Spindle and encoder

So it looks like the speed controller is quite non-linear
(and needs commanded speed of about 140 RPM to start turning)

So you can either live with it and set the RPM to values
that work for what you are doing, or attempt to fix it with say the offset
component or better but more complex to setup, the lincurve component.
 

What's the offset or lincurve component?  I might be willing to give it a try.

Take a look at the manual pages:

man offset

man lincurve

I referred to the LinuxCNC_Manual_Pages.pdf and it appears the lincurve is a function. The description looks minimal. Is it a function to optionally enable or implement? Is some low-level and high-level programming required? Is a specific description of the term 'offset'? Because the word is used enormously in the manual.

Doe this forum.linuxcnc.org/10-advanced-configura...m-output-for-spindle involve the process to setup lincurve?  It makes some sense but lacks information.  I may require some instructions.

I probably have a lot more than four points to compensate for the spindle. Does lincurve operate with variable speed (g96)?

First, very accurate spindle speed is not required in most cases
so you might just try what you have now unless you need very low
speeds.

G96 is not directly affected by the commanded spindle speed
since like most spindle synchronized moves, G96 uses the actual spindle
speed (not the commanded one). This means the feed per rev value will
be correct regardless of errors in the commanded spindle speed.

For the offset manual, just use the man page:

man offset

That's a good point about g96. However, the g96 is working and acceleration is smooth and constant but might not be calculating properly with the spindle visa versa. t1 x5.265 home and g96 s300 m3 and tachometer 67 rpm but should be 217 rpm. t1 x2.00 and tachometer 353 rpm but should be 573 rpm (300sf * 12 / 3.14 / 2.0 = 573 rpm). t1 at a 2.0 diameter is 220 rpm too slow. Also, the digital readout is jogging radius mode. Is toggle for diameter mode?

I would of expected the commanded spindle to match the rpm because the Tormach 8L probably uses a brush motor.  I operated the 8L for awhile and checked the pulley with a tachometer and within a few rpm.  PathPilot II had a electric low and high gear.  I had an issue in MDI mode with low and high gear accelerating beyond the max rpm and might of had something to do with toggling low and high gear and inputting more spindle speeds but aside that the rpm was accurate.  Could my configuration require a low and high gear (electric tier)?

I had forgotten that there is a bias pin on the PID component so you
should be able to partially correct the drive offset with that:

[SPINDLE_0]
P = 0.0
I = 0.0
D = 0.0
FF0 = 1.0
FF1 = 0.0
FF2 = 0.0
BIAS = 140 <<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
DEADBAND = 0.0
MAX_OUTPUT = 2800.0
ENCODER_SCALE = 4000.0
OUTPUT_SCALE = 2800
OUTPUT_MIN_LIMIT = -2800
OUTPUT_MAX_LIMIT = 2800

 

Maybe the speed controller accelerator dial should be maximum potential that the commanded spindle is less than the tachometer rpm. The dial is only slight 1/64 from maximum. Rather set the max spindle rpm with LinuxCNC. The ini file altered the BIAS to 140 before launching LinuxCNC but the spindle still a lot slower than the tach. For instance s500 about 400 rpm and s300 about 250 rpm. I'll have to do some research about calculating the BIAS. How did the 140 become a constant? Should I experiment with some BIAS amounts?