Hello. Can you correct me if I am wrong? If a Shapeoko 4 Standard has an ideal (perfect) calibration than the homing process moves the Gantry into its starting position which would be a distance equal on both sides from the Shapeoko 4 frame.
Suppose calibration has a problem where one belt is not equally tensioned compared with the other belt. Now the machine goes to its starting position. Is it now likely the Gantry is no longer on both sides an equal distance from the frame?
Suppose everything is ideal except there is a bent axle on one of the gantry stepper motor(s). How is the homing process impacted by a belt gear that is now wobbling along + or - a millimeter or so?
I am asking this because I lack enough experience and training to see the connection between the homing process and proper squareness of the gantry to the frame. I believe the homing process sets a starting point. But if the point is far enough out of alignment the gantry is out of square with the frame.
First, the machine origin is top right rear where the switches are, so everything is relative to that.
If the Y-axis belts aren’t evenly tensioned then what will happen is the side with the belt with less tension will not move as much as the other side which has more tension — but it will pull on the belt so as to increase the tension on the end the motor is moving towards, then it will have to overcome the mechanical structure of the gantry, warping that before it will get out of position — in practice, uneven Y-axis belt tension increases backlash, but doesn’t usually cause other problems.
A bent axle on a pulley will constantly increase/decrease belt tension by a tiny amount, but won’t otherwise be an issue — until it breaks, which usually happens in short order (the shafts are quite hard and tend to break rather than bend).
Hello. I’ve been studying the article about belt tensioning.
I need to see the connection the homing procedure has with effects of loose belts, a bent stepper shaft, or unsquare base frame.
(paraphrasing) You said the right rear limit switch sets the point the gantry begins. If a belt is loose, as the Gantry moves along the Y rail(s) the loose belt side will not move as far as the tighter belt side.
Now, assuming the belts are ok and instead the problem is a bent stepper shaft on the gantry. Starting from the home position, as the gantry travels, the belt gear wobbles and belt tension is periodically tightened and loosened again. So then the gantry, on the side with the stepper with bent axle, wobbles back and forth as the belt is loosened then tightened - the bent axle has a belt gear wobbling up and down +/- a single millimeter.
Now assume everything is okay except the base frame is unsquare. The homing sets the gantry at a position relative to the right rear switch you mentioned - and now the gantry is out of square with the frame? After all the frame is unsquare. And homing, for the gantry, just responds to the location of the Y limit switch.
Could you provide insight into what you found and tips that would help anyone else? I just got curious and realized my gantry doesn’t hit the rear brackets at the same time anymore, but it did when I set up.
Hello. I found out that tensioning the belts seems like an art. I loosened one of my belts. Now both seem looser than I intuitively guess a belt should be. I used a caliper and a post-it note pad to gauge distance gantry was on both sides and finally they were almost the same by about maybe within 1 mm.
My post it note pad was used as an adjustable feeler gauge. I add or remove sheets as needed to adjust thickness. Then I attempt to wedge the paper between the gap for the gantry and rail end plate. If the fit is a nice tight fit…I then use the calipers to measure thickness off the pad. Done for both sides and compared to arrive at a difference in width of gap on either side.
The belts are very tricky. I need to replace my Gantry stepper because it is no longer with in tolerances of the Shapeoko design. It wobbles and I can see that the belt loosens as it wobbles and I say no to this. What if the belt backlashes upon loosening?
Anyone care to share how they measure the gantry distance from Y-rail end plates?
I’m getting my baggage scale from Amazon tomorrow and will be able to measure my belt tension or lack of it.
Does anyone here really just pluck the belt with a feel for it’s tension? How did you develop this feeling for belt adjustment? By building desktop CNCs?
There is some need to come forward with fact that the Shapeoko 4 really isn’t for beginners unless they realize what tools they may need. That happens when you read articles about belt tension - those prove that you need more than just a feel for it which is a trait of the person lacking experience building belt driven desktop CNCs.
If i suspect my gantry isn’t square, I pull it all the way forward until it touches the two front plates. As for adjusting the belt tension, I do it as described in the link above, using a guitar tuning app on my phone.
It’s not rocket science, and these machines were intended for woodworking. Being able to cut other materials with it accurately is a bonus, takes patience and learning by trial and error.
You are cutting wood and I need more accuracy than this for routing printed circuit boards - e.g. my preview in carbide motion of a simple circuit pathway comes out several millimeters, far enough to leave no trace of copper for a pathway.
Most people must be cutting wood.
I can say that you can’t get the Gantry to touch the end plates unless it is a step taken before attaching belts. Once the belts are on the gantry separates from the y rails some. In my case one gap from gantry to end plate measures 3/4 mm the other is about 1.5 mm. The belts are somewhat loose but of the closest to squareness tension I have been able to get.
On the topic of gantry to end plate gaps, I would not consider this to be a particularly reliable measure if you’re looking for real squareness on the machine. There’s some movement possible in how the X axis attaches to the end plates and the frame of the machine is not necessarily as square as your target calibration.
I fix down some waste MDF in the four cutting corners of the machine, bore an 8mm hole in each corner, insert dowel pins and measure the distances between them. I’ve found a rod slightly shorter than the pin distance plus a caliper is a good way to get accurate comparative measurements.
The four corner measures allow me to test whether left and right Y are matched and then to test the diagonals to see if the machine is square. Some years ago I put set-screws at the rear of my machine so that I could push the gantry back on power-up to start up square, this allows me to be square within one full step of the stepper, which is about the lower limit without dual Y homing.
Thanks for the pictures Steve and info about squareness Liam.
When I have the machine off and I manually pull my gantry back, if the gantry touches the y rail end plates it “flexes” off of the plates to rest at 3/4 mm and 1.5 mm as stated before.
Its spring pressure from the belts that causes this flexing. Will try loosening them some more. Tonight I should get my baggage scale and can tighten with enough precision so it isn’t so tight it pulls my gantry out of square…but not too loose or unevenly loose per each side.
About checking squareness, I used a 72" strait edge. I picked up my Shapeoko 4 and placed it on a table in the middle of the floor of my shop. Back when I did this I noticed my machine was out of square by over an inch. I eventually made adjustments following numerous times removing the floor of the machine and loosening y rail shoulder bolts. To determine my frame to be square, I measured diagonally. I was careful using the marks on the straight edge. Carefully made sure I was computing/reasoning correctly.
As for using a well planted piece of wood and drilling holes and measuring, I have tried a similar approach but for a smaller set of calipers and my “cut file” was a carbide create equilateral triangle. My measurements were always worse for the side of my triangle affected by y axis. My problem can’t be measured with the CNC router and it’s drilling of holes into work pieces because of my belts.
You might want to try dropping the Y belts completely and then checking to see if the gantry rides ‘square’ in the machine. If you’ve determined the Y axes are sufficiently close to square, then you’re looking for the X to be normal to those. I had to shim the ends of my X rail to get the plates square enough to not tend to pull the gantry out of alignment.
That being said, once the machine is powered up, there’s more than enough holding force in the steppers to pull one end of of the X axis a few mm back into square, more than enough to correct your < 1mm offset.
Whatever you do, you may well find that pushing the machine to the back (or front as others do) in Y against the end plates or against some set-screws to give adjustment is the more reliable way of starting up square.
Of course, starting up square doesn’t help unless the Y belt tensions match and both belts are in decent condition.
I found that measuring the belt tension by tone was the most consistent and effective method, the big upside is that on the two Y axes you can put both belts on blocks and just listen to the tone difference between them to get them even.
I see above you mention a bent stepper shaft, as you say, nothing’s going to work properly until that’s fixed, it makes the rotation eccentric and therefore the distance per step vary.
As I’m stingy, I use four small offcuts of wood in the corners of the machine, simply to maximise the distance over which I’m measuring square to increase the signal to measurement noise ratio a little.
I never gave the X rail much much thought because how can it be out of square with the gantry? It is mounted on the gantry by v-wheels. And so you say it can become crooked somehow and you can measure this?
My baggage scale came in the mail and my belts are fine now. The only problem causing issues now is my crooked stepper shaft which should be fixed soon.
I found the bicycle belt app to be cumbersome. My scale worked smoothly, no questions asked. I have 5.8 pounds of tension on one side and 5.3 pounds in the other. Gantry is tight up against the end plates on each side now but only on the front. Not quite so on the back where the side of the machine with damaged stepper motor sits one mm off of the end plate, the other side tightly rests directly on the end plate.
I tried drilling holes using a 250 mm x 250 mm square as a cut file generated from Carbide Create. The right side of the square corresponding to the side of my machine with damaged stepper is 242 mm. The other side(s) are almost precisely 250 mm but I don’t have with me what I wrote down for a measurement.
William Adams; Do you use a scale or a the Gates belt frequency smart phone app? From your videos I never hear you mention the tools mentioned in the tutorials you have on auto fill. Did you get a feel for many things related to calibration because you’ve built so many Carbide 3d products?
I used a guitar tuner for the Z-axis for a while on my SO3 — since then, it’s just a matter of feel based on experience (we are not going to tally how many time I tore down and rebuilt my SO1 upgrading it, nor my SO2 writing the instructions).
