Servo Spindle motor encoder Mesa 7i76e

Hi all,

After making good parts with my emco 220, i was able to finally try threading.

Unfortunately the thread (or should be) looks rubbish.
Did some searching and i think i setup some things wrong. Every start of new cut is shifted some how, so its off the starting point.

encoder scale in the .ini file is -1024, same as in the motor drive. could be increased if its better.

In hal scope:
Phase A works
Phase B works
Phase Z works, but have a hard time to find its spot by rotating the spindle by hand, but i get a pulse though.

Can some one walk me through it to find the problem?

Are you sure the spindle encoder scale is correct?

An easy way to tell is to mark the spindle and hand turn
it say 10-20 turns, the spindle encoder position pin should
advance by exactly 1 for each spindle turn

it is always better to stay positiv in all main setups and turn a wire to get direction change
as the New interpreter is in someway getting troubled / confused on calculation
My personel expirience on the Motor setups

PCW wrote: Are you sure the spindle encoder scale is correct?

An easy way to tell is to mark the spindle and hand turn
it say 10-20 turns, the spindle encoder position pin should
advance by exactly 1 for each spindle turn

so i checked hm2_7i76e.0.encoder.00.position in halshow. Marked the spindel, did one rotation, value was 0.9560***.

so, its off. if i continue o get the value to 1 its like 10mm between the marks on the spindle.

That's an odd number, sounds like its miss-counting
is the number consistent? That is, do you get back to the mark
if you rotate backwards to 0?

If the count is not consistent I would suspect:

1. Electrical problems: Encoder input mode on 7I76 does not match encoder signals/wiring, wiring issues, bad levels
2. EMI issues (electrical noise from drives etc) possible fixable by adjusting the encoder filtering
3. Bad or dirty encoder wheel/optics
4. Slipping mechanics

PCW wrote: That's an odd number, sounds like its miss-counting
is the number consistent? That is, do you get back to the mark
if you rotate backwards to 0?

If the count is not consistent I would suspect:

1. Electrical problems: Encoder input mode on 7I76 does not match encoder signals/wiring, wiring issues, bad levels
2. EMI issues (electrical noise from drives etc) possible fixable by adjusting the encoder filtering
3. Bad or dirty encoder wheel/optics
4. Slipping mechanics

No when i rotate back, its sets to 0 just fine.
i also checked the .00.counts. one full rotation has a value of 975.
I set this in the ini file.
00.position value is now 1 at full rotation.

now 00.rawcounts value should be the same as 00.counts right? value is 1971.

for the information. W10 W11 and W13 are in differential mode. should these be in single?

975 is a very odd number so I suspect something is still wrong

The 7I76E encoder jumpering must match the electrical interface on the encoder
Differential encoder interfaces will have 6 signals instead of 3:

Differential:

A, /A B, /B Z, /Z

TTL/single ended:

A,B,Z

PCW wrote: 975 is a very odd number so I suspect something is still wrong

The 7I76E encoder jumpering must match the electrical interface on the encoder
Differential encoder interfaces will have 6 signals instead of 3:

Differential:

A, /A B, /B Z, /Z

TTL/single ended:

A,B,Z

oke, well that confirms that the jumpers are right.

i set the encoder scale in my .ini file back to 1024. in program run its wait for spindle reach.

this is really weird what cause this.

To check if its a bad encoder, you could rotate at a constant speed and plot the
encoder velocity with HALScope. A bad /dirty encoder wheel will show up as a repeating dip in velocity

PCW wrote: To check if its a bad encoder, you could rotate at a constant speed and plot the
encoder velocity with HALScope. A bad /dirty encoder wheel will show up as a repeating dip in velocity

Found it!

Nothing in linuxcnc or drivers.

There is a difference in the poly v belt wheels. A deviation of 5%
I get the same offset in the drive. So i was questioning my self why also the drive counts where off.

Now, I might add a encoder on the spindle its self (as it was original but one puls).
I would like to have also a quadrature encoder on it, so i can also ridgit tap.

Or, i replace the v belt wheels to timing belt. Might be more work, but then i can use the servo motor as i intended.

Any suggestions?