Poll: For Those with a Festool MFT...

[Holmz]

No fence and rarely the mitre job.
Just dogs in the back to index the wood to, and the flopping hinged rail holder has the Bosch rail on it.
I use dogs behind it to make sure it is square to the dog holes, and there is ~1/4-mm between the dogs and the rail.

The are also long(hot) dogs that can go into the rail to use the MFT holes, that go into T-Slots in the rail.

 

[Patrick Cox]

This was very helpful [member=1674]Peter Halle[/member], thanks for sharing this.  I started my setup over and followed your process exactly, other than I used my 50mm spacer for setting up the guide rail, and I feel like I was able to get the setup as you described.  I ran out of time to make a test cut but I plan to tomorrow and will report back.

Also, I like your green cool aide bottle! 

OK, well I made another test cut and unfortunately I am not seeing any improvement.  I performed the 5 cut test again and my measurements of my 5th cut were...

25.46mm at one end
23.08mm at the other end of a 426mm cut

So I decided I was going to try to cut using dogs only as my fence and before doing do, I placed three dogs into three holes in the MFT3 and I lined up a straight edge against the dogs and what did I discover - not all of the holes in my MFT3 are parallel.  And this MFT3 is only about two months old with not a lot of use.  After checking further, I am finding that some combinations of holes are parallel but others are not.  So I am not sure what to do from here.  I am not sure what the tolerances are for these holes but I am not getting anywhere near the precision that Peter is getting in this video...


So any further suggestions would be appreciated.  I purchased my MFT3 from one of the larger online Festool dealers.

Thanks.

 

[ear3]

The top is CNCd so it's highly unusual that there would be these errors.  Do the dogs have the same fit in all the holes, or at least between those that are good and those that produce the offset?

 

[peter halle]

If the holes are not in the same plane then this will be the first post I have ever read here since the beginning of the forum about that issue.  Weird - not questioning your observations - just weird.  Although Festool doesn't proclaim anything as to the accuracy of their grids you should probably reach out to them and send them some pictures because I am sure that they would be interested world wide.

Peter

 

[Patrick Cox]

Ok, please forgive me for the video and audio quality here.  I am not really setup to make videos.  So I strapped my GoPro to my head and made this video.  The sound is not great but here is what I did...

Test 1. Tested a row of holes using my Festool fence which I believe to be straight
- installed 4 Qwas Dogs in a row of holes
- placed a post it note between each dog and the fence.  (sticky side was up in the air.)
- pressed the fence up against the row of dogs and then attempted to remove each post it note.  (Using post it notes is how Brian Sedgeley at Festool showed us how to check squareness of the guide rail and the fence in the class that I took.  We would use a woodpecker square and then position the guide rail and fence until post it notes were "locked" in several locations and then lock everything down.)
- As you will see in the video, there are two hole locations where the post it note moves very easily.  There is certainly no contact with the dog and the fence.

Test 2. Second test was to line up my Woodpecker Rule against 3 dogs and try to move the rule.  With the first 3 dogs there was clear movement.  On the 2nd 3 dogs there was no movement.

So I would love to get feedback on these tests and take any suggestions for further testing.  Thanks.

 

[Klemm]

Patrick, I feel for you. One way to get a rock solid reading on "accuracy" is to take the top of your MFT-3 to any substantial machine shop in your area. Have them use their CMM coordinate measuring machine or non-contact laser to quickly map out the actual physical location of your hole pattern and match mark the table top so you can correlate data to the exact hole.
You'll either know that you have a dead accurate table or a strong case to have Festool give you a new one.
Cost to do this will be surprisingly reasonable because i takes no time for a QC department to do it.

If you can't find anyone in your area, PM me and I'll see who I can find for you.
Hans

 

[McNally Family]

Question.....are the results the same if you flip the guide rail around and use the other side?  Just trying to eliminate the guide rail or ruler as being out of alignment (unlikely, but a variable you need to eliminate).

 

[Patrick Cox]

Question.....are the results the same if you flip the guide rail around and use the other side?  Just trying to eliminate the guide rail or ruler as being out of alignment (unlikely, but a variable you need to eliminate).

The other side has the groove for securing the fence so it is not easy to make that test.  I have been planning to order a Veritas straight edge anyway so I am going to do so and use that to test as well.  I guess I will also call Festool tomorrow.  [sad]

 

[Stone Message]

Before anyone goes rushing off to complain about their MFT hole accuracy take a look again at Peter Halle's post. I agree with him.

I have never seen any claim by Festool that the pattern of holes on the MFT3 is done to any tolerance of any sort. Even the holes are of a nominal size and we know that they vary.

The MFT3 holes were designed as a clamping and jig layout aid.

If you need any proof of this then try the 4 (or 5) cut test on your MFT3.

Peter

 

[Patrick Cox]

Before anyone goes rushing off to complain about their MFT hole accuracy take a look again at Peter Halle's post. I agree with him.

I have never seen any claim by Festool that the pattern of holes on the MFT3 is done to any tolerance of any sort. Even the holes are of a nominal size and we know that they vary.

The MFT3 holes were designed as a clamping and jig layout aid.

If you need any proof of this then try the 4 (or 5) cut test on your MFT3.

Peter

I think you are correct Peter on Festool's position regarding the accuracy of the holes in the MFT3 stop.  But I continue to see people rave about how accurate they are and it seems that at least some of my holes are out of alignment with other holes to some degree.  What I am not sure of is how significant these differences are to my cutting accuracy or inaccuracies in this case.  Since I am very new at this, it is very possible that the dog holes are not my primary issue.  But I do plan to make the five cut test using dogs only with the cutting rail and see how that test turns out.

As always, thanks for your support.

Patrick

 

[Holmz]

Patrick, I feel for you. One way to get a rock solid reading on "accuracy" is to take the top of your MFT-3 to any substantial machine shop in your area. Have them use their CMM coordinate measuring machine or non-contact laser to quickly map out the actual physical location of your hole pattern and match mark the table top so you can correlate data to the exact hole.
You'll either know that you have a dead accurate table or a strong case to have Festool give you a new one.
Cost to do this will be surprisingly reasonable because i takes no time for a QC department to do it.

If you can't find anyone in your area, PM me and I'll see who I can find for you.
Hans

Or just measure diagonally NE to SW and SE to NW and compare the numbers...
It should be:
= (32-mm * N holes * SQRT(2)) - 20-mm
or if you measure to the outside of a dog (which is easier)... Then add 20-mm instead of subtracting it.
And should be the same on both diagonals.

 

[bertv]

It does not really matter if there is some variation in the holes. Once you have set up the fence and the rail with any of the methods described, you perform the 5-cut test and then you can compensate for the error by using feeler gauges to correct the positioning of the stops in the MFT-profiles.

How its done: Determine the error (compensate for the length of your test piece, as explained in the video's about the 5-cut method). With the front and back plates of the rail butted tightly to the stops, tighten them. Loosen the stop that has to ‘give’ (depending on minus or positive compensation), stick the required number of feeler gauges into the space between stop and plate, press stop onto that, tighten. Loosen plate and slide back to stop. Repeat procedure until satisfied.

 

[Birdhunter]

I get perfect 90 degree cuts using QWAS dogs, SlopStop, and Woodpecker framing square. I've never used the MFT protractor.

I now have the Woodpecker MFT square and will shift to that the next time I true up the MFT.

 

[Patrick Cox]

It does not really matter if there is some variation in the holes. Once you have set up the fence and the rail with any of the methods described, you perform the 5-cut test and then you can compensate for the error by using feeler gauges to correct the positioning of the stops in the MFT-profiles.

How its done: Determine the error (compensate for the length of your test piece, as explained in the video's about the 5-cut method). With the front and back plates of the rail butted tightly to the stops, tighten them. Loosen the stop that has to ‘give’ (depending on minus or positive compensation), stick the required number of feeler gauges into the space between stop and plate, press stop onto that, tighten. Loosen plate and slide back to stop. Repeat procedure until satisfied.

This is a good idea, thanks.  I tried to re-square my setup again and I am going to make some more test cuts today and if I am still off I will work on adjustments to the guide rail feather key stops in an attempt to get more square cuts.

I am also going to perform the 5 cut test with just dogs to really test the squareness of the dog holes.

Thanks again!

 

[Michael Kellough]

Patrick, it's good to read that you are aware that there could be several ways to skew the results.

Even though Festool doesn't claim the hole pattern in the MFT surface is extremely precise I'd be suspicious of the fence first. If that is not straight all else is futile.

It's easy to determine if the fence is not straight.
Just draw a line with it then put the fence on the opposite side of the line and draw another line.
Throw that wood-clinched pencil away, at least until you finish testing, you need a .5mm or finer line for this.
(You could use a wood-clinched pencil hard lead pencil like a 4H if you sand it to a sharp point with 320 grit or finer abrasive.

If the two overlapping lines form an oval or opposite shape then the fence is not straight. Replace it.

However, you should be able to get square cuts using those dogs, unless the hole pattern is truly off.

As others have said, the diameter of the holes is not extremely precise and you also need to check the diameters of the dogs, above and below the surface.

 

[Patrick Cox]

OK, after trying to setup my fence and guide rail all over again using the dog holes for alignment, below are pictures of my setup and the list of cuts I made and resulting measurements.  Bottom line, after 1,650mm of cuts, my last cut was off by 1mm over 434mm in length and my cumulative difference was 2.8mm over the full 1,650mm in length.

I am not sure if the totals are significant or not and to be honest, I don't really know how to interpret how off I really am.  It seems that my guide rail needs to be moved to the left slightly at the beginning of the cut but with such a small difference, I am not sure how to adjust the rail to correct this.  Just trial and error?  I also don't know why my 2nd cut was closer to being square than the others.

So further comments/suggestions will be appreciated!  Thanks!

Picture of setup after final cut...  (This is not positioned as the cut occurred.  The final off-cut was to the right of the kerf.  I took this picture after I measured everything and did not place back as it was cut.)

Cut list with measurements...

 

[Michael Kellough]

Is the fence straight? If not, fugidaboutit.

You should review the 4 cut process. You only need to measure the last off-cut. Any off-cuts prior to that are irrelevant so you only need to remove about twice the kerf width of material (just enough to trap the dust).

Just in case you don't know, you also need to keep the fence completely free of debris. A small bit of hard dust between the wood and the fence during any of the cuts spoils the whole procedure.

 

[Patrick Cox]

Is the fence straight? If not, fugidaboutit.

You should review the 4 cut process. You only need to measure the last off-cut. Any off-cuts prior to that are irrelevant so you only need to remove about twice the kerf width of material (just enough to trap the dust).

Just in case you don't know, you also need to keep the fence completely free of debris. A small bit of hard dust between the wood and the fence during any of the cuts spoils the whole procedure.

Michael,
I tested the straightness of both sides of my fence using a 0.7mm pencil and both sides appear to be straight to me.  So I don't think that is the issue.  I will test again while checking for dust in between each cut and then I have also ordered some Parf Dogs and I plan to test just using the dogs to see if I can narrow down to my setup in some way.

Regarding the 4 cut process, I have seen how to execute the process and how to calculate a number by dividing the measurement difference by 4 and by the distance of the cut.  However I have not seen what that number really means and what would be an acceptable metric for woodworking.  I will google some more and see if I can find more info.

Thanks again for your suggestions.

 

[Michael Kellough]

1mm difference on the fourth cut means your angle is 1/4mm off at 430 or so mm.

That is an angular difference of only 0.03 degrees. Considering that the best digital angle finders are only accurate to 0.1 degrees you are more than close enough for woodworking.

Get a bigger scrap of plywood and cut it at 90.03 degrees and save it for re-setting the fence if it gets out of wack in the future.

You can take a break from angle finding and make something.

 

[richk]

The Kapex Supplemental Manual p 24 describes the method: https://www.festoolusa.com/media/pdf/kapex-ks120-supplemental-manual.pdf