AS a rule you should keep a backup copy of the XML stored away somewhere safe (off of the computer) just in case you need it

Actually I have the entire drive 'ghosted' to a second drive in the machine, so I can get the XML right from the Mach3 dir on the backup drive. Thanks for all the good info!

PS:  YEP AC servos RULE (;-)

Not sure if I mentioned it, but ultimately the Granite Drive's 'Position Reached' signal turned out to be non functional. Something that they advertise, but does not actually work. However 'It might be a good idea for the next firmware version'. This MO unfortunately has proven to be typical for most if not all of the 'hobby level' drives as well as similarly positioned products like smoothstepper, etc.  Anyway, as I did with the 'swap axis' function, I implemented the needed feature in hardware. This time with a little 'sniffer' that snoops on the pulse stream and locks the 4th axis when the 'A' axis pulses stop for more than a (programmable) certain amount of time and then unlock it at the next pulse. This 'auto lock' can be enabled or disabled via 5V TTL from Mach and the lock can still be engaged via script or a manual button (for changing collets, etc). That's all wokring fine.

Now I have just installed the Mitsubishi J2S drive and it has a functional 'position reached' that even has a selectable proximity to trigger the signal. Works great running an led (so far)! Not used it with the lock yet, but will try that soon. It also has an externally variable torque limit so I could drop back on the torque during the lock if needed. Should be fun to dink around with it. The nice thing is that (so far) everything this drive is supposed to to, it does . .  a refreshing change.

Yes AC servos are the mutts nuts, auto tuning is good but I find I can still get things sweeter with a further manual tune. Not sure of your drives software but the scope feature in mine makes tuning a lot easier, can monitor up to 4 things at a time although I usually only monitor the three.

With your experience, tweaking to perfection is an option. For me at this point, well . . .   not so sure 

Actually in my application, I could have a heavy chuck with a heavy workpiece one minute (and I am currently making a trunnion table as well) and then a piece of 1/2" 2011 in a 5C collet spinning 3,000 RPM the next, so tweaking, even if I was competent to do it (in anything approaching a reasonable amount of time), would be quite a task indeed. I have so far tried two AC servo drive series. The first auto tuned over a 'period of time'. Now I am working with the Mitsubishi J2S series and it tunes within maybe 6 moves. Pretty friggin' amazing really.

With your experience, tweaking to perfection is an option. For me at this point, well . . .   not so sure 

My experience with the tuning is just my usual, alter a number and see what it does The real time scope in the software really helps with that.

Actually in my application, I could have a heavy chuck with a heavy workpiece one minute (and I am currently making a trunnion table as well) and then a piece of 1/2" 2011 in a 5C collet spinning 3,000 RPM the next, so tweaking, even if I was competent to do it (in anything approaching a reasonable amount of time), would be quite a task indeed. I have so far tried two AC servo drive series. The first auto tuned over a 'period of time'. Now I am working with the Mitsubishi J2S series and it tunes within maybe 6 moves. Pretty friggin' amazing really.

Auto tuning for me on the axis wasnt too far off the mark but for the spindle on the lathe it was a different story. All seemed well until I tried to do some hauler plates, they were a 19" Dia but where the normal ones are cast with hollow backs these ones were made of solid steel and were about 40mm thick, so quite a weight. Due to the slow speed I was facing at I was seeing the spindle overshoot and fall back and overshoot..........   messed about trying to get them done but took about 2hrs on the first one so decided to give the tuning a mess with. I got the tuning done and it was really sweet but was fully expecting having to detune again so I had exported the previous drive profile, ready to import it again,  but to my amazement that setting has been good and its still used for all work.

As for the Auto Tuning, my drives just turn a  user set amount of encoder counts one way then back and thats it, its amazing how good it can get it with such a short move.

Hood

FYI, the solution to the RPM readout is going to be a new computer. The problem developed rather suddenly and seems to be only with that particular computer. I re-loaded Mach from scratch to an empty folder and re configured from scratch . .  same thing.

Putting MACH on a couple of other computers did not duplicate the problem, so it seems to be a goblin that took up residence in the CNC computer. I've been around computers a long time and never seen anything quite like this. It could be a variety of things, but I'm not inclined to dig into it. New computer is the quick fix.

My experience with the tuning is just my usual, alter a number and see what it does The real time scope in the software really helps with that.

Mistubishi has tuning software also with all manner of cool stuff. I have not played with it yet as I need to make up a special RS232 cable for which the special connectors for the drive just arrived yesterday. I followed your lead and converted to differential at the BOB. You know . . any excuse to make another little PC board. I'm still fascinated by that whole process . . . .  

Yes, great thing about these industrial drives is the amount of I/O and the ability to set that I/O to do what you want. The software is also great as everything is configurable and easily so.

I helped a couple of guys recently to set up their machines with Allen Bradley drives and mentioned that personally I would set them up with differential Step/Dir. Both guys decided to try single ended and all seemed to be fine but after a while there were issues, making up a small line driver solved their issues so its definitely the way to go I am a firm believer that if Industrial type hardware has a feature like differential inputs  it is there for a reason.
Hood

Final update. New computer went in today and all is good . . . .  better than good, actually.

I wanted to re-use as much of my existing stuff as possible so I did a budget build using an ASUS motherboard P5G41-M LX2 with an Intel E6500 dual core processor. I re-used the memory, PCI parallel port card, hard drives, case, etc. from the old computer. The new computer is getting along fine so far with Mach3 and is running stable at the full 100k pulse speed.

Typically I use Intel motherboards for business and Gigabyte for everything else, but I have always heard good things about ASUS. This is my first ASUS and it's a quality board and everything went together perfectly.

is running stable at the full 100k pulse speed

excellent
This is promising news.

Amir

I got the tuning done and it was really sweet but was fully expecting having to detune again so I had exported the previous drive profile, ready to import it again,  but to my amazement that setting has been good and its still used for all work.

What has worked well for me is to tune the drives to be as stupidly sharp and tight as possible and then control the resulting 'bang' (with a light load) using Mach's acceleration settings. I find it easier to tweak with the acceleration than to retune the drives. And changing the acceleration in Mach has almost no effect on overshoot, so it is 'safer' to play with.

Incidentally,  are you running the Allen-Bradley drives using pulse step and direction out of Mach3?

Incidentally,  are you running the Allen-Bradley drives using pulse step and direction out of Mach3?

Yes, differential Step/Dir

Hood