Is it reasonable to make a board to sit an ESP32 Development Board with two 19 pin 2.54mm pitch headers at 25.4mm centres over XBee headers comprising two 10 pin 2mm pitch headers at 22mm centres?
It would leave 1.7mm between the header centres with pins down to the XBee socket and socket up to the ESP32 development board. At an absolute minimum I’d need to connect VCC, Tx, Rx, GND which are pins 1,2,3,10 on right side.
The ESP32 antenna needs to overhang the end of the XBee board.
Let’s reverse the sense and the direction of the question to elicit more data: Which aspects are you worried might be unreasonable?
If the pins of the different headers are 1.7mm apart will that cause a problem with through hole soldering or board operation? Headers will be soldered from opposite sides of the board.
So that’s 1.7 mm between the headers? I think you may have footprint overlap. Try it and see on pcbnew.
Obviously that’s my concern. My background is software. How would I know? Can you explain why you think it “may have footprint overlap”?
That distance is pin row to pin row. There will be pads and the headers will also occupy space. Just try it and see by placing them in pcbnew and running DRC.
If this was for production, I would probably use a 0.1" header for the ESP32 footprint.
https://duckduckgo.com/?q=0.1"+%2Bsmt+%2Bheader&t=h_&iax=images&ia=images
I assume the XBee module has castellated pads on the side? If so, then another solution is to use single row 2mm headers for the XBee module, but place them closer together on the ESP32. With oval pads and relatively big holes you can proably get these headers upto a 60 degree angle (90 degree would be the normal to the PCB)
You do not have to solder a castellated PCB directly to SMT pads of another PCB. The soldered version described above is a variant I saw on EEVBlog #592:
Another possibility is to move the XBee module 5mm to the left or right. This requires that the pins of the ESP are flush or below the surface of the PCB. For isolation I would put a double layer of kapton tape over the ESP header pads to prevent shorts.
No, your assumption is wrong unfortunately. It’s just an adapter for existing XBee female headers and the existing ESP32 Development Board male headers. A castellated ESP32 works nicely (pinout) but they cost too much compared to a dev board and I don’t have the expertise to build one.
This discussion and the image I posted above has been really worthwhile. I’ll work towards mounting the ESP32 Dev Modules elsewhere in enclosures and use flywires.
Now it gets confusing.
First you post:
and then you link to a module of that size with a ridiculous price and ESP32.
I have long given up on Zigbee myself, because I bluntly refuse to pay more then a few EUR for such a simple radio device.
Another thing that confuses me that you work with existing modules, yet want to make something in KiCad. So I don’t really understand what is which.
If you’re designing some custom THT footprint with narrow margins, then you can place the pad offset to the pin to increase clearance on one side, while still having some pad left to solder on on the other side.
We’re not disagreeing, that module is a ridiculous price. I thought I’d use standard modules (one with XBee Headers and an ESP32 Dev Board) with an adapter board (designed in KiCad). Yes, I considered offsetting the modules but I’m now of the opinion I’m better to keep them separate and wire them together - Vin, Gnd, Rx, stuff the rest.
It’s usually just a handful of wires. Power and I2C or GND.
And you have to connect it all to a power supply and other circuitry too (sensors?), so no need to make it more complicated then it has to be.
Also, have you considered other zigbee modules?
One of the “big stores” popped up in DuckDuck and they have them in lots of sizes, including apparently much smaller.
Have you considered using surface mount pin headers/sockets instead?
Maybe I’m missing something, but to me, “castellated ESP32” means a WROOM or WROVER module, which are quite cheap.
Hi flywire,
Your requirement is not possible as:
1.The center-to-center pin clearance of the two headers is really low. (As standard headers have a pad size of 40mils). Pad to pad clearance comes down to 2.9mils, whereas the required minimum is 8mils of clearance.
2. Since pins are really close to each other, the width of the header body becomes a concern.
You can also add the (3D) STEP files of the headers and then place them on the board to see if the headers and pins are overlapping each other.
This adapter looks worth developing:
This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.
