[RRRoamer]

Folks,

I picked up 6 Sanyo Denki P5 servos for a VERY reasonable price, three 100W and three 1kw servos. They appear to be almost brand new, which matches the description of their previous life from the seller to a tee. The smaller servos have grease in a gear reducer attached to the output shaft and the grease looks absolutely brand new.

The only odd feature of these servos is the encoder. All six servos have the same encoder on them: Kawasaki M21KA6Z0100B encoders. They are fairly small diameters encoders (case is about 1 1/2" in diameter).

The very odd part of them is the wiring. They only have four wires coming out:
Red
Black
White
Blue
and the wiring is done with what ALMOST looks like CAT5 twisted pair cable with red/black and white/blue twisted tightly and then the pairs loosely twisted together.

I pulled the case off of one encoder and it very much is an optical encoder with very tightly packed SMD components. In the quick inspection I was able to give the encoder, I did see a MAX417 (dual opamp) near the cable, but I did not have time to see how this thing is wired.

Anyone have any experience with these encoders? From looking at them, it almost looks like they are either outputting a differential resolver signal or possibly a serial type of communications. Needless to say, any help in figuring these things out would be greatly appreciated!

[samco]

my guess would be - they are just single ended quadature encoders.

power and A and B channels

black and red is power
white and blue are A and B.

(Just a guess)

sam

[Al_The_Man]

Also if not 0 to 5v on the A & B then check for a low signal sine wave output.
This type simulate a resolver in function.
With opamp fitted, this could be the case.
The other possibility to check if low signal is being single ended they may be open collector, if so they would need a pull-up resistor.
Al.

[vpcnc]

I think they are similar to this type of encoder.
http://www.usdigital.com/products/en...y/shaft/sp-16/

[Al_The_Man]

Those appear to be simple low res. encoders. 16 counts/rev?
If TTL out then you should not need a pullup and it will be 0-5v.
Al.

[RRRoamer]

They are definitely NOT low resolution encoders. There is definitely a high resolution glass encoder wheel in there and a LOT of SMD electronics packed on both sides of the board. None of that would be required for a low resolution encoder.

I should be able to do some more tests this afternoon and see what the scope shows me. That dual channel opamp makes me lean towards "resolver" more than quadrature encoder. Hopefully, it will be easy to find out!

One other thing: Is it safe to assume that 5V is the correct voltage (once I verify that power and ground leads to know chips...)? Are there very many odd ball parts that require lower voltage or higher voltage? The date stamp on the servos is 1996, so I can't see it requiring 3.3V or anything like that, especially given they are industrial servos.

[Al_The_Man]

I have never seen anything lower that 5vdc even for the Heidenhain sine wave out, which is either 1v p/p or 11µa.
Al.

[RRRoamer]

I dug a bit deeper into it. The servo has an integrated brake, so I had to apply 12V to that before I could even (easily) turn the shaft. The encoder itself is a three track unit with some very nice looking optical components above and below the disk.

There are also three major ICs on the bottom of the PCB:
1) 2S210-102 637100 that appears to be from Nikon. No datasheet found, but it appears to have something to do with decoding the encoder disk signals.
2) 2S237-024 as above
3) A TI 6LB176 ic that appears to be an RS-485 transceiver.

I did hook up 5V to the encoder and give the shaft a spin, but I wasn't able to get anything out of the white or blue leads. I really need to hook another motor up to this thing to keep it spinning at a decent speed so I spend more time scoping and less time spinning.

[RRRoamer]

Stranger and stranger. I did a bit more probing and found a couple of other interesting bits:
1) There is a voltage regulator on board. The MAX417 is getting 2.8V (within spec of 2.5V min) at it's input from an input of 4V to 6V at the red wire on the connector.

2) The outputs from the MAX417 do NOT go to the blue and white output wires at the connector. While reading the datasheet for the MAX417, it lists one of it's applications as a photodiode pre-amp...

3) I appear to be getting nothing on the output. I have pretty much ZERO experience with RS-485 drivers, so I have no idea what to expect. I was able to verify that the TI 6LB176 RS-485 driver IS the output driver. It's two output pins (6,7) check out as the white and blue wires in the connector.

Next steps to test this thing?

[Al_The_Man]

Strange, as RS485 is usually associated with differential output, and there is not enough outputs to implement the usual 2 outputs per pulse, i.e. each having the complement of the other, in an encoder without a 1 rev marker this would be at least 4 outputs.
But a single 6LB176 has only one output channel (and one IN).
So if the two encoder outputs were from a single 6LB176, you essentially would only have one signal and its complement??
Al.

[RRRoamer]

Strange, as RS485 is usually associated with differential output, and there is not enough outputs to implement the usual 2 outputs per pulse, i.e. each having the complement of the other, in an encoder without a 1 rev marker this would be at least 4 outputs.
But a single 6LB176 has only one output channel (and one IN).
So if the two encoder outputs were from a single 6LB176, you essentially would only have one signal and its complement??
Al.

A single RS485 stream can carry all the information needed to fully decode that wheel, even at 4500 rpm. It would definitely have to be sent as data instead of raw bit levels. So yes, it only has the + and - differential signals for a single RS485 stream. That also explains all the other high pin count ICs on the back of that PCB. There is a lot of processing going on inside there! I just hope someone comes along that has a bit more information on what was going on inside this encoder. Otherwise, figuring this out is NOT going to be quick or fun.

I suspect I will be looking into RS485 protocols in the near future..

[mike_Kilroy]

perhaps u would do better looking at sanyo danki for the answer? I googled and quickily found:

http://www.sanyodenki.eu/-200V-P5-SERVO-MOTOR-.html

the motor manual pdf download there describes a 'reduce wiring incremental encoder" which sounds like what u have. check it out.

[Al_The_Man]

According to the link, the encoder still has 8 conductors?
Al.

[mike_Kilroy]

ya, I saw that. figured RRRoamer could do some more digging now that we pointed out a new potential source and find his answer; we all may have been sidetracked with the encoder mfgr part number and not finding any info on it, rather than looking at the motor mfgr who has specs on their motors and encoders more readily available....

[RRRoamer]

Mike,

Thanks for the slap up side the head! I did a LOT of google searches, but I never even THOUGHT of going straight to the source and seeing what they had!

[RRRoamer]

While digging around that web site, I found a TINY bit of information on the PA035 absolute encoder. The picture is VERY poor quality, but the picture of the top of that PCB looks EXACTLY like the PCB on my servos. Down to those two magnets glued to the top of the board at 90 degree angles to each other...

I wonder how much information Sanyo would be willing to part with about these encoders?

[RRRoamer]

I have been able to dig out more information on this thing. As per the back label, it is a Kawasaki M21 serial absolute position encoder. They are 8,192 count/rev (13 bits of data) with a shaft revolution count from -32,768 to + 32,768 (16 bit signed of data).

So far, that is about all I have been able to find on these encoders. I've ordered some TI SN65LC176 RS485 samples (same part as on the encoder) so I can build an RS485 board to interface it to a microcontroller and see what I can make of the raw bit stream. Hopefully, these don't require any (or much) initialization from the host before they start outputing data.

If anyone has any clues where I could find more information on these encoders or just the serial data format used, I would be VERY grateful!

[vger]

I found a bit of info on those withing a document about an interface that connects to them (and others) here
http://www.mitchell-electronics.com/...20(9-2006).pdf

In one of their other documents...
http://www.mitchell-electronics.com/...20(1-2006).pdf

There is reference to the signal names for the 4 pin M21 encoders that is a bit confusing, perhaps the signals with the * are not used. But they are labled "SD" and "REQ". Perhaps SD is Serial Data and the REQ is request. not sure.

Steve

[RRRoamer]

Steve,

So far, that's the only bit of information I have found also. I think I will try sending an email to both Mitchel-Electronics and Kawasaki Robotics to see if anyone is willing to share some data on these encoders. I'm sure at one point this was all confidential protocols, but given that they are coming up on 15 years old, hopefully someone will throw me a bone.

[PBr]

Could you take a close photo of this encoder?
Is there a little battery in the encoder? If yes, i'ts for sure an absolute encoder with some kind of comm. protocol, which can be used with a proper driver.