I relate to this. I had similar thoughts on forking until I saw there was a v3. Overall, I’m happy with FluidNC but knowing you could improve something can drive a software dev crazy.
1 Like
Fluidnc’s webui v2 is “ok” but we can’t say it’s really complete or “great”
I always felt like it was missing some quality of life features, even pretty simple ones…
- can’t reload the files list without reloading the page… and you can’t do that if you’re in fullscreen mode (which apparently also has a mind on it’s own on my phone…)
- tablet mode is nice but misses homing, macros speed overrides and probing, I constantly find myself switching between modes …
- jogging is not that comfortable to use, I mis-clicked or sent the z axis up 100mm too many times
- lack of keyboard shortcuts, why would I add a full keyboard and only use 5 keys? Why can’t I use it to jog?
- workspaces only set in gcode, I’d like to have visual cues about the workspaces I’m in…
- and so on…
Those are all “not mandatory” features, but there’s certainly room for a bette UX
Duet Web Control in CNC mode is a great exemple of what can be achieved on modest boards without an additional SBC
Glad to see there’s a V3 coming, I’ll have to test it
1 Like
Didn’t see your reply before posting, but totally agree, adding an SBC for web control is probably the way to go, although DWC proves we could probably do better on the esp32 
Maybe some kind of dual esp32 setup? Like one for fluidnc and one for the webui? Could keep the cost down…
I’m gonna try it.
I can. I do it all the time. Make sure you are in a browser not just the sign in window thing.
That is Mitch’s pet project, he made it for himself mostly. Mike looked at it and you can make them yourself. MAslow runs on a custom page on top of fluid.
You can mace a custom button for your macro, zoom in, or try the V3 UI version, it is different. How do you feel it compares to MArlin LCD joggin?
I have not looked at V2 but V3 has them all and you can customize them to your hearts content.
At all times there are two displays of location. Marlin has zero indication of workspaces at all.
Please understand I am not a programmer, I did not write Fluid. It is above my skill level to even help edit. If you have the ability I am sure they would appreciate some help.
I do really like it.
Chinese manufacfurer Mellowfly has a port of DWC that runs on an ESP, and is available as an add-on.
Hmmm. I thought this was an ESP8266, but the description says it’s an ESP32. It has a serial UART connection and seems to use the 10 pin LCD connectors as well for the SKR Pro. Of course it’s developed for RRF firmware, I wonder if the DWC could be developed for FluidNC/GRBL. (It’s a really nice interface.)
1 Like
So I have what is good news, I think, regarding webUI V3. I entered the string of commands in the action box, and that doesn’t work. I mentioned this in the dev channel of the Discord and found out from Mitch that the action is supposed to be the name of a text file, not the string of commands itself. I think this is good news as the text file method to me seems to make more sense.
I have my string of commands working for LED’s on my UI v3, at least the one Mike made. I have several macros, pretty sure Mike even made a release with a few built in macros.
The format is a little different, it is semicolons and a single line no returns needed. I can boot it up and double check if you need me to.
1 Like
Yeah here it is for my preferences file.
"shownotificationspanel": true,
"opennotificationsonstart": true,
"notifautoscroll": true,
"showmacrospanel": true,
"openmacrosonstart": true,
"macros": [
{
"id": "cp5yk6pai",
"name": "Wait LED",
"icon": "Eye",
"action": "M62 P2;G4 P0.7;M63 P2",
"type": "CMD",
"key": ""
},
{
"id": "hxj4ojdl0",
"name": "White LED",
"icon": "Sun",
"action": "M62 P2;G4 P0.5;M63 P2",
"type": "CMD",
"key": ""
},
{
"id": "plldxoso3",
"name": "Resume",
"icon": "PlayCircle",
"action": "~",
"type": "CMD",
"key": ""
}
],
Maybe the branch I tried out is different, or maybe Mitch was mistaken. I’ll get it sorted out.
I had already tried string with and without semicolon at the end.
This is the meaty conversation I joined this community to have.
I’m a noob here, but not to CNC or software.
My terse thoughts are:
I’d love to have a board that focused primarily on motor control
That same board, or one closely tied to it can be focused on motion control
That same board, or one tied closely to it can be focused on program control and UI
Basically, we’ve got enough compute modules these days that you don’t need one thing to do it all. If I had a choice of where to spend the silicon on the jackpot board, I’d enable the curve commands, and embed some .svg handling possibly.
I’d make it super simple to stream commands from one of the interfaces (wifi, uart, i2c)
I’d have a rudimentary UI on it, “just enough”
Then, I’d have all the super duper UI running on something else. Whether raspi pi or whatever, it doesn’t need to be on the board, it can just communicate. Just like lcds, and whatnot.
My two cents. That’s what I’m interested in building anyway.
As for the core board itself, I’m looking for one that’s based on a Teensy instead of the Espresif silicon. I’ve seen a few that look promising.
Here’s my perspective on the user experience of a jackpot over SKR+Touchscreen.
its easier to insert a card, home, zero, start job from the SKR& touchscreen than jackpot and a phone.
Needing to have a phone or tablet, which always wants to turn the screen off, or lock itself, open, refresh, not able to display during moves having to scroll up and down instead of big button navigation — these things kill the fun for me. That combined with trying to thread a micosd into its little slot make it a frustrating experience compared with taking a nice big easy insert SDcard - in the side of the touch screen, tap tap tap, go and watch the progress bar.
WiFi upload solves the microsd slot issue, and the pendant, I feel, is a better local control experience. If I didn’t have that my 7” raspberry pi touchscreen would be used as a permanent controller and I have an adapter to enlarge and remote the microsd slot.
Either way I wouldn’t consider going back to marlin as I can get the jackpot to do everything I want. I don’t regret getting it at all and prefer it to the skr&touchscreen I had.
3 Likes
Wait until you see the 8 years of Marlin issues in the backlog…
Honestly, there are a lot of people getting tripped up on networking and config. That seems pretty predictable. Most people only poke at their network long enough to make it work. Anytime they have to do something other than “connect this new device to the Internet”, it can expose a lot of underlying issues and workarounds. I’m surprised there aren’t more issues. This is a big reason a lot of IoT companies connect their device to the cloud and have an app that interacts with the cloud. It skips a lot of the local lan issues (and people like me hate the cloud, so the fight rages on).
The configuration is picky. It needs to be. You’re commanding a computer and they are dumb. People who are used to editing word docs have a hard time editing yaml. People who code think it is a delight. It is somewhere in between. I has strict whitespace rules and characters are critical, which trips a lot of people up.
You’ve done a good job wrangling the endless possibilities into some good standard advice (yellow brick road). Using AP mode and copying your config makes it a lot easier. People wandering off the YBR are going to make topics about web uploads and configuration of a pendant and 4 axis milling and we will all enjoy the journey.
If things were flawless, I’m not sure you’d see anything. You know how many jackpots you’ve released on the world. If you are not getting a ton of private support emails, I’m guessing most people are having a positive experience and just not saying much about it.
I really wish the firmware in my fridge (my coffee machine too) was open source. The number of bugs that made it to prod are astounding. I complain to my wife and she looks at me like I am nuts.
Just to be clear, I’m not trying to say the Jackpot or FluidNC has any problem and you should do something about it or ship the Jackpot with any other firmware or modification
The pre-loaded firmware on the Jackpot is good, the provided default configuration is sensible, nothing wrong with all of this…
I’m not comparing to marlin either, as I never really used the lcd on my printers or LR2, I always added some sort of web interface through an rpi (octoprint, cncjs, …) and never cared about clumsy lcd and scrollwheels
TBH, I find FluidNC the easiest to configure of all firmwares I used (marlin, smoothieware, klipper, …) as it has the same simple config file approach as klipper but without too many features…
It’s just that I find the state of the open souce project WebUI (which is indeed a whole other project than FluidNC, although it’s closely tied to it) a bit lacking, or behind what you can find in projects like DWC
The interface works, but that’s it: “it works”, and I can understand people saying they’d like something a bit more elegant and usable on some points
Again, those are all “quality of life” requirements, the OG WebUI just works as-is and none of those features are really mandatory.
But then again, if everyone had stayed with the “good enough” state of marlin, we’d never have had smoothieware, klipper or fluidnc in the first place.
I briefly tried DWC and a klipper firmware/mainsail UI on the LR2 before building the LR3, and I can say it’s a whole other experience, much smoother, and easier to control (aside from klipper not really playing great with cncs)…
If we can bring home a few ideas from those experiences, that’s a win…
I think the goal here would be to:
A - Learn a bit more about FluidNC and WebUI features we might not know or use correctly, which you adressed through your answers (thanks!)
B - Learn about alternatives to WebUI, like DWC on ESP32, adding an rpi with CncJS, or even directly on the Jackpot (eg: GitHub - karoria/grblTouch: webUI optimized for 6 to 10 inch touch displays for grblHAL. , I tried but it didn’t work)
C - Maybe raise interest in new forks or open source projects for those requirements
I thought I’d add to the OTA SD card update NEED.
My uSD card is essentially inaccessible because the board is nicely tucked into its case. If I’m going to be using the card for uploads, I’ll need a different kind of case, which maybe has a little flip up door on it.
It’s a quality of life thing. I have two prusa printers. The mini does the network upload, the MK3 doesn’t. I tend to favor the mini, even though I walk down there to verify the print is going well in either case. It’s just much easier to deal with.
I love the jackpot, and have minimal/no experience using Marlin on an CNC, and don’t want that at all.
I’m using the V2 version of the webui, and put my LW3 on its own virtual network as discussed around. I had upload problems around the time I upgraded firmware recently, but a couple of reboots later its working again.
Looking at the firmware code, I’m thinking to dive into that to see if I can’t help make the SD card reading more solid. At the very least, maybe add some more file system support beyond FAT32, or more stability if possible.
Also, on the tokenizer, both for config and command processing, I’m looking at whitespace management.
It’s open source after all. so, I’m not complaining for what I’m getting for free. Maybe I can lend a hand to making some of the quality of life improvements we all want to see.
1 Like
I am hearing all this more clearly now. Thank you for the different points of view.
I guess I have to add my point of view, is it is so much better than Marlin on a CNC I let the little things that bug me go. To some of the above points, I agree (same with the new klipper printer build).
I have also been messing with V3 of the UI since right after launch of the jackpot, things moved quickly. Now that it is in Fluidnc beta that has slowed down. So for me I have just kinda been in a holding pattern. I know FluidNC currently has many commits that have not been in a public release so I am just waiting for the next release to sort of dive in deeper. V3 is better, not perfect but substantially better than V2. Mike got us over that hurdle and got Mitch, and Luc to add some changes to let us play with V3. Once we get an updated release from that point all you talented devs can choose the battle and see where you can be of the most help. We will need testers, ESP3D UI devs, and FluidNC / network help. Also V3 can be themed to an extent, so some of you might very easily be able to make us a custom UI screen that everyone loves. The current UI is 3D printer biased as always…but can be more easily changed now.
I whole heartily agree, this seems like it is on the verge of being amazing with what seems like some minor changes. I think a lot of you know these small things are much much larger problems (file transfer). Can this all be fixed to do what we want or will we have to keep searching for something else, I don’t know. Depends on the devs and if the hardware can actually handle it.
3 Likes
The current hardware may be limited. But, hardware always gets better. I’m sure we can make the currently hardware work optimally, and then along comes the next version of the chip family. Stronger, better, faster, more memory.
The only real option I’ve been exploring is teensy as a core instead of espressif, mainly because I like those boards.
1 Like
You just have to remember there is a giant mountain of code behind this. I have no idea if fluid is set up to easy add in other boards. Marlin and GRL obviously are but last I checked fluid is not fully working on the newer C8 or whatever esspressif. Being able to get monster processors for cheap doesn’t help if it takes 6 months to get the current code to work on it. Fluid’s Dev Mitch is monster at understanding how all the parts are intertwined and why something that seems innocent is a horrible idea, and what edge case it might break. The way code is ran, and the plethora of machines it runs on is very insane.They want newer fast more gpio chips as well but it is not always very easy. ( I think maslow runs on the newer chip if I am not mistaken…or plans to be)