7i76e and opto ancoder TCST2103

PCW wrote: The encoder signals will not look good because they are not representative
of the actual signals (A and B are 15 KHz signals at 15 RPS so you
will just get aliased signals when you sample them at 1 KHz)

You would need a real oscilloscope to see the encoder signal at this speed

at 15 RPS you should get a quite accurate RPM reading

Unfortunatly i no have real oscilloscope.

Yes you are right with NEMA encoder at 15 RPS display is very accurate !

Can i ask what is for you the best option ?

Or is there any other opto sensor know to work ?


PS : i can think about ordering some professional encoder but the (unretroffited for now) mill have the same TCST2103 encoder...

My guess is that you have some kind of signal fidelity issue maybe
related to the change in load resistor values. Unfortunately, without an
oscilloscope you are working blind...

Hello

I dont' know how to thanks enough.

At the end i have found some acceptable adjustement for opto position, not perfect but fluctuating range is less than 80tr min, i think the original encoder position is not a quadrature that make difficult to adjust (because one sensor is originaly connected to spindle controler and the second is connected to motion controler)

ASAP i plan to wire the encoder from the mill to the lathe card for check halscope maybee it can be better (it seem to be a "original quadrature" fixture plate but i'm not really sure)


Before doing this i need to investigate the index-enable and index pin, i think does not work fine for now.
-voltmeter checking is ok and same result for A/B (5v outside of sensor position, 0v aligned to sensor position)
-setp [HMOT](CARD0).encoder.00.index-invert 1

Never see the position reseted or the index-enable reverted to false after set to true and rotating spindle.

From Halscope i show nothing that change.

edit : maybee i have miss something and the reset index need to use some synchronised motion and is not reseted by hand rotation or simple velocity

Br

Hi
Today i have wired the mill encoder to the 7i96 and the result is something similar (brand new tcst2103, shielded harness, 150 ohm resistor for led)
I see in the TCST2103 datasheet other version with more tiny gap what did you think about that ?



And i have also do some test with some synchronised motion but the move does not start because the index-enable is never changed.

Test with multimeter www.mouser.in/datasheet/2/698/fn7529-1301979.pdf

regarding to position A
B1 = 0 or 5v A1 = 1.6v or 1.6v RO1 = 3.3v or 0.5v


regarding to position I
B2 = 0 or 5v A2 = 1.6v or 1.6v RO2 = 0.5v or 0.5v

Unfortunately i think the IC is damaged (i hope nothing other) but i don't know the reason, only one seller in china have this in stock i found nothing over the traditional shop.

Your velocity scale is set to 200m/div and we see 7 divisions, so +/- 0.7rps noise.

What is the test speed?

Bear in mind that a lot of the bad data you see is due to low-frequency sampling in Halscope. And at low speeds the effect of quadrature variability on measured velocity is larger.

What do you plan on using the velocity output for? Can you just low-pass filter it?

Hi

All speed does some noise i see +-80rpm at some middle speed

I don't know much about low-pass but i will check it asap, thanks for suggestion !

Velocity output is for => spindle.0.speed-in => pid.spindle.feedback

I'm afraid about this for synchronised motion.

Br


This picture if from 450rpm

alkabal wrote: Velocity output is for => spindle.0.speed-in => pid.spindle.feedback
I'm afraid about this for synchronised motion.

PID will naturally smooth out the variation,
Synchronised motion uses position not velocity.

Do not low-pass those inputs, but you might want to low-pass any on-screen display that you add for your own viewing.

My own viewing is not really important.

For the test i have bypassed PID

you says do not filtre spindle.0.speed-in => pid.spindle.feedback
If you think this is enough with some noisy signal like i have = job done and go to PID tuning ?

Br

I think so, yes.