I performed an experiment to see if I could scratch build a CNC/Printer/Laser-Engraver/Plasma-Cutter/EDM-Mill. I started with Mega/RAMPS/GRBL, a cumbersome pile of parts from at least two boxes of parts and junk, and verified that I can perform all five functions reasonably well, with about 10 minutes of Z-parts switch over, including recalibrations. I spent $115, but my junk boxes are impressive. It's a mess of ugly welding and chewing gum, but it works. I even made it WiFi with an ESP8266 and the ESP3D code.

The only real problem is that the CNC function can only cut wood (NEMA17 is what I had). So, I bought a WorkBee frame as the base for a second, more powerful machine. Its beefy X axis can handle multiple Z frame configurations, the bed frame can easily handle bed management, or a rotary axis etc. This is the right frame to use, given my goals and stingy budget.

Add a frame, ya gotta add a controller and drivers, right?

All of the controllers are a bit weak for one process or another process. Marlin or GRBL do printing and CNC OK, but Laser engraving and EDM are a small or giant pain, respectively. Moar resources are needed for what I want, so a 32bit board makes sense. The Duet3D-2 is the obvious best hardware currently, based on just its hardware capabilities. However, it only runs RRF, which is an ugly limitation. Being a programmer has its perqs, so I bought one.

Question: All DIY-anything controller users beat every conceivable modification to a greasy pulp within a few hours of opening the shipping carton, so why is the Duet3D still chained to reprap firmware after all this time? Did the higher price scare off the scientists? Am I beating a dead Haas??

If there are already ports of other firmware to the Duet, complete or otherwise, I'd appreciate a link.

Wow. I didn't say I had anything against RepRapFirmware. That was all you. Is that a habit? Ugh. I don't have anything against RepRapFirmware for devices that do just the one thing. In fact, it appears to be perfect for 3D printing, though I haven't even started building my new beefy print tooling yet.

GRBL doesn't care what you're doing, as long as you're trying to get stepper motors to ... step. Marlin only cares a little, but can be easily beaten into submission. I never tried others after I found two that do what I want. It will take longer to get RRF to do the things I want (TRIGGER ALERT: THINGS THAT YOU ARE NOT DOING).

You're wrong. Both GRBL and Marlin have a WiFi and web interface if one adds a $2 ESP8266-01, connects 4 wires, and spends 45 seconds burning firmware to the ESP. Gluing the ESP to something so it doesn't flop around is a nice aesthetic touch, but optional. Macros are a function of the gcode injection mechanism (an ESP8266-12F? on the Duet I believe) so GRBL and Marlin have all the macros your favorite gcode injection tool supports. The ESP mod I mentioned does, and Repartier provides for six macros of any length. Some others more, while others less or none. I wrote one in Qt that stores macros in an SQLite database, so that only macros which pertain to possible operations for the current tooling setup are visible. Eh? EH?? Sweet! Oh... Oh yeah.

I'm going to guess that your machines do one thing: additive machining. That's terrific, and you should beam with pride. I'll print a trophy when I have the tooling done, but that's only part of what I'm trying to do, and I went to considerable effort to explain how what I'm trying to is likely very different from what you're doing. You're offended by that. If someone cares, you should totally hang out with them sometime.

Again, I don't have a huge problem with modifying other firmware to my DIFFERENT AND WEIRD usage case, but it's a lot of work, it makes merging version upgrades difficult, and I'd rather do something else with all that time if more flexible firmware will run reasonably well on the Duet. So, try to control your emos, bruh, and mash the Reply button if my question needs further clarification (??), or if you have a possible answer.

@dc42 Yes. No matter what, RRF will do what I need within the physical limitations of an OpenBuilds frame. I know this because I bought the Duet and RRF is its only player piano roll. I'm confident I can make it control a multi-rotor heart-lung machine if need justifies the investment. Using RRF is my current Plan A. However, it's the only plan available and I'd be a slacker if I didn't look at all options. I'm a newb to Duet, I don't know what the options are, and I haven't followed all the Duet (or RRF) development threads.

You may be right about the cause of Duet having so few firmware options, but I'm hearing a great deal of box-turtle style thinking in this forum, and even some in this thread. It smells like ... FanBoyz, at least compared with e.g., the RAMPS forum. That's never a good sign.

Why is it, you ask a question, and a horde of zombies descends to criticize your question, and your reasons for asking the question. Is it to hide or discourage potential answers? Is it to wisdom-signal other forum members ("LOOK. LOOK. LOOK. I CAN WRITE TOO! ME! ME!")?

Q: "Who has a bread recipe for rye and spelt?"
A: "YOU ARE BEING SILLY. YOU SHOULD ONLY USE WHEAT TO MAKE BREAD. YOU ARE A FOOL IF YOU DON'T USE WHEAT. I USE WHEAT, HE USES WHEAT, YOU SHOULD USE WHEAT. WHEAT IS LOVE. EMBRACE WHEAT."

How did that become a common thing?

@Phaedrux Probably, yeah. I don't like people, and I make no effort to hide it. No one has complained within earshot since I was 17 or 18, so I can't know with any certainty. I don't care either way.

One big difference between a forum discussion and face-to-face is that asking a question to someone's face: they are extremely unlikely to attack your question or imply that they know what you want to do better than you. Every boy learns why that's a bad idea early in life. OK, most boys. A question face-to-face, you get an answer or a shrug.

I asked a very specific technical question and received only attacks on the question itself. I answer responses to my posts. I already figured out that you don't know the answer, yet, here you are again, with no answer, and no value-add counter-question. Why?

The original question is pages up now, so I'll repeat the gist:

Do you have a link to, or a lead on an existing alternate firmware project for the Duet3D-2 boards? PLEASE DO NOT ANSWER IF THAT ANSWER ISN'T A LINK OR LEAD.

PS: The answer is "Yes", Marlin 2 is in beta state, at least for the Duet 2 Maestro (SAM3X8E MCU). All Marlin features are implemented, and the current focus is on adding support for RRF gcode features that many Duet users rely on.

@deckingman

Yes. Running a single firmware is an ugly limitation for Open Source hardware.

The limitation, if there is one in my statement, is Duet's. I double-checked my phrasing and sentence structure. It is correct. It reads one way. This is why so many people want reforms to public education.

@Phaedrux

Thanks. Yes, there are many expensive machines that can (or could) do what I want to do, and more. OSU has some interesting additive milling gear in Columbus that uses ultrasonic friction to print in aluminum. They do tours! It's only a few hundred thousand. But, please recall, I already did 99% of what I want for $115 on a pair of Open Source boards that have nowhere near the power of a Duet. I'm going to do the same thing on a Duet. No additional cost, other than my time.

Why am I looking for alternatives before I even use RRF? Two reasons:

1. I use Marlin on the other system. I added a function and some macros that take an M666 Mn command, which allocates some GPIO pins, the extruder and fan drivers to perform tooling operation based tasks e.g., start EDM tank circulation, increase EDM current, frequency, etc. This also changes how, for example, a Spindle ON command is interpreted, what the max feed rates are, what the PID controller is really doing, etc. This allows me to use Fusion 360 to pretend that an EDM wire is really just a milling bit, or that a (velocity-sensitive) LASER cutting operation is really just a plain old drag knife fuggidabouddit. It simplifies and speeds up the process of making lots of things that a 3D printer won't yet do. Ideally, I'd run Marlin 2 on the Duet and fix the Paneldue display code. There's loads of memory on this beast, so expanding functionalities is no longer resource limited.

2. It took considerable time to design, write, and tweak the code to do all these things well, and I'd rather not do it over again. I also don't want to be locked in to using one set of firmware if I can avoid it. On anything. Ever. Why? Remember when Microsoft Windows wasn't a subscription-based service? (Oops, spoiler alert!)

I didn't ask for suggestions that involve a different board or a different frame. That might be of value to someone, but probably not in this thread. I already knew much of what was popular and possible before I asked my question.

I would like suggestions for alternative firmware, complete or in alpha state. I don't even need to hear if you aren't aware of any: I'm already not aware of any, thanks anyway.

Is there a way to probe an irregular (not flat) surface within a specified Xmin, Ymin, Xmax, Ymax rectangle, using specified probe spacing, to obtain a height map grid, and save that as a file (CSV, STL, OBJ etc)?

(I think this is basically auto-bed-levelling with more UI and control. So, it might already be there.)

I don't see a way to do this out of the box, but surely, someone else has requested this, as it's extremely useful for printing or CNC.

Why I want this: that map can be used to determine Z relative offsets, allowing 2D g-code to work correctly on irregular shaped surfaces, streamlining most 3D workflows. Mach 4 does this, but it's the only Mach 4 feature I want.

If it doesn't exist, and if I'm willing to add this functionality, is anyone else interested, or should I keep it as a private branch?

Looks like the 1.04 boards add a fuse to the A4403 supply circuit. Or, at least the circuit is powered from a fused supply.

@Phaedrux Purchased from Filastruder in March of '19, but not put into service until March of '20.

This A4403 buck converter is rated for 3 amps. That seems a little lightweight for all the 5V systems on this board. Is that really enough to power the PanelDue 7i, plus all the other 5V logic, plus the 3.3V drain of the controllers?

A replacement is $1.78 from Digikey, and some nightmarish hot air work down in a deep, dark parts canyon.

Maybe 1.10 could replace U3 with a bank of socketed T03 7805's on a big aluminum heat sink.

With thumb screws.

Thanks for the schematic!

My vision isn't that great, but I stuck it under a scope. Ahh HA!

U3 appears to have a volcanic vent obscuring the label. I'll have to dig through the schematic to find out what U3 is, and what might cause it to fry like this. It looks to have done some smoking too, but there was none of the usual magic smoke aroma...

Configured as a 3-axis CNC, housed in a PVC utility box with two fans, and a filter on the intake. It has around 140 hours of milling on it. I shut it off on Thursday, now it won't power up. Only the Vin (24.16VDC) LED is lit, PanelDue 7i stays dark.

Removing the board, plugging it into USB, only the Diag LED lights (dimly), and R23 (I think) gets very very hot. My fingers are too low-tech to say with confidence that the frying pork smell is coming from R23, but I think that's the one.

Is there a common area to check first? A known weak link, like an underrated linear regulator, etc? There's probably a schematic for this somewhere, so maybe some nice person can share a link.

@giblte535 Note for trying to use bossac:

The -e option in the instructions will fail if the firmware was erased with the button.

But here is what I was talking about with the bossac problem. Verify fails and the screen stays blank.

$ ./bossac.exe --port=COM13 -w -v -b PanelDue-5.0i-7.0i.bin
Atmel SMART device 0x289b09e0 found

Write 168276 bytes to flash (329 pages)
[==============================] 100% (329/329 pages)
done in 4.156 seconds

Verify 168276 bytes of flash
[==============================] 100% (329/329 pages)
Verify failed
Page errors: 297
Byte errors: 145433

If I dig into the source and figure out why that is, I'll update this post. Thank you again!

That solved it. Thank you sir!

@dc42 Thanks. I did not check those boxes but will try that when I'm done typing. I followed the instructions to the letter with bossac, and I see no instructions for using the GUI on the blog, presumably because of the visual issues of the GUI.

I'd MUCH rather use command line anyway.

Thanks again.

That makes sense.

I only press it as part of the firmware update though. It's part of Crocker's instructions.

I bought a Duet3D version of the integrated PanelDue 7i back in February, and saw that newer firmware was available from DC42. I followed the instructions on David Crocker's Solutions blog, used the 1.9.1 Bossa program to burn version 1.23.2 firmware (a 164KB image named PanelDue-5.0i-7.0i.bin).

It appears to have worked (Bossa writes and verifies the image successfully), but the screen is blank. The LCD's backlight is on, but the screen is dark. The Erase button does nothing when pressed or held. The reset button appears to reset the USB comms. It behaves the same whether it's on USB alone, connected to a Duet, or powered standalone from a bench supply.

The Bossa program doesn't adapt to non-standard font size, and the UI is difficult to use as a result, so I tried Bossac at the latest release (1.7.0) but it fails on verify. Apparently, that's a known issue.

I tried the two most current firmware releases, both logo and nologo.

What else should I check?

Is there a more correct way to update the firmware?