@Forum-user we've looked into this and the limiting factor is that the MCUs we use on expansion boards only provide a single quadrature decoder. We've looked at using an external circuit in combination with a counter on the chip to provide a second decoder but that wasn't possible to do with sufficient performance.

The MCU we use on the 6HC and 6XD main boards can provide 2 or 3 quadrature decoders, and there may be STM32 processors that can too; so we'll keep this under review.

Another possibility would be to use linear axis encoders that support BISS protocol instead of quadrature.

I'm still curious if there is any chance of this happening.

If no official release from Duet is forthcoming, is there anyone in the community who can help me create this? I'm willing to do the design work, but I need guidance on how to proceed, as I have no experience with designing electronics.

@fcwilt the motors run better at 60V when you push it

@T3P3Tony true, the board being 24V max is what keeps the size down (and the board is pretty big already)

one solution might be a double sided PCB and perhaps moving away from the "sticks to a stepper" design (which is weird anyway since most of the nice new extruders use pancake steppers anyways...)

This thread was useful when debugging some issues I was having using this exact pendant with the Arduino Nano as the microcontroller. I'm not making use of the PanelDue passthrough, just the pendant straight to a Duet3 6XD. For some reason when using the resistors rather than a level shifter on TX to bring it down to 3.3V, I was having issues with awful noise between the 6K8 resistor and the Duet 3, both when transmitting anything on TX and when just sitting idle at 5V. The resistor was on the pendant side of the long cable, and the noise was present on both the Duet and pendant sides. I spent a while figuring out what I'd done wrong but wound up using a level shifter (can't find a receipt but it looks the same as the 4 channel variant here) on the Duet side instead and that worked.

Perhaps it was due to me leaving the Nano unmodified, but my interpretation of the instructions here was that that was only necessary when using the PanelDue passthrough.

@TLAS the standard mini thermocouple connector is nearly 17mm wide, so fitting two of them on the thermocouple daughter board would not be feasible.

There are some smaller thermocouple connectors available but I haven't seen any in widespread use. Most thermocouples I see advertised are wire ended, not terminated in a connector.

The Duet3D daughter board works with several different types of thermocouple, not just type K. We have users who extrude very high temperature materials and use other types, and for them including a type K thermocouple connected would be a disadvantage.

So in short, we don't have any plans to do this. The screw terminals seem to work well with any type of thermocouple. If you connect the wires the wrong way round then it is quickly obvious, because the temperature reading goes down when the heater is turned on.

We could do a custom run if sufficient quantity was requested. The minimum quantity that was economical to make would probably be a few hundred. Alternatively you could do your own variant starting from our open source design files.

Hi everyone,

Just wanted to let the Duet team knows that this is also something of interest to us.

But we also are at an early stage of our work regarding this matter.

@Th0mpy the 3.5-dev branch of the Duet3Expansion project includes support for the RP2040.

@dc42 said in Support forPWM Conversion:

@jay_s_uk RS485 is something we may support in future. We're already looking at making a daughterboard to support it.

FYI
Supporting RS485 also opens the door to controlling the output of a plasma cutting system from the cnc controller.
Here is the programmers refererence for Hypertherm RS485 implementation

@Diamondback not yet...

@Marcossf I don't think an active tool distribution board of the sort you envisage is possible without radical changes to the CAN protocols that we use. The CAN bus does not support man-in-the-middle devices, because the ACK but has to be set in real time by the receiving nodes.

We are looking into integrating RFID readers into Duet 3 systems. One possibility is to connect a RFID reader with UART interface to a UART-capable port on the Duet. Another is to connect it to a UART port on a custom expansion board based on a SAMMYC21 or possibly RP2040. If several RFID reads are to be read (e.g. one per tool dock) then RS485-enabled RFID readers could be used.

@sonderzug et al

We have been considering a SAME70 CAN-FD control board designed to use CAN-FD expansion for most functions for a long time now, but need to consider the total stock of these processors available to use before adding another product that uses them.

If we did it Pi hat is one form factor that could be used but may not be optimal.

@ctilley79 just a thought ... if you are good enough to solder on wires, maybe desoldering the existing connectors would be an option for you.

@jrow I know an electronics tech whose day job is fixing production errors in dense PCBs that hires out her spare time. I'd love to mod a 6HC to 50 or 60VDC.

@3mm the BMG with a pancake nema 17 (in my case a Moon's) weighted in about 290g (75g for the BMG, ~200g for the motor, plus bowden coupler, silicon hose, bowden tube, ..).

The orbiter is 135g, so less than half the weight. It however isn't the weight on the flying extruder I'm concerned about, it's the yanking of the bowden tube (in my case was ~150mm) on my 380mm bed delta that I suspect is the source for artifacts I saw.

I can run 5k acc/600jerk/no input shaping/100mms-150mms, 0.015 PA, 0.6mm retractions with the direct mounted (on a 7cm tall, petg printed stand and tilted) Orbiter V2 with almost none / negligible artifacts.

I wasn't able to do that with my flying BMG.. had to run 3k/300jerk/~70mm/0.4 PA/1.5mm retractions for worse quality than on the current direct drive solution.

@nikscha I think it could be useful for some of those things but I would caution against replacing the strain gauge as the has been tried before for Deltas with the HE280 hotend from SeeMeCNC.

The strain gauge is more accurate and has worked better than accelerometer only. It could be great combined but I wouldn't go entirely over.

@pcr I think you still should have a diode on the input right?

Looking for a single header that will power the inductive probe and accept the signal generated. As well, it should be flexible to offer a few different voltages. I thought that a diode would be required to fully satisfy all the requirements.

@dc42

To be honest: Nothing will prevent me. This is my current setup. But see it like that:
3 wires for X and Y endstops
8 Wires for the two steppers
3 Wires for my enclosure temp sensor
2 Can wires (to the toolboard on the hotend
2 Power for the toolboard
= 18 Wires in total.

That can be reduced to 4 fires from the Mainboard to the gantry (the new tool board) and 4 wires forward to the toolboard on the hotend.
But this is not my only usecase, see things like the carrot feeder which need more than one stepper. And the 3HC is really a big boy for driving two nema14 plus a servo at the end.

I had the idea to bild a "poor door opening thingy" for my v2. My "brain internal" solution would need to steppers, three sensors too.
I could think about a camera moving thingy, like a gimbal which keeps the focus on the hotend. Which would need more than one servo too.

I think that there would be a lot of usecases for a "intermediate" board. The 3HC fits perfectly into the 6HC range. They are perfect if you need big beefy boards with a lot of power. But the Mini 5+ road seems to be a bit under supported here.

@o_lampe brought up the next good usecase: An IDEX printer can profit from such a board too. You can drive X and U with that board and save a lot of wiring

Do not get me wrong David, I understand fully that you can not create a board for every single usecase which somebody brings up. That is logical and can not be done by your time and by the financial impact. But see it from my site:

I buy a Mini5+ for some good reasons: The price and the almost "best" solutions for printers like a IDEX, Voron v0/2.4 etc. The steppers are fine and big enough, I can attach a toolboard or two. Wifi included. A very nice board for a fair price.

I attached the 2 driver daughter board to the Mini5+ because I need 4 steppers for Z and 2 for X-Y. I have one left. (Toolboard on the hotend)
I have now the case that I want to build something like the carrot feeder. What should I do? Use two toolboards? Use one stepper from the expansion and add an toolboard?
Or shall I use the 3HC? That is almost the price of a Mini 5+? That does not feel very right to me to be honest.

And I can not be the only one with this kind of struggle ...

Cheers, Chriss