Mitsubishi SSCNET

yep, go for it. if you grab some data i can decode it and quickly say if it looks like sscnet or not.

hello , whats your opinion on mds, i can be done ?

yes, i think i will do it.

Burtski has set up a logic analyzer on his MELDAS M500 control and captured lot of logs. they revealed that the protocol used with MDS is neither SSCNET not SSCNET II. but functionally it appears to be very similar. but there're few things in the protocol to understand, e.g. MDS-A-SP spindle drive sync tapping control mode.

i have started working on firmware. still waiting for the MDS drive i bought to be delivered to begin testing. i expect a beta version of the MDS version of the controller in a couple of months.

imgur.com/a/WXTjL let me know if you need some ladder pics or schematics from a leadwell

a couple of months turned into 4, but i finally got it spinning. here's the story...

got the MDS-A-SVJ-03 drive, bought a compatible motor on ebay. but the motor couldn't make it thru a mail from singapore for two months. got a refund. found another possibly compatible motor locally - HA-MH43. replaced encoder connector and luckily, the drive didn't complain. but when tried to spin it - it jerked and vibrated. i blamed misaligned encoder or wrong resolution. made a gadget to plug between the encoder and the amplifier intercepting the serial encoder protocol in attempt to understand what's going on. spent a week debugging. tried to modify encoder resolution on the fly - instant runaway and overload alarm. tried to shift encoder phase - no good either. then it came to me - maybe the drive is malfunctioning? looked inside, checked electrolytes - all fine. next, two HCPL-7806 current feedback isolation amplifiers. no datasheet, not available locally. desoldered, the silkscreen underneath says: HCPL-7800. great! plucked 7800 out of broken boards laying around, soldered in. and, finally, the damn thing works!

so, MDS-A-SVJ drives are working in incremental mode with 3.55ms servo cycle and up to 7 drives on the bus. most things in the protocol are clear like status bits, alarms, position/velocity/torque feedback, torque limit, special op mode. for now the firmware works with the same sscnet-ii linuxcnc driver. it needs more work to properly implement parameters setting (now they are just saved in flash on the controller) and monitoring. desperately trying to find some time to rewrite the driver and unify sscnet ii, MDS and upcoming sscnet i and iii. and, possibly, have a look at mechatrolink as well...

next thing is the MDS-A-SP spindle. will be a bit tricky - it doesn't report encoder position in velocity control mode. meldas control switch it to position control mode for rigid tapping... will report progress

Hi, just joined this form because of this. I have a mighty VMC500 that i am getting ready to retrofit because i cant get it 100% functioning, it wont spindle orient, tool change, or rigid tap. Since i cant find anyone that knows anything about the controller and both Mitsubishi and mighty are useless and seem to know nothing about the machine or the electronics they make. The drives are MDS-A-SVT-10 and the spindle drive is a mds-a-sp 75 with a mds-a-cv-75 linked. Looks like the card you are developing is just what i need. As soon as you have it ready i will be buying one for my project. i can not wait to get this machine to be more than just a large paper weight.