Angular readout decimal places

[MidniteMachinist]

Would it be possible to add more decimal places to the angular readout? The linear readouts go to 5 decimal places (in inches, 4 places in metric), but the angular only goes to 2, or 0.01 degrees (36 arc-seconds)

With linear resolution to ten millionths of an inch, I don't understand why angular readout is limited to half a minute of arc. For reference, the analog vernier scale on my rotary table has 10-second divisions, and one could probably interpolate 5 seconds.

If, for example, I wanted to cut a 127-tooth gear for my lathe (to cut metric threads with an imperial leadscrew), each tooth is 2.8346 degrees apart. 2 decimal places of resolution isn't enough to cut an accurate gear.

I've got a rotary encoder on my rotary table that has a (theoretical) resolution of 1 arc-second, which is 0.000277 (repeating) degrees. I realize that I'm probably an edge case with the kind of resolution, but I'm sure I'm not the only one who wants to use TouchDRO with a rotary table... And working to less than half a minute of arc is not that uncommon.

 

[benmychree]

Dividing for 127 teeth is best done by differential indexing, no accumulated error.

 

[MidniteMachinist]

Dividing for 127 teeth is best done by differential indexing, no accumulated error.

Yes, if one has a dividing head capable of differential indexing. I do not. Nor, I suspect, do most hobby machinists.

One might not need to divide for gear teeth. A straight cut that's tangent to 2 circles, or something like that. Or, perhaps, someone has a rotary encoder on their lathe compound, and wants to cut a Morse taper.

There's any number of things that require very specific angles, that have fractions smaller than half a minute of arc. Yes, there are may methods of setting angles and dividing circles that don't involve a DRO... One can also machine things using handwheel dials and accounting for backlash, touch off tools every time, and use an Indicol with a DTI to find the center of a part. That's not the point.

I'm just hoping that, if it's not too much of a pain to add, we could have more decimal places in the angular readout.

 

[ycroosh]

Yes, if one has a dividing head capable of differential indexing. I do not. Nor, I suspect, do most hobby machinists.

One might not need to divide for gear teeth. A straight cut that's tangent to 2 circles, or something like that. Or, perhaps, someone has a rotary encoder on their lathe compound, and wants to cut a Morse taper.

There's any number of things that require very specific angles, that have fractions smaller than half a minute of arc. Yes, there are may methods of setting angles and dividing circles that don't involve a DRO... One can also machine things using handwheel dials and accounting for backlash, touch off tools every time, and use an Indicol with a DTI to find the center of a part. That's not the point.

I'm just hoping that, if it's not too much of a pain to add, we could have more decimal places in the angular readout.

It's not that hard to add more decimal places. That said, I try to not add features to the app just to have them, and to be honest, I can't think of a practical scenario where 0.01 of a degree on a rotary table is not already an overkill.

Here are a few things to consider:
1. Regarding "my rotary table that has a (theoretical) resolution of 1 arc-second, which is 0.000277" - unless you are using a direct-reading rotary encoder with 1,296,000 steps per revolution, you probably don't have that resolution (or anywhere close to that).
If your encoder is measuring the revolutions of the worm screw, the variability in the oil film thickness between the worm gear and the screw will throw you off by multiple arc seconds. It can ranger from under a micron to several microns, depending on temperature, cranking speed and load. If your worm gear is 4" in diameter, one arc second will equal 0.24 um (0.000009").

2. One arc minute (not second) of angle will equal 0.00145" at the circumference of a 10" circle. One hundredth of a degree will be 0.00087". You'd need to be using a jig borer/jig grinder that is sitting of 2ft concrete slab in climate controlled shop for that amount of error to matter (your cutter likely deflected by more than that while cutting the gear).

The reason I added 5 digits for linear scales was because I was testing my laser interferometer, and forgot to comment that code out before the release. The interferomenter I own has theoretical native resolution of 1/4 wave length of HeNe laser. This is about 158 nm, or 0.00006". Fun fact - my 20 lb miniature Schnauzer stumping around the garage (with concrete floor) can make the last digits twitch. The whole setup sits on a 300LB granite plate, which in turn, sits on a set of vibration absorbing pads. If I run it remotely, and the temperature in the garage doesn't fluctuate too much, I can get decent measurement if I get lucky.

I realize that the following statement can start a holy war in a hurry, but 99.9% of us can't even reliably (and repeatably) measure to 0.0001", let alone hold those tolerances on metal cutting machines. It takes a LOT of effort to get to +/- 0.0005" on a surface grinder, unless I give everything time to stabilize, run flood coolant, the wheel is dressed and well balanced, and I'm feeding the part with consistent speed.

Regards
Yuriy

 

[RJSakowski]

The .01º resolution is equal or better than the ability of most of us to machine. The rounding error will be at worst 1/2 the resolution. However, if I were cutting a gear or similar, I would calculate absolute position of each tooth rather than using an incremental advance to the next position to avoid error stacking.

 

[MidniteMachinist]

Yuriy: Your second point is the perspective I needed... I wasn't seeing the forest, for the trees. I was hung up on the angle, and completely blind to what that means for the actual positional error. Focusing on the hypotenuse and base leg, and ignoring the height of the 3rd leg.

Thanks for helping me see. Apologies for harping on it.

(And yes, on paper, I have 1,296,000ppr... Also yes, the encoder is on the worm shaft. Further yes, I know this isn't ideal, and subject to myriad factors. Hence the "theoretical resolution" lol)