I decided to fully reverse engineer the Poundland GU10 lamp because it has an interesting LED driver that is based on discrete transistors.
I'm not 100% sure of the purpose of what appears to be a charge pump with a zener as a threshold arrangement. I can only guess that it's either part of the regulation or to protect against open/short circuit scenarios. Note that the associated capacitor seems to be pumped negative to the main transistors emitter, so if the voltage on it is too high it will effectively prevent the main transistor being turned on.
I'm not 100% sure of the purpose of what appears to be a charge pump with a zener as a threshold arrangement. I can only guess that it's either part of the regulation or to protect against open/short circuit scenarios. Note that the associated capacitor seems to be pumped negative to the main transistors emitter, so if the voltage on it is too high it will effectively prevent the main transistor being turned on.
I am not an expert but I can see only one diode at the entrance so not fully rectifying arrangement. Is it a pulsating circuit? How much electrosmog it produces? What differences with Gu10 for halogen? Do the firms want to kill electrosensitive people, with new developments all electropolutans?
Nice video
Life is much easier if circuits are run through LTSpice. It's free and PDG. Not a substitute for design skills but first step before breadboarding. There may be problems with sim of this circuit though (too lazy to work it out) because the oscillation might be furnished by saturation kick , like cheap flash gun circuits of old. It can be done though. But slap the circuit into LTSpice with transformer coupling of maybe 0.9 and it may fly.
Nice method of doing the schematic, easier than using a loupe 🙂
Strange that the 3K resistor is there though. I'd understand having a discharge resistor across the 400 volt cap (something missing from the circuit; it appears to rely on the driver circuit to drain the cap) but not across the 25 volt one.
I'd say you were right with regulation AND open circuit and also stability for the Zener diode part.. Really fascinating circuit! Keep finding them 🙂
Got one of these the other day and removed the 240v driver then put a dc-dc boost board which seems to be running 10V at 80ma and it lights quite well , I dont know what the maximum these led's can take.
Am I right in saying that I could adapt the LED board for automotive use by using an LM7810 and a current limiting resistor? (and of course a heat sink).
I would suspect the 3K resistor is to kill 'chirp' mode.
That is – if you have a parasitic few hundred microamps through capacitance of a long mains wire with a switch on the end – without this, the LED will flash – even when 'off'.
I have over the last few years bought a lot of Chinese GU 10 12 watt 4 led and now single led lamps. They all have a terrible failure rate, probably 30% within a few months. In your opinion which component do you think is failing. Sometimes the light blinks off and on for a few days before finally failing.
Dear Clive, thanks a lot for the effort to reverse engineer this circuit, your videos are always interessting! Proposal: Using an isolation transformer for safety and an osziloscope to measure whats really going on, you could understand and explain these kind of circuits even better, what do you think?
I've noticed that your lamp there uses the same style of LEDs as those used in those rotating disco lamps (the egg-shaped lamp you did a video on awhile back). I'd love to use some to make a warm-white disco lamp.
Great video. They really do teach me quite a bit