Flush Trim Router Bits Not Cutting Flush - What's Going Wrong?

[SeventyFix]

Thank you for taking the time to read this post.  I'm trying to build the stacked form that Marc Spagnuolo uses to make the bent arms for the Morris chair project.  He uses paper templates to create the first plywood template.  That first template is then used to cut the next piece using a pattern/flush trim bit. As additional pieces are added following each router trim operation, the bent arm lamination form takes shape.

 

That's where I'm having the problem.  When I use a flush trim router bit to trim the next piece onto the existing template, it cuts the new piece proud of the template (the second piece is held to the template with glue and brad nails).  About 0.013" proud to be exact.  It's enough that it's certainly noticeable.  And if I reference off each additional piece, the error grows.

 

Imagine the layers as follows:

Layer 1 - the original template - offset = 0"

Layer 2 - first addition - offset = 0.013"

Layer 3 - second addition - offset = 0.026"

Layer 4 - third addition - offset = 0.039"

... you see where this is going

 

I have tried router bits from 3 different manufacturers (a cheap one that I originally bought in a set, a new WhiteSide bit (3020) and a pattern bit that I bought today from Rockler (37060)).  I have tried mounting in a router (Festool OF1400) and in a router table (Porter Cable 7518).  I have seen a similar offset regardless.  

 

What on earth am I doing wrong?  I thought that these bits would cut flush with the template.

 

Confused.

 

 

 

 

[..Kev]

Well, that's weird!

 

Check the diameter of the bearing against the cutting edges and see if that difference matches up to what you're seeing on the wood.

 

I bought a Whiteside and didn't have that issue.

[Tpt life]

Two things:

This likely should not happen so take comment 2 with a pinch of salt.

The process of duplication is flawed whenever you register off of more than one area. Any variation at all is problematic like you saw. It requires a perfection not needed if you can always reference from a template.

[SeventyFix]

The bearing on the WhiteSide bit is 0.749".  The largest reading that I can get on the bit cutters is 0.744".  Although that is off, it doesn't seem to be enough to account for the discrepancy that I'm getting with the wood.

[SeventyFix]

C Shaffer - not a bad idea - use the template to cut each of the individual pieces.  They should all have the same offset and thus all be the same size.  The drawback of this approach is that I then have to assemble them back together and get the stack perfectly aligned.  The method that I am using now glues the pieces together first and then flush trims.  It takes that second alignment step out of the equation.

[Brendon_t]

Referencing off of one template is the only thing I can think to put a bandaid on it. As far as the root cause. . Hummudnnooww

If it is happening across multiple bits, and multiple routers, that leaves bad calipers and operator error.. finely repeatable operator error. <not likely to occur with .013 accuracy. .

[wtnhighlander]

Any chance you are getting buildup on the bearings? I can buy the cutter radius being a couple thou shy of the bearing, since these bits can trim laminate without leaving a scratch, but 13 thou seems a lot.

[SeventyFix]

Brendon_T - I agree - I always assume first that I'm doing something wrong.  

 

wtnhighlander - these bits are new - this is their first use.

[Strasberry]

Is there any chance the router isn't sitting flat perpendicular to the work surface? If it was at a slight angle it would allow you to pick up the .013.

[woodbutcher74]

Is the edge of your pattern cut at 90* to the surface that the router rides on? If it is tapered then the bearing is riding on the taper and won't coincide with the edge of the pattern. Hope you understand what I mean.

[..Kev]

Is the edge of your pattern cut at 90* to the surface that the router rides on? If it is tapered then the bearing is riding on the taper and won't coincide with the edge of the pattern. Hope you understand what I mean.

 

That's a very good point! 

[ponderingturtle]

Is the edge of your pattern cut at 90* to the surface that the router rides on? If it is tapered then the bearing is riding on the taper and won't coincide with the edge of the pattern. Hope you understand what I mean.

But the second cut by the pattern bit should be square.  So that could make a workable pattern for future ones if that is the issue.

[Brendon_t]

But the second cut by the pattern bit should be square. So that could make a workable pattern for future ones if that is the issue.

Only if the two angles are cancelling each other out. Say turning the router 180 degrees

[ponderingturtle]

Only if the two angles are cancelling each other out. Say turning the router 180 degrees

That would mean the router is out of square.  Given the multiple bits and routers being used that seems like it might not be likely.  The edge of the pattern not being square to the face would mean that the bearing references off of a different line than the edge of the board.  So the router would leave a square face but there could be the step we see in the pictures.

 

So if the problem is the pattern and not the routers then toss that one piece and you have a fine new square pattern.

[Brendon_t]

Ooh I got you now. I was visualizing it wrong.

[Coop]

Can we assume that you are using the bearing that comes with each bit, and not the same Whiteside bearing on the different bits?

[Tom King]

Can you post a picture of the router with bit down beside the piece to be trimmed, and the template?

[ncfowler]

do a stright line test. make sure the template is flat, wide, and and test edge is square, and at least 3' long, and if possible on a table, this will be a sure test to see if your bits are truly off, also check your router to see if there is any bearing play, don't assume just because it runs it's right,

[SeventyFix]

After a lot of fiddling, there were several problems:

(1) I tried to make a sanding block like Marc uses in the video.  It turned out pretty well.  But my technique was probably lacking and I was slightly rounding the edges of the template.

(2) The router bit travelled along the rounded edge which threw off the sizing.

(3) There is a slight variation in the size of the cutter and the size of the bearing.

So the errors accumulate over time.  I overcame this by

(1) Using MDF so that there were no splintered edges and voids as there were in the plywood.

(2) Sanding was kept to a minimum.  I used spray adhesive to attach sandpaper to a piece of 2x4 and held it vertically in my bench vise.  This held the sandpaper perpendicular to the bench top.  Then I slid the template back and forth along the bench time, guiding it against the sanding block.  This kept the edges perfectly crisp.

(3) Use a very high quality bit.

 

[TerryMcK]

Just to add my two cents. The flexible sanding strips that I make out of 1/4" or thinner plywood I size so they are only slightly wider than the edge I am sanding. So if the edge is 3/4" wide I'll make it about 7/8" to 1" wide and that minimizes any tendency to round over the edges.. I also use self adhesive sandpaper 3M gold from StewMac which comes on rolls..

Marc shows flexible strips in his videos that are 2" to 3" wide. That is fine for general usage as long as you are aware to keep pressure within the centre of the edge you are preparing otherwise, as you have found, you end up rounding over the edge.

[J0hn]

I ran into this problem when I was using a 1/2" flush trimming bit when I was trying to make a MFT style top for a Paulk workbench

What I found out is that several manufacturers deliberately oversize the bearing so that you do not cut into your pattern, but I don't recall it being out as much as what you are reporting.

I ended up talking to someone at Whiteside - and he actually went out to the production floor found a bit and bearing that matched exactly and sent it to me - how is that for customer service?

Before that I had good luck with a Freud 1/2" bit, bit that was getting a little worn and I was looking for a replacement

[Dknapp34]

I ran into this problem when I was using a 1/2" flush trimming bit when I was trying to make a MFT style top for a Paulk workbench

What I found out is that several manufacturers deliberately oversize the bearing so that you do not cut into your pattern, but I don't recall it being out as much as what you are reporting.

I ended up talking to someone at Whiteside - and he actually went out to the production floor found a bit and bearing that matched exactly and sent it to me - how is that for customer service?

Before that I had good luck with a Freud 1/2" bit, bit that was getting a little worn and I was looking for a replacement

I've noticed this with a Whiteside flush trim bit--the bearing being slightly larger than the bit.  I could see how its a problem when working with a pattern, but I actually don't mind it for trimming solid wood edge banding on plywood, as it gives me some comfort that I'm not going to tear through the veneer on the plywood and it's easy to fix with a card scraper.