Following on from Zigbee ConnectPort Options, I had purchased a second hand Digi ConnectPort X5.
I immediately encountered a few issues:
- The antenna adapters, that I had ordered were incorrect
- The AMPseal TE Connectivity 23 pin IP67 connector proved more difficult to assemble than I had anticipated
- The serial port, when connected to a PC, via a CP2102 USB-TTL interface, caused the Vin to drop from 12V to 3V, turning the X5 off and eventually causing the PC to crash and shut down.
A table listing the connections:
A diagram showing the connections available via the 23 pin connector
I had purchased two of the following:
- New 2.4GHz 16dBi RP-SMA Male WiFi Wireless LANS Antenna Cable Linksys
I had purchased the following item, SMA Series RF Coaxial Male Female RP-male RP-female Adapter Connectors Convertor: [RP-TNC Female To RP-SMA Male]. However, I had made an mistake in the original blog, and so the genders were incorrect.
On the ConnectPort X5, the antenna connectors are:
- 3 x female RP-TNC, and;
- 1 x male TNC
So, the adapters (X5 to antenna) need to be:
- 3 x male RP-TNC to female RP-SMA, and;
- 1 x female TNC to ? (depends on antenna)
Or, the antennas need to be:
- 3 x male RP-TNC, and;
- 1 x female TNC
I then had issues with the IP67 23 pin AMPseal connector.
Firstly, instead of using the SWG 16-20 wires (0.5 to 1.25 mm2) recommended, I use some old mains wires that were too fat to follow the TE Connectivity AMPSEAL Crimp Contact, Female, Crimp, Tin Plating 16 → 20 AWG pins, which fit into the holes at the rear of the AMPseal connector.
The wire is also supposed to follow the connector through the black holes and into the green rubber housing.
Because of this I had trouble making the female pins of the AMPseal connector mate with the male pins of the X5. The pins would not push through far enough into the body of the connector, such that, when the AMPseal connector was plugged in to the X5, the female and male pins could mate and power could flow to the X5.
Looking at a couple of explanatory videos on youtube, one issue that I found was that, while the videos state that you you should push the pins in until you feel a click, I felt a number of clicks at different depths, so I wasn’t sure how far to push them in. Also, the videos only show the pins being pushed in from the rear of the connector, but do not show the pins on the mating side, so it is hard to gauge how far they should go in.
This video states that it is important to check, after insertion, that the pins are flush, but flush with what?
So while the X5 powered up if I manually connected the pins, without the AMPseal connector,
the issue is that when I used the AMPseal connector, the X5 would not power up, due to the pins not mating. I obviously had not pushed them home, far enough. I had gone to various depths, ranging from the connectors just inside the rubber, to pushing the through so that they protruded from the back of the front by 2 to 5 mm. In exasperation, I had removed the red locking mechanism in order to see where exactly the pins were, and how far they were protruding.
After it became obvious that the mains wires were too fat to push through the rubber, I then used some thinner wires, but these were too flimsy to be able to apply enough pressure to push the pins in any further than the crimped end of the pin reaching the rubber seal. Also, the added heatshrink didn’t help matters, but even with that, subsequently, stripped off, the pins would not go in any further. The pins obviously needed to go in further than they currently were.
In order to be able to apply more pressure, yet have a thin enough wire to enter the rubber housing, I then soldered some female connectors into stiff copper wire, like so,
but still I was not able to push them in far enough.
It became clear that the small double locking tabs, poking up from the internal base of the connector, need to engage with the ridge on the connector.
However, it is extremely difficult to exert sufficient pressure of the wire, from behind to push the connector through far enough for the tabs to engage with the ridge.
This issue is still ongoing.
It was clear that at least one additional wire to each of the power pins would be required. For example, additional wires to chassis GND (pin 9) are required for some of the other connections, such as the RS232 serial port and the CAN connector:
- Serial port (Chassis GND)
- Ethernet (Chassis GND)
- CAN (??Vin and GND)
as well as [possibly at least] one additional line for logic
- Digitial I/O
- One spare
GND is optional, to the shielding, if any.
I wired, and soldered female connectors to, an old Ethernet cable that I had lying around like so:
- orange/white: TX+
- orange: TX-
- green/white: RX+
- blue/white: N/C
- blue: N/C
- green RX-
- brown: N/C
- brown/white: N/C
These differ from the diagram above as the cable appeared to have been wired up unconventionally. Note that the green is not on the left, and nor is it paired, adjacently, with the green/white wire. The green/white is between the orange and blue/wire wires.
However, the pin out is the same.