|
|
Post by Derek the Admin on Jun 5, 2017 2:24:52 GMT
I'm working on converting an old early 1980s, 4000 pound plus CNC mill to modern control electronics. I thought about running a thread on it if anyone was interested. If you've got an interest in following the project let me know and I'll document some things here as I have time.
|
|
|
|
Post by aforww on Jun 6, 2017 11:35:36 GMT
Well of course I am!
|
|
|
|
Post by Derek the Admin on Jun 6, 2017 14:01:00 GMT
Alright. I've got a ham packed next few days but after that aill see if I can start getting some pics up with some info on the process.
|
|
|
|
Post by mp4doggie on Jun 10, 2017 19:32:31 GMT
Pics and info when you have time , I have seen this on youtube , where people are taking a harbour freight mill and converting it to a cnc milling machine . I look forward to the progress post
|
|
|
|
Post by Derek the Admin on Jul 7, 2017 4:52:11 GMT
Sorry for the lack of attention to this. It's been very busy around here. I'm a little short on the pics right now, but basically here's the jist of it:
Hurco KMB1. It had an old style console controller that originally ran a tape deck to load firmware. (Think Grbl of the 80s). The tape deck had long since been bypassed and supplanted with a serial communication board to allow the firmware to be loaded. The memory board batteries were dead and the firmware had to be loaded every time, which was a huge time waster. The controller itself did not run g code, but rather used what they call conversational programming. As more of a CAD/CAM guy, this was not useful for me.
So, out with the old and in with the new. The machine has AC servos and encoders for position feedback, a VFD that required 0 to 10V control signalling, as well as a bunch of relays that had to be controlled, this needed a pretty advanced controller. My options were Dynomotion, Linux CNC using Mesa cards, or a newer industrial panel. Each of those options is pretty expensive (as in more than an upgraded M3 just for the boards! Eeek), but I settled on a Dynomotion Kflop and Kananlog for a few different reasons. I gutted the bottom part of the control cabinet and started figuring out all the mapping to get all the signals where they needed to be. I kept the relay panel, servo drives, limit switch interface board, transformers, limit switches, and encoders. I ran into a problem where I wasn't getting signals out of an encoder that the previous owners had just replaced. That threw me foe a loop (pardon the pun) for a bit, but I found that adjusting some pots on the encoder improved signal strength. I also got held up a while by my encoders being single ended whereas the Kananlog was looking for differential signals. The port I first switched to when I connected directly to the KFLOP had termination resistors that were attenuating the encoder signals, but I finally got passed that.
Got her up then I had to tune the servos. They work off feedback from the encoders so I had to get a hold of runaway issues on a couple axes and oscillations mixed with runaway on another. (Runaway is when you command it to position and it not only overshoots but it continues uncontrollably in the commanded direction). The Dynomotion software has diagnostics to allow for the determination of PID loop parameters. Got the axes under control and mostly tuned up and started running work on it.
The machine is relatively slow because the spindle has limited RPM. It's powered by a 5 horsepower 3 phase motor that maxes at 3600 RPM that consequently is extremely torquey. The spindle motor itself is probably a 150 pounds by itself! The machine has iron dovetail ways, huge ball screws, etc. Because of the low rpm, you can't get surface speed with little end mills, but this thing was made to pummel big chunks of metal. The table itself has a 500 pound capacity, so it was intended to be used with some really serious tooling, not little 1/16 inch mills. I have other ways to handle fine work, but this thing is a real boss for doing heavy work. The MRR capability is pretty impressive.
I recently moved it to another place in the shop and haven't wired it back in, but when I do I'll see if I can get some good action vids.
|
|
|
|
Post by activereality on Jul 7, 2017 5:35:37 GMT
Derek,
Very nice. The older control system methods were very odd. However, if they ever got them to work, it would take a freight train to break them. Encoder feedback loops for motor control is awesome- once tuned. Another big turn off is if there is any kind of frequency harmonics with other electrical circuits- most call it "electrical noise". Depending on the encoders though, since these were using a pot to act as a signal amplifier??, they may have either some dip switch settings or even be wired differently than the originals (I've run into it when a company changes their parts, but leaves it the same part number and everything- even the same serial schemes). But it sounds like you are making some very steady progress.
I really like VFD controls. So many ways to both control speed and insert safety into the system. I made a 2x72 belt grinder a few years ago, using an older g7 yaskawa drive. While I did not use an electronic interface to enter either a speed % or frequency command, I did have several OL settings for ESD situations. And just used the keypad for the vfd to change freq for motor speed. I was using an old motor from a treadmill that really didn't have a whole of... "strength" to it. So I set up a belt drive from the motor to be able to switch pulleys at both the motor and belt ends. It worked out pretty good, for a home build haha.
Keep us posted and can't wait to see some pics!
Aaron
|
|
