Lagun 3-Axis Retrofit

[millnturn]

Hello,
This is my first post on this forum, but I have spent a lot of time reading and trying to absorb and understand the vast wealth of knowledge and information that gathers here. I have been building up a home fabrication and machine shop as time and finances allow (sounds familiar, huh?), and recently, brought home a new toy. The machine is a Lagun 3-axis knee mill with an Anilam Crusader II control. I found it on Craigslist, and the listing said the controls needed some work. This ended up being a best case scenario, as the motors (3hp spindle, SEM and Baldor servo) all function perfectly, and the three glass scales check out fine. Mechanically, the machine is in excellent condition, just the normal minor issues to address. The machine was at a pattern shop for seven years, and never used. Moving it home was a bit of a production, as it is a solid machine! It has an 11 x 58 table, NT40 spindle, ballscrews, and a totally rigid, fixed head. The thing pretty much dwarfs my series one Bridgeport, and clears the ceiling by 4 inches.

After some disassembly, major cleaning, and a paint job, I am starting to turn this thing into a new machine. My plan is to lose the original electronics, except for the servo motors, glass scales, and servo power supply. I bought 3: Viper 200 drives from Larken Automation, a PC-2-route combo board from Campbell Designs, 3hp VFD from Automation Direct, and a 3 Ghz. P4 PC to run Mach 3. I have done quite a bit of searching on using glass scales for feedback to a servo drive, but have not found very much. I have tested each scale with an oscilloscope, and they are an A,B quadrature encoder with .01mm resolution. My thought is that because the machine is tight, and has virtually zero backlash in the screws and timing belts, that using the scales might actually work. Has anyone else actually tried this with any success?

I attached a few pictures of the machine in various stages of the project. It has been a fun project so far (except for cleaning), but now I think I need to consult the knowledge base here to help sort out the controls. I am sure I will have more questions in the near future.

Thanks,
Jon

[beezerlm]

Hi,

Looks like we have the same project here. I am planning to keep the motor, powersupply, and glass scales also. Using cnc4pc BOB and same viper drives you are using.

I think we can use the glass scales for position feedback and just eliminate the tach in the motor.

I will be anxious to see how yours turns out. looks like your a little further along than I.

[beezerlm]

pics

[beezerlm]

pic

[beezerlm]

Is the VFD you have for the spindle motor? How do you plan on using with the Low/High gears on the machine?

-Brandon

[Al_The_Man]

If the scale output is 5v TTL then it should be no problem to use them. With the right signal, the controller does not know it is a linear scale or rotary.
If there is zero, or almost zero backlash, then you should not experience the oscillation that can occur when using scales & backlash is present.
Some systems will give you the choice of the quadrature multiplication, but if Larken are x4 , then your resolution will be .0000984" as I make it.
Al.

[beezerlm]

unfortunately, my system is not as tight as Jon's. I think I have .005" ish of backlash in my ballscrews. Is that a no-go for me? Or will the viper drives permit me to allow a large enough amount of error to be ok?

[millnturn]

Brandon,
The VFD is for the spindle motor. My initial plan is to configure four pulley ratios in Mach 3. First will be low gear, lowest ratio, second will be low gear, highest ratio, third will be high gear, lowest ratio, etc.. I think doing it that way should work, and should be reasonably repeatable as you will pick the pulley in Mach, then select the gear, start the spindle, and crank the speed to either min or max. I could be way off course on that however. I did pull the head up from the main casting to explore the possibility of installing a spindle speed sensor. I found an area where you could drill and tap a 1/2"-20 hole and install a micro proximity switch mounted in a threaded tube. The proximity sensor would pick up the 7 spindle splines. I have an Omron prox. sensor that is roughly .250 dia x 1" lg. It has a rated switching frequency of 12 khz, and with 7 splines at 4500 rpm, the frequency would only be 525 hz. I think I will install the sensor now, and figure out how to integrate it later. At some point, it would be great to get the spindle under closed loop speed control with the software.

[beezerlm]

Souds like you have done your homework Jon! I am planning on leaving my spindle the way it is for now. Can't wait to see yours work though.

[Al_The_Man]

Originally Posted by beezerlm unfortunately, my system is not as tight as Jon's. I think I have .005" ish of backlash in my ballscrews. Is that a no-go for me? Or will the viper drives permit me to allow a large enough amount of error to be ok?

You may have to try it and see, 5 thou is getting up there.
I have not used these drives, so I cannot tell you. The systems I use have dual feedback (scale & motor encoder) and in this case they split the PID tuning up between two feedback devices.
If it turns out to be a problem then you may have to go with a motor encoder.
Al.

[jeep534]

Guys, I have a lagun mill for retrofit as well. Pictures posted here. http://jeep534.smugmug.com/gallery/1...kyseP#48828196
as for using glass scales for feedback that is a particulary bad idea. the backlash in the screws will cause the servo's to "hunt" and destroy the ball screw. you need to use encoders mounted on the end of the servo motors. Please keep us posted on progress. I have a Bridgeport boss machine that is in line ahead of the Lagun. I am planning to use the Larkin drives on the bridgeport machine keeping the existing servo's and encoders. and hopefully the power supplies.

[millnturn]

I understand the potential problems that may arise from having feedback which is not close coupled to the motor shaft. The machine has near zero backlash in the ball screws and timing belts. The machine originally had an Anilam Crusader II control, and used glass scales to close the servo loop. After all, the machine was in service for 20+ years, using scales.In talking with Larry at Larken, it sounds like it should work fine with a little tuning. I figure I have nothing to lose by trying the scales, and can always use US Digital encoders as plan B.


Jon