to replace the hall encoder and adapt a proper shaft encoder to the stub shaft.

Is this doable? Where do I start?

Thanks,
-Steve

[greggv]

You could mount the encoder, get a drive ( I would use an AutomationDirect GS3 ) rated for your horsepower, and set it up with the encoder input. It "should" be able to rigid tap, depending on your control program. I think you have to get an encoder board seperate

[mike_Kilroy]

Looks like pm dc brushless. If so gs3 wont run it.

[engnerdan]

I have a Hitachi WJ200 series VFD for my spindle and there is a setting for "Permanent Magnet Motor". It is function code b171 set to 03. But they never expand on it in the manual so I am not sure if it requires an encoder or the hall sensors be connected.

I think you could verify that it is a permanent magnet motor by hooking up a meter to 2 of the winding leads (there should be 3 like a 3 phase motor) and spinning the motor. You should see a voltage output from the motor. My guess is it should be at least a couple volts. Because of the permanent magnet the motor will act as a generator when you back drive it.

-Dan

[Al_The_Man]

I have tried BLDC PM motors on VFD with very poor results, Hitachi claim PMM but when I contacted them they said it is by special customization.
Al.

[engnerdan]

I have tried BLDC PM motors on VFD with very poor results, Hitachi claim PMM but when I contacted them they said it is by special customization.
Al.

That would explain why they don't have any information in the manual about it.

-Dan

[Al_The_Man]

Is this doable? Where do I start?

Thanks,
-Steve

I think you will find that is just a 2 pole motor, which is not the ideal for a spindle doing rigid tapping etc, the Hall effects on these motors are usually actual hall effect sensors and not optical like most.
If you short the stator winding's and try and spin it, count the number of 'Bumps' / rotation and this will tell you the pole count.
The A-M-C drives will work with these motors, but are pricey for that size.
Al.

[Windscreen]

I have a good 4 channel scope. If I record the signal on one or all three legs, can that be used to tell AC vs BLDC, and/or pole count? That would be quite an easy measurement for me to get.

-Steve

[Al_The_Man]

The motor is definitely BLDC per the Hall sensors.
If it is 2 pole then there would be one AC sine wave output per rev.
To scope it, wire a 1k resistor from each stator and connect the three resistors to create a ref point for the scope common.
Al.

[Windscreen]

Thanks Al. I don't have any differential probes, so my scope common reference will basically be a short to ground through those 1k resistors and the scope's ground traces. At the name plate 230 V, that's potentially 48 W through a resistor. I'm guessing I'm missing something here, as that's a big 'ol resistor. Does it work because you're doing this at no load?

Is there any reason I couldn't run one scope channel on each leg, put the 4th channel on some point to use for the ground reference, and do math between the first three channels and the 4th?

Thanks,
-Steve

[Al_The_Man]

I mean to measure with a scope, wire the resistors from each phase and connect the other resistor ends together to form a conceptual common to attach the probe to. do not power the motor, but spin it by hand or back feed it with a separate motor and you will see the wave forms on each phase as shown in the PDF, I use a DB scope to align the Halls to the winding as shown.
Al.

[Windscreen]

Ahh, doing that in motored operation makes much more sense! You can see where I was confused, thinking about doing it under power.

-Steve

[engnerdan]

Steve,
You have a scope, how good are you with electronics? I question because even though the PID control on that factory drive is poor, it might be a means to get you up and going with out spending more money. What I am thinking is there can only be about a half dozen wires running to the factory drive and "typically" there are only a couple common means of control. I would think there should be a good chance of figuring it out from what is there, with a scope and a multi meter. Additionally if you can post pictures of the drive and wiring, maybe its the same as another import drive sold by another company.

What are you using for a interface to the computer, break out board and Mach3, KFLOP, other?

-Dan

Thanks Al. I don't have any differential probes, so my scope common reference will basically be a short to ground through those 1k resistors and the scope's ground traces. At the name plate 230 V, that's potentially 48 W through a resistor. I'm guessing I'm missing something here, as that's a big 'ol resistor. Does it work because you're doing this at no load?

Is there any reason I couldn't run one scope channel on each leg, put the 4th channel on some point to use for the ground reference, and do math between the first three channels and the 4th?

Thanks,
-Steve

[engnerdan]

Does this wiring schematic look anything like the drive for your spindle motor? If so it looks like pins 10 and 11 on connector X6 are your variable speed. o to 5 volt analog signal. Not sure on which connections are for forward and reverse, but that should be easy enough to figure out. If you went with the KFLOP and KANALOG you talked about in your build thread a year ago, you should have no problem interfacing to that. I can help with that if needed.

-Dan

[Windscreen]

Hey Dan,

Yes, I am building my mill with a KFLOP and KANALOG. I've made little progress on the CNC side, as I poured all my time and efforts into building a proper shop to house everything. Realistically, I'll be resuming the conversion this fall, and trying to have my ducks in a row is why I started this thread.

Interfacing with the factory Optimum drive definitely doesn't scare me. It's the horribly lazy spindle speed control that I'm concerned with. If you have the mill in low gear and spin the speed pot to minimum, the spindle will actually stop for a moment, and then go back to minimum speed. I was roughing some steel recently, and as the tool came into the cut, the speed would dip (audibly) 25% or more, and then recover.

I'm worried the resultant high chip load is going to break or chip tools, or screw up expensive work pieces. It won't take many of those before a new drive or drive and motor would pay for themselves. Or, maybe I'm being overly cautious.

I suppose one could use the KFLOP to drive the speed demand signal via a PID loop on the KFLOP. But, then you've got a PID on top of a PID, and those can be very hard to tune, based on my engine control systems experience.

I'm pretty good with electronics, particularly around data acquisition and interconnecting of complex systems. Where I fall down is troubleshooting complex circuits or building circuits beyond simple R-L-C systems.

-Steve

[engnerdan]

Hey Dan,

Yes, I am building my mill with a KFLOP and KANALOG. I've made little progress on the CNC side, as I poured all my time and efforts into building a proper shop to house everything. Realistically, I'll be resuming the conversion this fall, and trying to have my ducks in a row is why I started this thread.

Interfacing with the factory Optimum drive definitely doesn't scare me. It's the horribly lazy spindle speed control that I'm concerned with. If you have the mill in low gear and spin the speed pot to minimum, the spindle will actually stop for a moment, and then go back to minimum speed. I was roughing some steel recently, and as the tool came into the cut, the speed would dip (audibly) 25% or more, and then recover.

I'm worried the resultant high chip load is going to break or chip tools, or screw up expensive work pieces. It won't take many of those before a new drive or drive and motor would pay for themselves. Or, maybe I'm being overly cautious.

I suppose one could use the KFLOP to drive the speed demand signal via a PID loop on the KFLOP. But, then you've got a PID on top of a PID, and those can be very hard to tune, based on my engine control systems experience.

I'm pretty good with electronics, particularly around data acquisition and interconnecting of complex systems. Where I fall down is troubleshooting or building circuits beyond simple R-L-C systems.

-Steve

Sounds like you have similar experience/skills as I do. I don't design board level electronics, but I can probe the hell out of a board and attempt to figure out how it works.

I get what you are saying about the crummy tuning, I was thinking maybe there was a way to get the drives PID loop to a usable state and let the KFLOP take over from there.

Is this a common problem on the BF46 machine or is it just your particular drive that is crappy?

If I remember correctly BLDC motors are not very high on torque and people were taking them off the older Novakon machines in favor of a AC induction motor.

-Dan

[Al_The_Man]

If I remember correctly BLDC motors are not very high on torque and people were taking them off the older Novakon machines in favor of a AC induction motor.

-Dan

Not quite true with the right drive and controller, The are used extensively for servo's, where they excel as a spindle is where there is a PID loop involved as with a servo, they have issues if used in open loop velocity mode.
Which I assume the OP motor is and the issue aggravated if 2 pole.
I have used large Fanuc servo's as spindle motors using and encoder.
Al.

[engnerdan]

Not quite true with the right drive and controller, The are used extensively for servo's, where they excel as a spindle is where there is a PID loop involved as with a servo, they have issues if used in open loop velocity mode.
Which I assume the OP motor is and the issue aggravated if 2 pole.
I have used large Fanuc servo's as spindle motors using and encoder.
Al.

I have forgot about the high output Fanuc motors being BLDC motors. Fanuc has managed to squeeze a lot of power into a small package with their permanent magnet motors.

-Dan

[Windscreen]

Dan:

The electrical schematic you attached, earlier, does match that of my machine. FYI - two of the limit switches you questioned are for a tapping mode, where in the spindle automatically starts (at half speed) when you crank down the spindle sleeve, and then reverses when the sleeve reaches a settable depth stop. It also has a similar drilling mode, where only the upper prox switch is used to start and stop the spindle. The third switch is for a spindle safety guard, which got the deep six about a day after I unboxed the mill.

I haven't heard anyone else complain about Optimum having poor speed control on their drive, but there isn't much forum discussion on Optimum mills, either. I'll have to do some asking around on that one.


Al and Dan:

I'm a bit confused by the discussion regarding open loop velocity mode, when there is a hall-effect sensor in use. Isn't that for feedback control? Or, in this case, is the hall sensor merely acting to trigger some sort of passive commutation circuitry, which the drive microcontroller doesn't actually have control over?

I ran out of time tonight to do any probing with the scope. Hopefully I'll get some data tomorrow night, and report back in; maybe even post up a wave form or two.

Thanks for the help,
-Steve

[engnerdan]

Dan:

The electrical schematic you attached, earlier, does match that of my machine. FYI - two of the limit switches you questioned are for a tapping mode, where in the spindle automatically starts (at half speed) when you crank down the spindle sleeve, and then reverses when the sleeve reaches a settable depth stop. It also has a similar drilling mode, where only the upper prox switch is used to start and stop the spindle. The third switch is for a spindle safety guard, which got the deep six about a day after I unboxed the mill.

I haven't heard anyone else complain about Optimum having poor speed control on their drive, but there isn't much forum discussion on Optimum mills, either. I'll have to do some asking around on that one.


Al and Dan:

I'm a bit confused by the discussion regarding open loop velocity mode, when there is a hall-effect sensor in use. Isn't that for feedback control? Or, in this case, is the hall sensor merely acting to trigger some sort of passive commutation circuitry, which the drive microcontroller doesn't actually have control over?

I ran out of time tonight to do any probing with the scope. Hopefully I'll get some data tomorrow night, and report back in; maybe even post up a wave form or two.

Thanks for the help,
-Steve

Al has more experience than I do so he can most likely elaborate some. From what I know it is not possible to run a BLDC motor without the hall effect feedback, because they take the place of the commutator of a brushed DC motor. Now with that the drive needs some form of PID loop to handle changes in speed or load. So by default a BLDC motor is a closed loop system of its own. In this case it is a "velocity" or "speed" system. So you give it a voltage input on pins 10 and 11 and it scales that against its internal settings max and min speed. Its whole goal is to maintain a set speed. Now as to why yours acts so erratically is beyond my experience, as I have not tuned very many motors.

What I do know is if the hall effect sensors are not positioned where the drive thinks they are, it will not run right. From my little experience I know it is possible for the some drives to do a detection routine to find where the halls are (this is done with the KSTEP) but I think the sensors are are typically at a fixed position and the driver just knows where the sensors are. Now that I say that, it makes me wonder if your problem might be related to the timing of the hall sensors.

To understand the hall sensors a bit more, check out this PDF http://www.pmdcorp.com/downloads/app...nsorConfig.pdf

Have you tested the speed POT to see if that is working fine?

-Dan