I was soooo close. My Last board of the kit and I've got a single string (30x negative, for what it's worth) that don't light at all, not even dimly. Every other LED, all 475 of them, behave themselves.

I've double checked that the LEDs are in the right way, there's only one jumper, and I can't find any duff solder joints. I've give the rest of the board a once over and can't see anything obviously wrong

Now I've looked at the schematics and can't think what else could be wrong (as everything else is working), but I did make it obvious to Windell how likely I might be to be waaaayyyy off here

Is there anything else I should look for before I desolder this string? What's the likelihood I've trashed a LED? is there any physical change I can check for if I have?



-plik

Let me get this straight-- The 30x positives are working correctly but the 30x negatives are not? Then that does lead to the conclusion that something is wrong in those six locations. We know that the chip is doing its job correctly, for example.

If they are really off, that means that no current is flowing through them. First, there could be no voltage across this set of components if there were bad or missing solder joint on one or the other end of the chain, or if there were a short circuit from one end or the other to ground or another place that it should not be. If you have a multimeter, you could probably check the voltage in the middle of the chain versus ground, and see what it's doing.

The second possibility is that you do have a bad component-- there usually isn't much way to tell. If a component has failed, it has failed in the manner of an "open circuit" leaving no connection between the two pins. It will not create a big hazard to run it with one fewer LED for a moment, so you could test each connection by shorting out the pins of each of those LEDs (and the jumper) one at a time-- if one of them is bad, that should point out which one. (Equivalent test if you have a multimeter: Check the voltage between the two pins of each LED. If there is a "large" voltage across the leads of one component, that one is the problem.)

Quote by: Windell

Let me get this straight-- The 30x positives are working correctly but the 30x negatives are not?

Yep that's right

Then that does lead to the conclusion that something is wrong in those six locations. We know that the chip is doing its job correctly, for example.

Phew, that means I've not forgotten everything I know about circuit diagrams

If you have a multimeter, you could probably check the voltage in the middle of the chain versus ground, and see what it's doing.

..you could test each connection by shorting out the pins of each of those LEDs (and the jumper) one at a time-- if one of them is bad, that should point out which one. (Equivalent test if you have a multimeter: Check the voltage between the two pins of each LED. If there is a "large" voltage across the leads of one component, that one is the problem.)

I'll give this a go after work and let you know.


Thanks for the help.



-plik

Quote by: Windell

If you have a multimeter, you could probably check the voltage in the middle of the chain versus ground, and see what it's doing.

nothing. So I went along the chain measuring between LED and ground, all nothing except the last point in the chain.

It will not create a big hazard to run it with one fewer LED for a moment, so you could test each connection by shorting out the pins of each of those LEDs (and the jumper) one at a time-- if one of them is bad, that should point out which one. (Equivalent test if you have a multimeter: Check the voltage between the two pins of each LED. If there is a "large" voltage across the leads of one component, that one is the problem.)

also nothing, which leads me to believe there may be two duff components in this chain, would this make sense?

I some how came to the conclusion that 305 was the issue (I thought it was 300) and so started to remove it. I realised it wasn't 300 shortly after the head of the LED snapped off which leads me to my next question; got any good tips for removing tiny bits of legs from PCB? soldersucker doesn't seem to be working too well. thinking off pushing from the top side with one of the thousands of trimmed legs I've collected but don't want to make things worse.

Removing dead parts of LED legs is not easy... Yes, you can push it out, and sometimes that's best. I would try to use the solder sucker to fully loosen it before pushing, though.

Things to check, all with board powered off:
1. Exactly ONE pin of DN300 is connected to R310 (i.e., ~ 0 ohms resistance between them).
2. Exactly ONE pin of DN305 is connected to +24V
3. Resistance across each LED is high, not zero.

Also, does your multimeter have a "diode" setting?

Quote by: Windell

1. Exactly ONE pin of DN300 is connected to R310 (i.e., ~ 0 ohms resistance between them).

yep

2. Exactly ONE pin of DN305 is connected to +24V

yep, well what's left of them

3. Resistance across each LED is high, not zero.

yes with the exception of DN3032 which is about 50ohm, is this the smoking gun?

Also, does your multimeter have a "diode" setting?

yep

yes with the exception of DN3032 which is about 50ohm, is this the smoking gun?

Probably so.

If you use the diode test feature on your multimeter (with the circuit off) you should be able to *very faintly* light up each of the LEDs in turn to check them.

sorted!

the diode setting highlighted an issue with an LED, but after replacing it it still wouldn't work. tested again and all seemed well with all 5. Continuity testing *between* them showed a broken track, from the freshly replaced LED, so I probably wrecked it when removing the previous one with extreme prejudice.

soldered a 0 ohm between the pins and I now have a fully operational final board. It's not the prettiest of solutions, but noone will ever see it.

once I track down the right size bolts I can mount them on the board I have, then it's just a simple case of building a table to put it in!


Thanks for all your help, I'll be sure to add more pics to EMS aux as I progress.