Remora - Rpi Software Stepping Using External Microcontroller via SPI

freemoore wrote:

Also, with apologies for splitting the thread topic, is there anything like a pre-rolled config to use with RPI 4 -> STM NucleoF411RE -> arduino cnc shield, and if not where should I start?

I dont have a nucleo 411 to test with, but it works with the blackpill
github.com/cakeslob/Remora/tree/nucleo_s...areBin/NUCLEO_F411RE

here is the docs

github.com/cakeslob/Remora-docs/blob/nuc...rdware/nucleo446.rst

first of all, thanks you for making remora

i am currently setting up remora on skr 1.4 (LPC1768) rpi4  and HBS86H(hardware-cnc.nl/images/PDF/ACT/HBS86H_English_Manuel.pdf) as the step motor driver.
so far i manage to get the remora_xyz.hal running and the step pulse show up on oscilloscope.
 

but i have not yet test the setup with  HBS86H driver yet since the driver require 5v-24v input so i am waiting for a logic level shifter

currently my only concern right now is about the direction change and step pulse timing

as in the timing diagram from datasheet, it shown that direction state need to be toggle at least 6 uSec before the next active step pulsing.
so my question is
"with my current setup,will remora toggle the direction pin at least 6uSec before next step pulse?"

with the oscilloscope i can see that one step pulse took 25 uSecs but i am not sure about how much early direction pin will be toggled when changing direction. i cant really test it because i have only one probe. 
accoding to sccota's reply from a post in Jul 2021, base thread took at least 8us to execute therefore direction pins will definitely be toggled at least 8us before next step pulse right?

There is no configuration setting for step and direction timing in the original Remora. I've used it like this for many years without issue. Remember, that on a direction change the velocity will be zero or close to it. So the likelihood of a step pulse being generated is very low.

Hurray!!!! After 8 days I've managed to get a version of Raspberry Pi OS, the prempt-rt patch, and Linuxcnc and linuuxcnc-dev on my 3B......Was picking at it daily while trying to read and absorb the documentation on it and Remora.....

So I guess the next step is to install the Remora stuff on Linuxcnc?

How's those cheap as chips boards coming along that you've been working on? Are they ready for a thick headed pensioner to make a version yet ? And what do I order as far as one of these "pill" things?

Cheers from sunny Canada,
Lewis

Lewis wrote:

How's those cheap as chips boards coming along that you've been working on? Are they ready for a thick headed pensioner to make a version yet ? And what do I order as far as one of these "pill" things?

so far so good. nothing blew up in my face. both boards send steps/io over ethernet/rpi. so far a success. Few small errors with the silkscreen, I labeled some pins wrong. Ill get your details in the discord

Good to hear ! I've never labeled silkscreen pins wrong.....(insert muffled laughter here).

scotta wrote:

There is no configuration setting for step and direction timing in the original Remora. I've used it like this for many years without issue. Remember, that on a direction change the velocity will be zero or close to it. So the likelihood of a step pulse being generated is very low.
 

hi scotta.
    i did a bit of testing after i got my second probe for the oscilloscope.
it seem like step pulse does happen when changing direction.

i create this result with by joging from right to left then left to right
luckily i can just change active pulse edge on HS86H driver to use the second rising/falling edge instead

as of right now,I'm planning to use Remora for other projects involved drivers like DM542 or tb6600 (both require direction to be change at least 5 uSecs before step pulse and they can't change the active edge.)
i dont really know how to correctly add a delay to the stepping pulse before changing direction. can you maybe explain how does stepgen.cpp work?
i mean i could try adding for loop count from 0 to some number as a delay after set directionPin statebut i am not sure if this will effect how remora works.

thanks,
gittawat

If you plan on using it for other projects, i suggest migrating to an stm32 board.


replace your stepgen with the one here

github.com/Expatria-Technologies/remora-...-OS5/modules/stepgen

I think i have this in the stm32 version, but looks like i commented the dir stuff.

fuuuuuuu
 



Thats what I get for assuming the encoder pins would be the same, and not verifying. So hardware encoder needs a board revision, but software encoder should work.


In the meantime , I did a github for my board thing.

github.com/cakeslob/RRW_LAB