[GME]

This is what I use but not on most wood projects. 11HC-12-16R 12" Combination Square with Square and Center Head

- - - Updated - - -

I have a set just like the one you linked to. Mine is probably 30 or so years old. I have the one Starrett and a couple of other sets as well. I don't use any of them that often. These days, I use fixed blade precision squares. Here's one set I have: I have a 4 piece set (2", 3", 4" and 6") of Groz that are square to 16 microns (.00063"). I use them pretty much everyday I'm in the shop for wood, metal, setting my table saw blade perpendicular to the table - you name it. I also have a 12" machinist square that gets some use. It's pretty heavy, though. For bigger stuff, I use my trusty Woodpeckers 18" triangle. Way too pricey, but I've found it worth the money. https://www.woodpeck.com/onetime-too...ngle-2013.html It was one of their "One Time Tools" and no longer available. They's brought it back by popular demand 3 times since I bought mine, so it may come up for sale again.

Here's another triangle that's very similar to Woodpeckers, and a few dollars cheaper. https://tsoproducts.com/tso-precisio...asic-triangle/ It's accurate to within .001" over the full 18". Pretty good. They have some variations, but the price goes way up real fast for them. If I ever decide to buy another 18"er, I probably go with the TSO Products model.

The triangle is great for setting up the Festool circular saw guide rails for square cuts. TSO also has a product that clamps onto Festool guide rails and works nicely for setting up square cuts. I have the GRS-16. https://tsoproducts.com/tso-products...-rail-squares/

FWIW, I used my Woodpeckers square to do the rough gantry squaring when I mounted the gantry on my first machine and on my current one. My first one was within a few thousands on a 71" gantry. Beats the heck out of just guessing. I'm not too confident on this machine, though, I wasn't able to set up the triangle the way I wanted. That's okay. My squaring routine should get it dialed in nicely.

Gary

[wmgeorge]

Gary we are All looking forward to the projects your going to make. Hey, and Happy 4th of July!!

[GME]

Gary we are All looking forward to the projects your going to make. Hey, and Happy 4th of July!!

Thank you, Bill. Happy 4th to you too!


- - - Updated - - -

[GME]

Three days ago, I posted about setting steps per for my X axis using 123 blocks and a dial indicator. Here's a question. Since the X and Y axis steppers are all the same, and the R&P drives and racks are all the same, shouldn't I just be able to copy my steps per from my X motor setting and paste it into the steps per for Y and Slave?

Common sense says the steps per should be the same for both X & Y. However, they are not for this machine, and were not for my first machine. They are close for both, but definitely not the same. I don't know why they are different.

Here's a photo of my setup for the Y axis:



The setup for the Y axis is very much like I used for the X axis, except I had to bridge crossbeams, rather than laying the 123 blocks on them. I used one of my Festool guide rails for my track saw. First, I aligned my precision triangle with the edge of the front crossbeam, and then aligned my guide rail off the perpendicular edge of the triangle. I then used a couple of F clamps to hold the guide rail in position. Note the raised cross section just right of the rightmost green strip. The 123 blocks laying flat are pushed up against the raised part. What you can't see in the picture is a very small profile edge (raised less than 1/16") to the right of the large raised part. I used the small raised part as a guide for the 123 block that's on edge. Coincidentally, it nicely locates the 123 block that's on edge.

In the photo, you can also see how I attached the dial indicator to the gantry. The steel part behind the indicator magnet is one of the parallels I use in my milling vise. I tried sticking the magnet to the lower linear rail, but wasn't satisfied with it's stability. It stuck tightly, but didn't take that much to get it to move around - too little contact area. The parallel did the trick.

The rest is done just like the X axis, except I found a couple more 123 blocks, so I was able to extend my measurement range by 6 inches.

When I measured X, and using the Mach4 calculator, I came up with 2037.5132 steps per. When I measured Y, I came up with 2038.2710 steps per or a difference of .7578 step per inch.

Gary

[ger21]

I've never used any 123 blocks, but is it possible that you are getting some deviation when you stack 10 of them together.

Other issues can be manufacturing tolerances with the racks.

And how precise is spring loading the pinion into the rack?

I would think the only way the steps/unit would be exactly the same, is if you used high accuracy precision bound ballscrews, with zero backlash.

[GME]

I'm closing in on getting my spoilboard assembly put together. However, because I have dados (rebates) to cut in one stage, and slots to cut in another, I thought I might just as well tram the spindle and square the gantry before moving on the the spoilboard installation This post is about tram.

I used a piece of plate glass sitting on the machines crossbeams as my reference surface. I recognized that after surfacing the spoilboard, I may need to recheck and maybe adjust the tram a bit.

I was surprised by the amount of shimming necessary for the front-to-back tram - .018". I had to shim more than I expected. I cut steel shim stock I keep on hand - a piece of .015" and a piece of .003". .003" cuts easily with scissors. .015 is too thick for scissors. I learned way back that serrated aviation snipes don't work well at all. They leave rough edges that take a lot of work to rehabilitate. I have a pair of long handled straight cut sheet metal sheers (not the duck bill shape) that work pretty well. I used a carbide scribe to mark the cut and cut away. Almost no curl, even though I was cutting a 6" long piece. Just a very light touch up with a diamond stone on the cut edge and it was good to go.

Long story short, I was able to tram in both directions to near perfection. Here are photos:



This shot is from front to back (side-to-side tram).



This shot is from side-to-side (front-to-back tram).


Awhile back, in someone else's thread, I commented that it's best to get the front-to-back tram done before the side-to-side. The reason was that doing the front-to-back will mess up the side-to-side. I was close to right, but not completely right. You have to adjust both in each test setup. You start by shimming the front-to-back, and moving directly to the side-to-side. I found that one direction affects the other, which makes it a very long, drawn out and fussy job. Even with the tramming gauge, a whole lot of trial and error. I worked on the tramming project for 3 to 4 hours before I was satisfied.

Gary

[GME]

Just a little organizer. I added a hook to keep my collet wrenches close at hand. It's made from a 5/16" U-bolt. Here's a picture of it.



I hope they don't rattle too much when the machine is cutting.

Gary

[wmgeorge]

Having two dial indicators and two mounts is not the way I was taught to setup a mill vise. One indicator reading both sides is more accurate, thats what I was told by my instructor who had years in the business.

[the4thseal]

Not sure about using the 1 2 3 blocks like that. Even with blocks the "wring" you might be asking for trouble. Though you should use what you have to the best of your ability. Good work so far.

[GME]

Having two dial indicators and two mounts is not the way I was taught to setup a mill vise. One indicator reading both sides is more accurate, thats what I was told by my instructor who had years in the business.

Oops! You misread my post, Bill. The double indicator setup is not for setting up a milling vise. I don't have any idea how you would go about setting up a vise with a double indicator. I learned to use a single indicator for setting my vise, and don't know of any other way to do it (except having a machined plate that locates the vise properly). I use a test indicator.

The gauge I used and photographed is for tramming only. You can use it to tram a CNC router, a mill, or whatever. Its only purpose is to get the spindle precisely perpendicular to the bed.

Gary

[GME]

Not sure about using the 1 2 3 blocks like that. Even with blocks the "wring" you might be asking for trouble. Though you should use what you have to the best of your ability. Good work so far.

I understand the term 'wringing' as it applies to 123 blocks and parallels. I've employed it. Here's a picture demonstrating the phenomenon, for those who may be unfamiliar.



Okay, I know what it is, but I don't understand your comments, "[n]ot sure about using 1 2 3 blocks like that," and "you might be asking for trouble."

Please explain.

Thank you,

Gary

[the4thseal]

When stacking block that do not wring, there is not an insignificant error that can be induced. My mentioning 1-2-3 blocks was that they do not wring, as well as not being "accurate" enough to be used as gauge blocks like this. In general.. I completely understand using to the maximum what is available to you. My concern is that you may be assuming that by adding the lengths together you will have an overall accurate dimension. This is difficult with gauge blocks that wring. So my comment is that you have 2 things against you with this set up.

[wmgeorge]

I would think stacking blocks end to end would introduce some error, vs using a one piece ruler. But where do you get an accurate ruler, accurate to .0001 over 3 feet? Its going to be interesting to see what Gary comes up with. Really when your trying to get a wood router to be this accurate, and I get the steps each as .00049 of an inch, is that correct and your off .7578 of one step?

[GME]

When stacking block that do not wring, there is not an insignificant error that can be induced. My mentioning 1-2-3 blocks was that they do not wring, as well as not being "accurate" enough to be used as gauge blocks like this. In general.. I completely understand using to the maximum what is available to you. My concern is that you may be assuming that by adding the lengths together you will have an overall accurate dimension. This is difficult with gauge blocks that wring. So my comment is that you have 2 things against you with this set up.

Thank you for sharing your concerns. I agree that 123 blocks lined up are probably not 100% accurate, especially given that I am not using the $100+/pair high precision blocks. And no, I did not make unrealistic assumptions about accuracy. I knew going in that I was not going to get perfection, but believed that using the blocks was likely more accurate than using a steel rule and some sort of pointy object mounted in the spindle, or other alternatives. Wringing was not in the cards, even if the blocks I used could be wrung. There were to many of them. In any event, the method I used was certainly more accurate than using a tape measure, which I seen others use. FWIW, I have an I-Gaging 36" steel straight edge with rule, but elected to use the 123 blocks. I thought I would get more accuracy with the blocks. With a rule, assuming it is very precise, you have to get both zero and the end length dead on. I have no confidence that I could do it well, even with a magnifying glass. I am not capable of making visual judgments in thousands of an inch.

I could have also used a dial indicator and measured over an inch distance. It thought about and rejected it. One problem with the method is the accuracy of the indicator. They are commonly rated at +/- .001", which can introduce a good deal of error over longer distances. Certificates on some of the expensive ones show an error in the .001" range at 2 dial revolutions. At 36", you could be off as much as .036". A huge error. It gets even more complicated, because the indicator's spring pressure can and will move the holding arm by around .0002" at 1/2 inch of movement - even for good brands of arms, like Noga. Of course, as spring tension increases. as would be the case with a 1" measurement, so will the holding arm movement. Assuming .0002" of movement, you would be off .0072" at 36", exclusive of any error in the dial indicator itself.

As I believe is evident, I gave all this a lot of thought. At the end of the day, I had to pick what I hoped was the the best of imperfect choices.

Having identified potential inaccuracies, which I truly appreciate, do you have any suggestions about how I might do it more accurately? As much thought as I gave to the process, I could have overlooked something obvious, or not so obvious. If you have a CNC, how did you do it?

Thank you,

Gary

[GME]

I would think stacking blocks end to end would introduce some error, vs using a one piece ruler. But where do you get an accurate ruler, accurate to .0001 over 3 feet? Its going to be interesting to see what Gary comes up with. Really when your trying to get a wood router to be this accurate, and I get the steps each as .00049 of an inch, is that correct and your off .7578 of one step?

Good points, Bill. FWIW, the math on the R&P works out to a theoretical 2037.18 microsteps per inch. For X, my measurement technique came out to 2037.5132 steps per. For Y, it came out to 2038.2710 steps per. So, X was off of theoretical by only .3332 microsteps or about .00016". Y was off of theoretical by 1.091? microsteps, or about .00055". Pretty darned close to theoretical in my never-to-be-humble opinion.

I'm guessing that there was some error canceling going on with the 123 blocks. I doubt it's reasonable to assume that any error is going to be all plus or all minus. So, some may have been long, others short. The steps per I got measuring over 36" are close enough to theoretical to support some amount of error canceling.

I also agree with potential inaccuracies in any rule one might use. No way to judge what you have with any certainty. Mine didn't come with any certifications, so error tolerance is an open question. As I mentioned in my reply to the4thseal, the bigger problem for me is the practical use of the rule to measure. You have to use some sort of pointy thing mounted in the spindle, and have to perfectly align the the point at the beginning of measurement and precisely read it at the end. My rule has 1/64" graduations (.015625?" separations), which is pretty darned small. Even assuming the rule is 100% accurate, how does one make accurate visual determinations in thousands of an inch? I sure couldn't distinguish between .015625" and .013625", or much greater differences, by eye - even with a magnifying glass. What about the effects of parallax? I have zero confidence I would have a better result using a rule.

I'm hopeful that the4thseal has an alternative technique or two in his bag of tricks. I don't have another alternative in mine.

Gary

[wmgeorge]

I am getting a different set of numbers than you. For Y, it came out to 2038.2710 steps per When I divide 1 by 2038.2710 I come up with .000490612 which in round numbers is .0005 or 5 / 10,000 of an inch?

Perhaps I am thinking wrong or my calculator is off? I think your trying to obtain clean room accuracy. With my Mitutoyo which reads to 10,000 a sheet of paper measures .0035.

Added, I went out and measured my 1,2,3 blocks one was 3.0015 the other 3.002 but these are not expensive $50 each ones. So my guess if you measured yours you will find the same, the accumulated error over 10 blocks could be noticeable.

[GME]

I am getting a different set of numbers than you. For Y, it came out to 2038.2710 steps per When I divide 1 by 2038.2710 I come up with .000490612 which in round numbers is .0005 or 5 / 10,000 of an inch?

Perhaps I am thinking wrong or my calculator is off? I think your trying to obtain clean room accuracy. With my Mitutoyo which reads to 10,000 a sheet of paper measures .0035 or .004.

i just quickly punched the numbers in my calculator. I could have gotten a number wrong when I punched them in.

I want to get the best accuracy I can get going in, knowing that an extrusion-based machine will probably not cut to that level in practice. In my mind, the better I can set it up, the better the ultimate outcomes, which won't be perfect. It take little more time to shoot for perfection. It's worth it to me to maximize practical accuracy. I know I won't end up with anything like "clean room accuracy" in day-to-day practice. Nevertheless, I will end up with better result than if I took less care in the setup. I think of it this way. Garbage in = garbage out. However, perfection in = better results out. I suspect there must be some point where diminishing returns make further efforts are effectively meaningless. Unfortunately, I have no way of assessing where that point might be.

No question that Mitutoyo makes quality instruments. However, high resolution (1/10,000) does not necessarily imply high accuracy. I have a couple of cheaper indicators with 1/10,000 resolution, but I wouldn't trust them to that degree of accuracy. I do what I believe most do. In our back brains, we know there is some minor degree of error, but to keep our sanity, we treat them like they are 100%. To do otherwise is to leave us perpetually in doubt and ungrounded.

Gary

[wmgeorge]

This what I use https://www.amazon.com/Mitutoyo-500-.../dp/B01LTCRASC and its the Made in Japan one, not the Chinese knock off.

[GME]

This what I use https://www.amazon.com/Mitutoyo-500-.../dp/B01LTCRASC and its the Made in Japan one, not the Chinese knock off.

Makes my point. Even a great set of $235 calipers are +/- .001".

[wmgeorge]

Makes my point. Even a great set of $235 calipers are +/- .001".

Thats worse case, the calibration Certificate that came with it, was more detailed and concise. Makes you wonder what the cheaper ones are at?