I was quite surprised to find that this fairly standard looking battery pack supplied with a set of Halloween LED lights was actually boosting the voltage up so it could run traditionally wired 24V LED strings. It's quite a neat little circuit inside, but has a very annoying flicker at about 50Hz. I'm not sure if it's firing the internal inductor in bursts gated at low frequency or if it's actually just running with 50Hz pulses.
The 50Hz guestimate is based on my probing with a high impedance speaker and getting a distinctive 50Hz buzz. I've not got an oscilloscope where I am at the moment so I can't really probe further to see if it is bursts of higher frequency.
Quiescent current draw seems to be around 50mA.
Once again the chip appears to have a pinout matching a PIC12 or some Chinese clone. I suppose it's an easy job for a PIC to bang out a drive signal of any frequency and mark-space ratio you like. The transistor is a common NPN one and while I initially thought it was a MOSFET with the drive capacitor providing fast turn on while the resistor held it on, I'm guessing that maybe the cap is to drive the transistor with a very short high current pulse with the resistor to maybe discharge it?
Quite a neat little circuit though. I could be tempted to make my own version for fun.
Followup:- I scoped the circuit with a digital storage scope and the output from the chip is a series of fast 8uS on 1uS off pulses gated at 3.5mS on and 4mS off. It's the gating that causes the visible flicker. The on to off ratio was different to what I was expecting. I was expecting it to be the opposite with a higher off time than on time. I guess maybe it's to do with holding the choke on long enough to cause saturation and the short off time lets it dump a spike into the smoothing cap. If the chip is a PIC12 or clone as suspected then the timing fits in with the likely instruction cycle throughput with a rough internal 4MHz clock. 1 MIPS meaning the 1uS off time is probably a BCF command followed by a BSF command. At those speeds the rest of the drive routine is probably spent counting down before returning to the gating routine.
I bet this would work as a constant 50/50 cycle with the following software routine:-
START BSF GPIO,2
CLRWDT (Clear watchdog timer - 1 cycle)
NOP
NOP
BCF GPIO,2
NOP
GOTO START (2 - cycle jump)
The 50Hz guestimate is based on my probing with a high impedance speaker and getting a distinctive 50Hz buzz. I've not got an oscilloscope where I am at the moment so I can't really probe further to see if it is bursts of higher frequency.
Quiescent current draw seems to be around 50mA.
Once again the chip appears to have a pinout matching a PIC12 or some Chinese clone. I suppose it's an easy job for a PIC to bang out a drive signal of any frequency and mark-space ratio you like. The transistor is a common NPN one and while I initially thought it was a MOSFET with the drive capacitor providing fast turn on while the resistor held it on, I'm guessing that maybe the cap is to drive the transistor with a very short high current pulse with the resistor to maybe discharge it?
Quite a neat little circuit though. I could be tempted to make my own version for fun.
Followup:- I scoped the circuit with a digital storage scope and the output from the chip is a series of fast 8uS on 1uS off pulses gated at 3.5mS on and 4mS off. It's the gating that causes the visible flicker. The on to off ratio was different to what I was expecting. I was expecting it to be the opposite with a higher off time than on time. I guess maybe it's to do with holding the choke on long enough to cause saturation and the short off time lets it dump a spike into the smoothing cap. If the chip is a PIC12 or clone as suspected then the timing fits in with the likely instruction cycle throughput with a rough internal 4MHz clock. 1 MIPS meaning the 1uS off time is probably a BCF command followed by a BSF command. At those speeds the rest of the drive routine is probably spent counting down before returning to the gating routine.
I bet this would work as a constant 50/50 cycle with the following software routine:-
START BSF GPIO,2
CLRWDT (Clear watchdog timer - 1 cycle)
NOP
NOP
BCF GPIO,2
NOP
GOTO START (2 - cycle jump)
A similar circuit is commonly used to generate high voltages for Geiger counters, although a higher voltage transistor is used. Sometimes driven by a 555 or an arduino.
This is the perfect inductor for your garden light. 🙂
How about using a joule thief??
Out of interest, where did you buy these? 🙂
i love what you do
If you want to find out the frequency there's a downloadable frequency counter/oscilloscope which is quite handy on a PC if you can hook it up to the sound card (via a resistor!) It's called "musical tuner by W A Steer" it hasn't been updated since 2001 but it's very useful and surprisingly accurate.
Well thats an odd way of doing it.. Pretty neat! Pretty odd that it's running at such a low frequency, as it would be much more efficient at 20+kHz with these little components.