520 Extension Table/Fence issue

algale · Post by **algale** » Sat Jun 23, 2012 6:54 pm

dusty wrote:Very Interesting.:(

To test that theory:

Nice test solution but I am left to speculate: is that a reflection or daylight I see between the rail and the the bottom 3-4 inches of the square in the last photo?

dusty · Post by **dusty** » Sat Jun 23, 2012 7:10 pm

algale wrote:Nice test solution but I am left to speculate: is that a reflection or daylight I see between the rail and the the bottom 3-4 inches of the square in the last photo?

That is a little bit of both. That framing square has been around for many years and the edges are less than pristine.

The significance of the method is in having the table tilted. The framing square can be manipulated much more easily.

You can do the same test based on the side bosses (protrusions) with the table flat but it takes all three hands.

When you are all through with this there is one thing that you can well proclaim. The ribs on your table top are all parallel and perpendicular to the front edge of the table. Many do not trust that but they are.

algale · Post by **algale** » Sat Jun 23, 2012 7:41 pm

JPG40504 wrote:Been away from this all afternoon and the videos are 'enlightening'.

1) They illustrate the apparently uncontrolable fetish with the table top grooves. Yes from a visual perspective(pun intended) they should be consistent between the main and the other tables, but really are not relevant to the objective of the fence being consistent regardless of which table the fence is attached to.

JPG, I think you may have missed the point of why I spent time talking about the grooves in the tables and how the fence aligned to them, probably because I didn't explain myself as well as I should have.

We know from geometry that If A is parallel to B and C is parallel to B then C is also parallel to A. If C is not parallel to B then it is not parallel to A.

In this case you can think of A is the grooves AND the miter slot in the main table. Think of B as the grooves in the floating table. C is the fence. Since I have established A (the miter slot and grooves in the main table) were parallel to B (the grooves in the floating table) by putting C (the fence) on the extension table and demonstrating C wasn't parallel to B, I proved that C (when put on the extension table) wasn't parallel to the grooves or miter slot on the main table (A). Since the miter slot on the main table is parallel to the blade, this proved the fence was not parallel to the blade when on the extension table.

I thought this would be grasped and that is was more accurate than trying to measure from both ends of the fence back to the blade.

I need to cogitate about the rest of your post, for which I thank you.

Al

JPG · Post by **JPG** » Sat Jun 23, 2012 7:48 pm

dusty wrote:That is a little bit of both. That framing square has been around for many years and the edges are less than pristine.

The significance of the method is in having the table tilted. The framing square can be manipulated much more easily.

You can do the same test based on the side bosses (protrusions) with the table flat but it takes all three hands.

When you are all through with this there is one thing that you can well proclaim. The ribs on your table top are all parallel and perpendicular to the front edge of the table. Many do not trust that but they are.

I really think they should be and have no reason to think they are not! It is just that I think they are a distraction to the task at hand.

I predict that when this is resolved, they will indeed be so, but are not part of the solution nor a relevant reference!;)

While 'predicting', I'll hazard a guess that the inside top surfaces of the main table rail are not positioning the tube parallel to the front fence reference surface. I am not entirely comfortable with that considering the direction of the tube skew and the existence of two clamping screws on the main table.

Maybe a burr on the back surface of the top of the right end of the main table rail. That would cause the tube to both skew in the apparent direction it is, and, explain the interference with the aux table left end.

Wish I could get my hands on the recalcitrant tables!:(

JPG · Post by **JPG** » Sat Jun 23, 2012 8:15 pm

algale wrote:JPG, I think you may have missed the point of why I spent time talking about the grooves in the tables and how the fence aligned to them, probably because I didn't explain myself as well as I should have.

We know from geometry that If A is parallel to B and C is parallel to B then C is also parallel to A. If C is not parallel to B then it is not parallel to A.

In this case you can think of A is the grooves AND the miter slot in the main table. Think of B as the grooves in the floating table. C is the fence. Since I have established A (the miter slot and grooves in the main table) were parallel to B (the grooves in the floating table) by putting C (the fence) on the extension table and demonstrating C wasn't parallel to B, I proved that C (when put on the extension table) wasn't parallel to the grooves or miter slot on the main table (A). Since the miter slot on the main table is parallel to the blade, this proved the fence was not parallel to the blade when on the extension table.

I thought this would be grasped and that is was more accurate than trying to measure from both ends of the fence back to the blade.

I need to cogitate about the rest of your post, for which I thank you.

Al

That additional elucidation was not necessary.

Indeed it summarizes the problem quite well.

My point re the grooves, is that are not directly relevant to the problem. They are however a visual indicator of the problem.

Some of the details have started to sink in and drown some earlier brain storming.

Let me try to enclose my understand of the symptoms etc.

It has been well established that the fence(by way of the front reference surface of the main table rails) is essentially perfectly aligned parallel to the miter gauge slot(the blade alignment is a different adjustment!).

The same fence incidentally also aligns to the grooves in the main table top.

Also the same fence when secured to the aux table independent from the main table also aligns with the grooves in the aux table top.

It is when the aux table is attached to the main table with the connecting tubes and the fence then attached to the aux table rails(fully or bridging to the main table rail) that the fence no longer aligns with the grooves and by inference(?) the main table miter gauge slot.

One point I am unclear on is the video showing the aux table grooves being parallel to the main table grooves. At that time, how was the aux table referenced to the main table? Sides of the tables, or the tubes and rails?

Forgive me for beating this drum repeatedly, but the sides and the grooves are really not relevant(although visually IMHO should be parallel to the slot).

algale · Post by **algale** » Sat Jun 23, 2012 8:25 pm

JPG40504 wrote:
2) The starrett square edge against each of the main and aux table(I will not distinguish between extension and floating table here as the extension table only differs by the inclusion of the mounting tubes which is an independent alignment adjustment) was informative, but did not illustrate what is relevant here. What 'matters' is the straightness from the left end of the main table rail all the way to the right end of the aux table rail. That 'straightness' will illustrate the parallelism of the two rails to each other.

The test was done against each rail separately to remove the small possibility either was bent (on the outside) and then across them to demonstrate there was a small misalignment. This small misalignment, which I suspect is in the main table, may be the source of the problem.

JPG40504 wrote:
3) No attention is being paid to the straightness of the connecting tube. None of all this will work if the tubes are not straight. The tubes must be tight against the reference surface(s) on the inside of the rails. The 520 screws force the tubes against the top of the inside of the rails. This differs from the 510 which forced the tubes against the inside front of a pipe(rail). It is crucial that the inner top surface of the 520 rail extrusion also force the tube into a position parallel with the outer front surface of the rail extrusion.

The tubes are not discussed in the videos, but as I said in a response to this issue earlier, I did re-check the rails. They are very much like the rails in Nick's sawdust session on aligning the extension table on either end of the Shopsmith, which is to say they aren't perfect. As Nick demonstrated, I rotate them in the miter slot until they sit in the slot with no rocker, then mark that spot on the tube with a Sharpie and orient that spot "up" in the fence rail. This should ensure that any arch in the tube is oriented up and is not deflecting from side to side.

JPG40504 wrote: I have mentioned the tube straightness, but we must also pay attention to more localized defects(burrs, dents, out of round. . .).

The rails need to be examined for similar defects, but the most likely place is the ends of the rails. Probably the most incidious would be a slight movement of the front of the rail bent inward thus skewing the tube relative to the front.

So for current lack of any better places to 'look', the tubes and the top inside of the rails.

There are some minor burs in the bottoms of the tubes where the thumb screws have been tightened and left minor impressions; however, even when the tubes are slid moved so that the screw engages an unburred portion of the tube, I have the same problem. Since the bottoms of the rail don't reference anything or affect the reference of anything unless they are the point where the screws are touching, these burrs are inconsequential and would likely effect only vertical reference rather than side to sidel.

As for tubes bending inward, because I have oriented the bend in the tube in the vertical orientation, this is not the issue either.

Another fact: when I mount the extension table in the base arm with no connecting tubes (i.e. isolating the extension table) and put the fence on it and lock it in place, the fence sits at the same slant relative to the table grooves (there's my fetish again) as on the floating tables when the floating tables are connected to the main table.

I think this proves that the rails on the main table are skewed relative to the main table and that the fence has been aligned to compensate for this skewing. As long as the fence stays on the main table, all is well. Once the fence is moved to any of the auxiliary tables, this skewing reveals itself.

algale · Post by **algale** » Sat Jun 23, 2012 8:32 pm

JPG40504 wrote:

Also the same fence when secured to the aux table independent from the main table also aligns with the grooves in the aux table top.

False. When the fence is placed on the extension table set by itself in the base arm with no connection to the main table (i.e. no table tubes, no touching of table surfaces), it locks in the same skewed position relative to the grooves in the extension table.

This can only mean that the table rails on the main table are skewed relative to the edge of the table and hence the grooves/miter slots in the main table. The fence is parallel to the miter slots and grooves of the main table because the fence has been (inadvertently) adjusted with the exact complementary angle to the angle of the fence rails, thus compensating for the skewed main table rail. All is well as long as the fence remains on the main table but the skewing is revealed when moved to an auxiliary table which has rails that are not skewed.

Now I just have to prove this hypothesis.

algale · Post by **algale** » Sat Jun 23, 2012 8:37 pm

JPG40504 wrote:That additional elucidation was not necessary.
One point I am unclear on is the video showing the aux table grooves being parallel to the main table grooves. At that time, how was the aux table referenced to the main table? Sides of the tables, or the tubes and rails?

Forgive me for beating this drum repeatedly, but the sides and the grooves are really not relevant(although visually IMHO should be parallel to the slot).

In the video, the floating table was touching the main table side bosses.

There's nothing to forgive. I deeply appreciate all your input. It is helping to refine my own thinking and may ultimately be correct.

JPG · Post by **JPG** » Sat Jun 23, 2012 8:55 pm

Perhaps it is time to attempt to discover things which do not align correctly two parts at a time. This may be more difficult, but start with the front main table rail and a tube. Insert the tube and secure it oriented as you would if you were going to attach an aux table. Using a straight edge(longer than the combination square) determine if the tube is indeed parallel to the front outer surface of the rail. Repeat using the same tube oriented as before but secure it to an aux table(only) and again check for parallel to the front outer surface of the aux table rail. Look at the spacing between the extended part of the tube and a straight edge held tightly against the rail front surface. If parallel the spacing should be constant along the total extended portion of the tube.

To put some perspective on this, that 1/16" deviation front to rear of the fence to the grooves is 1/3 of that across the front rail.(0.020") Bear in mind any 'vertical' differences will be the result of the 'bowed' tube.

This needs to be done on both the front and rear rails.

Come back with some observations.

algale · Post by **algale** » Sat Jun 23, 2012 9:15 pm

Ok. The front fence rail on the main table most definitely was not perpendicular to the edge of the table and hence not perpendicular to the miter slots or table grooves. Proof here: http://www.youtube.com/watch?v=K8vcg6jsgp8