More thought required: The simplest designs aren’t so simple
I'm in the process of designing an I2C system that uses telephone cable to route the signals around. It uses standard four-conductor cables with RJ-11 type plugs on the ends. For my system, it is important that there be no twists in the cable. Getting the signals reversed (clock and data) is not too bad. It will just prevent the system from working. What is bad is getting the power and ground reversed. It will significantly warm the cable.
Now, it is possible if your eyes are good and you squint to just give the connectors a visual inspection to check the wring. If the colors are in the same order, you are fine. Well, I don't like squinting. Besides, at my age, I don't squint as well as I used to. So building a simple tester is in order. Six LEDs, three resistors, and two RJ-ll jacks. How hard could it be? So I did a quick design, laid out the board, and had three produced by OSH Park. Here is the schematic.
After the order was in process, I noticed something on my schematic. If the wires going to pins 2 and 3 or those going to pins 4 and 5 are shorted together, a bad cable with shorts between those adjacent pins will still show good. Well, that is less than optimum. So I did a redesign. Got boards produced. Built one up. And started testing. Here is the schematic for that version.
And then a funny thing happened. I plugged in a commercial cable with only one pair of wires (I found out a little later) and nothing lit up. This is not bad, since such a cable is useless for my system. But the tester does not give all the information easily available. So on to design three. Here is the schematic for that version.
This design is a final. It will not show all possible faults, but any opens will show up as shorts between any two connector pins, reversed wiring, and if there is only one pair of wires available, it will show that as well.
I also found out something interesting. All the commercially built cables I tested were wired-reversed. That does not matter at all for modern telephones which don't use E and M signaling, where a temporary reversal of polarity is used to send information down the wires. But if you have a very old telephone (rotary) connected to a Bell System, it may make a difference. Why don't they make the cables the "right" way? Probably because it would raise the cost of the assembly machine.
And the moral of the story? Even the simplest designs can require more thought than you might think.
If you are interested in building a tester, the parts list, parts layout, and board schematic can be found here. You can get the building and operating instructions at Space-Time Productions. And OSH Park sells the boards for $2.45.
M. Simon's e-mail can be found on the sidebar at Space-Time Productions.
Engineering is the art of making what you want from what you can get at a profit.