[dirtdigger257]

Hi I.m try to hookup a set of Ditron DC10 linear encoder to a kflop board. I have the encoders wired up to match a pinout off of the Ditron website but I.m not getting any Voltage change when i test with a multimeter(the tester is an analog type with a needle) i have a 10k resistor on channel a and 2.2 on channel B with 5v going in. Has anyone used these encoders or know the proper wiring Attachment 221522Attachment 221524
do i need to link the encoder to the axis with c code if so what is the wording in c i.am using I/O 0 1 pins 7 and 8 on the jp7 i also should note these are running threw my c10 board pins 10 and 11 witch are wired to pins 7 and 8 on jp7 on kflop. any help would be great

[Al_The_Man]

They appear fairly straight forward, you should be getting a transition when you move the head, ever so slightly, are you using the 0v for common? The GND is for earth ground.
Al.

[dirtdigger257]

No 0v not hooked up ground is just hooked to the c10s ground. I'm not sure where to hookup the 0v. It does make since that could be the problem.

[Al_The_Man]

0v is the +5v common, and GND is for the shield and is connected to earth ground connection, it should work with this unconnected.
0v has to be connected however.
Read the Fig title at the bottom!
Al.

[dirtdigger257]

Got it picking up changes with the multimeter thanks. Now i'm tring to get tuning going the red position line is moving now but it but the spike are small compared to the command

[dirtdigger257]

I had the input channel set wrong i have position and command overlapping. The movement isnt as smooth as it should be i think the sample time of the driver and encoder arent messeing well so i need to figure out how to tune it. Any ideas.

[dirtdigger257]

Thanks Al

[TomKerekes]

Hi dirtdigger257,

Please describe what your system is and post some screen shots of the Step, Config, and filter screens and we can try to help.

Regards

[dirtdigger257]

Attachment 221698Attachment 221694
the encoders are .005 resolution and i'm running 2000 microstep on the driver 5mm pitch ballscrew
so 1 rev =5mm =.19685039. for 5mm of travel the encoder should be .005/2=.0025*2000=5
1 step =.00098425 the reciprocal is 1016.002032 and 1 cycle of 4 step is .00098425*4 = .00393701
the reciprocal is 253.99986284 is this the number i put in for inv dist per cycle
when i move the y or z (they are running open loop with no encoder ) they move some what smooth. the x move smoother one way than the other ( x has the encoder) when the i move z or y and it make the crunchy sound the x moves (kind of osculates)

[TomKerekes]

Hi dirtdigger257,

You forgot to describe your system. You didn't show your Filters Screen.

It is better to grab screen shots rather than photos. Use Alt-PrintScreen then paste into paint.exe and save as *.png files.

Your Screen settings are different than your C Program settings. Please watch this flash video to help understand different parameter locations and how to copy/move them. Initially just use the screen settings for testing. Showing C program settings that are different and not being used doesn't help us.

Change Dead band range to zero (gain of 1 has no effect anyways, the gain is 1 inside and outside the range)

Remove P Gain setting.

Remove Feed Forward setting.

It seems you are badly missing steps or the encoder is not counting correctly because after the 400 step move the Output Correction is ~150 steps. Did the motor move back to the same place?

Regards

[dirtdigger257]

my system is kl-5056d drivers set to 2000 microsteps 1/10, c10 board ,1.8 degree step angle 2 phase motors. I'm not using a filter. the p gain you referred to is that the proportional? i watch the flash video it made sense. its going back to the same point when i hit the move button confirmed with an indicator . i changed the settings the way you described. I will post a screen shot as soon as can figure out how to make it not so grainy looking.

[TomKerekes]

Hi dirtdigger257,

Yes P gain is Proportional Gain. Set it to zero.

First get things working open loop before attempting to run closed loop. It is important to check if the steppers are moving correctly and the encoder is measuring correctly and that they have matched resolution. On the Step Response Screen the red and blue plots should match with small errors when running open loop.

When you say encoders are 0.005 is that millimeters? I can't follow your math but it seems the steppers are 400 steps/mm and the encoder might be 200 counts/mm.

invDistPerCycle is not used for Step/Dir mode. Use InputGain0 to match the encoder counts to steps. Possibly with a value of 2.0. If the value is wrong you will be able to see the factor by looking at the Step Response Screen.

Regards

[dirtdigger257]

[TomKerekes]

Hi dirtdigger257,

You are still running closed loop. Please read the previous post.

The bottom portion of the Filter screen is what is used by KFLOP. We can not see it.

You have two programs flashed and running in KFLOP. Please remove/erase those by Flashing "New Version".

Regards

[dirtdigger257]

[TomKerekes]

Hi dirtdigger257,

Also leave the Input Mode set to Encoder so we can see what the encoder is indicating.

Regards

[dirtdigger257]

Attachment 221752
it is over shooting or going passed about .005 inches in open loop

[dirtdigger257]

I flashed the new version zeroed step response screen and my dial indicator it returning to zero
After 20 time of hitting the move button it was of by .005

[dirtdigger257]

i think i have something figured out. when i unplug the encoder from the c10 board the travel is smooth in open loop mode using mach3.

same problem i was having with the limit switches. if i plug in anything other than just the x y z driver i start making the grinding noise.

do you think it`s the drivers (kl-5056d) i dont think it`s a grounding problem never had one before the system has been running on mach3 only for a year with no problems.

i might try putting the .047 uf cap in place if you have no other suggestion.

thanks taking time out of your weekend i do appreciate it.

[TomKerekes]

Hi dirtdigger257,

Good progress, but it seems there are still a few issues.

Regarding:

Grinding - Connecting the encoder should have no effect on open loop stepping. There must be some type of wiring, grounding, or supply problem. If you post specs on the breakout you are using and how everything is wired we might be able to help.

Encoder off by 0.005 - You may have a Direction setup time issue where the drives sometime step in the wrong direction on direction reversal. I believer those drives are technically incompatible with KFLOP's 4us setup time.

There is a line of C code that can be used to change to max pulse and setup time.

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

Are you doing this? If not, that may help with the two issues above.

It looks like the encoder scale factor is close but not quite right. On a 400 count move it is short by ~20 counts. It is hard to say if that is mechanics or the InputGain0 ratio. Make a much bigger move and it will be more obvious.

HTH
Regards