Brushless Servo Hall Sensor Output

[keebler303]

I am working on deciphering the pinout for an encoder on an AC Servo. I have found the quadrature signals. What type of signal are the hall sensor outputs for commutation? Square wave? Phase Relationship? Anything else?

Thanks
Matt

[Al_The_Man]

The three pulses are typically square wave, they used to be actuall hall effect devices, hence the name, but now are commonly tracks on the encoder disk.
Here is a Renco diagram, the S1, S2, S3 show the commutation and shows the pattern for 1 electrical revolution, How many electrical revolutions per mechanical revolution depends on the pole count of the motor.
http://users.tinyworld.co.uk/flecc/4...otor031102.swf
Al.

[keebler303]

Thanks Al

I have the standard incremental encoder signals figured out( differential pairs for A,B and I).
Then I have:
1 differential pair with 18 pulses per mech. rev.
1 wire with 9 pulses per mech. rev.
3 wires with 3 pulses per mech. rev.

Edit: 3 wires with 3 pulses are out of phase, can't tell exactly what is going on as I only have a 2 channel scope and no good way to rotate the motor at a constant speed. Using my fingers I can't really figure out the phase relationship as the pulse width is too jerky based on my finger movements.

1 wire with sine output.
1 wire with cosine output.

So I would assume that the 3 wires with 3 cycles per rev. are my hall sensors? Does that mean I have 3 electrical revolutions per 1 mechanical revolution?

What about the 9 and 18 pulse signals?

Thanks Again
Matt

[Al_The_Man]

3 outputs with three pulses/rev point to a 6 pole motor if these are the commutation pulses, what else you have going on there I am not sure, the sine/cosine resolver signals are usually for AC sinusoidal motor, these sometimes have hall effect to synchronize the resolver after one motor rev.
What make/model is the motor?
Al.

[keebler303]

Panasonic AC Servo
AMKB200B10LAK
100 V
200 W
.637 Nm

Can't find jack about it online. I bought several from a guy on ebay. He initially had 500 motors and has sold around 300 already. Seems like there would be more online about them unless people have bought them and are not using them because they don't know how to hook them up.

Same as discussed here: http://www.cnczone.com/forums/showth...ight=panasonic

I plan to use Granite Devices drives with them and I have found the incremental encoder portion of the output, which is all I really need. However, it would be nice to know which are the hall sensors, if thats what they are, so I can connect them as well. The VSD-E can then skip the initialization on powerup where it has to sync the encoder with the windings so it knows where its at. If you feed it hall sensor signals, this is not necessary. Not a big deal but nice if the signals are there.

The sin/cos outputs seem pretty high frequency, much higher than one cycle per rev. Maybe around 10-20 cycles per rev? Haven't paid them much attention.

Matt

[jdgoguen]

Matt,

I just bought several of these motors. Would you please share the pin out of the encoder connector as you've determined it? I have a four channel scope and should be able to work the whole thing out with a little of your info as a start.

Thanks!

Jon

[keebler303]

Jon

I'll share on the condition that you post anything further you discover. Such as which hall sensor is which, which wire is which phase for the motor, etc.

I started by popping off the cover of the encoder. I found a 26LS31 quad line driver inside. I googled the part to determine the power and ground pins. Then I used my multimeter to find which wires those were. I hooked up 5V and ground to those and then started poking around with the scope. Here is what I found:

5V = white/red str This is a white wire with a red stripe
GND = white/blue str

The quadrature encoder is 1000 ppr. 4000 counts per rev with 4x decoding.

Channel A = red/black str
Channel A\ = red This is the channal A differential signal
Channel B = yellow/blk str
Cahnnel B\ = orange
Index = blue/blk str
Index\ = yellow

18 Pulses per Rev = brown/blk str
18 Pulses per Rev\ = green

9 Pulses per Rev = blue

Hall Sensors (not sure which one is which)
All three are 3 Pulses per Rev.
1 = pink
2 = gray/blk str
3 = gray

Sin/Cos signals
orange/blk str
green/blk str

The final wire is the shield for the cable.

Matt

[Al_The_Man]

If you wish to use a double beam scope to confirm the pole count and which hall is which, connect a 10k resistor to each motor power phase with the other resistor ends connected together.
Use this star point as a common probe point for 1 channel. The Ch1 live probe would go on one of the motor phases.
Supply power for the encoder/halls , and with the 2nd channel probe, detect which hall output coincides with a rising AC output, you need to rotate the motor shaft slowly, either manually or by some other means.
For example, phase number one would have the hall effect go high simultaneously as the generated AC occurs.
Also the amount of times the AC rises/rev indicates the pole count.
Al.

[jdgoguen]

Matt,

I haven't had a chance to put the scope on the motor yet, but I do have some more information. I got a connection diagram showing the motor installation wiring from the Ebay guy. It doesn't show the connector pinout explicitly, but does give the color code and signal designations, and matches well with what you've found. I worked out the pinout to the connector (had a hard time resolving the colors until I put the damned thing under a stereoscope). The 18 pin connector has the even numbers on one side and the odd numbers on the other. Pin 17 is the empty one. Signal designations and color code are as follows:

Pin# Function Colors
1 GND white/blue
2 +5V white/red
3 A+ red/black
4 A- red
5 B yellow/black
6 B- orange
7 Z blue/black
8 Z- yellow
9 Abn. U gray/black
10 Abn. V gray
11 Abn. W purple (looks pink)
12 X blue
13 Abn. Y brown/black
14 Y green
15 A SIN green/black
16 B SIN orange/black
17 NC
18 Shield black
(motor case)

For the power connector:

Pin Function Color
1 U red
2 V white
3 W blue
4 G green
(Motor case)

This seems to solve the phase/commutation issue: pins 9, 10, and 11 provide commutation for phases U, V, and W respectively. (Anybody understand the Abn. designation for the commutation signals?)

The mystery signals you identified are the Abn. Y (18 ppr), Y (18 ppr) and X (9 ppr). Could these be used to provide quadrant information for the sine interpolation?

I probably won't have time to play with the motor before Sunday. (I'm shipping my daughter off for a semester abroad in Kenya on Saturday.) Please let me know if you have a chance to test a motor via this scheme.

Cheers,

Jon

[keebler303]

I hooked up the scope as Al described. The posted labels for U,V,W are correct for the hall sensors and motor phases when turning clockwise as viewed from the encoder end. Attached scope trace is motor phase U on ch.1 and hall sensor U on ch.2 This is the proper relationship for motor and hall sensor, correct?



The motor is 3 cycles per mechanical revolution. Ch.1 is motor phase and ch. 2 is the index pulse. That means a 3 pole motor?



I think everything important is now known. The other wires are not necessary for any drives I plan on using. It might be a while before I test the motor because I am waiting on the high voltage version of the VSD-E from Granite Devices.

Matt

[Al_The_Man]

Originally Posted by keebler303 The motor is 3 cycles per mechanical revolution. Ch.1 is motor phase and ch. 2 is the index pulse. That means a 3 pole motor? Matt

The number of electrical cycles per 1 Mechanical revolution x 2 = the pole count.
So it would indicate you have 6 pole motors.
Al.

[digital_life]

hi all, i have bought 1 unit of AMKA100b10LAK and one of AMKB200B10LAK, also from ebay. it was time ago. im about to need to use them, and found that there was no datasheet or information avaible. also im planing on use GraniteDevices drivers too.
is the information posted all i need to hook them up an use them?
iwill