Grbl tests 2022

Updates and News
1

I finally felt the need to do some more GRBL testing.

TLDR - GRBL is more capable of data throughput for raster Laser image burning. It can handle approximately 4x as much speed/data as Marlin. For CNC use, I see absolutely no difference, so you can use either.

Why Not - Since day one of this CNC thing everyone has been telling me to use GRBL I did initially test it a long time back. At that time 3D printing was a major use case for the MPCNC so GRBL was a no go for that reason alone. Then with Bart Drings ZenXY / laser board I started to use it a lot just because of the board, real world use there was no difference between it and the ramps I had on there.

Another reason was no displays were available, you had to be tethered or use a Pi. Last I test V1Pi was not athing yet. That was a no go, I did not want a computer in the dirty CNC environment. Currently, it looks like a ESP8266 or better would be a good solution (good solution for Marlin as well).

Findings -
So for the pretty basic raster tests I have been running with the Rambo Marlin could get to about 22mm/s before it would hiccup and leave artifacts in the image. Slightly faster with a 32 bit board at 28mm/s. GRBL on a Rambo let me do flawless raster at 81mm/s, but I went all the way up to 150mm/s and it would hiccup but NOT leave any artifacts in the image. Something about the way it handles speed, acceleration, and laser power is far superior to Marlin’s current system (9/10/22).

Again in terms of CNC cutting I do not see any differences. Cut and travel moves are far far less data than a Laser Raster burn that changes info every .19mm at high speeds. In CNC we move at slow speeds and change data in terms of many millimeters to meters. Plenty of time for processing and buffering.

I do not believe there to be any acceleration differences (other than laser/pwm) between the two anymore. GRBL used to be slightly better, I think they actually use the same equations for XYZ these days.

Wanna try?

GRBL on a rambo is simple - John Boiles made a flawless dual endstop port of GRBL pre-configured for our Rambo’s, so flawless you can flash between GRBL and Marlin in a minute or two and not have to change a single wire.

GRBL on an SKR - I have not tested it yet but it looks like it is also pretty easy as well. I need to verify the wire parity, settings.

To Do
Figure out a headless solution in addition to V1Pi, If you have a Pi do this for GRBL or Marlin. I would like to figure out a way to get the extremely inexpensive ESP8266 or ESP32 worked out to work with both Marlin and GRBL, on the Mini, Rambo, and SKR Pro.

12 Likes
2

I’ll chime in here too that the major selling point for me of GRBL was that you could immediately pause or stop it mid move. Marlin seemed to want to finish the current move before stopping or pausing (though maybe this is better in Marlin now, let me know if it is)

2 Likes
3

Marlin finishes the movement que before things pause/stop. And with laser it sometimes stop with the laser activated. Not optimal… I guess lightburn might have an option for this, haven’t checked it out. GRBL seems very interesting for a number of reasons!

2 Likes
4

I have attached a bigtreetech display to a nano based laser machine running GRBL in the past using the Tx- Rx pins. I wouldn’t say the two serial inputs worked without issues but it did work, and it did allow movement control and file printing from the SD card slot. IMHO the main problem with this is still the lack of a dedicated laser icon and instruction set.

1 Like
5

I think there is an option now, I think it is in the parser options. I might even have it enabled in the nightlies.

I am going to dig into the headless stuff a bit more, if we can get that on the marlin and grbl sides that makes something like the Rambo and ideal board. Have either “OS” you want and bring the price down a bit. I imagine that would have people using their devices, or having a good use for an old device, or even buying a super cheap tablet. Solves the problem of wanting a pendant as well.

John I have said it a ton this last week but that is just fantastic work, I wish I would have know about the laser differences earlier. This has really made me start digging into hiring a little help around here again. I spend too much time packing boxes, I should be upgrading the user experience instead. You made this possible 3 years ago…I am very embarrassed I did not dig deeper then.

4 Likes
6

Thank you for the kind words @vicious1, I had fun doing it! You’ve made a really fantastic low cost cnc platform in the MPCNC, and it’s been a blast to build it, use it, and hack on it!

3 Likes
7

I have been reading the github for the skr pro grbl over and over again but after some changes i did to my machine i would like better to get another board ( basically one of Bart Dring boards) but i would love to give this a try on the skr.

1 Like
8

What do you expect to gain? Barts board is awesome, but if you already have a skr pro I am not sure if there is going to be any benefit.

1 Like
9

Had a quick look at this MKS DLC32 and TFT TS35 combo. Mainly because MKS seem to have been busy developing both cnc and laser based firmwares that run on the DLC32 board which has a dumb tft(ts24 or ts35). Looks interesting - esp32 based - GRBL - decent icon and instruction set… I might just have to splash the cash!

Video here
Fluid_nc install here

While these videos are laser based, MKS do have cnc based versions

2 Likes
10

I would not want to give up auto squaring with that board.

I think we are getting close to finding the right board, >8bit, 5-6 drivers, onboard sd (or maybe the esp32 cam board could work), Marlin and GRBL compatible.

11

@vicious1 maybe you could make a guide to flash the grblhal to the skr pro 1.2. that github is practically a mess to read.

1 Like
12

The docs describe the “yellow brick road” approach. The V1 machines are very flexible and you can wander as far off if the path as you want. But we can’t document everything that is possible with them.

Grbl on an skr pro is very interesting. But until it gets a lot more mature, I wouldn’t think a doc would be helpful.

A post with some instructions is no problem (if someone wanted to do that). But the docs aren’t the right place (yet).

2 Likes
13

Yeah the flashing instructions are pretty rough for that one. If I figure it out I will post a bin somewhere so we can all test it more before we move any further.

1 Like
14

True enough… but it is close, the GRBL that is compiled in both the laser and cnc version have provision for 6 channels so the software is ready…it just needs MKS to produce a esp32 based 4/5/6 channel board.
I am awaiting one of these dlc32 boards to be delivered but have already had a poke around in the firmware as the tool provided in the firmware download package allows flashing any esp32.

The web interface is a simplified ESP3D with a ‘control’ page, an ‘SD’ page and a grbl ‘settings’ page. The flash tool is simple to use, the configuration file is uploaded via sd card and from what I can see from the video I posted the icons and screen look capable… time will tell.

The MKS Tinybee board has the required channels but the DLC32 firmware wouldn’t work on it as the hardware is different (to supply the extra channels) and the wiring to the display sockets is different.
Mk

15

Jeffeb3 knows more about this than I do but from what I understand the ESP32 is extremely limited in IO’s, so as much as I love them I have a feeling the board we will be looking for is a different processor in the long run. Bart’s board is one of the few that have broken out enough pins/multiplex/duplex or however he does it.

16

The 6 pack doesnt seem to be available at the moment, but since it is open source it looks like you could get a quote from SMT PCB Assembly in Shenzen to produce more 6 pack boards by sending the files contained in 6-Pack_CNC_Controller/versions/V1p8 at main ¡ bdring/6-Pack_CNC_Controller ¡ GitHub

The board uses 74AHC595 shift registers to multiplex io:

1 Like
17

Last I saw he was in the works to get them mass-produced. I am not really sure how I would go about funding and testing larger batches of boards. I definitely think about it a lot though.

18

Something like “6 pack controller is available for preorder on V1 Engineering, we need x orders to submit an order for a production run”

1 Like
19

I am very uncomfortable with pre orders. I have only really done it once successfully with the LR3, and that was just days in advance. Previously, any preorders I did ended up with me flooded with “when will it ship” “where am I at on the list” “can you change my address” “can you cancel my order”. Not worth the extra work.

If we decide that is the board to go with I would ask Bart what we could do together to get more made.

It is looking like that is actually not the way to go though. The SKR and Rambo both work with Marlin, GRBL, RRF. I would rather use these boards where the “OS” is optional.

20

Lot of Klipper fans around here, too. I’m getting more firmly into the RRF camp myself, so … yeah.

I do use GRBL for my laser, and I’ll probably move that to Fluidnc as soon as I get around to it, but I’ve never felt quite as confident that I’ll get 100% what I want. It seems to work well enough, though.

I think that the ability to have an “MPCNC board” is tempting, but it’s a deep rabbit hole, and much work to keep up as technology moves forward. There would be a string temptation to stop at “good enough” and not move forward until something major shifts, and then it’s a mad scramble to catch up. This works out OK for the Chinese manufacture who has a machine that they’ll sell for as long as it sells, but is a less tenable position for someone who is in it for the long haul. The MPCNC has been around a long time, and I hope will continue to be around for a long time more, so better (I think) to let other people take care of having the hardware/firmware keep up, and focus on the machine development as long as the electronics/firmware development isn’t the bottleneck. Doesn’t look like FDM 3D printers are going away anytime soon, so this isn’t likely to be a problem.

1 Like