Updated review: SuperBar by MasterGage, Inc.

[Ed Bennett]

Hi John,

That's funny, I remember it being an error factor, not an error quotient. I don't have my books here at the shop but I found this on the web:

seems to think it's a factor too. Hmmm...

So, I rigged up a 0.0001/div indicator to see what happens. I tilted it down 30 degrees relative to the table. This makes it 30 degrees off axis with a line perpendicular to the surface of the blade. I used a cone indicator point with an 80 degree included angle to avoid interference from edges. Then I used two gage blocks (0.1000 and

0.1050) against a 2,3,4 block to check the difference in reading. If there were no error, then you would expect the indicator to show a 0.005" change. Instead, it showed a 0.0058" change. The error did make the reading larger, just like your formula would predict. It's off by a little but then I didn't do anything to ensure alignment in the other direction).

The emperical data does support your formula. And, as I think about the geometry, it seems to make sense. Hmmmm.....

Ed Bennett [e-mail]

of the TS-Aligner

[Ed Bennett]

Hi John,

Are you sure? I don't have my books here at the shop but I found this on the web:

guy is saying it's an error factor, not a quotient.

You've got my curiosity going. I jiged up an indicator to see what happens and the reading was higher than expected. For a 0.005" change, the reading is 0.0059". That's pretty close to the calculated

0.005773, considering I didn't do anything to align the indicator in the other direction.

I tilted the indicator down so that it formed a 30 degree angle with the table surface. This would be the same as being 30 degrees off axis with a line perpendicular to the blade. Then I measured two gauge blocks (0.1000 and 0.1050) proped vertically against a 2,3,4 block (standing in for the blade). I used a cone shaped indicator point (80 deg included angle) to ensure that no edges were going to interfere with the measurement.

I'll check the books when I get home tonight. The emperical data does support your formula.

Ed Bennett [e-mail]

of the TS-Aligner

[John Cochran]

If you were off as little as 2 degrees in the other direction, the error angle would be close enough to 32 degrees to account for your reading. And frankly, 2 degrees is pretty darn close considering that you most likely only have about an inch or two distance.

I came up with my formula by looking at the geometry of the situation. Just looking at a right triangle, what you're attempting to measure is the change in length of side B. However, what you're actually measuring is the change in length of hypotenuse C. For a tilt angle of 0 degrees, you run into the happy situation of both lengths being the same. However, for any other tilt angle, the hypotenuse will always increase in length faster than the side.

B a .............................. ...... c. ...... . A ...... . C ...... . ... b

Now I agree that you always want to have a tilt angle as close to zero as possible. Not only does it give you a better absolute reading on the error, but I suspect that many dial gauges don't like having lateral loads on their spindles and in fact may have a fair amount of "wiggle" in their spindles when a lateral load is applied to them which would tend to make determining the exact tilt angle rather hard to calculate, predict, or measure.

[Ed Bennett]

Yep. It's not an elaborate setup (mag base, 2,3,4 block, calibrated eyeball, etc.). The result is close enough to say that it verifies your formula.

Good analysis. It's what I was thinking too. For the most part, this is the sort of situation that you read about but never actually experience. Most people don't deliberately tilt the dial indicator off axis with the object being measured. In all my years I've never had the occasion to do it. There's always some way to resolve the issue so that the readings represent the actual geometry.

Yep, you're absolutely right. Using the Offset Bar on the TS-Aligner products solves the reading problem but it does not avoid the lateral forces on the plunger. Some indicators have more "wiggle" than others but for the most part this isn't an issue. There is also increased friction in the movement which reduces sensitivity, but with sufficient pre-load this isn't an issue either.

Some might think that this setup magnifies the reading, making it easier to eliminate small alignment error. Well, not only is the error magnified, but any instability is too. And, blade alignment isn't the only procedure which benefits from having the stylus close to the table surface. Using a square or angle blocks to set the miter gauge benefits greatly as well.

In order of overall perference, I would have to say that using a blade replacement plate is the least desireable solution to the problem. You are far better off doing the alignment over the shorter distance, tilting the indicator, or (my favorite) using an Offset Bar. Marking a spot on the blade and rotating the blade so that all your measurements are with the stylus tip on that spot will always be more accurate, much more convenient (no need to swap out the blade), and much less expensive (just get a $0.49 sharpie).

Ed Bennett [e-mail]

of the TS-Aligner