New Omron servo; commutation

[Swede]

I recently picked up a neat little Omron AC servomotor of the highest quality. It has a standard quadrature encoder with compliments, and 4 lines for the power, 3 phases plus ground. What it does NOT have is any sort of commutation lines, such as hall. Is this typical of an AC motor? Once again, documentation seems scant. Thanks for everyone's help with my recent servo adventures.

[unterhaus]

I just bought a yaskawa that is the same. Don't know how I'm going to drive it. There are certainly drives out there that will do it, but I'm not ready to pay the money for them. Effectively, you need a sensorless drive.

We were playing with an impedance meter and a 3 phase induction motor yesterday. The impedance changes markedly with rotor position. I'm going to do the same with a PM brushless although I expect to see the same thing. What this means is that a drive can do an impedance check at startup, and then there is no need for commutation information. I figure that the impedance check could be performed at a high enough frequency that the motor wouldn't respond. A small bump would probably be required to see the relationship between all three windings.

[Al_The_Man]

This is an article that covers somewhat the sensorless servo control.
http://www.electricmotors.machinedes...px?artid=57684
Al

[ESjaavik]

Are you sure it is a positioning servo?
For a speed controlled motor the hall sensor is not needed. By monitoring the Back EMF, the drive can decide when the poles pass the commutation points. The problem is that this BEMF is small at low speed, and absent at standstill. So startup of these motors is not trivial. I guess it can keep account of the commutation points even at standstill after having been up to speed by using the encoder.

[unterhaus]

They probably were speed control motors. With the proper drive, they certainly are capable of fine positioning. Mine came with a 1:64 gearbox on it. I haven't really figured out why you would have an encoder but no halls, encoders are tons more expensive. On a PM motor that probably retailed for over $1k, halls are free. I guess the technique of using a current pulse at rest for positioning has been around for quite some time.

[steveg]

AC servos like the Omron and Yaskawa (I believe the Omrons maker) as well as Fanuc,Meldas(mitsubishi)
all use the keyed and aligned built-in encoder for the commutation.

these motors MUST be used with the drive that is built
for them, as the encoder connects to the drive for feedback

The same encoder provides all the needed feedback
position,phase angle(commutation) and velocity.

[unterhaus]

Originally Posted by steveg AC servos like the Omron and Yaskawa (I believe the Omrons maker) as well as Fanuc,Meldas(mitsubishi) all use the keyed and aligned built-in encoder for the commutation. these motors MUST be used with the drive that is built for them, as the encoder connects to the drive for feedback The same encoder provides all the needed feedback position,phase angle(commutation) and velocity.

Many Yaskawas come with an absolute encoder, but the motor I have came with an incremental encoder. If there are commutation tracks, there are dozens of drives that will work with it. There is nothing stopping someone from making a drive that will work with any brushless motor other than a lack of market. The motor I have can be driven by some drives available from AMC as just one example.

I also have come to believe that the Omron and Panasonic motors are made by Yaskawa. There seems to be many fewer companies actually making motors and drives than market them. I've often wondered if the Electrocraft DDM-xxx model drive was from Yaskawa, it looks almost identical to the Sigma-II drives. It seems that the Sigma II motors will work in incremental mode, but I've never really researched that. They go for too much money on Ebay for me to play that game. I do have some really nice little Fanuc motors that I haven't figured out how to drive yet. It isn't easy to come by the tech info for them. Me --> <-- Fanuc

[arvidb]

The Yaskawa servos, at least, have the commutation information in the third (index) encoder track. The advantages/reasons for this are no hall sensors needed and less conductors needed back to the drive.

The Z track on these motors is not a single pulse once per turn, but instead a square wave with varying duty cycle and phase (compared to the A track). After looking at this signal with an oscilloscope I think the duty cycle varies in fixed steps of 25, 50 and 75 percent, and the phase also in fixed steps (pi/2?). It should not be impossible to figure this out and build a converter to normal hall signals, if one really wanted to.

Arvid

[Al_The_Man]

Originally Posted by unterhaus I do have some really nice little Fanuc motors that I haven't figured out how to drive yet. It isn't easy to come by the tech info for them.

I have some Mitsubishi servo's I picked up a couple of years ago when Mitsubishi US was dumping their old stock of spindle motors,servo motors and VFD's for a song.
The servo's I have left over I believe use a similar feedback as Fanuc now use, some of these encoders have a resolution of 1,000,000p/rev. and are transmitted to the Drive in a serial fashion. There is only 4 signal lines for enc. and commutation.
I assume it is some proprietary method.
Al

[unterhaus]

Originally Posted by arvidb The Z track on these motors is not a single pulse once per turn, but instead a square wave with varying duty cycle and phase (compared to the A track). After looking at this signal with an oscilloscope I think the duty cycle varies in fixed steps of 25, 50 and 75 percent, and the phase also in fixed steps (pi/2?). It should not be impossible to figure this out and build a converter to normal hall signals, if one really wanted to. Arvid

That sounds a lot like the electrocraft F and H series motors. They have what's called an "abs" (absolute) track, but also have "halls" tracks. I've attached a diagram of the absolute track from the manual. I've always assumed that they meant that the signal was PWM, because it's got to be difficult to get an analog voltage from an encoder any other way.