Mini 5+ with 3HC Expansion for XY steppers in a CoreXY printer?
I am building a CoreXY printer and would like to use a Duet 3 Expansion 3HC to drive large XY steppers. I'd prefer to use the 3HC expansion with Duet 3 Mini 5+ rather than the Duet 3 6HC because I'd like to have built-in Wifi and I only need two high current drivers.
I am new to Duet hardware. So I am asking: Is there any reason that this configuration wouldn't work? I saw something about CAN bus latency but I don't know how much it might affect my application.
If this is a bad idea, I'll buy the 6HC and a long Ethernet cable.
@mjwaters There is no significant difference between step rates achievable between the 3HC and the 6HC. If the combination provides you all the IO you need an motor connections then it will work well.
Wifi vs Ethernet depends on your wifi signal and contention from surrounding networks. the Wifi modules used on Duets are 2.4Ghz (not 5 Ghz) and deal less well than modern phones and laptops with congested environments and weak signals.
@mjwaters yes that should work. Two points to note:
- Stall detection is not yet implemented on CAN-connected boards, so you will need to use endstop switches for the X and Y axes
- Currently it's best to drive the Z motor(s) from the main board, to get good results from a Z probe
See https://duet3d.dozuki.com/Wiki/Duet_3_firmware_configuration_limitations for the current firmware limitations when using CAN-connected boards. We are gradually removing most of these limitations.
That's an excellent point since I am in a WiFi congested area.
- Limit switches were always part of the build plan since nobody here likes the startling noise of stepper stall!
- Not a problem.
Do you think the CAN latency will affect the XY accuracy if I use the 3HC expansion? What is the latency? from a print speed, we can estimate a position latency, right?
BTW, the RatRig team also is curious my proposed configuration so I'll be forwarding them the details.
@mjwaters the moves are synchronised with the clock on the mainboard by recovering the timing signal from the can bus. they are sent ahead of time and then step generation is done on the expansion board and mainboard off the can bus clock signal. The CAN bus limits how many different boards can be addressed and what data rate, but not the synchronisation of the steps in a move in a latency sense.
That's very cool and sounds like it could have taken a fair amount of time to implement.
Alright, I am going to put in an order with your resellers today.
Thanks for all your help!