Welcome to LinuxCNC.org

I received the manuals last night, it looks like the motor fields are hardwired for 4 poles already. I was able to find out the model number of the spindle drive, its a Yasakawa VS-626MT III.

I found the link to the factory manual here:

http://www.yaskawa.com/site/dmspindle.nsf/link2/TKUR-5EKSVU/$file/NTR0010.pdf

I would like to swap out the spindle drive, mostly because from the manual this thing seems like its just WAY to complicated (headache.) The only thing that has me thinking is that the MAX voltage on the motor says 200V where as if I used an automation direct VFD, the output voltage is the same as the input voltage, so if my installation is 240V it'll try and feed 240V to the motor, at least thats my understanding. Do you think this would be an issue? I guess the other option would be to put a transformer on each phase going into the VFD and step the voltage down to 200V. Honestly that makes me a little nervous because power transmission is not something I am too familliar with. What do you all think?

If I did keep the Yasakawa drive, it wants +-10V for the speed reference, I assume I could just plug the spindle reference input to one of the 6 +-10V output channels from a 7i77, same as with the servos?

Frank

If you use a standard analog servo interface like the 7I77, you have a +-10V analog signal available for the spindle

(the 7I76 step/dir interface does only have 0-10V)

For what it is worth.....Yaskawa gear is very durable,its fail rate is pretty low.The 626 is a very good drive,If you remove it,you will have more then just a drive to replace,as the drive and motor are an integrated system.

As of now,all that is required for Emc to run your spindle is a voltage and an enable and/or a removal of inhibit signal.If you replace the existing drive,you will need a drive and some sort of braking unit minimum.If you go cheap,say a Hitachi SJ700,you are looking at about $2500, 1800 for the drive,200 for the encoder card and about 500 for the braking unit and resistors.Yes it is a new drive,but I doubt you will see any difference in the performance of the spindle.A new drive will need to be programmed(about 100 parameters).

I have the same drive on my Matsurra's,but with a 10hp motor and the spindle is unstoppable!! I have old MX-1 controls on them,so rigid tapping isnt a option.

Admittingly,fixing yaskawa stuff isnt cheap,but it doesnt fail much.You said you have Siemens motors,these to are durable.You seem to have lucked out in not having Siemens drives,these do suck and Seimens is BRUTAL on repair costs.

I found this out the hard way,all electrolytic caps in old drives go bad,when they go,they take out other components.Replacing ALL the caps in all your drives is super cheap insurance.

Ill buy the 626 mkIII if you are really gonna boot it,let me know....

Now that I have the manuals, I've spent the last few days taking the machine apart, cleaning it and doing some research on the different motion control components. This is what I have been able to figure out:

On the servos and drives:

The servos are MSI 7Kw AC servos, however the encoders on the servos are analog units, they out put a reference pulse, a sin voltage and a cos voltage relative to the position. The motors are connected to the ballscrew through a 1:1 belt drive to the ball screw unit. There is a digital encoder attached directly to the end of the ballscrew, this encoder is fed directly to the Dynapath Control.

The drives are also MSI 7kw AC drives, the analog encoder output from the motors are fed directly to the MSI drive.

From what I have been able to find out, there are no parts or support for these drives and servos. I can't even find manuals to be able to know what kind of inputs they want.

For the X,Y and Z joints, I decided I am going to swap the drives and servo motors for some (newish) used Mitsubishi motors and drives of the same capacity using +-10V velocity mode. I may or may not continue to use the existing external digital encoders since, I would rather have my external position reference from linear scales attached directly to the table and spindle assembly (assuming I can fit everything inside the way covers.)

JR, I understand your point, however, I have a question (for anyone): What would be the difference between the current Yaskawa spindle drive or the Hitachi drive you mentioned and say an Automation Direct GS3 drive with Brake resistor run V/H, +-10V control so that the spindle encoder is fed directly to linuxcnc and linuxcnc can take care of the closed loop operation? Is this not a viable option? The one thing I can see is that the current spindle motor is rated for 200V max and the VFD's have seen (and also the Hitachi inverter drive you mentioned) all say the output voltage is the same as the input. I guess I would just run a transformer to the spindle drive to bring the input and output voltage of the spindle drive from 240V to 200V to not fry the spindle motor? To clarify, I would like to be able to keep the option of rigid tapping on this machine, though when possible I'd probably be thread-milling.

Frank

The encoders you speak of are resolvers.They are being fed back to the drive as a velocity signal.There is no real advantage to removing the encoders on the screw unless you are going to direct drive the screws.

I googled MSI and they seem to be a distubter,can you post some pics of the motors and drives.I suspect they are yaskawa.The drives will definetly be analog.a Dynapath20 isnt new enough to be digital.

My experience with the cheaper VFD drives is they are really weak in the low end.Im not familar with the automation direct drive.I used a Hitachi sj300 in my Cinci lathe and it works well,but it is a Vector flux drive.I know for a fact the 626 drive has low end grunt.You can run the encoder to an encoder input in Emc right out of the drive.You would have the feed back to the drive for exact spindle speed and to the control for position.You can run your spindle drive open loop.

Im presently playing the spindle drive game with my Fadal,the Mitsubishi drive died on saturday.The machine never had any balls,so I replaced the drive with an AMC 10 hp drive and a 10hp Fanuc 6s 6000rpm red cap motor.Its gonna snow aluminium !!!

Some things are definately worth replacing,old wiring,belts ,bearings old dc spindle motors....sorry for all the misspelled words,im beat!!

dangercraft wrote:I would suggest leaving all this system intact. You can't simply replace the encoders with incremental digital units as AC servos need an absolute encoder to allow for the motor commutation.
The existing encoders are actually Resolvers. Resolvers are extremely good devices, very accurate, very reliable, very tough, very expensive.
To use the existing motors with a different drive you would need to add Hall sensors to the motors, or swap the resolvers for some form of absolute encoder, or fit 6-channel "commutation encoders". (There are ways to run a brushless motor with an incremental encoder, but it's not the right thing to do with the axes on a high-spec machine).
You ought to be able to leave the motors and drives untouched, and control the drives with a +/- 10V signal from LinuxCNC-connected hardware, and use the incremental encoder on the screw for axis feedback to LinuxCNC. It is very unlikely that Ye Olde Dynapathe used anything other than +/-10V as the control signal.
I don't see much sense spending a lot of money to replace what are probably perfectly serviceable parts.