My Squaring Adventure

[johnwillyn (JohnL)]

As requested by a few readers here, I am posting my process for achieving and testing “perfectly square” alignment. Posting will be in a couple of parts.

My Squaring Adventure - Part 1

Let me preface all of this by saying this is specific to the OF WW Pro, and not the Elite.

This first post is mostly background, but thought it might inform a few folks about

Up front, let me say that I am probably a bit more picky than most when it comes to precise machine setup. I have spent hours to days tweaking fences, blades and tables on woodworking equipment over the years (most of which I no longer own).

I kind of stumbled into this at first with my OF, but then it quickly became an obsession. I tried to create a waste board that had 20mm dog holes at 96mm centers along the X and Y axes, and offset 1 inch in from the origin of each. When I completed this cut on my waste board, I found the dog hole axes were slightly out of square (as measured by my framing square, which is pretty accurate).

This is when I realized that my OF WW Pro needed to be adjusted, and I started researching and reading the forums for advice and ideas on how to align the X and Y axis so they were “perfectly square” (which is a relative term for many folks).

The OF Forums have a number of postings on the topic of squaring your OF machine, and I believe most folks will point back to some of the older threads such as this one:

Is your router frame a parallelogram? Mine was!

I’m on day 2 of CNC setup and I just discovered something that I do not recall seeing on this forum so I thought I’d share. I mounted my machine per the instructions in the manual and YouTube video and it seemed to go well. Everything was jogging smoothly and no binding was evident. I expect most other people have the same experience (good job on design Onefinity!). I was in process of making my spoilboard (t-nuts) and got ahead of myself. I mounted the board to the table and surfaced it, tram…

From there, you can link to all sorts of threads on the topic of squaring, and I read most of them.

I tried to follow the suggestions and techniques in some of these posts, and since my goal was to have a set of dog holes aligned to the axes of my waste board, and I have a Festool MFT as one of my primary work surfaces in my shop, I decided that testing my dog hole carves against the Festool MFT would be a good “acid test” of square. The Festool MFT is very precisely machined, and I confirmed this by laying a Bora Centipede work surface with dog holes over it, and they aligned perfectly.

That test turned out to open a rabbit hole that played right into my obsession with precise machine setup. After all, I thought here is a machine that is essentially “digital” and supposed to have precision to at least 0.1mm, or even better on a linear basis, but getting to “perfectly square” seemed to be somewhat elusive.

Many of posts on this topic contain dimensions for the various components (Y rail feet, etc) of the OF machines, and while these seemed useful, many of the posts state that “your dimensions may vary”. What this told me was that the machining and assembly tolerances of the OF were not as precise as we might believe, and that small deviations would translate to larger errors by the time they translate to the axis of the spindle.

I did follow one recommendation that suggested to measure the diagonal of the machine where the upper tubes entered the Y axis (and B axis, or right side Y axis) mounting blocks. This seemed reasonable to me as this would at least eliminate the mounting feet and bolt hole variations that might be present. I set up a fairly simple platform and “measuring stick” apparatus (which I highlighted in another thread), and set my diagonals to be as close to equal as I could.

Squaring jig hack

I have been reading through all of the threads on squaring a 1F, and opted to use the “equal diagonals” technique. I came up with a variation of a “bar gauge” that might be of interest to others. In one of the threads, it was suggested that you measure the diagonals where the top Y rails enter the foot, which sounded reasonable to me (since it is closest to the X rail and eliminates a lot of variables of the mounting parts). I had a couple of pieces of angle aluminum around my shop, and I wou…

Unfortunately, this did not provide very good results (more on that later), and the resulting test cut of my dog holes was quite a bit out of square. I could feel the dog hole misalignment with my finger tips. Also, it required quite a bit of tweaking of my X and Y asis mounting blocks, and the X axis in particular were no longer aligned on the Y axis mounting blocks.

I decided there must be a better way, and decided to sleep on it.

[johnwillyn (JohnL)]

Here is Part 2, which outlines my process.

My Squaring Adventure - Part 2

After thinking on the approach to squaring a bit, I decided that one of the other suggestions on one of the threads was probably a more direct approach.

Namely, use the machine to make “marks” at the four corners of the cutting area, and measure the diagonals between these corners and make adjustments from there.

This made the most sense to me since this would be measuring directly at the spindle axis, and would remove any variation of the mounting blocks or rail tube ends.

There was some advice on how to make these marks, but I decided to use a 30d engraving bit in the router and make my marks as small as possible; roughly a 0.5mm or so.

I also decided to make a calibration overlay for my spoil board. I wanted to make something that would register against my waste board so I could place it consistently after and between adjustments.

The other thing I quickly determined as more than a “nice to have” is an accurate (metric) measuring device. I chose to purchase this one from Lee Valley:

https://www.leevalley.com/en-us/shop/tools/hand-tools/marking-and-measuring/rules/115380-veritas-1m-shop-rule?item=05N0426

I had a metric tape measure on hand, but I felt the precision I was seeking required laser etched markings and a more rigid flat body.

At this point, I decided it was a good idea to start from scratch, or at least establish a base line to start from, and go back to the machine assembly instructions and follow these directions precisely. I decided to follow the written directions rather than the online video (which suggests a different approach to jogging the axes). I paid particular attention to aligning the X-axis feet to the Y axes mounting blocks, which the instructions highlighted as critical.

I should point out (as will be evident from the pictures) that I have a QCW frame and mobile base under my machine. In addition, I have mounted a waste board above the QCW slats since I decided to use double-sided tape as my primary hold down, and not the T tracks that a built into the QCW.

(Your configuration is probably different, so I just suggest you follow the instructions from OF to set your initial alignment.)

I made my squaring template to fit this configuration, and mounted an alignment block under a piece of 1/2 inch MDF to register against the lower left corner of my waste board.

Since I anticipated that would be making multiple iterations and adjustments, I decided to put down some blue masking tape at the corners, which also provided a higher contrast mark.

I made the marks very carefully by jogging the machine to each corner, and then lowering the bit until I could see a “dot” in the tape. I learned (the hard way) to do this on the slowest jog setting. I also made sure the screen showed the proper coordinates at each corner (e.g. X = 0 and Y = 0 for lower left).

(If you make a larger mark and go too deep, you can fill it in with wood putty and put down a new piece of blue tape.)

At each corner, I embellished the mark with a 0.5mm mechanical pencil, which also provided a nice measure of the target mark size.

As I said, I have a OF Woodworker Pro, so my soft limit cutting area is 816mm x 816mm, which implies a (theoretical) diagonal of 1153.998mm, or 1154mm, which meant my 1m ruler was going to be a little short for my measurements. I played around with using my metric tape measure for the last 154mm, but it seemed too limit accuracy. I wound up using my digital calipers and measuring as accurately as possible.

I also used an additional straight edge to make sure I was measuring the true diagonal between the marks.

The caliper was nice because the pointed tip would “slot” into the mark (v hole) at the corner, and then I tried the best I could to align to the middle of the 1000mm tick on the ruler.

My initial measurements turned out to show an error of about 1mm or so; 1153.2 for the lower right to upper left and 1154.3 for the lower left to upper right corners. This was fairly good for a first pass, and might be acceptable to some, but my test cut in 3/4 MDF failed the MFT alignment test. If you do the geometry, the X axis is off by about .5mm at the far right end.

I measured the distance between the corners along the axes, and they were all 816mm on the “dot”, so my rails were forming a paralellogram, not a square.

In my mind, I treat the left Y axis as “fixed”, and want to make adjustments to the right Y axis (or B axis as some call it). Based on my measurements, I needed to shift the right Y axis “down” (or toward the front of the machine) just a bit to adjust. I loosened the mounting bolts on this rail, as well as those holding the X rail to the Y rails (all 8 of them), and “nudged” the B rail just a bit by lightly tapping the back foot with a rubber mallet.

Before tightening anything back, I ran the X rail to the front of the machine, and then to the back to make sure the axes were not binding (somehow).

I tightened everything down again and re-ran the measurements. It was closer, but still off by .5mm or so. I did another adjustment (with a smaller nudge), and the diagonal measurements came very close to each other. They came in around 1153.80 to 1153.85; the slight delta to the theoretical can be attributed to measurement error.

At this point, I ran another test cut of my dog hole axes, and it was a success (IMO)! I had to use my dog hole reamer tool to properly size the new dog holes, and I (finally) managed to get longer dogs to push down through the dog holes in my test piece through my MFT.

Ah, homing! That is a topic unto itself, but I found a workaround that seems to work for now. The OF Pro series uses stall homing, which is noted as being inconsistent, so I needed a way of making it consistent.

What I found is that if I re-home the Y axis multiple times from the “zero point” (already homed), it seems to yield a fairly consistent result. I determined this by using a set of feeler gauges and measuring the gap between the Y axis and the rail end, and doing the same on the B rail. In my case, the B rail offset is slightly tighter than the Y rail by about 0.04mm (as measured by the feeler gauges), but it lands there pretty consistently after several homing cycles (of the Y axis only).

I hope this helps others with their squaring adventures. I will (obviously) not be doing this each time I startup my machine. I expect it to be close enough for many, many cutting operations, but when I need the extra accuracy (to cut a grid of dog holes when I replace my spoil board), I will probably go to the trouble to at least check it to see if it stayed “perfectly square”.

[Aiph5u (Aiph5u (not affiliated with Onefinity))]

Hey John,

now that you are satisfied with rectangularity (machine being “squared”), you can proceed to checking and adjusting coplanarity (machine not being twisted) with the fishing lines method.

[AndyP (Badger)]

I am not the most fastidious of people, more bish-bash-bosh if there is such a definition. I have squared up my 1F a number of times as i move benches and applied upgrades. To date, touch wood, i have not experienced any issues with the things i have made; Aluminium electronics containers (200x200mm) and server units which snap fit using 2-3 mm thick Al, drawers/cabinets in bamboo (600x800x400) and dovetailed units in oak and ash.

This has made me wonder whether there is an acceptable tolerance documented somewhere that is greater than we generally feel we must achieve and may be dependent on usecase?

Maybe this should be another thread, sorry, but it is related.