Faceting machine head assembly.

[Faceting Frank]

Thanks for the complement RC.
 I now understand the logic behind your average reading of the encoder RC, there are to many variables going on with the lap. Have your code on averaging put where I think it should be, but still // out. Not going to bother figuring out the code needed to reset the 90deg with a push button if I make mods to the machine. Took only two minutes to change the code on the PC and download it. Thinking of having a light flash for a second when the encoder hits the target angle, that would save staring at the screen. Have lights already wired and coded in, so should be easy to do.

[MakkyBrown]

Well done Frank 
I'm not running any smoothing on mine, I like the flicker just before the angle going solid and use it as the point of a solid angle as the stop.

[Faceting Frank]

Got the light to come on at the target angle, if this turns out to be very precise I wont bother with the averaging thing.

float anglestop;
.
  anglestop = (90+ (1.0*encPos-enc90)*degPerStep);    // The actual angle in degrees
.
if (anglestop <= eStoredAngle) {
 digitalWrite(RED_pin,HIGH);
  delay(500);}

eStoredAngle being the variable name for the target angle.

[RoughCreations]

Hi Frank,.
the code will add a half-second delay to every loop, this will be like having a slower micro-controller in effect, and probably lessen the need for smoothing somewhat

[Faceting Frank]

Hi RC.
If your talking about the 1/2 second delay in the bit of code I have posted it will not effect the loop. It will only read the line
 {digitalWrite(RED_pin,HIGH); delay(500);}  if (anglestop <= eStoredAngle) is true. Otherwise it will ignore anything in the {}
Not that I know much about this stuff, but the YouTube vids I've watched over the last 6 months on how different functions work has helped me a lot.

The encoder has picked up a fault with the round linier bearing that I suspected was a problem, when moving the quill from side to side the bearing wants to climb up the shaft about 0.1mm one way and then drop 0.1mm.

[RoughCreations]

That's true, fair enough.
I guess in the situation I find with my cutting and laps in practice is that when you get close to the eStoredAngle, the unevenness of the laps usually causes the anglestop to be greater than eStoredAngle for a brief instant, even when you may not be truly at the 'defintion' of when eStoredAngle is actually reached. A 500mS delay would be triggered quite often when you were close to eStoredAngle. That was where I thought the smoothing algorithm should come in, it defines when you are actually at eStoredAngle.

RC

[Faceting Frank]

That’s right about the lap flutter or any variation in the side to side sweep setting the light off RC and your averaging code will help with that.

 Got a stone to finish off and then will fix the round linier bearing problem. Any inconsistency with that is going to stuff the encoder accuracy.

[MakkyBrown]

Frank I was looking at the pic of the encoder on your machine. Make sure you retain all the wires somehow so there is no movement on the wires where they plug into the encoder. Just having them pluged in and hanging caused me issues and almost wrecked my encoder housing as it all became wobbly. Look at the pics of the cable ties on RC machine, mine is worse. You might be able to make some nice saddles to retain the wires.

Also worth checking to see if you get EM interference from your machines motor, I have ferrite magnets on the 240 power lead and smaller clip on ones on the encoder cable. Look for random flickering or artifacts on your screen at high rpm.
 
MB

[Faceting Frank]

Now have a permanent setup as far as the cable goes MB, also a guard over the encoder. Not that there is anything it can bump into, but I have a habit of holding the quill support at that point.
Have not detected any interference on the screen, so all seems good that way. Will take the delay function out on the light, only need to have (RED_pin,HIGH) if the angle has passed the stored angle.
Should have the round linier bearing issue fixed shortly, but still being able to lift the quill assembly up to eye level to look at girdle meet points.

[Faceting Frank]

Back on track with the faceting machine and cutting again. Not going to bother with the averaging code, if the lap speed is kept the same then any variations in the lap is the same for all facets in that row. Also keeping the 0.5 second delay with the light, the light coming on at the target angle is a signal to stop cutting. So, it does not affect the loop speed.

[Alaskan]

FFrank, is this 'flutter' you and we are all experiencing - (me on the dial indicator) - coming from the machine or the lap?
My current cutting is going slow - new spectacles were ground wrong or the material is faulty - segmented and blurry.
Went to the competition who arrive every other week for a 2nd opinion - new prescription on order and this guy dialed me in - awesome.
Just finished the Pavilion on AB's Supernova in yellow zircon - epoxy cured by now - saw and separate tomorrow for crown after 48 hours.
The green sapphire that finished at .85 cwt was encouraging ~ 25% nice clue as to expected loss.
Where did that go?

[Faceting Frank]

The "flutter" on the dial gauge is from the laps. My reply 137 on this thread goes into some detail on how I minimize it. The finest of your diamond cutting laps is the most important one, it will be the lap you are getting all your meet points finished.

[Alaskan]

Corian laps for my final polish of 3 - 8 - 14 then 50k with Yamaha saxophone oil - on smaller facets start with 3000/oil on 8" Corian. Using a tried and true 600 then 1200 or 1500 for final preforms.
Thanks, Frank.