Question Pros / Cons of Umbilical for Trident

[J-V]

Printer Model

Trident

Extruder Type

LGX

Cooling Type

Stealthburner

I'm curious who has switched from cable chains to umbilical with Trident and what are the pros/cons of doing it?

On a related note, it seems toolhead boards with CAN starting to become common. Does this also change the equation?

 

[120decibell]

I haven't done umbilical myself, but toolhead boards with CAN make umbilical a lot easier because there are only 4 wires to run instead of 14-ish.

 

[DerDackel]

I just finished my first Trident, completely skipped the cable chains from the get go and set up an umbilical with an EBB36. The common wisdom would be that this allows for a lighter X/Y axis and higher speeds, but my guess is that for these gains to actually manifest, you have to really push accels for this to matter. I would love for someone to actually produce some data on that!

 

[ah2049]

I've switched to a CANBUS umbilical on my existing 2.4, and I have a CANBUS umbilical going into the Trident I'm currently building. I had (silicone) wires fail in the 2.4 drag chains after a couple hundred hours and I never want to deal with replacing them again. I'm half expecting the same to happen to my 0.1, but I don't remember what kind of wires I ran in the stock umbilical, haha.

Going CANBUS, umbilical, and PTFE wires should significantly cut down on the frequency of failure, and the frustration of replacement.

 

[IronBrain]

From a systems engineering standpoint, and as an electrical engineer of many years, I would be skeptical about the reliability improvement of using an umbilical with several connectors.

The CANBUS is approach is nice in that it reduces the number of wires, but moves the failure from hardware to software. How reliable it is, or how failure safe it is will depend on the software. I would do this if the wires and connectors are well spec’d (kept to a minimum count, properly sized wires, good strain relief on the connectors), the temperature of the microcontrollers can be kept within operating temp, and if I was confident in the software management of failure modes. I’ve seen hot microcontrollers do some pretty wonky things.

Connectors are a major weak point in any electrical system.

Conventional wisdom I read from hobbyists here and elsewhere over-specs wire gauge for 3D printers and the recommendation to use silicone wires is misplaced for use in cable chains.

You want to use the thinnest wires that can carry the current the signal requires within the voltage drop and temperature limits of the wires specified for each signal. This is because bending stresses goes up with r^2 for the same bend. Repetitive bending stress will work harden the copper eventually leading to a crack which will then propagate through the stranded wire bundle fairly quickly. The sheathing acts as a strain relief, essentially keeping the bend radius large which minimizes the stress.

Silicone is meant for high temperature but has high friction. The high friction limits movement in the wire bundle, increasing stress on the inner and outer wires, making the bundle act like a big fat wire. And silicone is also very flexibile which actually hurts because the flexibility means that the stresses are carried more by the copper. PTFE wire is less flexible, and generally has thicker sheathing which acts as a better strain relief, but has lower friction as well.

There is no reason you shouldn’t be able to get 10,000+ hours from the right wire in a motion system like a 3D printer

 

[J-V]

Today my PTFE tube rubs against the led lighting. It isn't an issue (so far). I'm hoping there is a nice 3part PCB CAN design to be dropped with tap in the next few days.

Keep us posted @ah2049 on how it goes with the Trident. And would love to see pictures when you get further along.

 

[MrTeliP]

I am running CAN with an umbilical in my Trident. The main issue is how to route/secure the umbilical without much Z clearance above the gantry. I route the wires up to the A motor mount and place a cable gland there and then tie off the cable with a ziptie on the horizontal rear extrusion up by the exhaust port. I'm using the PG7 cable glands that seem to be popular to help provide a large bend radius around the ends of the cables at the toolhead and the gantry connection. This is most important at the toolhead as that's the part of the umbilical that will see the most bending. The problem is that the gland height on top of the extruder. It's tall enough that it touches the top of the enclosure.
Overall, I think it is totally worth doing. The CAN protocol is a proven and reliable form of communication. If it's good enough for electrical systems in a car, it should be suitable for my printer. It was super easy to set up too.
Also, from what I've seen it seems that the cable chains provided in most kits seem to be improperly sized for the wad of cables they are carrying. People tend to pull the wires through to tightly, they add stuff to the toolhead, there are a lot of different types of wires and sheathings in the bundle, friction, it just seems like a problem. That's not really my area though, I bE a MecHanIcAl DesIgN EnGineEr. I used to do the electromechanical design thing back in the day, but I was always super conservative when it came to cable routing. Inevitably there would end up being way more crap at the end of the wires than what was originally spec'd. Feature creep is real and very apparent in this hobby where people like to add all the cool stuff to their printers. You buy a Formbot kit to save some money and a month later it looks like Doc Brown's DeLorean. Anyway, that's enough of that...I'll try to add pics of my setup if I remember.

 

[MrTeliP]

I am running CAN with an umbilical in my Trident. The main issue is how to route/secure the umbilical without much Z clearance above the gantry. I route the wires up to the A motor mount and place a cable gland there and then tie off the cable with a ziptie on the horizontal rear extrusion up by the exhaust port. I'm using the PG7 cable glands that seem to be popular to help provide a large bend radius around the ends of the cables at the toolhead and the gantry connection. This is most important at the toolhead as that's the part of the umbilical that will see the most bending. The problem is that the gland height on top of the extruder. It's tall enough that it touches the top of the enclosure.
Overall, I think it is totally worth doing. The CAN protocol is a proven and reliable form of communication. If it's good enough for electrical systems in a car, it should be suitable for my printer. It was super easy to set up too.
Also, from what I've seen it seems that the cable chains provided in most kits seem to be improperly sized for the wad of cables they are carrying. People tend to pull the wires through to tightly, they add stuff to the toolhead, there are a lot of different types of wires and sheathings in the bundle, friction, it just seems like a problem. That's not really my area though, I bE a MecHanIcAl DesIgN EnGineEr. I used to do the electromechanical design thing back in the day, but I was always super conservative when it came to cable routing. Inevitably there would end up being way more crap at the end of the wires than what was originally spec'd. Feature creep is real and very apparent in this hobby where people like to add all the cool stuff to their printers. You buy a Formbot kit to save some money and a month later it looks like Doc Brown's DeLorean. Anyway, that's enough of that...I'll try to add pics of my setup if I remember.

I may need to run some sort of stiffener through the cable sleeve to keep it from kinking at the top of the cable gland. It's not as bad as it appears in the pic, but I don't like it.

 

[shiftingtech]

recommendation to use silicone wires is misplaced for use in cable chains.

I'm pretty sure all the Voron Design BOMs have been putting FEP/PTFE in cable chains for a while now. (yes, that is a change, but not a terribly recent one at this point)

 

[Alan B]

Prusa puts a piece of filament (Nylon?) into the MK3 umbilical to stiffen it, and minimizes connectors. It lasts quite well but is less fun to repair as the umbilical has to be reworked with each toolhead component replacement.

 

[Sanity Agathion]

Prusa puts a piece of filament (Nylon?) into the MK3 umbilical to stiffen it, and minimizes connectors. It lasts quite well but is less fun to repair as the umbilical has to be reworked with each toolhead component replacement.

Fortunately with CAN, you run only 4 wires Replacing those is way easier, especially as they all use single connector (Molex Microfit or XT30 2+2).

 

[Alan B]

CAN is a good improvement. Saves tons of copper in cars. For reliability smaller, finer stranded wires are better. Multiple parallel wires for power and ground might be the way to go. The thin flat cables are good also, and again parallel paths are used to handle the current while providing high reliability.

 

[Lab4450]

Prusa puts a piece of filament (Nylon?) into the MK3 umbilical to stiffen it, and minimizes connectors. It lasts quite well but is less fun to repair as the umbilical has to be reworked with each toolhead component replacement.

I'm thinking about doing the same. Prusa places a 2.8mm nylon filament to give some stifness to the wire harness. I use trimmer line that is also nylon and super cheap (hardware store look for grass trimmer machine line).
plus my mini trident will have the gantry 25mm lower to allow wires to move more freely

 

[garrettwp]

Here is my V1.8 running with umbilical for a year now without issues. I use a Keybak to tension the umbilical cord to keep it from flopping around.

 

[DerDackel]

I had no piece of nylon filament around and didn't want to buy an expensive spool of 2.85mm that I would take <1m out of to use this way, so I put a length of 4mm PTFE tube inside the umbilical to give it stiffness in my V2.4. Inside my trident, that would be too thick and I haven't really found anything yet that gives it the right amount of stiffness.
In the 2.4 I have the Umbilical going down to the A motor mount, with lots of space above to form a nice stiff arc.
On my Trident, I'm going straight out the back and don't have that much z-space, so the umbilical currently pretty much curls up behind the toolhead.

 

[garrettwp]

This is what I use to hold the cord up and out of the way of the tool head. I have it attached to the back top left corner of the frame.

 

[ah2049]

Today my PTFE tube rubs against the led lighting. It isn't an issue (so far). I'm hoping there is a nice 3part PCB CAN design to be dropped with tap in the next few days.

Keep us posted @ah2049 on how it goes with the Trident. And would love to see pictures when you get further along.

All that's really left is the electronics bay wiring at this point. I'll be dealing with the enclosure at some point in the future if I decide to print anything other than PLA on it. I'll definitely post a picture of it.

Umbilical rubbing on the upper parts is something I haven't really looked into yet. My mind has been in a "I'm just building a slightly different 2.4" the whole time, so I didn't take into consideration that the toolhead and umbilical will always be sitting at the top like the 0.1. I will probably end up designing a 90 degree strain relief "noodle" to aim it toward the back of the machine instead of the current vertical relief.

Supporting the umbilical slack will be an interesting thing to deal with. I do like the solution that @garrettwp posted, and I might just order another keybak to do it. But I could also try to modify the silicone bungee method I used on my 2.4 (inspired by/copied from cryzzgrantham on reddit: his 2.4).

 

[Alan B]

Anyone considering a single horizontal cable chain from the printhead to the back or side wall? I've seen this somewhere and it appeared to be a compact elegant solution.

 

[MrTeliP]

Anyone considering a single horizontal cable chain from the printhead to the back or side wall? I've seen this somewhere and it appeared to be a compact elegant solution.

This is actually a really good thought and I feel ashamed that I didn't think about or explore this to see what it would look like. Of course 80% of mods like this are based on how cool it looks so if it doesn't look cool it's totally out, lol.

 

[ah2049]

Anyone considering a single horizontal cable chain from the printhead to the back or side wall? I've seen this somewhere and it appeared to be a compact elegant solution.

I believe some RatRig V-core printers are using horizontal drag chains. There are also some Annex Engineering printers that are using horizontal FFC setups to achieve the same thing.

If I'm thinking correctly, using drag chains in a horizontal fashion would possibly cause issues in Voron machines. In order to have enough chain to go from the back of the machine to the front diagonally, you'd probably have too much chain sticking out on a back left to back right move to use a standard enclosure on the printer. And assuming omni-directional drag chains are an actual thing, the flexibility and weight would cause ringing issues from it flopping around in any direction during quick moves, not to mention possible issues of it kinking/stacking up and preventing moves.