Fanuc servo drivers with fiber optic

probably you've seen that. TODX283: elcodis.com/parts/6164179/todx283.html
availatble on ebay, has a TTL interface. should do the trick, imho

So, I made some next steps: I asked Toshiba support for TODX2910 datasheet and they answered it is a custom part and they have no informations about it. Also no info about wavelenght. But I found, it is visible red. I bought some Toslink cables and splitter and tried to connect. Toslink connectors are a bit different, but it is possible to grind them and connect with Fanuc stuff. I also made simple adapter to connect toslink cables to the fanuc connector. Machine still work after I insert my cables and adapters between controller and optical cable to servodrivers. Unfortunatelly, when I also insert passive splitter, machine doesn't work. When I compared light strenght before and after passive splitter, it is possible to see too high loss, visible by eyes. I also found some active toslink splitters, so my idea is to order one to continue.

hm. the cheapest toslink cable connectors fit fine into hitachi transceivers without any grinding. and i'm surprised about attenuation.. mitsubishi drives work even without properly seating the connector - it is enough to shine a light from the end of the fiber into the hole from a few millimeters away.
i doubt an active splitter will work - i think it is electronic and doesn't have enough bandwidth.
what kind of cables and splitter you've got? have a picture?

the wavelength. i've seen pictures of some random fanuc stuff using TODX283 and DC9325. both are 650nm and 50Mbps. unless they have multiple incompatible standards - good chance that TODX2910 is the same...

This Fanuc FSSB uses probably optical connectors JIS F07 (duplex) it was necessary to grind plastic part of connectors a bit, metal insert with the fibre match OK. I used these cables:
www.krup.cz/default.asp?cls=stoitem&stiid=927
Cables was OK, machine works with these cables.
I also bought this passive splitter and it doesn't work, it has really too high loss:
www.krup.cz/default.asp?cls=stoitem&stiid=5145
Now I want toorder this active splitter:
www.krup.cz/default.asp?cls=stoitem&stiid=5749
I hope there will be toslink optical receiver connected with two transmitters - and then should be possible to sniff the data here
When I will open machine next time, I will take some photos.
BTW: Fanuc PCI card at your picture looks interesting.What is it? Maybe, is it possible to use it?

probably your cables should fit if you cut off the molded PVC connector shell with the strain relief...

toslink receivers/transmitters are rated to 15Mbps - i doubt you will sniff anything with them.

fanuc pci card - no idea that it is, i just googled for "fanuc pcb" images

passive splitter which worked for me looks like this: www.nej-ceny.cz/647295/2portovy-digitaln...cerny-vlasp2502.html

Hallo, I have new informations. I ordered passive splitter like at your picture, but is it similar like I had before, Fanuc doesn't work with it. I think, optical power is divided half and half to each outputs and 50% is not enough for receiver, I think. I also ordered active splitter, but as we estimated, it doesn't work at all. There are some SMD ICs inside (amplifiers? Schmitt triggers?) but all signal ways are capacitor-coupled, so it will be frequency sensitive. I think it is the time to order 50Mb receiver and transmitter and make my own "sniffing amplifier"

huh! looks like fanucs make somewhat picky hardware...
FYI, mitsubishi drives re-clock the signal when passing it down the chain, it's not just direct receiver-transmitter connection...
btw, you probably can plug into the last amp output to at least have a look at the controller commands...

dm17ry wrote:

packets are XOR'd with a pseudo-random sequence and then transmitted using a slightly modified 8b/10b encoding.
yurtaev.com/yssc3p.html

Can you help to supply more info for this?
1. Packets was encoded 8B/10B before or after scrambling (XOR'd with a pseudo-random sequence) ?
2. What is scrambler polynomial? p(x)= x7 + x4 + 1?

1. scrambled then encoded with 8b/10b
2. i couldn't find a way to generate this sequence with a single lfsr. instead i'm using eight 12-bit galois lfsrs with f(x) = x^9 + 1, each gives one bit of the byte sequence

dm17ry wrote:

1. scrambled then encoded with 8b/10b
2. i couldn't find a way to generate this sequence with a single lfsr. instead i'm using eight 12-bit galois lfsrs with f(x) = x^9 + 1, each gives one bit of the byte sequence

Thank you for answer.
1. I did group 10bit from NRZ data on QD75MH (using logic analyzer) but have some group is "0111111100'' (LSB - MSB) and it is not 8B10B encoding, it is SOF and EOF?
2. Look like it very difficult to conver to original data byte from scrambled data byte, can you take an example ?
3. Can share original data sample from controller and driver?