Encoders best CPR

What is the recommended CPR for milling machine encoders? I assume too high a count slows the rapid speed? Too low makes oval circles?

I recommend that you make your calculations so that you could achieve a theoretic resolution of 0.0001mm or less (I use this to obtain 0.001 real position, where your drive could work out with 10 encoder positions to make its internal calculations and therefor achieve the 0.001).
You need to make your calculations based on the screw pitch and reduction to get the encoder resolution.
It can lower your step rate depending on the hardware that you choose.

I an getting some weirdness with the scale factor.
I put three different encoders on my machine and scaled so I have .100 inch of travel per motor revolution which is what I require.
Encoder 1 has 250ppr and needs a scale factor of 10,000
Encoder 2 has 1000ppr and needs 40,000
Encoder 3 has 1024ppr and needs 20,480

Encoder 3 is only getting half the counts it should.
They are all the same BEI ABZC indexed encoders.
I can live with the lower count but does anyone know what happened?

Also a spare encoder is open collector not line driver. Is there a way to use it with a converter chip of some sort or should I get rid of it?

I would check encoder 3's output signals (sounds like an encoder or wiring fault)

Its possible to use open collector encoders but you may need pullup resistors
to 5V on the A/B/Z wires (most mesa encoder inputs have these pullups)

greeder88 wrote:

What is the recommended CPR for milling machine encoders? I assume too high a count slows the rapid speed? Too low makes oval circles?

It depends somewhat on the technology of the servo system.
I have 20,000 counts/inch on X and Y and 40,000 counts/inch on Z on my
Bridgeport mill, with analog velocity servos. This works well.
However, on my minimill with PWM servos I have 128,000 counts/inch,
and wouldn't mind having more.

The only time encoder resolution is likely to be a speed limit is
with software generated step pulses and step-servos like the
Gecko G320 drives.

Too low resolution won't make oval circles, it will make stairsteps
on certain moves where one axis is moving slowly while the other
one moves fast.

As for open collector encoders, just use a resistor to pull all
outputs up to +5 V. Generally, 1 K Ohm resistors should work
fine.

Jon

I learned that if you get your encoder B,B_ sequence reversed you loose half your counts unless you also reversed the A,A_ then your good.
Also learned you should hook up the 0v first (While you mind is less confused). I found out the 0v wire is the killer of encoders if you get that wrong.

I changed the jumpers on the 7i77 to single ended and my open collector encoder worked!
With the jumpers switched would I still have a Differential encoder since the encoder has the A, A-.B,B-,Z,Z- ? Or does the jumper take the (not ones) out of action?

You only use the differential encoder if you use both the A A_ B and B_
By using A_ and B_ you are only providing half of the information, like using A and B only.
The use of the complementary information (A and A_ for example) is to provide better immunity for noise. If your system is not electrical noise active (for example if it does not have a VFD) then you do not need to use the complement, although it is recommended.

You cannot use B and B_ only, like you cannot use A and A_ only.
Wiring 0V first then 5V after or the way around makes no difference (I believe that you are wiring them with the power disconnected) !