[TomKerekes]

Hi Tim,

Why not look?

1600 steps/rev would be 8X microstepping. That should be fine for the test.

BTW before running the test you will need to push "Move" on the Step Response Screen at least one time so to set all the Configuration in KFLOP for the Axis. As well as run the C program to set the pulse time.

Regards
TK

[timbochung]

Hi Tom,

I ran the experiment with each my drives and it looks like the KL-8070D is KFLOP compatible whereas the KL-5056D...not so much.

Going to get in touch with John at Keiling - hopefully, he has either a solution or let me exchange the drives.

Attached are pics of the setup + results

Tim

https://www.dropbox.com/s/yxpx0dxfew...Test%20Vid.MOV

https://www.dropbox.com/s/ng48cas5ey...Test%20Vid.MOV

UPDATE: some of the pics for the KL-8070D are mislabeled as KL-8056D...

[Stigoe]

You don't have a KL-8056D, you have a KL-8070D, so which stepper motor do you have? You have set both types of drivers to max current, which for KL-5056D is 4A RMS, while on the KL-8070D it's 5A RMS, so you may overdrive your motors. It may not have anything to do with your problem, but something to be aware of.

[timbochung]

Whoops , it is the KL-8070D driver mated with a KL34H295-43-8B, 906 oz-in NEMA 34 Stepper Motor

[TomKerekes]

Hi Tim,

Strange it would be different.

Inverting the step pulse can provide more or less Direction setup time. You might try inverting the pulse to see what happens. To invert the pulse change:

FPGA(STEP_PULSE_LENGTH_ADD) = 63 + 0x80; // set polarity and pulse length to 4us

to

FPGA(STEP_PULSE_LENGTH_ADD) = 63; // set polarity and pulse length to 4us

Another option to try would be to add a 0.047uF capacitor between Step Input of the C10 board and GND on the C10 board. That should add about 2.5us of delay to the step and put it in spec.

Regards
TK

[timbochung]

Hi Tom,

I inverted the pulse as you suggested and it worked like a charm!

Tested both drives and increased the # of cycles to 1000 to verify performance.

Tim

[dirtdigger257]

// address of 6 bit pulse length 0-63= # 16.666MHz clocks,
// bit6 muxes generators 4-7 from JP7 to JP4 and JP6,
// bit7 reverses polarity

// Geckos work well with this and "open collector" outputs

FPGA(STEP_PULSE_LENGTH_ADD) = 0x80 + 63; // set inverted and to max 4us

will i have to use jp4 and 6 or just keep using jp7 and if i do have to change it what pins will the generator output it to. i see i am supposed to delete 0x80+ i'll try this and see what happens

[TomKerekes]

Hi dirtdigger257,

Don't understand what you are trying to do. But see the Step/Dir Generator documentation here.

Regards

[dirtdigger257]

if i run the code bellow does it change the pins it out puts it to. example for z axis jp7 pins 19 and 20 i/o 12, 13 change it to i/o bit 32 and 33 i thought those were only 3.3v capable. and what is mux 0 and mux 1?

[TomKerekes]

I don't see the code you say is "below".

BTW The Z axis is not necessarily the Step/Dir Generator #2 even though that is often a logical choice. The are two levels of configuration: the DefineCoordSystem can map any Gcode Axis to any KFLOP Axis Channel. And a KFLOP Axis Channel can be configured to use any Step/Dir generator.

But yes the Mux=1 option will move Step/Dir generators 4-7 from JP7 to JP4 and JP6 as you describe for one of them.

Read the description of how to control the Mux option in the link I sent under the heading:
Global Register sets Pulse Width, Polarity, Multiplexor

All of KFLOP's IO are 3.3V. But can be used to sink signals to ground. This applies +5V to an opto if the anode of the opto is connected to +5V. To turn off the opto the output can float (open collector mode) to about 3.8V. Because IO 32 and 33 have 150 ohm pull down resistors open collector mode can not be used with those IO. However IO 34 and 35 can be used in Open collector mode.

What type of signals does your Step/Dir drives require and what are the timing requirements? KFLOP can only be set to a maximum Step Pulse of 3.78us with a maximum of 4us Direction setup time.

Regards

[dirtdigger257]

command type pul/rir or cw/ccw active edge rising or falling direction def high or low. mine is set default because i dont have the cable to tune them.
Default setting is PUL/DIR mode and rising edge active (NPN, and PNP control signal is on the contrary).
The KL-5056D can accept differential and single-ended inputs (including open-collector and PNP
output). The KL-5056D has 3 optically isolated logic inputs which are located on connector P1 to
accept line driver control signals. These inputs are isolated to minimize or eliminate electrical noises
coupled onto the drive control signals. Recommend use line driver control signals to increase noise
immunity of the driver in interference environments.
I cant find the step pulse requirements. i have kflop set 5v open collector i'am not sure what pnp mode is or if kflop supports it. i was thinking of using a pig tail from the rj45 to the rj45 of the drive and wire it accordingly. i am using a c10 board right now but might pick up a 26 pin bob from winfordhopefully i can get this to work because my linear encoder should be here monday and a hole new head ache will begin.

[TomKerekes]

Hi dirtdigger257,

I Googled and found this:

http://www.kelinginc.net/KL-5056D.pdf

Unfortunately those drives require a 5us Direction Setup time which KFLOP doesn't support (max=4us). See page 3. Not having enough direction setup time may cause the drive to step in the wrong direction sometimes whenever the direction reverses. Most drives don't have such a long setup requirement. We may change KFLOP in the future to support these type of drives as they are becoming very popular. But I can't say when or even if we will do this. One workaround is to make a circuit to delay the step pulses by 1us (Resistor + Capacitor + buffer chip).

I'm not sure if you are willing or capable of going through the extra difficulty. If you want to return your KFLOP we can arrange that.

Regards

[dirtdigger257]

Ok, thanks. I did get all three axises working moving the proper amount and in the right direction after adding the extra setup up time. But now when i try to run with limit switches it acts funny, crunchy sounding. There is also a 7sec delay from when i adjust feedrate or hit the stop button witch displays a message unexpected motion buffer starved. I'am running kmotion 430 and mach3 R3.043.066 versions. i turned sticky keys off and it didn't help. kmotioncnc wont work on my laptop but worked on an older computer i was using Thanks for all the help

[TomKerekes]

Hi dirtdigger257,

funny, crunchy sounding

No idea what that might be. Any issue with the limits to KFLOP I would expect to just stop/disable the axis. I would not expect a change in motion smoothness. Possibly a noise/grounding problem affecting the step signals to the drives?

There is also a 7sec delay from when i adjust feedrate or hit the stop button witch displays a message unexpected motion buffer starved. I'am running kmotion 430 and mach3 R3.043.066 versions.

In later Versions of Mach3 (like R3.043.066) the message to plugins for Stop seems to have been changed. Our later Test Versions should handle the new (and old) Stop messages and Stop immediately. Please try our latest Test Version here. FeedRate override will have a delayed effect due to buffering. You can minimize it by reducing the Windows Buffer Time in the Dynomotion Plugin Configuration.

KMotionCNC should work on any Laptop. But I thought you are using Mach3?

Regards