Questions about umbilical + CANBUS on a v2.4

[YaaJ]

The first question is about the X and Y endstops.  Three options.

- sensorless homing. Not a great fan...
- keeping the "stock" XY endstops on the XY joint, with a skinny 10x10 drag chain
- relocating the X and Y endstops somewhere on the toolhead or the X carriage, and connecting them to the EBB36 dedicated endstops ports ; didn't find any example on the web...

The second question is about the umbilical. Wiring a twisted pair, or using CANBUS over USB ; not sure it can be done by Klipper on the RasPi, but there's a dedicated USB A port on the Octopus (also on the U2C I will do anything to keep out of this project). Tested Raspi to Octopus over UART at 2 Mbits/s, seems to be working, and should be enough...

Is there a reason why we don't see the EBB36 or 42 using the USB C port for CANBUS ? Moreover, USB C cables are rated for 48V/5A. Why don't manufacturers offer full CANBUS over USB C, including power for the hotend ???

I'm asking because decisions have to be taken for my build. A LDO v2.4. I'll probably have to design a breakout PCB for the umbilical, and I'd like not to forget some circuit or do something that cannot work !!! I'll be reusing the 14 wires from the toolhead harness, I'd like not to cut them to the wrong length for example. (BTW, de-pinning the connectors was much easier than expected)

[YaaJ]

I answer to myself...

Endstops : I read here : https://mpx.wiki/canbus_guide "X Endstop : From 2.4R2, vorondesign add X endstop mount at toolhead" WTF ? The 2.4 r2 manual describes the installation of "the pod" ; LDO kit comes with a pod PCB. The 2.4r2 assembly STEP shows the pod. The X and Y endstops relocation is undocumented.

Looking at the X carriage, there are two little holes, with the exact spacing for a standard microswitch, and a recess that matches the switch. Now have to slightly modify the right half carriage for a pair of M2 heatset inserts (where I live, self tapping screws are forbidden ; they even are anticonstitutional). Hopefully there's enough meat, or some can be added. Also have to make a mount for the A drive. (seen a couple available for download, now understand why !)

I thought the EBB USB C connector could be used for CAN over USB (like USB A connectors on the U2C) ; after some more RTFM, it appeared it isn't. It is regular USB. Meaning USB can be used for a virtual UART over USB, between the toolhead and the Pi. The way the Picobilical is connected to the Pi on the V0. Just another MCU to be declared. Definitively worth testing as it makes things much simpler, removing software + hardware layers. Have to find a good data USB C - USB A cable, with angled USB C. Also maybe make a breakout USB + Power on the A drive for easy replacement when it will fail. (USB cables being Murphy approved).

[ChicagoKeri]

I suspect that one could use the USB port on the CANBUS toolhead for data instead of CAN.  You certainly would have a much greater range of available USB cables to choose from than CANBUS  cables.

But getting the CAN bus to operate is probably the better option. After all, that is what it is designed for.  So far the CAN bus has been the most reliable part of my V2.4 build so far... being a pain to configure but very reliable and trouble-free once running. This is with the much-maligned U2C and three different CANBUS toolheads, 2x BTT and 1x Mellow FLY. The U2C was the least trouble to configure. 

I just use the standard endstop pod and full size cable chains which works well enough.  If I was doing it again, I would probably go with sensorless homing as the very rigid construction of a Voron makes it well suited to sensorless homing.. I'm surprised this wasn't common much earlier.

 

[YaaJ]

On 11/24/2023 at 4:45 AM, ChicagoKeri said:

I suspect that one could use the USB port on the CANBUS toolhead for data instead of CAN

Yes ! It is what I'm doing.

I found one example of someone using it this way (Reddit ?), answering he solved his problems with CANBUS just stopping using it, and connecting the EBB with USB.

There's no difference between the toolhead controller as a USB peripheral, and using two printer motherboards. After all, Klipper considers them as an extended GPIO, whatever is the communication infrastruture... (if I understand correctly Klipper !!!)

I'd be more interested in CANBUS if we really could use is as it is intended, chaining all devices (sensors and effectors), the industrial or automotive way (played a lot and had a lot of fun with its ancestor, MODBUS, using Arduinos, STM32s, variable frequency drive and MAX485 ICs). There's somewhere an example of someone doing this (more or less) by using a EBB36 to control a chamber heater or a exhaust fan (can't remember)

It seems BTT mechanically designed the EBB36 to also be used this way. Depending on how the PCB if bolted to the extruder stepper, it is either the USB port or the CANBUS port that will be oriented upwards. Pretty sure it is intentional. The Stealthburner was wired so the EBB36 will be used with USB in the first place (yesterday). You can see here that USB is in line with the extruder and the nozzle ; BTW, it is the simplest (just 2x 20 30 mm standoffs) and cleanest cable management that can be made, as well as the better driver cooling. I bet it is a optimal solution.

Looking at the schematics, we can see JP3, between VBUS (USB power) and +5V. A 24V->5V buck converter could be installed there, and 24V be injected at the other end... F2 D5 and C43 could have to be replaced or removed, shorted, etc.

I'll maybe order a pair of EBB36. One will be for the v0, and the second one as a guinea pig for 24V on USB C.

But for now, I've a little problem : the lack of room at the back, when the toolhead is at Z Y max : interference between the provided heatsink and the Z drag chain. Will probably have to machine a custom heatsink. Currently working on this issue...

[EDIT] just ordered two EBB36 ; black friday on Amazon, good pricing relative to what I payed for 2 months back. Still have to source a USB C breakout board (don't want to design aa PCB and solder such a connector on a DIY laser transfer)

[EDIT 2] oh dang ! Forgot USB-C cables that deal with high voltages / currents are active cables ! Not for tinkerers. The problem with the EBB mounted reversed and the heatsink is solved. There's so few copper to route that a 7x7 Z drag chain is more than enough. But milling a custom low profile heatsink still is an interesting project : a 1 mm or so plate, as large as the EBB, with a small island protruding 2 more mm, plus a 0.5mm or so silicone thermal pad. Also, there's a mistake : the standoffs are 30mm, not 20mm. Tested 25 and 27.5 : doesn't fill well because of the JST connectors.

Edited November 26, 2023 by YaaJ