SIL asked me to carve designs in 18x poster-sized linoleum sheets this weekend. Setting up for the first job, I thought it would be best to fix bottom and left fences for the obvious reasons. To initially align the bottom fence, I used a framing square indexed off the inside of the right rail. After locking the fence down and manually jogging a .25" EM across the edge to ensure true alignment with X-axis travel, I found the machined face “significantly” out of square with the rail. Greater than.125" over a foot which leads me to the …
First question: I realize my machine is probably out of square, but if it wasn’t, can I/should I rely on the powder coated rails to be true references for precision indexing and accurate stock alignment?
Coincidentally, the very next day, my vac hose gets caught under the right side drag chain while the gantry is moving to the front of the machine. The drag chain jams up the carriage and the right side loses steps while the left side keeps moving causing the gantry to get cocked between the rails. I hit the E-Stop, turned off the controller, and tried pulling both sides of the the gantry all the way to the front hard stops, but the right side keeps springing backwards .25-.5". It feels like I’m forcing the stepper motor to be between steps when I pull the gantry up against the stops. And now the …
Second question: Why doesn’t re-initializing re-square the gantry since there are limit sensors on both Y-rails? And what is the proper way to fix?
Sounds more like the rails are out of square & the torsion on the gantry is what’s making it spring back. If there is no power to the steppers, they are dead & loose. Besides, “between steps” is microns.
If the gantry is springing back at the front, it’s logical to think it would do the same at the back where the sensors are. Sounds to me like you need to re-square the rails to the base.
Relying on the rails to square things up is reasonable, I think, if the machine is properly squared.
Also depending on the tolerances you are looking for. “designs in linoleum sheets” doesn’t sound like high tolerance work, so I think you’re safe. I machine some industrial parts on my HDM with some tolerances less than 0.001", so I only rely on XY positions when I set up a part.
Not sure how the “2 sensors” on the Y axes works as it’s new to the 5. But I would also assume that both Y axes are homed independently, so dialing in the sensors would be the way to fine tune the squareness.
The rails should theoretically be a good reference edge, but ultimately the best gauge of the direction of a machine’s axis would be to have the machine cut a little groove or line and match it up against that.
The Shapeoko 5 does square itself of the sensors in the back of the Y rails, BUT in the event of a big crash you should also check that the X/Gantry extrusion is sitting snugly against the reference notches on the Y-carrier plates. Sometimes when the X axis gets twisted, the gantry gets shifted off the reference position.
The Y-limit switches are designed to be adjustable in position so that you can dial in the gantry to be square, we don’t have a software compensation or anything to offset one side more than the other. It’s just based purely on when each switch is triggered. Also, from a safety perspective, the gantry needs to be “close” to square before homing. If the switches don’t trigger within a second or two of each other, it stops. That’s to prevent a failed switch from causing the machine to try to rip the gantry apart.
It is best to get it as square on the diagonals as close as possible. They did update the YouTube assembly video to cover that.
Rather than trying to use the front, side, or some other machine physical plane as a reference, It is far better to cut a 90 degree angle with material bolted down to the machine. That way your X and Y references are true to the actual movement of the axis, rather than to some plane of the machine.
And thanks for confirming my thinking regarding accuracy from indexing off the rails (and other part of the machine). I understand they might be good references for basic woodwork, so from a best practice perspective, I’ll start working off a machined face or reference line on the spoil board aligned with axis travel.
But first I’ll pull the slats and bed extrusions off and square the machine back up measuring across the diagonals.
WRT the Y-axis limit switches, my rational brain says alignment/calibration is probably not required, but at the same time it makes my irrational/OCD side super uncomfortable. If I’m going to mess with it, what is the process? What do I use as a reference? I have some thoughts but wanted if y’all have better ideas before I throw mine out for all to see.
There are many ways. I start by machining something that I can accurately measure. My calipers are 8", so I machine a 6" square with 0.5" chamfers on the corners. (~7.7" corner to corner).
Measure the 2 corner to corner dimensions, subtract the difference & divide that by 1.414. That is how far off it is over 6". Calculate the comparative difference over the width of the machine at the points where the adjustment will be made. Make the adjustment, then retest.
You can use the same square to measure the width & length to dial in your steps/mm.
The big thing with calibration is that you want the measurement tools you are using to be 10 times more precise than the standard you are calibrating to, and the difference between good and better is not nearly so well-defined as that between bad and worse.
The best consideration is how the machine cuts, and a quick/easy and objective test of squareness is to cut two circles, then stack them, then rotate one.
If I’m going to mess with it, what is the process? What do I use as a reference? I have some thoughts but wanted if y’all have better ideas before I throw mine out for all to see.