How Good Are The Homing Switches

[Bob La Londe]

I have found mine (linear) to be within a thou. I was actually surprised they worked that well, but I still touch off. As one poster stated, if you are doing parts start-to-finish from raw stock using a setup from Fusion-360, a thou or two (or even 5) is close enough. Still, touching off just once after powerup takes so little time and gives some piece of mind. Especially if your setups change often.

All perfectly valid. Right now I often take stock from the saw to my big mill, square it to dimension (faster for heavy cuts - 1 pass rough cut, 1 pass finish cut). I never touch off for X/Y when I do that. It doesn't matter. I just home and set the offset for that vise. G54 is working. I have center of top jaw set to G55 for one vise and G56 for the other vise. Then I'll transfer it to my little light weight high speed machines for machining or finish it in place depending on the succeeding operations. If machined in place I can move to the reference point for the project, switch back to G54 and then set the work coordinates. I still only have to set Z height. Since its a knee mill with 6 inches of Z quill travel I leave home, safe Z, and tool change position set at 2 inches and use a 2 inch tool height setter 99.999% of the time. I just raise or lower the knee. On the little high speed machines I have to touch off and manually set Z height for every tool (ER16 or ER11 bed mills.) I'm thinking next time its time to have spindles rebuilt I'll switch over to ISO 20 spindles on those machines. Then I'll only have one touch off per job.

Speaking of ISO20 spindles. I wonder if it would be worth while to consider an ISO20 spindle for the Tormach's high speed spindle instead of the ER spindle they offer. It could even be ATCed with a slight change in the tool forks. Maybe the 15iss acceleration limit would just drive me bonkers if I tried to use it for my detailed, short move, small cutter 3D work. LOL. Seriously though.

Sorry, my brain is going a mile a minute this morning.

[mountaindew]

. Still, touching off just once after powerup takes so little time and gives some piece of mind. Especially if your setups change often.

I do small multi-sided, multi-offset parts all in one program. My first part is setup with stops and I touch off and set each offset and side as the part progresses. Then when part is complete the machine is setup and ready to basically rinse and repeat for the rest of that day. As for accuracy, If I do everything right my Tormach has an x-y accuracy I trust more then my modest quality measurement tools I use to Inspect the part. Z is a different story, If I chase any accuracy its on that axis and its easy to see why. I would like to try using an on machine tool setter to possibly improve this a little, Thinking some error is creeping in using a surface plate, tool holder, and height gauge system .

[Bob La Londe]

Z seems to be an issue I chase as well. In addition to backlash gravity, friction, and counter force all play a factor. On my little machines they have gas springs which remove backlash in the wrong direction in my opinion. The heads have a net negative weight. I added lead bars inside the heads on those to put a constant slight positive weight, and Z axis accuracy and repeatability improved. The additional mass did cause me to have to turn down acceleration slightly. Eventually I'll remove the gas springs (and lead bars) and replace them with air cylinders I can regulate to a suitable pressure for less than head weight amount of lift at peak pressure before relief and I'll be able to turn the acceleration back up. Basically allow the weight of the head to eliminate Z axis backlash. Of course if my air system blows a line somewhere I'll be in trouble. LOL. I already have air to all those machines for the spindle seals anyway though. The steppers on those machines have more than enough torque to deal with 20 or 30 lbs of vertical positive weight. I'll let them.

The servo, ballscrew, and quill elevator on the Hurco don't seem to have that problem, but its really only lifting the quill and elevator. Not a whole cast iron head like the little mills. The ballscrews on that machine are also an order of magnitude better than the ones on the little mills, and there is no counter weight or counter force. Gravity is always pulling it down. I changed the air operated oiler on the Hurco to an all electric one. I could run it without air. It does have an air brake, but if I wanted to leave the head open I could use a wrench. I may incorporate a a manual lock of some kind in the future "just in case." I have a smaller compressor I can hook up though if my compressor dies, and a new one can be had at a half dozen places locally the same day, so its really not a huge issue.

I can repair airlines pretty darn quickly, and I have a spare refrigeration type air dryer on the shelf still in the box. So the requirement for air on the ATC and the PDB for new Tormach isn't a huge issue for me. I like that it has an air safety interlock. I may have to install an air pressure alarm though just to be safe since my other machines all require air to some degree too.

Brain still going a mile a minute. LOL.

[TurboStep]

Typical ballscrews here on cnczone are c7/c5 on accuracy, meaning about 0.1 mm max positional error over 300-500 mm.

Just for the record, Tormach 1100 and 770 machines use p4 grade ball screws which, according to their PCNC Design Whitepaper , are specified as follows:

"P4 allows 16 um error within any 300 mm. In imperial units, this is 0.0006" per foot. Cumulative error at 900 mm is limited to 22 um (0.0008" at 35")."

My Tormach is WAY more accurate than any of the values you quoted for you machine.
Step

[soymilk]

Tormach uses a Tend TM1308-1 for X and Z. Tend TZ7311 for the Y axis. They are unshielded.

https://www.tormach.com/store/index....show&ref=31860
https://www.tormach.com/store/index....show&ref=30461

I have rehomed and tested it multiple times in the past. Generally speaking, its usually under .002" for me normally, but I have had them jump .020" or more due to swarf. Because of this, it is in my honest opinion to always double check at the beginning of the day. I generally try to make a feature in my fixture that I can easily indicate off of. Honestly it doesn't take longer than 2-3 mins usually. In my experience, every time I think " ah it should be good " is always when something bad happens. Ymmv


Most people don't use G52 or even know about it. I don't think people understood your reference to using G52 to "reset" the 4th axis. Please note that G52 is not available in linuxCNC, and I don't expect it to be in pathpilot. I'll try to make a macro to "rewind" the A axis when I get up in a bit. Seems linuxcnc doesn't have all the parameters fanuc macro b has.

[Bob La Londe]

Tormach uses a Tend TM1308-1 for X and Z. Tend TZ7311 for the Y axis. They are unshielded.

https://www.tormach.com/store/index....show&ref=31860
https://www.tormach.com/store/index....show&ref=30461

I have rehomed and tested it multiple times in the past. Generally speaking, its usually under .002" for me normally, but I have had them jump .020" or more due to swarf. Because of this, it is in my honest opinion to always double check at the beginning of the day. I generally try to make a feature in my fixture that I can easily indicate off of. Honestly it doesn't take longer than 2-3 mins usually. In my experience, every time I think " ah it should be good " is always when something bad happens. Ymmv


Most people don't use G52 or even know about it. I don't think people understood your reference to using G52 to "reset" the 4th axis. Please note that G52 is not available in linuxCNC, and I don't expect it to be in pathpilot. I'll try to make a macro to "rewind" the A axis when I get up in a bit. Seems linuxcnc doesn't have all the parameters fanuc macro b has.

Very good and informative post. Thanks. Yeah swarf sucks, but its better than no swarf. LOL.

[TurboStep]

As to the rotary: If its just a matter of "unwinding" it may be addressable in code. Just insert a line of code that changes the offset to match what you want it to be. Change it from 362 degrees to 2 degrees or as needed.

You might find what you're looking for here:
http://www.cnczone.com/forums/tormac...ml#post2080704
There are a couple of similar variations.
Step