I designed a USB to RS-232 adapter this evening. There's no particular reason I did. It was just sort of "because I could." It was actually sort of tricky to route, since the RS-232 level converter chip wasn't conveniently laid out - like with RS-232 signals on one side and TTL/CMOS on the other.
My first cut wound up being a 2.08 x 1.25 inch board. It seemed like it could be smaller than that. Most of the space was taken up by the through-hole DB9 connector, so I swapped that out for an edge-mounted DB9. The only thing attaching the connector to the board will be the soldered pins, so it's quite a bit weaker, but in principle such a thing would be mounted in a plastic case designed to relieve any strain. I got the board down to 1.8 square inches. I used an SOIC RS-232 level converter instead of an SSOP one, so there probably are still gains to be made. In principle, it's conceivable that one could attempt to place two SSOP-28 packages on opposite sides of a much smaller board. That's probably how they achieve a USB-to-RS-232 adapter that fits entirely inside of a DB9 shell. I don't feel a need to go quite that far. Heck, I don't really have a particular use for this anyway. I'm not sure we have an RS-232 connected peripheral left in the house.
I'm not really sure why I set down the road to make one of these. But I've got 3 on order, and I can sell them for $15 each (case and micro USB cable not included), given the price of the boards and BOM.
Hello Nick,
ReplyDeleteDo you have a BOM for this project, however tiny it may be? Based on the design, I'd say a FT232 with a MAX 213, but you used quite a few more capacitors compared to the reference design.
I came across your design on OSH Park and this would save me quite some tracing.
Yes, it's an FT232 (though that is now a Cypress CY7C65213 after FTDIGate) and MAX213IDBR. It's pretty much exactly the reference design. There are four caps for the charge pump on the MAX213, there's a bypass cap on Vcc for both the UART and MAX213, a cap and ferrite bead on the incoming USB bus power and finally a 10 µF filter cap.
ReplyDeleteMaybe some of that - particularly the last two - isn't completely required. It just seemed like a good idea at the time.
Thanks for confirming my assumptions, I just ordered a couple of boards with OSH Park.
ReplyDeleteFTDIgate is pretty much what made me look for an adapter where I could control which chips are used, since my local suppliers only have adapters with either a fake Prolific or FTDI-chip on board.
Oh, well, it's too late, since you've ordered some, but I can quite easily add this board as a product to my Tindie store. I never bothered before because I sort of thought the market for USB->RS232 (or even USB->TTL) adapter boards was saturated.
ReplyDeleteGiven the low volume, I'd probably have to charge around $15 for them, but at least they'd be a known part that uses the CDC protocol.