It was finally time for me to work on fine-tuning the tramming on my OF, and I tried using the published guide for adjusting front-to-back adjustment by tightening/loosening the adjustment bolts on the X axis rails - I bungled the job apparently. At the moment, I still have useable but poor front-to-back tilt and most distressing for me, a different amount tilt when at Y=0 inches than when at Y=32 inches, and some of the adjustment bolts are already as deeply seated as they will go.
Can anyone offer advice on how I can restore all eight adjustment bolts to a “neutral” position before I retry the front-to-back adjustment again?
Thanks in advance!
That almost sounds like either your rails have a bit of twist or your table does. My table had a twist that would bite me depending where the weight of the axis was positioned… When I was fully forward or back everything layed flat but when the y was centered there was a done3formed in my table top. While flattening this would show as a tramming issue as I approached and left center of my table.
If I completely loosen all eight bolts and then by eye/revolution count run each opposing set of two the same amount inward until all are gently snug, does that get me to a good place from which to try again?
Front to back tramming: if I understand you correctly I achieved this by shimming with varying folds of tin foil behind the zslider.
I needed a .03" shim before fine tuning with the tramming bolts. The tramming bolts will not correct the tilt if the error is too large to begin with.
Hey Andy, hey John @jnordyke, hey BJ, hey SkyKam, hey all,
I think John rather means the Guide Tramming - Front To Back in the FAQ category, is this right John?
I haven’t used these tramming bolts yet, but that sounds logical. It think (hope) it can be assumed that the machined holes holding the hollow shafts in the X feet are of good precision and that the screws are of identical length.
How do you measure? Do you have a tool to check perpendicularity between router axle and table surface like this, this, this, this or this (no guarantee for ascending price order) 
?
Or do you check your accuracy by looking at the patterns your surfacing bit makes, like here?
That sounds logical. The weight of the router/milling motor and of the Z Assembly, in conjunction with tolerance in the linear bushings and bending of the linear shafts both in X and then in Y axes might add to an error even if tramming bolts hold the X shafts in a centered position in their holes.
Did you check this: Support: Removing Twist From The Onefinity Rails?
Also did you do something to ensure flatness and coplanarity of your machine’s base (=table or QCW frame) before both the tramming step and the surfacing step?
Did you surface already?
1 Like
To answer Aiph5u’s questions -
Remember, I have had “satisfactory for a beginner” precision for quite some time, I am now trying to do more granular adjustments…
Yes, I was following the Tramming Front To Back from FAQ. I used both observation of the surfacing bit patterns and a simple homemade “swinging arm” of 15 inch length attached to router collet to gauge tramming. The Support: Removing Twist From The Onefinity Rails procedure I did use 9 months ago when I first installed my machine and using the metric of how well feet sit flat on table surface there is now no twist in the rails. My table surface is a torsion box constructed of 3/4 inch MDF with 4 inchx1/2inch MDF webbing. Flatness and coplanarity were judged with a 48 inch bubble level and a 36 inch steel machinist’s rule. I have already surfaced/resurfaced/replaced spoilboards 3 times in the last 9 months.
1 Like
Hey John,
this remains to be found out why. How much is it?
And how shows it up? When measuring towards the front?
The swinging arm of 15 inch length, doesn’t it overhang the surface towards back at the end of Y? And towards the front at the beginning of Y?
And where does stand X then when you measure a deviation of tilt on different Y positions? Did you measure different deviations when X is at left, at middle and at right?
For this of course there is a table with coplanarity necessary (no twist between the positions where the feet stood). I hope there was no error here.
That sounds reasonable to prevent twisting. Important is that a twist is not already in the box when assembling it. The torsion box’s rectangularity lives from avoiding minimal displacements of the parts which cannot be assured if the surface part of torsion box is lying on a surface that already has twist, i.e. deficient coplanarity, when assembling it. Parts may have true right angles, but twisted surface would shift them in a minimal way if torsion box is forced into a twisted shape obtruded by twisted underground.
With a bubble level, you can check coplanarity, which means the absence of twist in the surface, and 48 inch is very long, but bubble levels are not that precise. That’s why I didn’t mention them in my list of tools to check coplanarity of the surface (fishing line method which is simple and very accurate, winding sticks, water tube level and level (optical instrument)). A lack of coplanarity, which means a twist in the mounting surface, i.e. if one of the machine’s feet is higher than the others, cannot be removed by surfacing the surface. The spindle axle will deviate from a vertical position on a machine with a twisted surface and with a long enough measuring arm, error might become apparent because the existing surface deviates from the imaginary plane perpendicular to the axle. This would be one of possible reasons for why you measure different tilt on front and back of Y travel. Do you think you could use the fishing line method to ensure coplanarity again?
1 Like
Hey Aiph5u,
First a goof from not good proofreading when I posted - the differences in nod are at the extremes of the x axis, not the y…so sorry! And, those differences (which I detect visually from surfacing) are not unusual - they are mentioned in the posted procedure that one must check at both extremes as well as the middle. The swinging arm point of rotation is placed at x=16.025 y=16.025 so that the pin on the other end is still an inch inside the boundary of full surface. I used the top of my very large cabinet saw as reference plane when constructing my torsion box, so I am confident in that. Really, with the explanation that I confused x and y axis in my original post, I think we can agree I was just careless in performing the front-to-back adjustments and will re-execute the adjustment after bringing bolts back to a more neutral beginning position. Thank you so much for the investment of your time in responding.
1 Like
Hey John,
I agree, a saw table usually is a good reference
If you are sure that 1. your both Y assemblies are with no twist in their feet and 2. they sit on a highly coplanar surface, i.e. there is no feet higher than the others (if not, check with the fishing line method or with winding sticks), then what remains would be to put these X hollow shafts parallel and one over the other by the tramming bolts.
If you have a different tilt of the router axle when X is at left than when it is on the right of X travel, you could attach one vertical ruler that touches both hollow shafts on each end of the X axis. Each ruler would prolongate what you want to see in the direction of ceiling. Then you would sight from the side if both rulers are parallel or how much they deviate from being parallel. This is a workflow similar to using winding sticks: By prolongating the contact line between the lower und the upper tube using a ruler or a winding stick, the deviation between the tubes at left X end in comparison to right X end is multiplied (exaggerated) and with this, the tilt (or twist) becomes more visible. Cabinetmakers and Joiners used winding sticks for centuries (millenaries!) as method to eliminate twisting (see a few links below). I don’t know if you used that method already but I can say it is very effective and often underestimated. And the same way that it works for ensuring the coplanarity of a piece of wood or a table surface (with winding sticks large enough), you can use this method to ensure that your two hollow shafts on X axis are one over the other at both ends, left and right, by using the rulers as “vertical” winding sticks. I would make a drawing if I had more time but I hope you can imagine?
Note that in this step, it is not at first essential that to eliminate any tilt, just the difference between tilt on left X end and tilt on right X end should be eliminated. The general tilting that remains same over entire X travel could then, if possible, be eliminated either by using the tramming bolts with exactly counting every revolution on each side and appliying it identically on both sides or, if tramming bolts come to their ends, by shims under the Z slider, as @SkyKam mentioned here.
Further Reading
About Winding Sticks
|
|
Liam Rickerby, CC BY-SA 4.0, via Wikimedia Commons |
How To Use Winding Sticks - Youtube
How to make Winding Sticks (Part 1) - Paul Sellers - Youtube
1 Like
Thanks. I don’t know why I hadn’t thought of winding sticks before - I am very old, and an experienced woodworker who has used winding sticks in traditional manner for decades. When I was a young man, we were using rocks tied to sticks with rawhide as hammers 
5 Likes
Pro tip: When using the tramming bolts, it’s important to do each far end of travel individually. If attempted from the center of travel, you’ll be chasing your tail and never get a good balanced tram. In this pic, assume the red circles represent one end of the machine, and the green represents the other end. It will be nearly impossible to remove the twist unless done from each end of X travel. Things can really get complex if your table isn’t near perfectly flat though.
2 Likes
Hey John @jnordyke, hey Bill,
if you now proceed the way I suggested (with one ruler attached to both hollow shafts at the right end of X, and one ruler the same way attached at the left end of X, both prolongating the alignment of the shafts one over the other towards ceiling, and thereby exaggerating (=making more visible) the tilt of the two shafts at each end) and you use the tramming bolts to align them so that the tilt is identical both at the left end of X and at the right end of X travel, then you can repeat your measuring perpendicularity to the table with your 15 inch measuring arm in the collet, once at left and and once at right end of X travel. If then, you still have a different tilt on left side and on right side referencing to table, then you have proven that your table is twisted (i.e. is not coplanar, i.e. has one foot higher than the others).
1 Like
^ Yes, this. ^
(Thank you for that, sometimes I get tunnel-vision when thinking about how I do things on my particular machine, which does not take the bigger picture into consideration with the potential complications of a wood table)
As @Aiph5u pointed out, this will ensure planar alignment of the X rails (regardless of actual front-back tram at this moment). The issue with using an indicator to tram each end is not knowing the exact condition of the working surface. With an indicator, you could end up with an artificially good tram because its been trammed to an imperfect surface. Because my machine is built from aluminum extrusion with a flat steel plate as the work surface, I can tram with an indicator to the table with confidence. A woodworking machine has the consideration of the table.
1 Like
Hey Bill,
yes, Journeyman is incredibly big. Ensuring coplanarity of the surface and at same time resistibility against deformation or distortion on such a big table (or on a base like the QCW frame) is a challenge.
But I am happy that you immediately got what I tried to explain
1 Like
Sometimes my thought process is oversimplified, and you, more than once, have made me look at other factors & solutions. 
2 Likes