RS485 Breakout with 48V Passthrough Building Block Board — Part 3#

Today I sent the first version of the RS485 Building Block Board to JLCPCB for manufacture.

There is no real substitute for getting a board into physical form and running it through testing. Otherwise, it is very easy to keep redesigning the board forever, especially when there are lots of small choices to tweak. I know I have a tendency to do that, so when I recognise it happening, that is usually a sign that it is time to move to the next step.

This time around, however, I did something slightly different.

Rather than choosing every component independently and then checking whether JLCPCB could assemble it, I made a more deliberate effort to select parts that were already available in the JLCPCB parts library. I focused especially on commodity parts such as resistors, capacitors, and diodes.

That did make a noticeable difference.

With the first batch of regulator boards I sent for assembly recently, the parts setup costs stung a bit. It was not a disaster, but it was enough to make me pay closer attention this time. If these boards are going to become a regular part of my workflow, and possibly something I keep in small batches, that is not a mistake I want to keep repeating.

JLCPCB splits parts into a few different categories.

There are basic parts, which are the easiest and cheapest to use from an assembly point of view. Then there are extended parts, which are available but normally carry an additional setup cost. Finally, there are extended promotion parts, which are technically extended parts but, at least at the time of ordering, did not carry the same setup cost.

Those setup costs can add up quite quickly if a board uses lots of different extended parts. On a one-off prototype, it might not matter too much. But if the goal is to build small reusable boards that can be ordered, assembled, tested, and possibly stocked in small batches, then BOM discipline starts to matter.

For this board, I was not trying to optimise every last fraction of a penny out of the design. That is not really the point at this stage. The aim was to be a bit more deliberate:

  • use sensible, available parts where possible
  • avoid unnecessary extended-part setup costs
  • keep the design practical to assemble
  • reduce friction if I need to order another batch later
  • still keep the circuit robust enough for real-world testing

The result is not necessarily the final version of the RS485 board, but it is a version I can now test properly.

Why Build the Board Now?#

The design has already been through several iterations.

At one point I considered using RJ45 connectors for the RS485 and 48 V lines, partly because they are convenient and familiar. After thinking through the practicalities, I decided against that approach for this version.

The main issue is that RJ45 connectors carry assumptions.

If someone sees an RJ45 socket, there is always the possibility that they will treat it like Ethernet, even if the board is clearly labelled. Accidentally putting 48 V onto something that someone might plug a laptop or network switch into is not a risk I particularly want to encourage.

There is also a practical deployment issue. Using RJ45 would mean needing both screw-terminal wiring and crimped network-style cables, depending on how the board was used. For a small prototyping and test board, that felt like unnecessary complexity.

So for this version I moved to screw terminals.

The board now has terminals for the RS485 bus, 48 V input, and 48 V passthrough output. That makes the wiring more explicit and, hopefully, harder to misunderstand.

Board Intent#

The purpose of this board is fairly simple.

It is intended to provide a small RS485 interface board with optional 48 V passthrough for distributed control and automation experiments.

The logic side of the board uses a 3.3 V interface, making it suitable for use with many modern microcontrollers. The RS485 side uses an SP3485 transceiver, which provides the differential A/B bus interface.

The 48 V side is not there to power the transceiver directly. Instead, it is there as a convenient passthrough supply for systems where power and data need to be distributed together.

It is not isolated, regulated, or converted on this board. It is simply passed through, with the layout designed to keep the higher-voltage supply region distinct from the logic side.

That kind of arrangement could be useful in future projects such as distributed IO nodes, lighting controllers, small automation systems, or other devices where several boards may sit along the same cable run.

I thought it would be a handy feature to have for my own projects, and it may be useful for others working on similar systems.

As with the other building block boards, this first version is primarily about validation. The goal is to get real boards in hand, check the assumptions, test the protection and layout choices, and then decide what needs to change in the next revision.

Next Steps#

While the RS485 board is being manufactured, I intend to work on the next building block board. The likely candidate is an IO expander board, probably based around the 74HC165 and 74HC595.

I also need to start updating my websites and deal with a few other business admin tasks, which are necessary but not nearly as interesting.

If you are looking for reusable open hardware modules to speed up your bench testing, or need support designing custom breakout circuits, check out our catalog at Protoworks (NI).