[CNC-Dude]

It is kind of useless to do 0.0001 to measure backlash because most of the time the steppers doesn't even move. This is a Series II machine, so I wonder why isn't 0.0005" supported?

While enabling the network this morning, I noticed the MDI packs some secret messages. Of course, chances are these won't be divulged as this way we can't tinker with PP. Or have these instructions been made public somewhere? Maybe there is one to configure the step rate increments?

[mountaindew]

Select 0.0010 and select 50% jog
I don't have pp or use it but it looks like fun
md

[kstrauss]

I noticed the MDI packs some secret messages. Of course, chances are these won't be divulged as this way we can't tinker with PP. Or have these instructions been made public somewhere?

Care to elaborate?

[CNC-Dude]

Care to elaborate?

Well, when I used Path Pilot for the first times, I went through all the possible buttons and panes only to find I had access to Path Pilot stuff (as expected). There was no access to any other kind of configuration utility or Unix shell. There must be some, though, as PP runs on Linux, except that the programmers went through great extent to hide them away (as they should have!).

To configure the Network, however, you get to see one of these configuration windows if you type in the MDI "ADMIN NETWORK". It opens up the NETWORK configuration pane. In the same fashion, there must be a bunch of other access windows and this must be how they are accessed. "ADMIN THIS", "ADMIN THAT", etc.

I understand that Tormach doesn't want us tinkering with these access points and with all reason! If we start messing up PP, then they will get way much more calls than they can efficiently support. But I think it is a matter of time before this information becomes available. Maybe much after the platform is released officially.

[MichaelHenry]

It is kind of useless to do 0.0001 to measure backlash because most of the time the steppers doesn't even move. This is a Series II machine, so I wonder why isn't 0.0005" supported?

I'm not sure I understand this - my Series I mill moves in increments of 0.0001 inch if that step size is selected in Mach3. Or at least it tries to. Due to what I assume is slip-stick, it will sometimes not move for one 0.0001" increment and then move about 0.0002" on the next, or 0.0003" on the third. But the steppers always seem to be rotating.

Mike

[CNC-Dude]

I'm not sure I understand this - my Series I mill moves in increments of 0.0001 inch if that step size is selected in Mach3. Or at least it tries to. Due to what I assume is slip-stick, it will sometimes not move for one 0.0001" increment and then move about 0.0002" on the next, or 0.0003" on the third. But the steppers always seem to be rotating.

Mike

I think that 0.0001" is just an approximation. Just because the stepper is moving (or seems to be moving) does not mean it is moving what we think it is moving. Tormach has expressed that Series III machines (or any machine with the SERIES III upgrade) will have better resolution than the SERIES II counterparts. Considering the following, it can only make sense to me.

SERIES III basically means polyphase motors which is nothing other than a 3 phase stepper. Bipolar steppers have two phases and require 4 commutation steps to do 1 electrical revolution. Because they are built with 50 pole pairs, you get 200 steps per revolution which means a single step will displace the shaft 1.8 degrees radially. We add microstepping into the mix, and now we are dividing that step distance further (i.e. microsteps).

3 Phase steppers are made with 3 phases and require 6 commutation states per electrical revolution. With the same 50 pole pairs, you get 300 steps per revolution or 1.2 degrees per step.

Of course, the 0.0001" linear motion is a factor of how many degrees of microstepping are being employed and what is the ballscrew turn ratio. The Tormach ball screws are 5 TPI. SERIES 2 motion will give you 0.001" linear motion per full step (1"/5TPI * 1T/200STEPS), whereas SERIES III will give you 0.000666" linear motion per full step (Sounds micro-demonic? heh heh heh). So now you microstep and you get:

Microstepping SERIES II SERIES III
HALF STEP: 0.0005" 0.000333"
QUAD STEP: 0.00025" 0.000166"
EIGHT STEP: 0.000125" 0.000083"
1/16 MS: 0.0000625 0.000042"

And so on. Then add to that microstepping error and I won't blame anybody wanting to jump from a 10th floor. Notice that regardless, there is no way to move 0.0001", but Series III can get way closer. Also, 0.0005" is a natural distance for Series II machines (i.e. requires way less approximation than 0.0001").

BTW, I talked about SERIES II machines, but I believe the same applies to SERIES I machines.

[popspipes]

Well, when I used Path Pilot for the first times, I went through all the possible buttons and panes only to find I had access to Path Pilot stuff (as expected). There was no access to any other kind of configuration utility or Unix shell. There must be some, though, as PP runs on Linux, except that the programmers went through great extent to hide them away (as they should have!).

To configure the Network, however, you get to see one of these configuration windows if you type in the MDI "ADMIN NETWORK". It opens up the NETWORK configuration pane. In the same fashion, there must be a bunch of other access windows and this must be how they are accessed. "ADMIN THIS", "ADMIN THAT", etc.

I understand that Tormach doesn't want us tinkering with these access points and with all reason! If we start messing up PP, then they will get way much more calls than they can efficiently support. But I think it is a matter of time before this information becomes available. Maybe much after the platform is released officially.

Typing in just "admin" brings up a list on the main window, when status is clicked.

[MichaelHenry]

CNC-Dude - that all makes sense but runs counter to my memory of what I saw. I'll check again and report back my results are different from your calculations.

Mike

[pickled]

Why use the jog shuttle thing at all for attempting micro-steps? Just type in a coordinate at whatever distance you would like to travel (even at a specified dynamic feed rate) and then return to your origin. You will see your backlash. Heck you can even run a loop in groups of 25 movements to see if the backlash grows via a linear fashion vs. some odd quadratic function. I'm still trying to figure out everyone's recent laser focus on micro-stepping...maybe I am missing out on something? I am looking forward to getting some time to play with Path Pilot soon. I have to wait for the card to come in and for production needs to slow down first. It looks very nice.

[CNC-Dude]

Pickled, I don't think you can measure Backlash like that very well. You can measure lost steps, that's for certain. If you just move back and forth, the system will pretty much return to where it was at the beginning because you have the same backlash on both directions and it will cancel out.

The only way in which I can see the backlash is by moving one direction until the screw has engaged the table and right there reverse the table. Whatever steps are issued but cause no motion on the table, are backlash. When using 0.0005" distance moves, I get 1 lost step for motion. When using 0.0001", nothing makes sense. The table moves randomly and I pretty much can take a number from a hat and have a better chance at measuring backlash than with the 0.0001" steps and the aforementioned method.

One word of caution. When using this method to measure backlash, where you are on the stepper angular position may give different results. This is because of microstepping resolution and current regulation errors. For example, throughout most of the angular position on a per electrical revolution (4 full step sequence), I can see 1 missed 0.0005" step. On certain positions, however, I can see 2 steps missed. I think this is because not all microsteps are equal. I would need to go into a detailed microstepping science explanation but I would like to avoid most forms of cursing where my name might end up being interlaced within ;-)

[TurboStep]

...Notice that regardless, there is no way to move 0.0001"...

If I understand you correctly, you may want to read this:
http://users.ece.utexas.edu/~valvano...rMicrostep.pdf
Step

[CNC-Dude]

If I understand you correctly, you may want to read this:
http://users.ece.utexas.edu/~valvano...rMicrostep.pdf
Step

That's a very nice paper! Thanks for sharing!