Greetings from Geoff

Welcome Geoff! Pleasure to have you here.

What kind of hobbies are you working on currently?

Hi Aaron, thanks for the welcome.

In answer to your question...

At the moment I'm busy trying to fix something which I really should just replace as it was so cheap, but the challenge is calling to me.

I'll start with an apology, this is a long post!

I purchase a string of LEDs, about 50 or so wired in parallel and they came with a driver with a USB plug and a little circuit board with a chip and an IR sensor and a separate IR remote control. The whole thing only cost about £10.00. This was a year ago and it worked fine when I got it, there were lots of different patterns, controlled with the Remote control. 

Of course, when my wife asked me to dig them out this year and spread them through the bushes in front of the house for Christmas, they didn't work. I checked the voltage as it came out of the driver and it was totally dead. The controller is a multi layer board and there are no exposed tracks, so hard to diagnose.

I am more of a software guy than a hardware guy, so I am very naive when it comes to diagnosing hardware problems and creating circuits.

I cut of the driver and put 3 volts across the string and they lit up, a bit more playing about with the bench top power supply and it looked like it was happy at about 3.3 volts @ 35 ma.

For purposes of the description below, I am going the refer to one conductor of the LED string as leg A and the other leg B.

So I put together a breadboard with a couple of resistors and a transistor and drove it from a Raspberry Pi. I believe the Raspberry PI can only deliver 16 ma on its GPIO pins, hence the need for the transistor. I wrote a little C program to control the pattern and the start/stop time and was happy until I noticed that only alternate LEDs were illuminating. A guess and couple of tests later showed that the LEDs had alternating polarity so only half would light when they were wired with the positive to leg A and negative to leg B and the other half would light when the negative was on leg A and positive was swapped to leg B.

So I decided I needed two transistors one for each polarity which meant that the first transistor was setup with positive to leg A and negative to leg B and the other with positive to leg B and negative to leg A. Of course I now had a positive from one transistor on leg A and also a negative from the other transistor on the same leg A, not good. So I added two more transistors to break the connections from the emitter side to ground.

This didn't work. My first attempt to make it work was to swap out the transistors (BC547) for MOSFETS ( IRLZ44N) this didn't work first time, but I swapped some resistor values and it now seems to work.

I don't know if the BC547 didn't work because of the difference in the way they work or if I had messed up the resistor values on that setup.

Does my problem explanation make any sense to you?

If so, is there a much simpler way to make this kind of alternate polarity LED string setup work?

And as a bonus question, do you have any idea why the BC547 might now work when the IRLZ44N do? 

I will try to work out how to draw a circuit and post it with the text next time as it would make it much easier to describe 🙂

I should probably have put this reply in a separate part of the forum, but this is my first experience of using a forum and I'll do better next time 🙂

Thanks,

Geoff

@geoffa 

Check out Big Clive's video on this. He disects one with a battery case instead of USB but it's the same A/C lighting effect as you describe.

@will 

Hi Will,

Thanks for pointing me to that video. The circuit board on my set was almost exactly the same as Clive's but mine, as well as having the USB instead of batteries, also had an IR receiver. Had I seen the video first, it would have saved me a bit of time. I'll search around for examples of the H-Bridge that Clive alludes to, to see how that would work. I've seen them used for driving motors but never given them much thought. It will be an interesting diversion.

Geoff