Mitsubishi SSCNET

I was hoping either you or drdyna would reply. I don't see any more direct way to contact him...? We have plenty of drives and motors to experiment with!

...and btw, speaking on the subj: just got the SSCNET III/H working. the last version: at 150Mbps, with full 22bit encoder resolution and so on...

dm17ry wrote:

...and btw, speaking on the subj: just got the SSCNET III/H working. the last version: at 150Mbps, with full 22bit encoder resolution and so on...

Great work !!!

SSCNET III/H still using 8B10B encode and scramble string similar SSCNET III?
Communication cycle which you are using 0.2222ms?

Thanks

hehe, thanks!
yep, same encoding. slightly modified like now it has only 1 start-of-frame symbol instead of 3 and the scrambler is seeded differently.
i'm running linuxcnc servo thread at 0.22, but currently with 0.44 comm cycle. to have the second half-period entirely free to accommodate linuxcnc's real-time jitter and PCI transfers.

implemented a support for MDS-D-SVJ3/SPJ3 servo drives. it turned out that they are using a slightly different sync symbols than SSCNET-III... while the next gen (J4 based MDS-DJ/EJ) can understand both. anyways, was able to reflash firmware/modify MR-J3-B drives into MDS-D-SVJ3/SPJ3. the SPJ3 was interesting for me as it is based on the same hardware as MR-J3-B and can drive induction motors. so with a bit of fiddling with the hardware i was able to get the firmware running, but not enough to get a generic induction motor spinning with a OBA18 encoder. that encoder is not officially supported, original induction motor spindles use TS5690 PLG detector with 2/4Mcpr. the OBA18 is recognized by the firmware, however the resolution is downgraded by the firmware from 18 bits (262144cpr) to 260000cpr using software floating pont arythmetic (sic!). the spindle drive worked fine with PM motors, but with an IM it failed:



i played with it for a while trying to tweak all the motor constants, but without any significant improvement. got tired and angry, disassembled the firmware and removed all the encoder resolution checks. set the resolution by servo parameters to proper 18bits (262144cpr), and viola - it worked! an induction motor behaves perfectly under position control


 

Very interesting for spindle orientation I would think. Does it actually actively hold a position at standstill?

yes it does! feels pretty much like a PM motor to me... with MDS-D/E drives (MELDAS M70/700 M80/800 controls) they totally ditched a speed control for the spindle from the protocol. previously (M60/M600) the spindle was speed controlled and they only switched it to position for sync tapping/C-axis. tool changer orientation was controlled by a separate signal bit. starting with 7th gen the spindle is always position controlled...

actually, my next step would be trying to make my own mitsubishi serial-interface encoder and try to run a small chineese 24rpm induction motor spindle with a MDS-D-SPJ3-075 drive. the encoder is based on a TLE5012B magnetic IC. not quite 20+ bits like the original ones, but i hope it will work. will report how it does

 

This is good to know-I have a pile of Mitsubishi servos and the STMBL project is NOT going well...

MR-J4-70B controllers are about the same cost as building more STMBLs these days on ebay, so if I never get these dang things going I have an option!

Thanks Dm17ry!

a bit of a progress here... made an ABZ incremental encoder to mitsubishi serial converter. tried that converted MDS-D-SPJ3 drive with an induction motor and a generic 12-bit incremental encoder instead of OBA18. it spins, but kinda rough and noisy. the drive seems to expect much finer resolution so doesn't take much effort processing the feedback. gonna try to put it thru something like a kalman filter, but no idea if it will help...

HI
I tried contacting dm17ry via the email listed on his website This email address is being protected from spambots. You need JavaScript enabled to view it. for some more info on his Meldas pci card, but the mail came back?
Not sure where he is based, hope he gets this message

Im trying to repair a Mitsubishi MDS-B-SVJ2 spindle drive. To test is, you need to establish the communication link to the CNC control. Meldas M500 in this case. the Machine is 700km from me, so testing on the machine is not really an option.

Any ideas would be welcome.

Thanks