AMT102 Encoder Mesa 5i25 & 7I75 (or not ) makes me crazy

So today i went an bought an Oszilloscope ,..
It is clear we have an Wide A and B signal (which i can see at Linuxcnc ) but the widen of the Index signal is very narrow i would say ( and for sure not like mentioned here www.cui.com/product/resource/amt10-v.pdf )
But now my question my standard speed is maximum 4000RPM (or max 6000RPM) average will be around 1900rpm
my intended PPR setting would be 192 (or 512) so do you think it makes sense to use an Condensator to widen the Index signal ? and who can calulcate the size of it ?
or is it an stupid idea ?
thomas

The Index should be as wide as the A or B pulse. Well, depending exactly on how they do it, it COULD be half the width of the A or B pulse. If it is just a really narrow spike, there could be a bad connection, and all you are getting is the capacitive coupling across the bad contact.

Jon

Hi
even if i move it by hand moderate (lower than one rev / second ) it is very narrow
as i wrote in 2nd post " it doesn't show the Voltage when i turn the encoder by hand. The AMT10x Document www.cui.com/product/resource/amt10-v.pdf Figure 1 says Index Width is T=360/192 =1,875 Pulse Width = T/2 = 0,9375 Index Width = P/2 = 0,46875 this should be enough."
so my expectation it is 0,25 of A or B width but it is a very tiny ping even on the oscilloscope and it doesn't show on the mesa that is what it is all about.
And i tried it with 3 different encoders ,.. so my question is is there anything like an condensation to widen then index that mesa can see it ?
any suggestion would help
thx thomas

Jmelseon you are right
i bought an Oszilloscpe and found out that the spike is too small for the Mesa card.
I build then an NE555 based circuit to broaden the signal now it works !!!!

But now an much stranger behavior at G33.1 the spindle starts to turn but at the "bottom" it doesn't change direction and go back it stays at the bottom

now my question tho the knowledgeable people
What signal situation triggers the change of direction
is is reach of depth or is it waiting for the index ? or does it do an index-enable and then wait for the index ?

can i change it that it simply turns at the bottom and goes back to original z position ?

this would help me a big time ?

Index is only checked at the beginning of the cycle

I doubt that the index pulse is tooshort for the 5I25 to see if you disable the filter
with the filter enabled, pulses shorter than 450 ns are ignored, if the filter is disabled,
pulses shorter than 90 ns are ignored

That said, if the AMT encoder is not generating valid index pulses (minimum 1/2 A width)
I would consider it faulty

Hi,
it was exactly around 90 ,.. even if i measured direct at the encoder and i tried 5 of them (i use them for closed loop at the steppers as well ) but now i enlarged the ns and it works so at G33.1 it starts to turn in the right direction and it does it
well till the bottom

i have now the feeling that i might have an config issue thats why i asked here

Is it possible you spindle reverse is not working? that would cause a hang at the bottom of the tapping cycle

If the index pulse is really just 90 ns long, I would consider the encoder
design faulty since a TTL signal of 90 ns is easily generated by impulse noise

Right! Try running the spindle in forward (M03) and reverse (M04). It should operate both ways from MDI commands before you try rigid tapping. You have to enter a speed command on most versions of LinuxCNC, so the MDI commands would be something like :

M03 S1500

and

M04 S1500

M04 S1500 will actually set the spindle speed to a value of -1500, what happens with this depends on your HAL config. In some systems this will generate a negative voltage to a servo drive, in others, it will activate a different relay to command the reverse rotation to the spindle drive.

Jon