[Bruce Griffing]

I have used and understand conventional quadrature encoders. Fanuc has a class of encoders they call "serial pulse coders". Just what is the architecture of a serial pulse coder?

[Al_The_Man]

There is not many Fanuc encoders that are serial, the DC motors used commonly 2k and 2.5k pulse/rev differential as do the AC together with high resolution 10k+ p/rev.
There is also a low signal sine wave type.
The difference in the AC motor encoders is the commutation pulse which is a 4 bit code to produce a pseudo sinusoidal signal for commutation.
On the other hand, Mitsubishi have primarily used serial style encoders where two way communication is over a proprietary two wire serial link up to 100kp/rev.
Al.

[Bruce Griffing]

Al -
Thanks for your reply. Two points. First why do I want to know this?? Well, when I look at the 0 Connection Manual (hardware) on p177 it shows a table of connections for "axis control boards". Based on this table, I have a 4 axis control board that uses "serial pulse coders". This is due to the M1xx or M2xx connector numbers shown in the table that correspond to my axis board.

Second point. I want to run a 4th axis on my machine. I now have a rotary table, Pac Sci servo motor, a US digital encoder to replace the resolver on the motor and a Fanuc servo amp. I was thinking of using a PIC or the like to convert the 10k quadrature encoder to "serial pulse coder" format - that is if I could figure out what "serial pulse coder" format is exactly.

[Bruce Griffing]

Al
OK - I found evidence to support your point about "serial pulse coders" being rare. I looked at the only axis motor that is exposed in my machine - Z. It is an alpha22-3000. By its part number it has an alpha-I64 pulse coder. From what I read, this is an incremental encoder. So my axis board must be able to do incremental as well as serial. Now I just need to determine the details of the incremental interface.

[Bruce Griffing]

The incremental interface appears to have a 1-2-4-8 counter for the low order bits and an A and B channel for a quadrature signal. If that is the case, the I and Q are just a 4 state machine driven by the carry from the low order counter. So can I simulate the Fanuc interface with a 4 bit up/down counter counting the i and q transitions (+ and -) on the encoder with a little logic? Pretty easily implemented in a PIC or otherwise if that is the case.

There is also the matter of the index mark, but that seems straightforward so I won't say more about it.

[Al_The_Man]

If this is a AC servo motor, what is the part number or the encoder part number, see the pdf file I posted for the 4 bit date you mention.
Al.

[Bruce Griffing]

It is an AC motor - A06B-0148-B177. That is the Z motor on my mill. I used it to determine that the axis board will handle a Fanuc style incremental encoder. I want to use this motor (A06B-0373-B169-R) on my 4th axis. My intent was to replace the existing encoder per the discussion above and replace it with an encoder that would work with my control.

[Al_The_Man]

B169 is taper shaft with brake and is a serial encoder, I do not have any info on the serial form, however.
Al.

[Bruce Griffing]

Al-
My intent was to replace the serial encoder with my "homebrew" Fanuc compatible incremental encoder. I just need to learn enough about the Fanuc encoder to emulate it.

[Al_The_Man]

I have been using Fanuc Motors for spindle motors on small lathes etc, Because I was using AMC drives, I had to replace the encoder with a standard BLDC commutation encoder using Renco products.
If your intention is to still interface to Fanuc, you could look at replacing the serial with a standard Fanuc differential type if the mounting is the same.
Apart from the standard encoder pulse, you need the 4 bit commutation pulses as per the PDF.
This is the only proprietary odd ball part of the differential encoder.
Using any other non-Fanuc encoder means aligning the commutation pulses to the stator poles.
I have a couple of Fanuc incremental encoders, but unfortunately they are odd ball low voltage sine wave signal output.
Al.

[Bruce Griffing]

Al-
Thanks for all the help. I now understand what the C1-8 channels are doing. If C1-8 did not need to be aligned to the motor poles, it would be easy to generate them on the encoder with a small microprocessor like a PIC. With a small backup battery to keep the PIC alive, it could be done using the index pulse - but the encoder would then need periodic battery replacement. Also, it would be necessary to understand what the standard alignment of C1-8 should be wrt to the poles and the index pulse, so the index pulse could be correctly aligned to the motor when installed and C1-8 would be generated with the correct phase. I wonder if there is a market for Fanuc compatible encoders...

[Al_The_Man]

Jim Elson has made a cpal interface for the Fanuc commutation to BLDC, with the encoder passed through and used as-is.
The Longer pulse is directly under each pole.
Al.

[Bruce Griffing]

How do I get information on his solution?

[Al_The_Man]

http://www.cnczone.com/forums/showthread.php?t=86487
Al.

[Bruce Griffing]

OK Jim's making the Fanuc encoders work with BLDC drives that normally use three Hall sensors to control commutation. I now clearly understand that Fanuc uses C1-8 for the same purpose - to control commutation. So I have to figure out how to make a normal ABZ encoder output signals equivalent to the Fanuc "pulse coder".

[Al_The_Man]

I would be tempted to pick up a cheap motor just for the encoder, this guy sells alot.
ebay 290404891051
This one appears to be 2500 incremental.
You just need to watch as some are absolute and sine wave hi res type.
Al.

[Bruce Griffing]

Al-
Thanks again for all the suggestions. The idea of picking up a motor just for the encoder would certainly be the most expedient. And I probably will do that. OTOH, I try to develop a solution and sell them on the internet.