So I've just managed to snatch 5 minutes to take a quick video of the CatsEyes project I'm working on now. As I explained in my previous post it is a fun little project to test continuity in all eight conductors of an ethernet cable. If a shielded cable is plugged into the device the shielding is also tested for continuity.
The finished product will have three simple modes of operation, slow, fast and manual. These reflect the speed at which the LEDs change. Slow is 1Hz, fast will probably be 3Hz and manual will only change when a button is pressed. Manual will be a particularly useful mode when trying to fix a flaky connection on a single wire in a patch panel. The three modes will be toggled by way of a 2nd switch. With a 3rd switch used to turn the device on or off.
A difficulty I encountered early in the design stage was in creating a return path for the current from the far-end device. Online examples by other people were in short supply but of those I found the wires were merely being tested as four pairs rather than eight singlets. So the far-end would simply bridge wires 1 and 2 together, for example. If there were no breaks the LED on the near-end would light. The obvious short-coming of this design is that you do not know in which of the two wires a bad connection might be. I finally stumbled across the solution by connecting all the wires together at the far end with diodes and allowing the processor to sink the excess current back through it's output pins. If this is all a bit confusing the circuit diagrams should help clarify what's going on.
Not featured in the video is that the near-end should also have its own set of 9 LEDs. Preferably in series with the 9 on the far-end. In this way a break in a wire within the ethernet cable will fail to illuminate the LED at both ends of the device. The problem with this plan is it would appear to require at least 4v to drive two LEDs in series yet I am hoping to use a 3v CR2032 with holder I scavenged from a dead motherboard.
My other issue is placement of the LEDs on the stripboard. I would love to replicate the design of the far-end device, with the LEDs on the opposite edge to the 8P8C/RJ45 socket. But as you can tell from the circuit diagram the LEDs really want to be in the middle of the whole thing. I guess I'll be using a lot of carefully bent wires to hop back and forth?
Let me know what you think so far, any improvements you can suggest? Or problems you can foresee that I have missed through inexperience?