Duet 0.6 Setup questions

[DeltaCon]

I was able to get a Duet 0.6 for a reasonable price and have been "dry testing" it, without any printer actually connected.
It looks to work okay so far and the web interface is up. I updated it to DC's 1.16 (the latest stable) firmware.

- Since the update (needed a press on the Erase and the Reset buttons, after that I used Bossac. The Board works, the web interface works. But it strikes me that the Duet now manifestates as "Serial USB Device" in Windows Device Management, instead of "Duet - something" like it was before. Also it seems to shift up a Serialport each time, is that necessary?

- I read that since the recent versions the firmware can be updated through the web interface. But I can't find any guides for that.

- I In the guide here: http://reprappro.com/documentation/comm ... e_Firmware it says that the SD card files must be update to the same version as the bin. I did find the WebControl zip, but cannot find anything else. Shouldn't I update the other (SYS?) dir also

I am waiting for some connectors to fly in, so I will have to rest my experimentation for now.

[CRPerryJr]

Most of the info I needed to update my Duet was located here:
https://duet3d.com/wiki/Duet_Wifi_Wiki# ... et_WiFi.29
and here:
https://duet3d.com/wiki/Where_to_get_Du ... _and_tools

- I read that since the recent versions the firmware can be updated through the web interface. But I can't find any guides for that.

The firmware update helper file (iap.bin) is available at https://github.com/chrishamm/DuetIAP/tr ... er/Release
Place the iap. bin file into the SYS directory of the SD card, reset, and then you can update the firmware and DWC through the DWC.

Hope this helps a little...

[Polygonhell]

I would tend to agree the how to update firmware from the web interface is far from obvious until you've done it once.
Basically you just drag and drop the firmware file and the config files into the web interface, I think you can also use the upload button.
Mine hasn't been connected to the USB Port since I did the initial firmware update.

[DeltaCon]

Thanks guys. It is probably because the upload button does not give a clue on what kind of files to expect, and the term "firmware" is left out of the description. It's only talking about /sys and /www files. But you can as easily drop a .bin on there as it seems

Mine hasn't been connected to the USB Port since I did the initial firmware update.

But do you still use a display? I did not go for the paneldue, because the of pricetag, and I hear many people running duet solely on the webinterface. I probably have some older tablet lying around to mount to the frame as display

[Polygonhell]

I have a panel due, but it usually sits out of the way where I can't actually see it. Occasionally I'll use it to preheat and start a print I uploaded previously, but my usual workflow is

turn on printer
connect with web interface
set print temperatures
Slice object
upload object
wait for the preheat to finish if it hasn't already
Start print
close web interface

I use the Web interface almost exclusively, I'm usually near a computer, so if I want to check on the print I just pull up the web interface
But I rarely check on things mid print unless something fails.
Most tablets don't present the Duet web interface very well, it has exactly the wrong aspect ratio.

[mhackney]

The upload button can upload the .bin (firmware) files, .g (configuration) files - anything that goes into either /sys or /www including firmware itself. The UI could use a bit of clarification.

Like Polygonhell, once I started using the DuetWebControl with firmware update capability, I haven't connected my Duets to USB at all. I had to do that yesterday to prepare a special DuetWiFI and PanelDue for "someone special" and I had to go back to my notes to remember how to do it.

[DeltaCon]

I am studying the pinlayout of the Duet 0.6. I thought there would be an abundance of data available on the web about this, but it seems the duet only got popular after the release of the 0.8.5 version.

I understand there is only one controlled fan output, labeled as FAN0. My guess is that this one is to be used for the layer fan(s), and the peek-fan should be connected directly to the 12V PSU rail (think I'll attach a connector to the 12V power in screw-terminal). Is that correct? Is there an easy way to make the peekfan thermo-controlled outside of the bed? I'd like to get it to go off after dropping below 50C like the rambo does.

[dc42]

The best wiring diagrams for the Duet 0.6 are probably the ones at reprappro.com in the documentation for the Ormerod. The 0.6 has only one controlled fan output. You can pick up 12V from the double row connector next to the motor connectors.

I've had it in mind to support a second controlled fan on the 0.6 in firmware using an external mosfet, but not got round to implementing yet because you are the first person to ask for it.

[Mac The Knife]

Doesn't the RamBo use an extruder "channel" to control one of the fans?

[dc42]

Doesn't the RamBo use an extruder "channel" to control one of the fans?

Maybe, but the Duet 0.6 has only one extruder heater output. The later Duet 0.8.5 adds a second extruder heater output and a second controlled fan output. The Duet WiFi adds a third controlled fan output.

[Polygonhell]

Doesn't the RamBo use an extruder "channel" to control one of the fans?

That's an artifact of SeeMe's firmware copying the initial configuration work I did with Repetier years ago.
When I was wiring the hotend fan, the wires I had would not reach the second fan output without some messing around, so I set up the firmware to use the second extruder output. On the RAMBO the only difference between the Extruder and Fan outputs is the label on the Board,. The Duet Fan outputs are not the same as the extruder outputs, although if it had multiple heater outputs they could be used.

[dc42]

There is a thread on adding a second controlled fan to a Duet 0.6 here https://www.duet3d.com/forum/thread.php?id=630. Basically the same processor pin that the Duet 0.8.5 uses to control the second fan is brought out to the expansion bus on the Duet 0.6. So you can add an external resistor and MOSFET to drive the second fan.

[DeltaCon]

Still no connectors for leads, so I cut off everything and started knitting it to screwterminals as they are. I was already able to do a 7-factor calibration, so motion is all okay. Still a few questions though. But first a picture of my Max on the IC of my office

Looking through the config.g i see values for the stepper current are expressed in milliamps. How does this relate to the values most of us are used to in the Rambo firmware. I am using the original Wantai motors the machine was send out with by SeeMeCNC. The older SeeMe firmware worked at a value of 165 (1-255). For the more recent firmwares I use a percentage value of 65. But what would be comparable values for the DC42 firmware? It is working at 1000mA righht now, seems to be good but I'd like more detail.

I have been trying an autotune on Heater1, but I get a respons like "Auto tune cancelled because starting temperature is not stable". I don't know if it is normal, but the temp value in the GUI is fluctioating about 1,5C. My thermocouple shows a pretty stable temp with no fluct at all. If I engage the heater it seems to heat up pretty normally, and I do get 12V-ish at the heater connection. Why is my thermister so jumpy, and could that be the cause of this error?

Few days ago we debated the use of dead-time control, now I learned that there is no such control in DC42's firmware. But the default bang-bang works great on my setup, with a temp fluctuation of not more than a degree C. So that is as neat as dead-time did on the rambo. Some however say it is better to switch to PID, even on a SSRed bed (that cannot give partial current). My PSU however has a fan that starts without temp controll. It just blows above a certain current (in practice, only when the bed heats). I think PID would wear out that fan quickly because of the quicker switching rate. Does that make sense or am I over concerned about that?

I can't seem to get any extrusion done. The extrude and retract buttons in the webGUI are Greyed Out. I suppose I need to heat first to above a certain temp?

Thanks again!

[dc42]

Looking through the config.g i see values for the stepper current are expressed in milliamps. How does this relate to the values most of us are used to in the Rambo firmware. I am using the original Wantai motors the machine was send out with by SeeMeCNC. The older SeeMe firmware worked at a value of 165 (1-255). For the more recent firmwares I use a percentage value of 65. But what would be comparable values for the DC42 firmware? It is working at 1000mA righht now, seems to be good but I'd like more detail.

According to this post http://forum.seemecnc.com/viewtopic.php?t=7239, RAMBO 135 = ~0.75A, 185 = ~1A. If that is correct, 165 would be about 900mA.

I have been trying an autotune on Heater1, but I get a respons like "Auto tune cancelled because starting temperature is not stable". I don't know if it is normal, but the temp value in the GUI is fluctioating about 1,5C. My thermocouple shows a pretty stable temp with no fluct at all. If I engage the heater it seems to heat up pretty normally, and I do get 12V-ish at the heater connection. Why is my thermister so jumpy, and could that be the cause of this error?

My guess is that you have one of the older Duet 0.6 boards with 1K thermistor series resistors, and a low ambient temperature. The 1k series resistors give poor accuracy and resolution at low temperatures, which is why they were increased to 4.7K in later production. Unless you can change the resistors, try putting the printer in a warmer environment while you do the auto turning. 20C is probably high enough.

Few days ago we debated the use of dead-time control, now I learned that there is no such control in DC42's firmware. But the default bang-bang works great on my setup, with a temp fluctuation of not more than a degree C. So that is as neat as dead-time did on the rambo. Some however say it is better to switch to PID, even on a SSRed bed (that cannot give partial current). My PSU however has a fan that starts without temp controll. It just blows above a certain current (in practice, only when the bed heats). I think PID would wear out that fan quickly because of the quicker switching rate. Does that make sense or am I over concerned about that?

It depends on how fast the fan control circuit reacts to changes in current. If the fan sounds OK, it probably is.

I can't seem to get any extrusion done. The extrude and retract buttons in the webGUI are Greyed Out. I suppose I need to heat first to above a certain temp?

Yes, or enable cold extrusion by sending M302 P1. Also you need to have a tool selected so that the firmware knows which extruder to drive, so you may need to send T0.

[DeltaCon]

My guess is that you have one of the older Duet 0.6 boards with 1K thermistor series resistors, and a low ambient temperature. The 1k series resistors give poor accuracy and resolution at low temperatures, which is why they were increased to 4.7K in later production. Unless you can change the resistors, try putting the printer in a warmer environment while you do the auto turning. 20C is probably high enough.

That is correct, i do have the 1K version, and changed the config.g accordingly. My ambient temp is about 25C, so that should be warm enough though. My thermocouple is rock stable, but the thermistor value jumps a bit. What can I do about that? Preheat to 30?

Thanks for all the other answers. Will check tonight!

[DeltaCon]

I just tried a pid tune preheated to 30C using M303 H1 P0.5 S240. That is working. First I got an error about reaching the max temp. I lowered the P to 0.3 and with that it reaches a max of 202.1C. I copied the resulting M307 H1 A615.2 C174.3 D9.1 B0 to the config.g I guess the PID values are calculated from there? Now the hotend temp is pretty stable, and the thermistor reports within 0.6C correctly. I think I will leave it like this

Autotuning the bed is a nogo I think, Th 20V SSR setup only goes flat-out, so it will always hit the max I think. It looks pretty stable as it is though. I find that surprising because there is really no PWM commands in the config.g at all. Are there already some sensible defaults that work for onyxes? This autotuning is really working differently than for the Rambo...

[dc42]

I just tried a pid tune preheated to 30C using M303 H1 P0.5 S240. That is working. First I got an error about reaching the max temp. I lowered the P to 0.3 and with that it reaches a max of 202.1C. I copied the resulting M307 H1 A615.2 C174.3 D9.1 B0 to the config.g I guess the PID values are calculated from there? Now the hotend temp is pretty stable, and the thermistor reports within 0.6C correctly. I think I will leave it like this

Autotuning the bed is a nogo I think, Th 20V SSR setup only goes flat-out, so it will always hit the max I think. It looks pretty stable as it is though. I find that surprising because there is really no PWM commands in the config.g at all. Are there already some sensible defaults that work for onyxes? This autotuning is really working differently than for the Rambo...

Auto tuning the bed should be possible. The firmware uses slow (10Hz) PWM for the bed and chamber heaters so as to be compatible with all types of SSR. However, you can estimate the bed heater model parameters manually if you prefer, see https://duet3d.com/wiki/Tuning_the_heat ... s_manually.

PID parameters are calculated from the model parameters unless you override them with M301 commands. If you run M307 H1 then it will show the calculated PID parameters as well as the model parameters.

[DeltaCon]

Thanks DC, but I am confused. While tuning the hotend, I learned that (other then with the Rambo) you reduce power to the heater until it cannot heat anymore. Calculated PID values are based on the power, and the time it needs to go to the max temp. Since the SSR always gives 100% power in an open state, even though PWM dictates e.g. 50%, My Onyx will always go beyond what is considered healthy for it, no matter how much power I base the tuning on. Tell me where my logic is flawed

[Xenocrates]

Thanks DC, but I am confused. While tuning the hotend, I learned that (other then with the Rambo) you reduce power to the heater until it cannot heat anymore. Calculated PID values are based on the power, and the time it needs to go to the max temp. Since the SSR always gives 100% power in an open state, even though PWM dictates e.g. 50%, My Onyx will always go beyond what is considered healthy for it, no matter how much power I base the tuning on. Tell me where my logic is flawed

This is a issue of not applying pwm. If I apply full power for one second, then half power for 2 is equivalent. Pwm works on that, by switching fast enough you barely notice. I'm pretty sure that the MOSFETs on the duet are never run in the switching zone as it generates much more heat and would reduce the power the board could handle. All the peaks are the same height, it is the number of them or their width that matters for power delivery

[DeltaCon]

I understand what you are telling Xenocrates. In the PID tune for the hotend I had to set the P-parameter to 0.4 for the hotend not to exceed 240. In the case of the bed, due to the SSR, I cannot set the P parameter any why that it works, because the SSR is just closed or not. I understand that PWM can choose short power bursts instead of half power, but I do not think the autotune is going to work when the heater is setup this way, Do you have any idea how hot it will get at max without any power reduction? (20V / 29A). I don't think I want to find out...

Duet GUI tells me this:

Code: Select all

21:17:15	Auto tune of heater 0 with P=1.00 S=100.0 cancelled because temperature limit exceeded. Use lower P or higher S in m303 command.
21:13:02	M303 H0 P1 S100
Auto tuning heater 0 - do not leave printer unattended

It reaches 100C in just over 4 minutes during a tune. I would not want to to try what the max reachable temp is

EDIT:
To amuse myself I tried P0.3 in the above command. I thought that would not work, but it seems to alternate power on/power off like Xenocrates mentioned. So the PID tuning for the bed works differently than for the HE, because there it will just partial current to the heater. So I will let my bed cool down and give it another shot. Thanks!

[dc42]

PWM works by turning the heater fully on for part of the time, and fully off for the rest of the time. That's what the MOSFETs on Rambo and Duet do, for the hot end heaters and for the bed heater. It makes no difference whether you use a MOSFET or an SSR to do the switching, except that some types of SSR can't switch very fast. That's why we use a low PWM frequency for bed and chamber heaters.

[DeltaCon]

Thanks DC42 for your patience towards my pre-assumptions.
Sometimes we say: "assumptions are killing"

The board is working nicely now, and I am ready to take it inside of the base. Yesterday I printed a 20mm testcube in ABS. That looked already better than anything I printed before. Can't wait to get the news skates and balcup arms installed, I hear that is really a big improvement too.

I have an issue with KISS that causes the extruder to go bezirk and extrude way too fast. At first I thought that is was a firmware thing. But I tested the same cube sliced in slic3r and that goes fine. Odd, because I used the same settings on the Rambo before. But that's something for a new thread...