[Epiphany]

I am confused on the options I have for step and dir output and need some help. I have a KFLOP with 1 Snap Amp and a Kanalog board. I am putting a servo on the spindle of a lathe and need to use the step and dir for this purpose. The outputs will be multiplexed over to JP6. I have attached an image of the servo hardware...I would like to use the differential inputs in C3-1 to take advantage of the higher pulse frequency.

As a test I connected a differential encoder channel A to terminals 6 & 19 and channel B to terminals 7 & 20 and applied the 5V and ground, then set P035 to 2 and the servo shaft will follow the encoder shaft when I spin it by hand.

Attachment 261738

I have looked at the "Configure Step and Direction Outputs" in the help and assume I want quadrature outputs, how is this wired? Can this be done.

Thanks for looking.

[TomKerekes]

Hi Epiphany,

KFLOP doesn't have differential outputs for its Step/Dir Generators. If you need differential outputs you would need to add external differential drivers. Low cost 3rd Party modules can do this. They would be the same modules that convert single ended encoder to differential line drivers signals.

But you might use KFLOP's single ended open collector mode in Quadrature mode as you mentioned. The drive permits up to 100KHz cycles (400K quadrature counts per second in that mode). Is that fast enough to handle your top speed?

Regards

[Epiphany]

Tom,

Thanks for the reply. I think that will work. The servo is 10,000 ppr and at 2500 RPM it would be 416,666 pps....so 400,000 will work. I will experiment as it tells me I am limited to 200K in this mode on the servo end but that may be per channel as well. Thanks for the support. Happy Holidays.

[TomKerekes]

Hi Epiphany,

I think if you read it carefully the quadrature mode allows 400K quadrature counts/sec. It shows 10us for a complete cycle of 4 transitions. Quadrature mode is nice because it only requires one signal transition per count rather than a pulse which requires 2.

Regards

[Epiphany]

Tom,

You are correct it accepted the 400K quad counts per second. I wrote the new pulse train calculations and have it working perfectly through KMCNC. Thanks again.