Wow! That sounds like the beginning of the personal computer days at the Homebrew club, which I was not a part of. I didn’t assemble my 1st PC until ~ 25 years later. 👍🏾👍🏾👍🏾👍🏾👍🏾
Thanks for the video. I wasn't finding a good explanation of some of this stuff elsewhere, namely what to do about the endstops. I think I will give sensorless homing a try since I dont want to print anything just yet.
Thank you for a clearly explained video. As someone who has tinkered my way through an Ender 5 Pro. I am looking for clear information to expand my skills and knowledge. I am probably going to build a Voron 2.4 due to the amount of support available but I prefer the design of the VZbot.
Just an FYI, I just picked up the Mellow FLY 36 and after it arrived, they released a new one, so typical, it has a better power connector and magnetic encoder built in, which I assume is for fine extruder motor control.
Thanks for putting this video out. I would be careful when classifying an option as "traditional" especially when referring to a USB to CAN adapter. I would also mention that you missed the original method of can bus for the RPI, and that is using a CAN hat for the RPI, which uses a SPI bus to talk from the RPI to the CAN hat. I would also raise caution for the type of can bus being deployed. The FD version at this stage seems to have a features that are important to the use of deploying CAN on a printer, including greater bandwidth and flexibility with in the can bus. As well as hardware tx/rx buffers and in most cases with the higher end FD can bus built in error correction and timestamp bits. MCP2518FD vs the cheaper MCP2515 as an example.
Many valid points. I geared this video more toward the less technical which is why I concentrated on more commercially supported USB to CAN hardware, but there are several ways you can accomplish your build. - Allen
When you look at the extruder section in printer.cfg the UART pin definition is separated from the other extruder pins- just make sure you update all four pin definitions and you'll be set.
Excellent video with descriptions! Would you mind providing the ways and links to the STLs of how you mounted the X and Y endstops please. Going this route on a new 2.4r2 and would be extremely helpful. Keep making great detailed videos!
It's a lot more streamlined design choice to have a common bus voltage then have whatever component convert their required voltage locally. Many of the chips are using 3 volts not 5 so it's not you avoiding circuitry. Also running low voltage say 5 volts over from a power supply in the base to print head introduces a lot of room for voltage drop and noise. Probs not major but another thing people will have to look out for. Though with some current hardware options an additional 5v cord could make sense.
question- I'm building a custom IDEX machine with two independent hot ends that can print simultaneously. I've read that CANBUS can be daisy chained- is it possible to wire from the U2C to the first EBB36, and then from that one to another EBB36? or would I need to have two independent U2C boards controlling each EBB36 individually?
Does the canbus work with afterburner? I have the out of date printed parts, seems like a waste to not use new parts, but I like the idea of not having to thread a bunch of wires around and thru the cable chains and that, i'm coming up on a point in my part sourcing where theres choices to be made beyond just the size.
USB-C would just be complete over engineering in this case. - First USB-C has way more than just 4 wires (for Power Delivery Communication, USB2.0, USB3.2, USB3.2x2 which requires an additional USB3.2 pair, etc.) you would actually add wires back into the mix when the goal is to reduce them. - As @allenrowand6988 pointed out, USB-C connectors and wires aren't rated for this kind of constant motion - USB-C PD needs some negotiation which means both sides need extra circuitry to negotiate the power and converters to convert the voltage and current to the one supported by both sides and the cable (yes the cable also communicates with a chip), this adds even more power circuitry into the mix. - The current MCU generation will at best go up to USB2.0 speeds and not go into USB3.x territory anytime soon so it would be complete overkill having a cable supporting those speeds when you can't really use them. If you want to keep it simple in design while also retaining the reliability and real-time capability there are some protocols that are tried and tested in industrial usecases. CAN being one of them. There are also others like Profibus, Modbus, other protocols running on RS-485 or RS-422, or some protocols running on Ethernet (Ethercat, Profinet) although those also bring back some complexity and processing requirements.
I think that the CAN bus does not suitable for things like stepper motors. Because every step motor must be in sync with others.The CAN bus is not capable to transmit and receive data in straight order, so there mait be change in order which will couse unsyncking for shure. So CAN bus is not a good idea.
The protocol used by Duet boards (which use CAN-FD for higher bandwidth instead of old CAN) running RRF includes a time sync protocol to sync the boards to within a few microseconds. RRF on the main board sends the info for complete moves a little in advance of when they are needed, and the expansion board delays starting them until they are due. So CAN latency is not an issue for RRF.
Wireless connections are prone to interference especially in the ranges you can legally use without a license. Real time communication is also very tricky since you need to add extra margin for intermittent data loss and potential retransmission as well as a mechanism to detect and request a resend. With those margins added you might even run into a bottleneck where the mainboard can't communicate with the head fast enough and print speed needs to be reduced. Safety wise there are also a ton of issues, for example if the controller can't detect the endstops during connection loss, etc. In my opinion if for such a system where you want to keep the machine from rapidly disassembling itself and you need real time communication there should be no wireless communication in the critical data path.
2 years later and out of everything new i watched this video helped me the most thanks
After learning about and utilizing CAN on my master thesis project i will most definitely implement it on a future Voron build! Great video!
Thank you- I'm glad you found it useful!
I wish these things and videos were when I was in middle school. My career path would probably be much different. :)
Great video, Subbed! Can confirm, LEDs make printers go faster
Thank you!
Wow! That sounds like the beginning of the personal computer days at the Homebrew club, which I was not a part of. I didn’t assemble my 1st PC until ~ 25 years later. 👍🏾👍🏾👍🏾👍🏾👍🏾
Twisting the power cables will reduce the noise produced by these cables. Using multiple thin cables in parallel is better than just two thick cables.
Thanks to videos like this I’ve got it working… and even have updating Klipper over canbus
Thanks so much for clearly explaining this to normal non techies. Now I fully understand what can is and will be looking into it.
First time I subscribed to a channel with just one video :D kudos to the quality !
Thank you! Plenty of more videos are on the way.
Thanks for the video. I wasn't finding a good explanation of some of this stuff elsewhere, namely what to do about the endstops. I think I will give sensorless homing a try since I dont want to print anything just yet.
Duet3 platform already has CAN bus built-in for use with their expansion boards and toolboards.
Playing catch up👍
Awesome! This is the first upgrade on my list and your my goto shop for Voron parts! I may be asking you to help put together a shopping list.
Any time!! Thank you for your support!
Duet 3 with RepRap firmware has been using CAN bus for at least 2 years
Thank you for a clearly explained video. As someone who has tinkered my way through an Ender 5 Pro. I am looking for clear information to expand my skills and knowledge. I am probably going to build a Voron 2.4 due to the amount of support available but I prefer the design of the VZbot.
Thank you, glad you found it helpful
Thank you for making this.
Like from Spain because the TRUE REALITY OF SPEED!!!! MORE LEDS!!!!! Also thanks for the tutorial, its gonna be very useful, at least for me
Just an FYI, I just picked up the Mellow FLY 36 and after it arrived, they released a new one, so typical, it has a better power connector and magnetic encoder built in, which I assume is for fine extruder motor control.
Thanks for putting this video out. I would be careful when classifying an option as "traditional" especially when referring to a USB to CAN adapter. I would also mention that you missed the original method of can bus for the RPI, and that is using a CAN hat for the RPI, which uses a SPI bus to talk from the RPI to the CAN hat. I would also raise caution for the type of can bus being deployed. The FD version at this stage seems to have a features that are important to the use of deploying CAN on a printer, including greater bandwidth and flexibility with in the can bus. As well as hardware tx/rx buffers and in most cases with the higher end FD can bus built in error correction and timestamp bits. MCP2518FD vs the cheaper MCP2515 as an example.
Many valid points. I geared this video more toward the less technical which is why I concentrated on more commercially supported USB to CAN hardware, but there are several ways you can accomplish your build. - Allen
Hey dude, thanks for this video! Super interesting idea, I'm going to take a look at your website. Subscribed, please keep it up :)
Thanks!
Good LORD this video is SO SO helpful
Would you mind clarifying what you mean about the UART pin, and the steps involved in avoiding the issues that you had?
When you look at the extruder section in printer.cfg the UART pin definition is separated from the other extruder pins- just make sure you update all four pin definitions and you'll be set.
Excellent video with descriptions! Would you mind providing the ways and links to the STLs of how you mounted the X and Y endstops please. Going this route on a new 2.4r2 and would be extremely helpful. Keep making great detailed videos!
The X endstop is mounted directly to the toolhead, the Y endstop is in Minsekt's repo in the description. Thanks for the kind words!
So, what brand of LED do you recommend? Obviously I want the absolute fastest possible.
Very informative, much appreciated.
I think it makes sense to use 5 wires instead of 4, one with 24V and one with 5V, they are commonly used and no voltage reducer is necessary
It's a lot more streamlined design choice to have a common bus voltage then have whatever component convert their required voltage locally. Many of the chips are using 3 volts not 5 so it's not you avoiding circuitry.
Also running low voltage say 5 volts over from a power supply in the base to print head introduces a lot of room for voltage drop and noise. Probs not major but another thing people will have to look out for.
Though with some current hardware options an additional 5v cord could make sense.
Unless you have a specific high current draw device it’s much easier to use the 5v supplied by the toolhead board.
this is terrific! Are you guys considering adding CAN support via the leviathan to the micron+ custom/self source kit?
You got that right. LEDs make printer go faster. I have lots of LEDs. ;)
question- I'm building a custom IDEX machine with two independent hot ends that can print simultaneously. I've read that CANBUS can be daisy chained- is it possible to wire from the U2C to the first EBB36, and then from that one to another EBB36? or would I need to have two independent U2C boards controlling each EBB36 individually?
You can't daisychain the EBBs, but you can run from the U2C to each toolhead separately. The U2C has 3 CAN outputs.
Great overview video!
there's a spelling error- your missing an "i" in the middle of canbus...
Does the canbus work with afterburner? I have the out of date printed parts, seems like a waste to not use new parts, but I like the idea of not having to thread a bunch of wires around and thru the cable chains and that, i'm coming up on a point in my part sourcing where theres choices to be made beyond just the size.
You'd need an EBB42 CAN board to fit the CW1 extruder motor
Great idea if only we could find those coveted can bus chips
Can the stepper driver on the ebb36 handle being in a 50c chamber?
My machine run around 60º chambers with no issues
Can Ethernet PoE be good alternative to CAN? It supports high data speeds and cables are reliable by default?
I don't know if PoE can deliver enough amperage, and there are no Ethernet toolhead boards that I'm aware of.
how to setup input shaüing on a ebb36
Great video. Well stated
Thank you!
What are the pros of using CAN in this application vs USB
This existed before USB toolhead boards were widely available.
Why not use USB-C with Power Delivery instead of CAN?
Interesting question
USB wiring and connectors aren't rated for this kind of motion.
@@creatornaut503 Or potential heat load in an enclosed printer. Good point.
USB-C would just be complete over engineering in this case.
- First USB-C has way more than just 4 wires (for Power Delivery Communication, USB2.0, USB3.2, USB3.2x2 which requires an additional USB3.2 pair, etc.) you would actually add wires back into the mix when the goal is to reduce them.
- As @allenrowand6988 pointed out, USB-C connectors and wires aren't rated for this kind of constant motion
- USB-C PD needs some negotiation which means both sides need extra circuitry to negotiate the power and converters to convert the voltage and current to the one supported by both sides and the cable (yes the cable also communicates with a chip), this adds even more power circuitry into the mix.
- The current MCU generation will at best go up to USB2.0 speeds and not go into USB3.x territory anytime soon so it would be complete overkill having a cable supporting those speeds when you can't really use them.
If you want to keep it simple in design while also retaining the reliability and real-time capability there are some protocols that are tried and tested in industrial usecases. CAN being one of them. There are also others like Profibus, Modbus, other protocols running on RS-485 or RS-422, or some protocols running on Ethernet (Ethercat, Profinet) although those also bring back some complexity and processing requirements.
Can't figure out how to make your printer last longer? With CAN you can!
I like the red color of your cable sleeve. Where did you get it?
Thanks! Got it from DFH.fm
PLAIN LED DOESNT MAKE YOUR PRINTER FASTER, RGB does😀
thank you.
You're welcome
your voice is absolutely identical to Charly (moistCritical)
It is not. Charles White sounds different.
I think that the CAN bus does not suitable for things like stepper motors. Because every step motor must be in sync with others.The CAN bus is not capable to transmit and receive data in straight order, so there mait be change in order which will couse unsyncking for shure. So CAN bus is not a good idea.
The protocol used by Duet boards (which use CAN-FD for higher bandwidth instead of old CAN) running RRF includes a time sync protocol to sync the boards to within a few microseconds. RRF on the main board sends the info for complete moves a little in advance of when they are needed, and the expansion board delays starting them until they are due. So CAN latency is not an issue for RRF.
i'll wait for industrial ethernet for my machine ;P
LED up
turns out it was 10 minute ad for west 3d
Don't understand this. I think I smoked to much can-i-bus
Make CAN wireless, then only +/- are needed.
Wireless connections are prone to interference especially in the ranges you can legally use without a license. Real time communication is also very tricky since you need to add extra margin for intermittent data loss and potential retransmission as well as a mechanism to detect and request a resend.
With those margins added you might even run into a bottleneck where the mainboard can't communicate with the head fast enough and print speed needs to be reduced.
Safety wise there are also a ton of issues, for example if the controller can't detect the endstops during connection loss, etc.
In my opinion if for such a system where you want to keep the machine from rapidly disassembling itself and you need real time communication there should be no wireless communication in the critical data path.
0:40 me that have a car without canbus 😑
this is not a guide. You didn't describe what you entered and how - you redirected me to the documentation and it was weak.
I can't understand how LEDs makes the printer go faster... hmm. Joking
Useless for now . To much trouble with very little benefit but maybe that will change in a few years.
Fewer wires - not less.