Smoothie boards not good for CNC?

I have been listening to Makers on Tap podcast and they seem to think the Smoothie board is not appropriate for a CNC router. I have been thinking about building a small CNC router, about 2 foot by 2 foot size for personal craft projects. I have a Smoothie board that has 4 stepper drivers on it just setting around doing nothing as I really did not like it for my 3d printers. I was thinking of using the smoothie board. Why would it not be appropriate?

This should have gone in the electronics section of the cnc router part. Sorry.

That seems… broad. Do they provide rationale? Or do they have alternatives they suggest instead? That might give hints about their implicit rationale.

I think it had somthing to do with smoothieware itself. I think they said something about using an old gerbl board before they would use smoothie. I went to Duetwifi on my 3d printers and it is easier to configure that for 3d printing than the smoothieware was and the duet had the features I wanted easier to configure.

It is just that the last time I looked at smoothieware, it seemed more appropriate for a cnc router than a 3d printer.

1 Like

I run 2 routers on Smoothware and love them. They work very well.


Smoothie is used on thousands of CNC routers and mills around the world. Smoothie was coded for CNC mills before 3D printing was ever added. I think those people talked without being very informed, do you have a link to their podcast please ?


Makers on tap is the podcast. Don’t have a web address for them right now.

Here’s a link to the episode in question

This is Aaron from Makers on Tap.

For starters, we don’t dislike smoothie, we run a K40 laser at our makerspace that runs on a cohesion3d laserboard (which runs smoothie) and it works fantastically! I pinged Joe on our Slack channel to have him come and provide some more context.

1 Like

I might have misunderstood what was meant by the comment. I did listen to all of the mot episodes in 1 week.

1 Like

Hey guys!! I’m stoked someone listens and cares enough to bring it up on the internet ha!

So this is Joe, from MakersOnTap. I want to clear a few things up.

So my issues with smoothie are the same as my issues with GRBL. And @Arthur_Wolf if I’m mistaken, please correct me. I’m always open to learn especially if my knowledge is out of date.

  1. Stopping the machine quickly (like right NOW, reliably every time) is not possible without either A) cutting power to the motors/spindle and letting the control think its moving things or B) killing the buffer and loosing machine position. Sure I can home but depending on where I was in micro-stepping there will likely be a visible shift on the part.

  2. Backlash comp is non-existent. @Arthur_Wolf you and I have talked about this before. But without some sort of basic backlash comp you really can’t effectively utilize a Rack and Pinion or non-double nutted ball screw for your motion system. They all have a small amount of backlash that needs to be taken care of control side. I think GRBL recently added that. Not sure if smoothie has.

  3. It’s not possible to home a dual- driven gantry on a machine like the R7 or the OpenBuilds Lead machines as separate axis’ to allow for in process gantry squaring. @Arthur_Wolf again we have talked about this in the past. (this is called Gantry kinematics in LinuxCNC)

  4. Jogging in most of the G-code sending programs out there is done by sending distance commands at a percentage speed. It is possible to buffer multiples of those commands when jogging from the keyboard, there-by creating a possibly dangerous situation where the machine is traveling an unknown distance with no way to stop the machine quickly without loosing position (see issue 1).

NOW. All that said. You can totally run a CNC router from smoothie and have a satisfactory experience. I just think that using something like LinuxCNC is a better more reliable way to control a CNC machine. I can stop the machine reliably using hardware or software based methods without loosing position, I can jog in a continuous fashion without crashing the machine, I can add 1 line of code to the config to enable backlash comp etc.

I love smoothie for lasers I have 2 personal lasers that run smoothieware and have been helping ray test all of the cohesion cards for years. I wanted smoothie to be amazing for CNC but until at least backlash comp and the ability to stop without position loss become a possibility it will not be my go to.

Also it’s totally possible that some of my issues are due to my g-code sender and not due to smoothieware. I have used Chilli-peppr, bCNC, CNC.js, I was a tester for every version of laserweb all the way back to the clown-car-web days and so far I think bCNC has been the most reliable.

Totally open to feedback though. Id love to discuss this with people who know more than me.


We actually responded to your reddit post in this weeks episode and I covered some of the above. I also recently found the board I was testing smoothieCNC with had some hardware issues which I think was a lot of my connectivity issues I was having. But that wouldn’t change the above.

Thank you for clarifying this.


  1. With the right config, most machines can stop the way you want in under half a second. We could make it shorter with a lot of work, but it’s not a priority right now ( users who want real instant stop will loose their position )
  2. Not a planned feature, and not one we are asked for commonly. I have sold ( smoothie-powered ) routers with rack and pinion that do not require backlash comp btw. Most CNC folks I talk to agree that modern machines don’t require backlash comp as they are just designed to have backlash lower than practically matters.
  4. Smoopi allows for jogging the way you want, and it even lets you do that with a MPG nowadays. This is more a host issue than a Smoothie issue, and Smoopi was written to adress some of the limitations of existing hosts.