Tramming the Router


I’m very much still learning and a bish-bash-bosh sort of person (I’m no engineer), so I haven’t trammed using the pins/gauges etc. But my thoughts/opinions are:

  • being slight out in the tram can’t be a big issue unless you’re using large-diameter bits.
  • the only significant diameter bit I have is for wasteboard surfacing; I think it’s a 1.25 or 1.5-inch flattening bit.
  • I, therefore, used the wasteboard surfacing process to tram my CNC (I created a spiralling-out wasteboard flattening path using v carve). I had to shim the bottom of the zslider to get a wasteboard that, to my eyes and fingers, is perfectly smooth.
  • once you have a smooth wasteboard, I would get cutting (I find that I can use extra kindling this time of year :smile:)
1 Like

Unless you’re machining metal is go with “it’s good enough”. I’m willing to bet that any wood you use will change dimensions by more than that just from environmental conditions. Including your wasteboard.

It’s funny how we used to be good with the results we’d get using hand tools and sandpaper. Now that we have a machine with the capability of achieving measurably precise tolerances we chase them to 4 or 5 digits of precision :smile:


Brian, having built a couple of wood CNC routers, I can tell you are pretty darn close. Unless you have some inlay projects, you will probably not notice. The work on the table will average things out for you. The weather may have a bigger influence than what you are seeing. Start making some chips until your projects start getting complex. When you start making joints check them for square and then start to revisit tramming. As long as the machine is making circles and squares accurately, I think you are good to go.

1 Like

Hey Jim, hey all,

but this is different! If you do something with a hand tool and sandpaper (or as I rather would, with a cabinet scraper), then you will never get the same result twice. You’re having a bad day and doing a little more crooked than yesterday where you were in love.

But if you buy a robot that does the work for you, then if it does something crooked, it will do it crooked in exactly the same way every day, every minute, and at some point you’ll hate him for it. And then of course you will make sure that it does not make crooked, but perfect.


I make crooked things by hand every day, every minute. :grinning_face_with_smiling_eyes:


Hey Brian,

what exactly is the unit with which you give these measurements? Is it thousands of an inch? If yes, the difference between 0.040″ and 0.060″ is 0.020″, that is 0.5 mm. Whether one would tolerate that depends on what standard you apply and your preference.

At the moment when you measured this, it seems that your jig shows that front numbers are higher than rear numbers, and left numbers lower than right numbers, which would show that Z Axis is slightly tilted backwards and slightly to the left. But for the difference between the values at different measuring positions, the different heights may be a result of play in your measurement setup or of the unevennes of your wasteboard, or differences in the height of the points were your machine’s feet are located.

What you are concerned about may not be a matter of trimming the perpendicularity of the Z Axis (“tramming the router”). Trimming this will mainly have effects only on which patterns appear when surfacing or on the accuracy of vertical walls in deep pockets that you mill.

I ask myself, if you are crazy about trimming Z perpendicularity up to the point of not having started to use your machine since two weeks, i.e. if you are at the step of checking the perpendicularity of the Z Axis in relation to the plane that is formed by X and Y, you already have to be sure that this plane, i.e. the mounting surface, is flat, wich means both coplanar and even. You can ensure the evenness with the surfacing process, but not the coplanarity and thus not the flatness.

Machine could be twisted

If you have twisting in your mounting surface, it is difficult to measure it this way, because the twist will twist the X Axis too. And then your router will tilt according to the twist when moving it from left to right.

It is important that before doing 1.) is to have 0.): Ensuring Coplanarity of the worksurface, or better said, at first coplanarity of the four machine’s feet. Coplanarity means that all four machine’s feet are on one plane, which means to avoid that one machine’s feet is higher than the others, or better said, that your tabletop is twisted.

Many people go crazy for adjusting the rectangularity (“squareness”) of the Onefinity CNC (rightly because the Onefinity CNC, as its main weakness, does not ensure the rectangularity by itself) but they forget the crucial importance of coplanarity of the mounting surface.

Also many people believe that irregularities in their wasteboard will be eliminated with surfacing the wasteboard with a surfacing bit. But this is not true. If your machine’s mounting surface is twisted, i.e. if one of the machine’s feet is higher than the others, then your wasteboard will remain twisted, and so will every workpiece’s surface be twisted after processing.

To ensure coplanarity, Tom @TMToronto mentioned the fishing line method here the other day, which is simple and very accurate. Other methods are winding sticks, water tube level and level (optical instrument).

I would ensure this as the first step before assembling the Onefinity CNC on a table. Especially if your table has no diagonals to prevent twisting. This would mean that the coplanarity of your worktop depends on the evenness of the floor and, even after tabletop trimming, the coplanarity will go down the drain if you move the table somewhere else.

1 Like

Hey Rex,

I find this is a candidate for the “As precise as possible, as simple as necessary” prize.

I like solutions like this.

(yes I know, I like totally expensive measurement tools, but in my heart I like this.)

I want to thank everyone for the help and encouragement to keep going. I have trammed it to within 0.003" which is basically lean on the table tolerances. Hint for everyone who doesn’t want to go down the rabbit hole - use a piece of paper to get close not a dial indicator, the numbers may seem bigger on the dial vs just “I am one piece of paper off” lol.


level (optical instrument).

A rotary laser level would work well also.

1 Like

After a few months of use I finally trammed my Z and fine tuned the stepper motors’ travel. I was able to get within .001" left-right, front-back. Combination of foil tape on the Z mount and using the jacking screws on the X axis. Plenty precise enough.


Interested. Please send details.

1 Like

I’ll PM you the details.


More tramming experience - I followed advice from Alex McClure (thanks!) and moved my Z-Slider to middle hole. This allows tramming gauge to swivel underneath the Z-Slider. I zeroed the X-axis tram by loosening and v-e-r-y carefully retightening the two router mount screws while observing the twin gauges - tightening the top screw tends to shift the X axis one way, tightening the bottom screw reverses that. For the Y-axis tram, I ended up shimming the Z-slider at the top with a .020" feeler gauge as the shim; this corrected the “nod” of the router to within .0015" (it was .005"), which I’m calling done.


Just sent you an email. :slight_smile:

I had to remake the spoil board because of Z-Slider hole move. Resurfaced. Tested on .75" pone with surfacing bit. Not happy. So I pulled out the glass to recheck tramming. That’s when I decided to chuck the glass and order granite surfacing slab from Lee Valley. In for a penny, in for a pound, as they say. Stay tuned.

1 Like

I am in the midst of calibrating my new set-up, and ordered the same granite. I noticed a discrepancy between the advertised accuracy (.0001") and that on the certification card that accompanied the granite (.0002"). I have emailed them and they said they would let the appropriate departments know. They even opened a box of their own and confirmed what I found was also in their box.

Not a deal breaker for the price. They said I could return it, but I do not plan to. It should match the accuracy and repeatability of what I am currently finding.

1 Like

I wonder how flat a piece of granite countertop is compared to the surfacing slab.

The difference in flatness could possibly be inferred in that surfacing slabs are presented with flatness tolerances while the countertop slabs are not.

But now I want a kitchen countertop rated to ±0.0001 so I can achieve perfection in cooking… because obviously it’s the weak specs of the countertop to blame and not my overwhelming lack of cooking skills. :wink:


How do you guys know your surface is perfectly flat when you’re doing these tramming calibrations?

It’s typically determined by chucking a dial indicator capable of measuring flatness deviations to 0.001. You’d observe the needle moving on the indicator as the gantry is manually moved along the board (but hopefully not if the board is actually flat).

The low-fi setup employs something like two drill bits embedded in a piece of wood, one bit chucked to the spindle / one zero’d to the board, that’s manually rotated around the board along with use of feeler gauges to discern discrepancies in flatness.

1 Like