Tip:
Highlight text to annotate it
X
Hi Folks! I received a lot of questions and requests to test things
with my joule thief lighting a compact fluorescent lightbulb. So as
a follow-up to my "How to Make Joule Thief Light a CFL" video,
I thought I'd make this video. I even came up with a pretty cool
solar cell test to do with it.
The first thing I did was to desolder it all and convert it to
a breadboard for easy testing.
The first test is to add batteries and see if the light is
brighter. So I adapted a box with the lightbulb inside so that
it can be nice and dark. I put a hole in one side for my
camera to look in without letting in a lot of extra light.
The best way I found to connect the batteries without influencing
things too much was to use electrically conductive, neodymium
magnets. I used them to connect the batteries together and to connect
the alligator clips too, all of which are attracted to the magnets.
I first tried with two 1.5 volt AA batteries,... then a 3rd,...
and then a 4th.
As you can see, there's only a little improvement. Though the
more batteries you have, the longer it will run for. And this
was with everything on my camera set to manual and manually adjusted.
I was asked what effect the variable resistor or potentiometer has.
As you can see, decreasing the resistance increases the brightness and
increasing the resistance decreases the brightness. So less resistance
is brighter.
Here I'm using a 9 volt battery instead of the usualy AA batteries.
And as you can see it makes a big difference, though I've found that
without the potentiometer turned to a high resistance, the battery
doesn't last long.
And now for the oscilloscope to see the voltage waveforms. Here
I'm putting the probe in the scope's channel 2.
First I connect it directly across the batteries, so you can see
the batteries' voltage. It's a nice steady, flat 3.09 volts DC.
I turn on the joule thief,... and freeze the display.
You can see when it's on that the voltage at this point in the circuit is anything
but flat DC. In fact you can see the voltage is usually close to
zero with periodic spikes of around 504 millivolts,... and at
a frequency of around 17.24 kilohertz.
Zooming in, you can see that the spikes are actually
a dampened alternating wave with sharp rises and falls.
The household 60 hertz AC is also having a bit of an effect here.
Next I connect a probe to channel 1. I connect it up to
measure voltage at the normal joule thief output, across the
transistor's collector and the battery negative. Notice that there are spikes here too,...
and switching the cursors to channel 1,... that the peak voltage is
around 38.8. The frequency is also around 17.24 kilohertz,...
in sync with the spikes across the battery.
And now I move channel 1 to the output of the coil going to the
compact fluorescent lightbulb, while still leaving channel 2
across the batteries. When I unfreeze the display you can see that
the peaks of the output also also in time with what we saw
at the battery.
And lastly, I disconnect channel 2 from the battery so we're left
with just the output at the lightbulb. Bringing it up so we
can see the bottoms at least, I zoom in and you can see that this
is really an alternating voltage in the form of a dampened wave.
It's lowest voltage peak is around 228 volts. But scrolling up as far
as I can, I can't even get to the top of the highest peak. Since I've
scrolled down from here 500 volts and there's still 200 volts here,
that means it's over 700 volts.
But if I increase the resistance I bring the peak down, as well
as increase the frequency. Now it's around 700 volts.
Next I scroll back down and get the frequency. It's around
23.81 kilohertz with a 700 volt high peak.
And now for the solar cell test. It's unlikely anyone can tell with the naked
eye if the lightbulb is flickering as a result of the periodic voltage
spikes that keep it going. So instead I got out a solar cell.
The solar cell is powered by the light from the lightbulb. To get enough
light I'm powering the joule thief using a 9 volt battery, though it's
around 8.4 volts at this time. I close the box to eliminate outside
light, and turn it on. Channel 2 on the oscilloscope is the voltage
output of the solar cell. And channel 1 is the output
of the joule thief that feeds the lightbulb. As you can see there
is a correlation since their frequencies match.
Remember, this is really measuring the optical output of the lightbulb,
and given how well it matches the electrical output at the joule
thief, the solar cell's response time is fairly good.
Most of the time the solar cell is seeing a fairly steady amount
of light. These gentle slopes represent a gentle dimming of the light.
This frequency is around 31.25 kilohertz, or 31 thousand times a second,
which is too fast for the human eye or brain to register so the gentle
dimming goes unnoticed. This spike-like activity here is around
1 microsecond long, which is also too fast to register. It is
interesting that the voltage actually reverses here.
After turning it off, I open the box, and this is what
the solar cell output looks like in its dark corner, pretty flat.
That's normal.
This 2N3055 transistor is an NPN type transistor. I was asked
if a PNP type transistor could be used instead. From searching the
web, I knew it could with some modifications. So here's an
NTE219 PNP transistor. I put it in the circuit with the emitter,
collector and base going to the same places as the NPN transistor
was. But to make it work you just have to reverse the batteries,
switching the positives and negatives. And here it is in action.
Many people wanted to know how long the AA batteries lasted when
powering the joule thief. So with some brand new batteries in place,
I put a sheet of paper with writing on it in the box next to the
lightbulb,... and connected a volt meter across the batteries.
Before turning on the joule thief I noted that the batteries voltage was 3.11 volts.
I then turned on the joule thief and adjusted the potentiometer's
resistance while looking in the hole until the light was just
bright enough to read the writing on the paper. Checking the resistance
later, it was 34.3 ohms. I then inserted the camera and
adjusted the manual settings to match what my eye saw.
I kept track of the time that the joule thief was on, along with
when I took video of the light.
But the results weren't great. The light remained bright like this for 25
minutes running time. Then for the last 30 minutes it was dimmer
until finally after a total run time of 55 minutes, it dimmed and
then went out.
Also, every 10 minutes or 15 minutes the transistors and batteries would
be getting hot so I turned everything off and waited 10 minutes before
turning everything back on again. But I wanted better.
So I replaced my 13 watt compact fluorescent lightbulb with a
5 watt one instead. I also mounted the transistor on a heatsink.
I put in new batteries and tried again. This time the I could get the same brightness
with the potentiometer set to a resistance of 69 ohms,
which meant the current and the heating would be less to do
the same. Sure enough, the transistor felt cool throughout
and the batteries seemed okay. It remained bright for an hour
and a half and was out completely when I checked 15 minutes later
for a total run time of one hour and 45 minutes. It did seem
to die more quickly at the end so maybe there was some battery
heating after all. But given it was bright for three times as
long, it was a big improvement.
I didn't think measuring current would give accurate results
since all the points in the circuit are far from DC or sinusoidal
AC but a lot of people seem to do it anyway. So I tried it.
Unfortunately, the light wouldn't come on when I tried with either
my digital or analog meters.
I was asked if it would create interference with things. So here
are the things I had on hand for testing. First, I swept the entire
AM radio band and here's an example of the interference.
I couldn't pick up any stations on the shortwave band from
3.2 megahertz to 22 megahertz but there was noise added.
I didn't notice any interference with across the FM band, and that
was even with sweeping the joule thief frequencies using the resistor.
Here's an analog TV receiving digital stations using an
antenna, which are then converted to analog. But there's no interference.
And lastly, here's an LCD screen, also with no interference.
Well thanks for watching!
See my youtube channel, rimstarorg, for more videos like this.
That includes the one where I show step-by-step how to make this
joule thief circuit. Another on how to attach a transistor to a heat sink.
And for variety, how to get a piezoelectric crystal from a
lighter for easily creating sparks.
And a big thanks to Todd Harrison for reviewing a large chunk of this video.
Be sure to check out his great electronics and DIY videos on his channel.
And don't forget to subscribe if you like these videos, or give
a thumbs up or leave a question or comment below.
See you soon!