Hardinge CHNC retrofit- Another one on the way!

I have the resolvers for x and z plugged in they seem to be working correctly. I had quite a bit of fun using my axis screen as a sort of etch-a-sketch with the x and y resolver . Part of the problem I was having was that the resolver input scale was set at -40000, this was over driving the axis dro's. I changed the scale a .2 and the dro's hold steady. I still need to measure the ac voltage out, didint have time tonight.

On another note the rest of my transistors came in today, I should be soldering them in tomorrow and testing the amp.... wish me luck.

The spindle lives again! I never thought I would be this excited to see a peice of metal spinning, . After reinstalling the servo drive boards and applying a velocity command the spindle fired right up, even turns both ways if I reverse the voltage polarity, sweeeet! The servo drives had quite a few blown transistors. I replaced all of the feild transistors using 2N3773's and the drivers are NTE61's, there were a few smaller transistors whos part number I dont remember right now, I will post them a little later. The only tricky thing about these motor drives is that when they blow, it evaporates a trace on the board, and there is a non appearant open circuit because of this. I had to solder a jumper to replace the blown trace. the reason I know is that the other newer of the 2 boards already had this mod and I had to apply it to the blown board as well. Now that thats done, I can move on other items like building the operator panel, interfacing the mpg and fianlly connecting the pc to the lathe. I hope to have time this weekend to work on it. Anyway here are a few photos





The test rig waiting to be connected to the lathe

Next question, how do I scale the velocity command signal from emc? I need a max output of +/- 8.5 V for the G.E. drives. However i can only seem to get +10V out on the vel cmd pins., even if I turn the spindle feed rate over ride to 1%. Also, when I hit the spindle reverse, I dont get a -10 v signal, it just goes to 0 and sits there?

If the analog out is from a 7I49 you need to make sure the PWM mode for all PWM channels driving the 7I49 are set to 2 or you will get strange analog out behavior.

Probably the easiest way to limit the output to 8.5V is with the PID comps 'maxoutput' parameter (assuming closed loop control).

I did have the Maxout parameter set to 8.5v but was still seeing a min of 10v, however that was with the step gens set to 1 on a 7i49 . Also on the 7i49 is it possible to recover the unused resolver pins as general IO pins? Thanks PCW

Sure, the outputs are simply not controllable in PWM mode 1
Best to use another connector for I/O

Finally hooked the controll up to the lathe and had some success. Spindle works great, however the speed inside axis jumps around alot and is not stable when at speed.

Also I cant seem to get the resolvers scale correct. My lead screw is .2 in per revolution and the resolver is geared up 5:1 so I should be getting 25 turns per inch. It seems that emc is not using the resolver scale parameter at all. When I change this parameter the axis still reads the same value for 1 turn of the ball screw no matter the setting .

Another problem is that the axis run away when I enable the resolver pid feed back. With these lines comented out I was able to get the axis's to stablize using the servo amp balance pots, however as soon as I enable the feeback loop, they run away. unfortunetly the first time it happened the x axis crashed before I could hit the estop. I had the drive belts loose but it still broke the coupling between the resolver and ball screw,

Its likely that feedback direction is backwards so you get a runaway
This should be _EXPECTED_ when first setting up a servo system.
You should also have working drive enable and reasonable tuning ferror setting
so a runaway is stopped in no more than an inch or so of motion.

The resolver scale should be .040 (for inches) that is xxxx.resolver.x.scale should be set to.040
Note that the resolver.x.position is scaled by this scale factor

Appearantly I forgot that I had jumped the amp enable pins when testing them with a battery. Because of this Emc couldnt disable the amps during run away, woops! I also found that I accidently reversed the z and x vel cmd output. My resolver scale problem was because the resolver was refering to the input scale not resolver scale in the ini file. After fixing these 2 items, setting the resolver scale correctly and fussing with PID and scale values, I have axis motion! Sweet. They still have a little over shoot on position comands, and will "following error" out every now and again, but I think they are close enought to begin tuning? Spindle speed readout is still bonkers, it jumps form 400 to 20,000 rpm even though spindle speed seems constant, didint have time tonight to chase that problem down.

Thanks agin for the help guys.

Next I think will be the turret index.