Posts made by Thorpydo

@deckingman
Could you please provide a link or product name so that we can take a look at what you are referring to for your alternative material under heated bed?
Thanks!

@dc42 Great, thanks for the clarification!

I am in the process of designing and laying out the wiring for a printer that will have multiple Duet3 mainboard 6HC's (5 in total) used as expansion boards.

I would like to understand if any consideration should be given to which components are wired to which mainboards.
IE does it matter which mainboard is used for:

• bed heater output
• bed heater thermistor
• hot end outputs
• hot end thermistor
• homing switches

I would like to optimize for low CANFD bus load and therefore, my thinking is to keep as many components on the main mainboard as possible. Homing switches will only be used during startup when bus load will be load, so shouldn't matter which mainboard wired to.

I don't have a good grasp on what can be sent in a single CANFD message. If commands for all steppers and all heaters outputs can be sent in a single message, perhaps it's only important to keep the thermistors wired to the main mainboard to avoid a message from thermistor mainboards back to main mainboard.

Is my thinking correct here? Please ground my thinking and provide any insight you may have.
Thanks for your help and a great community!!

@dc42
Thanks, that helps.

I think I will use something like these RJ11 phone splitters to keep the junctions easy, clean, and (maybe shielded?)
https://www.amazon.com/Phone-Splitter-Telephone-Suitable-Machine/dp/B0B1D9775P/ref=sr_1_19?hvadid=409957000450&hvdev=c&hvlocphy=9031520&hvnetw=g&hvqmt=b&hvrand=13040326026356310989&hvtargid=kwd-11735544458&hydadcr=19108_11276318&keywords=rj11%2Bline%2Bsplitter&qid=1689094335&sr=8-19&th=1

@kuon
How has your project turned out? What did you do here? Any issues with CAN communication?

@T3P3Tony
As I understand, 2 conductors will need to go to each pin for CAN1_L and CAN1_H for the middle main board(s).

I've created a diagram as I understand the wiring should be. Is it correct?
How is it recommended to make the junction? Insert two wires into the RJ11 connector for each pin? They are made for 1... seems like it will be wonky or not work when inserting two.

https://imgur.com/a/CbCiCU1

(if it's helpful for reference, cross posting from https://forum.duet3d.com/topic/31190/increase-maximum-number-of-axii-to-30)

@Thorpydo
Hmmm, the RJ11 connector is made to accept one wire per pin, and it gets inserted instead of crimped.
How best to handle the junction here?

@DIY-O-Sphere
Perfect, that's really helpful.

It looks like my understanding of the CAN connection was incorrect (two separate buses vs two connections for one bus).
I've created a diagram to make sure I understand how to correctly wire.

https://imgur.com/a/CbCiCU1
Is it correct?

Thanks for the help here!

@dc42 said in Increase maximum number of axii to 30?:

When a 6HC is used as an expansion board, a single CAN message contains the movement information for all 6 drivers. So only 5 CAN messages would be needed per move even if all 29 motors move at once.

After reviewing the hardware configuration documentation and CAN bus interconnects, I was thinking that it would be easier to use x1 mainboard 6HC and 8x expansion board 3HC's because each expansion board comes with 2x RJ11 connectors and termination resistor is straight forward.

However, after re-reviewing this thread and DC42's comment above, it sounds like it I should optimize from a bus loading perspective and keep the number of boards to a minimum, thus 5x mainboard 6HC's, using both CANFD-0 and CANFD-1 connectors, and bypass the termination resistor on the middle 3 mainboards; first and last mainboards using CANFD-1 connector.

I see in the revision notes for mainboard 6HC v1.02a:
"improved the CAN-FD termination resistor bypass solder jumper design to make it easier to use the 6HC as an expansion board."

Where do I find documentation or detail on bypassing termination resistor for the mainboard 6HC v1.02a?
Does this all sound appropriate?

@o_lampe
Thanks!
I might have to clarify how to use the M595/look ahead buffer when I'm ready for that.

Currently, the slicer outputs extrusion in absolute mode but I would like to change to relative extrusion mode. It looks like that will be no problem and I can use M83 in start Gcode to do so.

Here's a single line gcode output; I believe this matches what has been discussed and aligned on in this thread:

G1 X556.90665 Y559.4645 Z559.4645 U556.90665 V564.15183 W564.15183 a556.90665 b559.4645 c559.4645 d556.90665 e564.15183 f564.15183 g556.90665 h559.4645 i559.4645 j556.90665 k564.15183 l564.15183 m111.49454 n113.41718 o113.41718 p111.49454 q111.36289 r111.36289 t-14886 E0:2712.35717:1611.15904:25136.476 F250

Thank you all for guiding me here!

@droftarts
Perfect, that jives with above. Thank you!

@o_lampe
I might not have made it clear before:
All motion systems and all extruders moving at the same time. Multiple extruders, moving independently, extruding at the same time. Four heaters, four extruders, four filaments being driven, etc. (all at the same time)
I understand there is currently no slicer available to be able to control what I am asking.
Writing the slicer is part of my scope of work. I am writing the slicer and that's why I need to understand explicitly the proper way to communicate with RRF and command with Gcode, all four extruders, and all motion axes, simultaneously.

It's still not clear to me if format should be:
G1 Xn Yn En0:En1:En2:En3
Or:
G1 Xn Yn En0:n1:n2:n3
(where n0, n1, n2, n3 represent feeding material in mm/s)

Based on your input, it sounds like the later?
G1 Xn Yn En0:n1:n2:n3

@lampe
Okay, thanks, I now understand nomenclature should be E0,E1,E2,E3.

What is the definition for a tool, as defined by the firmware?
Is the tool considered the XYZ motion system or is the tool considered something that can be attached or interchanged to the motion system?

@deckingman
It's great to hear that there shouldn't be further issues here; that moving multiple extruders already does work.

For this application, I don't want any mixing ratio applied.
I believe I want 4 tools, each with its own nozzle.

"... if the G1 command is in the format En1:En2:En3 then each extruder will move by those n1, n2 and n3 values ..."
Great, this is the desired output.
As I understand, Gcode output should be G1 Xn Yn En0:En1:En2:En3

@o_lampe Ahh, gotcha. Okay.

My modified understanding:

For extruders:
E1, E2, E3, E4
For motion:
X,Y,Z,U,V,W,a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p,q,r,t
(motion could use other lower case letters or A,B,C,D, but total number of axes cannot exceed 32 30 based on some internal limit)

Thanks for clarifying!

@o_lampe Hah! You and me both!

@dc42 I'd like to make sure I understand the appropriate letters to be used.
My understanding is not adding up:
all lower case letters are available; 26 letters.
X,Y,Z,U,V,W,A,B,C,D uppercase letters are available; 10 letters.
It seems this nomenclature allows for 36 letters which doesn't jive with your comments of 30/32.

I would like to use:
For extruders:
A,B,C,D
For motion:
X,Y,Z,U,V,W,a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p,q,r,t

Will this letter assignment be acceptable?

@dc42
That's wonderful to hear David!
I really appreciate your support in accommodating this feature so promptly.
I understand the change is still experimental and performance might be impacted.

I still have some work to do with the slicer and hardware, but will report back when I have some meaningful results and can comment on any performance impact.

Thank you! You guys rock!

Okay, I see.

Perhaps then, we can determine a strategy and all sections of code that would be need to be modified to increase the maximum number of axes past 30.

For reference, the same topic has been discussed previously in these threads:
https://forum.duet3d.com/topic/16023/maximum-number-of-axis
https://forum.duet3d.com/topic/26493/axis-limitations

In reviewing the previous discussions, lowercase letters were added to give additional axes. Since the software is not able to differentiate between upper and lower case, a single parenthesis is used to signify lower case, '.

In the same discussion, DC42 suggests that Q could be used in the same way as '.
"because currently letter Q isn't used in any commands that accept axis letters."

So, perhaps a simple approach could be to use both ' and Q, instead of just '.
This approach would yield 32 unique drive letters:
X,Y,Z,U,V,W,A,B,C,D
abcdefghijkl ('A, 'B, 'C, 'D, 'E, 'F, 'G, 'H, 'I, 'J, 'K, 'L)
QX, QY, QZ, QU, QV, QW, QA, QB, QC, QD

Can we simply replicate the same code that is used for ' to also work with Q?
Are there any potential 'gotchas' or problems with the suggested solution?

@o_lampe If possible, I'd like to avoid a custom hardware solution, which I assume the PLL clock generator approach would require. Is it possible to buy off the shelf components to make this work?

What all would be required to drive multiple main boards from a SBC with RealTimeClock?
This approach seams like it could be a good solution.

I'd like to avoid solutions that sync periodically per layer or otherwise, if possible.

@dc42 What is the limitation in supporting 30 axes with multiple main boards ganged together? It sounds like bus load is not an issue. I believe I read in a previous thread that naming convention was the primary issue but solutions were discussed like using an upper and lowercase letter in conjunction.

Thank you both for giving this some thought and providing input. I really appreciate it!

@dc42 yes, for independently running 6HC boards, they would need to remain perfectly in sync.

yes, pausing the print would require all of them to pause.

@dc42 yes, the 29 axes are for multiple 6 DOF motion systems and extruders. Yes, they will all be moving at the same time.

Great to hear that bus congestion won't be a problem.
Which do you think is the better or more feasible approach?

An alternative approach might be to have 4 independently run MB6HC boards with 4 separate gcode streams, 7 and 8 axii each.
Doing so would alleviate CAN communication congestion, controller load and enable infite axii to be added.
To enable this approach, each MB6HC would need to be able to be start the print at exactly the same time.
Is it possible with existing feature set today or would it need to be added?