As a former electronics lead on Team 4 ELEMENT, I’ve faced many challenges and problems with wiring Robots. I am now Designing a new board designed specifically for Swerve and would love to gather your insights. This concept aims to significantly enhance Swerve modules by offering a compact design that fits under the Power Distribution Hub (PDH) and above the swerve modules, while also simplifying the CAN network by reducing the number of necessary connections, and cleaning up the robot’s appearance by minimizing wiring clutter. While also making it hot-swappable using Molex cables for improved reliability and security.
What features are essential for a Swerve drive breakout board?
Are there other subsystems for which you’d find similar solutions useful?
Would your team or organization consider using a product like this?
Please share your thoughts, experiences, and any additional features you want. Your insights will play a vital role in shaping this potential product. I look forward to your valuable input and am excited about the possibility of bringing this project to life!
This is a really cool idea, but how do you plan on overcoming the 16AWG max wire size and 13A max rated current per conductor in those Molex connectors?
In the past, we’ve used a block of PP45 connectors to make modules swappable and it’s pretty good even though it’s bulky.
Wa2go! I like the modularity that it introduces, but you are now relying on three different connector families to do the same task? That seems problematic from a reliability standpoint, and is also a little more burdensome on installation/assembly technique and parts stock.
Since this design is not final. I was looking into something in automotive that can handle more. I am still searching for the best option but I wanted to get a feel for the Interest on it.
You’ve created a CUSTOM CIRCUIT for power wiring which is illegal under the current rules. You should focus on making something that might be housed in a custom housing, but uses only COTS wire and connectors rated for 40A+.
I don’t think different connector families are a problem if they’re justified.
My team has recently developed a PCB with Weidmuller push connectors that we use to make our swerve drive easy to configure. You can see them in our practice project robot on Team 1710’s Open Alliance thread. The Weidmuller connectors tend to be more reliable than Molex since there are fewer failure points and, when combined with feurrels, the wires don’t back out.
I would obviously go through First to get it to pass the rules and inspection. But I am just testing the waters. and also FIRST might consider it as a “passive
PCBs such as those used to adapt motor terminals to connectors are considered
connectors” since it is in fact passive and not “directly alter[ing] the power pathways” Also I was considering making the Anderson connector on the module side a solder point so then it considered as a connector.
I believe this is incorrect: A CUSTOM CIRCUIT is a active device, and as written, rules state CUSTOM CIRCUITS may not alter the critical power paths. In this case, it is not altering the power path, it’s simply following a static route.
By the regulations, a breakout board such as this should fall into PASSIVE DEVICE regulations.
A PASSIVE CONDUCTOR is any device or circuit whose capability is limited to the
conduction and/or static regulation of the electrical energy applied to it (e.g. wire,
splices, connectors, printed wiring board, etc.).
So, should be legal, assuming it’s sufficiently designed to handle the rated currents and heat.
Team 2252 will be releasing our Swerve Module Board that we used this season in the next couple days. Currently getting ready for Ohio State Championship this weekend.
Things ours does:
Connects canbus to a central location
Makes it easy to power the cancoder and puts a 100mA resettable fuse in line
Status leds for power to board and that the fuse to cancoder is good
120ohm resistor on board and jumpers to terminate the canbus network at any of the swerve modules
use 2 pin(power) and 4pin(canbus) molex 90 deg connectors to keep the board as short as possible
While I stand corrected here, I’m a little concerned about passing 40A through the PCB. I don’t know how you’d prove that it’s properly specced to handle the current, especially for multiple motors.
It’s of course absolutely possible, it’s the verification I’m uncertain about.
You size the trace to limit temperature rise (typical guidance for this is 10C rise at ambient). There are lots of calculators for this, essentially it’s a function of trace width and copper weight (thickness). It’s legal, but basically impossible for an average inspector to validate (even if a team provides PCB drawings, few inspectors have the knowledge or experience to independently analyze it), so it’s basically a “trust me” from the team.
Per the above Q&A, a separate approval is not required (an inspector could escalate to LRIs etc, but it’s not a COTS pre-approval type thing like what is required for passing through PWM control signals from the MXP port). Maybe teams who did this in 2024 could speak to their inspection experiences and what material they came prepared with to demonstrate rule compliance vs what was actually checked?
I don’t know if consolidating all the power connections into one PCB make sense if they are going to have to split at the PDH side anyway. It seems much more reliable to just run wire straight from the PDH to the krakens; two less connections that could go wrong (I’ve never used Krakens so there may be something I’m missing).
It seems like the major advantage is that you can swap out the whole module quickly. If this is the goal I would create a PCB solely for the signal connections, and 3D printed housings for 4 powerpoles and the signal connection.
If it were for our team, I’d probably just consolidate all the signal connections into one Molex SL connector on the module side (possibly without a PCB), and have a PCB to convert that on the robot side. This could increase the time to swap a motor/sensor with the module on the robot, though.
For our board design, the traces we used are good to 1A(I will verify the design and post it in the release)
At competitions, the inspectors didn’t worry to much about them since they are passive. Never was asked about traces and current but thinking about it, it should be documented
