Need info on Fanuc brushless motor

[KTP]

I have a Fanuc A06B-0512-B504 red cap brushless motor (new) manufactured in 1999 that I would like to use on my cnc mill to replace a brush motor. Does anyone have the pinout of the 17 pin connector on this series motor? I am starting to think that perhaps it does not have hall sensors, but just the pulse coder output. I was under the impression that all drives need hall sensors at least on startup to get the commutation right. Is this true? I took the outer red cap off and the pulse coder is A860-0315-T101 2000P. All the wires in the 17 pin connector go to the pulse coder...no evidence of halls coming from the rest of the motor.

ideas?

[unterhaus]

there are alternatives to halls. The most obvious is to simply have tracks that put out electrical signals that are just like halls. Similarly, you can have absolute encoder tracks that tell you where you are to greater accuracy. If a motor is designed to always be coupled with an encoder, it will almost always skip the halls because of the ease of combining the commutation function with the encoder. Discrete Hall sensors would be more expensive in that case.

[Al_The_Man]

The Fanuc AC servo has an encoder and a 4bit binary output for commutation.
Pin
A- A
B- B
C- Bit1
D- /A
E - /B
F- Z
G- /Z
H- SHIELD
J- +5V
K- +5V
L- Bit 4
M- Bit 8
N- 0v
P- Bit 2
R- HnA ?
S- HnB ?
T- 0v
Al.

[KTP]

Oh thank you thank you for the pinout info!

[KTP]

Ok, so using those pinouts (thanks again!) I probed Bit 1, Bit 2, Bit 4 and Bit 8 using my oscilloscope and hand rotating the motor shaft.

It appears Bit1 pulses 16 times per revolution, Bit2 8 times per revolution, Bit4 and Bit8 both pulse 4 times per revolution, 90 degrees out of phase with each other.

weird!

I was more expecting Bit1 to do 8 times, Bit 2 to do 4 times, Bit4 to do 2 times and Bit 1 to change once per rev. Or at least 16,8,4 and 2.

But anyway, I can work with what I get. I am going to feed these 4 signals into a little cpld test board I have and generate 3 hall type signals that I will use with a AMC BE15A8E-H brushless amplifier. This particular amp has selectable 120 degree or 60 degree phasing. The fanuc motor is 8 pole. Hmmm, I need to find some info to determine what my hall signals should look like with respec to this motor and drive...

I will post success or failure stories :-)

[Al_The_Man]

I am not sure how Fanuc use these signals and also the Ha Hb ? They do have a procedure for setting the 4bits up to orient the encoder, This is done by injecting a DC voltage into the windings for positioning the armature and the 12 bits have to read a certain pattern.
If you want this procedure I could post it.
Al.

[KTP]

yes please...any usefull info like that would be much appreciated.

My encoder *should* be correctly aligned to the motor shaft since the motor is brand new, but that alignment procedure might shed some light into how these signals are meant to be interpreted.

not real sure what the H signals are...

[Al_The_Man]

The exitation of the windings is done by connecting the V&W together and this is the -ve supply the +ve goes to U terminal. Fanuc state a DC current that should be passed depending on the motor, for 3-0S=1.2amp up to 20S~70S=20amps, I am not sure why you would need that just to orient the armature?
The bit value in this condition should be:
C1=1 C2=1 C4=1 C8=1or0
I have used a car battery with a suitable resistor.
Al.

[KTP]

Thanks Al, you the man :-)

I hooked V and W up to gnd and U up to a 5V supply and the motor shaft locked to a new position. The 4 leds I now have hooked up to B8, B4, B2 and B1 indeed read 0, 1, 1, 1 as you said it should.

Also, I mapped out the B8,B4,B2, and B1 signals as the shaft is rotated. Four times per motor shaft revolution, you get this pattern:

B8 B4 B2 B1
0 0 1 1
0 0 1 0
0 0 0 0
0 0 0 1
0 1 0 1
0 1 0 0
0 1 1 0
0 1 1 1
1 1 1 1
1 1 1 0
1 1 0 0
1 1 0 1
1 0 0 1
1 0 0 0
1 0 1 0
1 0 1 1


So it is like an absolute position indicator for 1/4 of a revolution. An 8 pole motor rotates 1/8 revolution for 1 cycle of U,V,W, correct?

Now I need to figure out the logic to translate those 16 table entries into 3 hall signals that would be correct for hooking this 8 pole motor up to an AMC drive with 60/120 degree phasing....

[KTP]

Success!

Maybe not perfect, but I now have the Fanuc motor running on a AMC15A8 drive. It seems to have just a bit more torque ripple CCW than CW at very low speeds (under 10rpm or so). This is all just using the built in test on the drive in current mode, with no velocity feedback. I still may need to shift the conversion of the 16 four bit values to 3 hall signals by one place foward or backward, or move which halls get short changed in timing because 16 doesn't divide by 6. It runs *really* smooth CW, even at 1 rpm..can barely hear the commutation. This was after I shifted the 16 value to 3 hall conversion 1 place foward. Before that it was running really smooth CCW :-)

Makes me wonder if you could change the timing on the fly to get smooth in both directions. Of course all of this may be moot when I hook it into my velocity/position controller.

[Al_The_Man]

I would be interested in seeing the way you have it hooked up, are you doing any further de-coding? Did you find out the pattern for the Ha Hb signals?
Al.

[KTP]

I hooked it up so it starts up reading the 4 values to generate a position, then it switches over to using the encoder to generate the 3 hall values. This makes things very smooth and got rid of a slight vibration that was evidently coming from the fact that 16 is not evenly divisible by 6.

Pretty happy.