Firmware 2.02 released!

dc42

G1 Z0 should place the nozzle just touching the glass if you have set the trigger height correctly in the G31 command. The printer will command the nozzle to the first layer height before it prints the first layer, for example it will send G1 Z0.2 if your first layer height is 0.2mm.

If you are using a config-override.g file, check that the trigger height isn't being overwritten by another G31 command in that file.

leif

Installed 2.02 today. Download and install of .zip was insanely easy. No issues here on a kossel mini with smarteffector. Homing et.al. worked as expected.

Thanks again for your hard work @dc42

Catalin_RO

Just upgraded my WorkBee to firmware 2.02 as I have to do some parts. There was no change required in the configuration files.

But, as I had to machine a cylindrical shape, I can clearly see a possible problem with the G2/G3 segment length correlated to the arc radius. The piece was milled at very low speed, so the cutting tool could not make such regular marks.

The whole GCode is made up of G2/G3 for the cutting. The circle radius varies between 15mm and 18mm and the facets are perfectly aligned between the two diameters.

Is there a way to control the relative segment length or chose between new and old behavior?

dc42

Perhaps I should change the firmware to reduce the segment length when the commanded speed is low. What would be a typical milling speed?

Catalin_RO

@dc42 It depends a lot on material and cutting tool. Standard WorkBee can't exceed 2500mm/min for screw driven, but it can be as fast as 10000mm/min for belt driven one. Also, CNC machining can have two passes (and most of the time it does so!) - roughing, when the bulk of the material is removed and quality is not important as there is some left over material for the second pass, and finishing, when just a fine layer of material is removed and the machining parameters are focused toward surface quality. The picture shows a piece half finished, with the remaining finishing to be done with a different setup.

P.S. Maybe adding some parameter to upper limit the segments length?

dc42

What would you consider a reasonable upper limit on segment length? 1mm?

Catalin_RO

@dc42 It depends on what needs to be machined. In the photo the facets are about 1.2mm wide. So 1mm would not be significantly different. Honestly I would consider something like 0.1mm a lot more suitable, but that is because the WorkBee accuracy is around that. If a CNC has much better mechanics, even 0.1mm might be too much. That is why the segment length upper limit should be configurable!

Right now I have an Al piece with corners machined with 3mm radius. I can clearly see the facets. A picture will come soon... once the machining ends and I can clean the Al part for a photo!

Catalin_RO

As promised
, a 3mm radius corner, with two facets properly visible.

dc42

Rather than making the upper limit configurable, I think it better to make it speed-dependent. The reason I made it radius-dependent in the first place is that users complained that in 3D printer mode, the maximum speed achievable for G2/G3 commands was too low. But speeds are much lower in CNC machines anyway, so I could leave the segment length at 0.1mm unless the requested speed requires a longer segment and the radius allows that without losing much accuracy.

Catalin_RO

@dc42 Sounds better! But it is important to know the threshold speed value. While roughing apparently is done at higher speed than finishing, it depends on many factors and it may be actually done at lower speeds (bulk material removal requires a lot more torque than just shaving off a thin layer). 4000mm/s is a reasonable upper limit for 0.1mm/s, if possible. The WorkBee that I have is capable of 2500mm/s, and it's not that rare that I need to set feed rates close to the top limit. If using larger cutters I would have to go faster, just that I know that the machine is not rigid enough for the associated torques and thus I don't go beyond a certain diameter.

DIY CNCs usually rely on lowest cost motion solutions - that is stepper. Considering a 300rpm limit for the stepper while still having meaningful torque, combined with the usual 1610 lead screw (16mm thick screw, 10mm lead), that gives 3000mm/min feed rate. That is why I consider 4000mm/min a reasonable limit for CNCs that use steppers rated for up to 3A. CNCs that use belts are imprecise at higher speeds because of the belts flexibility and the ones that use screws and can go faster have much larger steppers, impossible to use with the Duet 2.

Or make different rules depending on the controller mode if the threshold speed is too low - no more than or fixed at 0.1mm for CNC (whatever was before) and new algorithm for 3D printing.

Vlad

@dc42 have you stopped supporting DUETWIFI or something? I can't find any new firmware on that board, all firmware switched to DUET2WIFI, which is basically the same board, but the problem is that firmware won't install on DUETWIFI. Needless to say, that there is of course zero instructions about that.

Phaedrux

@vlad the duetwifi and duwt2wifi are one in the same.

Instructions for updating are the same.

https://duet3d.dozuki.com/Wiki/Installing_and_Updating_Firmware

Vlad

@phaedrux Thanks. Only if it was mentioned anywhere, as the firmware has different name and won't install, unless you rename it, which I did.

dc42

@vlad said in Firmware 2.02 released!:

@phaedrux Thanks. Only if it was mentioned anywhere, as the firmware has different name and won't install, unless you rename it, which I did.

I guess you didn't read https://duet3d.dozuki.com/Wiki/Installing_and_Updating_Firmware#Section_Upgrading_a_Duet_2_WiFi_or_Ethernet_to_firmware_1_21_or_later_from_1_20_or_earlier.

Vlad

@dc42 I figured that out on my own already. Not easy to find a documentation in your holly mess here David lol. I haven't seen such a mess since I saw my room when I was a kid. Impossible to find any required info. You should check adafruit website as a good example of how documentation should be organized. P.S: my board has failed, so I guess I won't need anything for a while.

dc42

And you should learn to use the search facility. Duet+RRF is 100x more complicated than anything Adafriot sells.

zerspaner_gerd

M291 no longer allows S0 and T0 together because that would create a message that can never time out of be dismissed

@dc42

I used it as follows

M291 P"Nozzle is heated" R"Loading Support" S0 T0

G10 S230; TEMP

G28

G1 X125. Y10. F6000

M116 H1:2 S3
M292 ; Hide the message
......

I see now no way to implement the equivalent after the change.

Maybe we could create a new mode that includes:

not blocking
Only with the "Cancel" button, to cancel the current process.

That would have been missing before me, because if there is a mistake (for example, unintentional confirm) was the one zigste way to end this with an emergency stop.

What do the other people mean?

Mugga

Thx very much for the release sadly I got massive wifi connection issues with this release.
Sometimes 2/10 I would say, I'm loosing connection during the print.

But I'm loosing always connection when idle for some time. I thought this firmware release fixed the reconnect issue?

Any ideas how to get around this?

Firmware Electronics: Duet WiFi 1.02 or later
Firmware Version: 2.02(RTOS) (2018-12-24b1)
WiFi Server Version: 1.22
Web Interface Version: 1.22.6

dc42

@zerspaner_gerd said in Firmware 2.02 released!:

M291 no longer allows S0 and T0 together because that would create a message that can never time out of be dismissed

@dc42

I used it as follows
M291 P"Nozzle is heated" R"Loading Support" S0 T0

G10 S230; TEMP

G28

G1 X125. Y10. F6000

M116 H1:2 S3
M292 ; Hide the message
......
I see now no way to implement the equivalent after the change.

Maybe we could create a new mode that includes:

not blocking

Only with the "Cancel" button, to cancel the current process.

That would have been missing before me, because if there is a mistake (for example, unintentional confirm) was the one zigste way to end this with an emergency stop.

What do the other people mean?

Why don't you just set a long timeout instead of no timeout? Or allow it to be dismissed?