ATC Project - Debug phase

There won't be a cumulative error, no. It is just a "window" that means that it doesn't have to see an exact count to register a new position.

Try 4 first, see if that fixes 50 speed units.

I am not sure what your step rate is at 50 units, but if you imagine that it only checks the step count every 1mS, if you have more than 1000 steps / second it is going to start missing a pocket sometimes if there is no leeway allowed.

OMG - progress.

I set width to 5 and it's stable up to speed = 150.  Also had to set the stepgen maxaccel = 0 & maxvel = 0; with those at 'soft' settings it would coast past a pocket (or 3) regardless of whether carousel was getting the counts right.

Since I cant really figure out what the speed units are, I used the stepper tuning software to monitor the RPM via rs232.

150 units  = ~100 motor RPM, 20 carousel RPM (5:1 gearbox).  So with a 4500ppr step count that's a pulse rate of about 7.5khz.

It was actually stable up to 200 units, which was proper fast.  Like Robodrill tool changes.

Thanks a ton Andy.

Now that hurdle is behind me, I'm happy to do more testing if you want to fiddle with the homing or pin sense functions.

PS - while I'm asking for favors, an additional nice-to-have would be the ability to unhome the carousel without having to shut down LCNC and restart it.

spumco wrote:

Since I cant really figure out what the speed units are, I used the stepper tuning software to monitor the RPM via rs232.
150 units  = ~100 motor RPM, 20 carousel RPM (5:1 gearbox).  So with a 4500ppr step count that's a pulse rate of about 7.5khz

The speed units are things-per-second, but that rather depends on how you have scaled the step generator.

If your step scale is set  to give degrees of carousel motion then it would be degrees per second. If it was set to one motor revolution then it would be motor revs per second. 

Probably

That would meat it's it degrees/second since I've got the stepgen scale set to 50.  Or at least it was until I changed the microstepping in the drive to see if different values would change the rotational issue sorted out yesterday.  This passes the sanity test since 150 degrees per second looks about like 20rpm to the untrained eye.

Next moves:

• Remove unused pocket pins & make a better index flag
• Remap M6
• Test sequencing and actually change a tool
• Connect various other remaps to Probe Basic GUI buttons
• ????
• Profit

spumco wrote:

• Remove unused pocket pins & make a better index flag

Hold off on that. I think that adding a latching move can make the work well. In principle I like the pins.

andypugh wrote:

spumco wrote:

• Remove unused pocket pins & make a better index flag
  
  
  
  Hold off on that. I think that adding a latching move can make the work well. In principle I like the pins.

Holding, Aye.

Easy enough to reconnect sense-01 back in hal; nothing was actually removed from the machine.