This one started out as a pulse motor build,but took a left turn when i noticed the battery voltage climbing when it went into self oscillation.
http://www.youtube.com/watch?v=2fvLwjvTA60
TinMan
So your running a load and very slowly charging a nicad battery with no additional input.....
Some folks call that looping? Other folks say Its "loopy"!!
Better get out the rain coat just in case ..... ^-^
Thx
Chet
Brad, that is very VERY cool, not just cool.
Chet, umbrella as well might be needed >:-)
Wasn't Steven doing something like this? I'm sure he will like this a lot, especially when I believe someone took a machine gun at him for his tests I also believe.
Nice instructive vid as usual Brad O0 the area of frequency is interesting, I find a lot happens between 3-8khz for some reason :-\
Mike 8)
tinman,
Have you tried with a light blocking hood over the LED?
LEDs may also act as photoelectric cells.
>>Edit:
I was still watching your vid when I posted the above.
You answered my question by turning your shed inverter off causing the lights to go out.
I had one of these that ran for several months with no external power sources applied. I checked the same possible interactions you did.
What I determined was the extra energy was be added by me when I had to jostle or disconnect-reconnect the circuit or turn the lights on to check it each day to keep it on track to charge the battery vs. drain it.
forgot.....
One LED is very unlikely to be a hidden source of charging power but if you go the route of other experimenters and 'load' your JT down with a huge array of LEDs you are actually just adding solar cell capacity as the charging source.
I bite my tongue when I see folks 'loading' their JT with huge LED arrays or seemingly useless diodes in the circuit in a well lit room. ;)
Glass diodes do a fairly good job acting like solar cells.
I had to play with mine every morning because the lights were out the previous night C.C
Don't forget that batteries can gain voltage capability in the short term by pulsed elimination of sulphation thus reducing internal impedance.
Also there is the battery warming effect. Tinmans increase is on the order of 2%. I suspect if left running until mostly desulphated, the voltage rise will then turn around.
I just connected one of those not quite so white white LEDs from a garden lamp my wife told me to fix to an AA from the same light.
It lights (barely) and the battery voltage started to climb under the bench light after a few minutes. The LED puts out 1.7V without the battery connected and held close to the bench light.
ION,
Yes. Another tricky part to these battery circuits and why most folks fade away when they replace the battery with a cap. Of course, some well known folks just claim a battery and a cap don't work the same way. They are correct ;) but not for the wild reasons given.
Those batteries the "1.5 Volt rechargeable's"
I believe they were chosen to stay away from the Lead acid battery " Issues/Anomolies" ?
thx
Chet
I often see that effect, the power draw doesn't drop the supply battery much but the charge battery can build a voltage in its capacitance or something while under the charge of the spikes so the two batteries in series show a raised voltage. The LED won't draw any current below a certain voltage which may well be higher than the charge battery voltage and is in series with the charge battery so the spike is clipped by the LED and the battery get's the rest. Those AA's can hold a lot of Milli aH charge and if the supply one is in good condition and the charge battery not liking to take a charge the supply battery won't drop much but the charge battery will raise voltage quicker if it
has certain problems (higher internal resistance).
Cheers
What's most puzzling is how the circuit is
able to oscillate. The scope waveshape is
that of the classic Blocking Oscillator but the
circuit diagram doesn't seem to support such
operation.
Any clues?
Hi all
I would like to comment on some of the thoughts above,and also give the readings from the meters after the cool joule was running all night in the dark.
Total voltage is now 2.487
Run battery-1.222 volts,a drop of .006 volt's
Charge battery-1.259 volt's,a rise of .154 volt's--This was since the last time i measured the charge battery voltage alone.
Both these battery's are new-first timed used-straight out of the garden lights.
WW-One LED is very unlikely to be a hidden source of charging power but if you go the route of other experimenters and 'load' your JT down with a huge array of LEDs you are actually just adding solar cell capacity as the charging source.]
Not possable,as the current flow from the LED would be in the wrong direction to charge.I also have a diode on the collector that would stop current flow from the LED in that direction.The transistor would also be open when and if the current could flow in that direction.I think you will find that all ss system's would not be receiving a charge from the LED's,as the current flow is in the wrong direction-and also because the collector is open when the current from an LED array could flow.Even when the C-E junction is close-the positive charge would be flowing straight to the negative of the run battery-as the emitter on the npn is ground(or negative).So i cannot see any way that any number of LED's could be adding to the system?.
@ION-Don't forget that batteries can gain voltage capability in the short term by pulsed elimination of sulphation thus reducing internal impedance
As i said above,the battery's are new-first time used.They are also nicad's-so no desulphation taking place here,only the conversion of potassium hydroxide to cadmium hydroxide as the battery discharges-so no sulphide involved here.
@WW-Yes. Another tricky part to these battery circuits and why most folks fade away when they replace the battery with a cap. Of course, some well known folks just claim a battery and a cap don't work the same way. They are correct but not for the wild reasons given.
The only problem i see is that although a cap;s parallel resistance is low(like a batteries),the series resistance is very high.As we have the batteries hooked in series,this may be a problem when switching to cap's-as the circuits resistance will rise?.I do intend on switching to cap's,but i will use super caps to try and bring the series resistance down as far as we can.There is also that lose you get when charging one cap with another-i believed Ex and poynt mentioned this a couple of time's.
I have been playing around with pulse motors and ss system's for many years now,and am well aware of battery effect's that can take place when using SLA's.I also see nothing special about this simple oscillator,other than i have never seen this happen with nicad's.The thing to remember here is that we dont have high voltage spike's charging a battery-the voltage the charge battery is receiving is only 2.42 volt's(after the LED has had it's bite at the voltage available).
The reason i like using nicad's is because to gain voltage when charging,you must convert one chemical back to another.The amount of this change determines the charge within the battery-unlike SLA's where you can have a surface charge form on the plate's them self.
Quote from: Dumped on 2013.04.28, 00:54:00
What's most puzzling is how the circuit is
able to oscillate. The scope waveshape is
that of the classic Blocking Oscillator but the
circuit diagram doesn't seem to support such
operation.
Any clues?
I am also a bit stumped as to how the circuit still oscillates with the two coil's(run/trigger)so far apart?.-as seen in the video.
Quote from: tinman on 2013.04.28, 01:17:32
Not possable,as the current flow from the LED would be in the wrong direction to charge.I also have a diode on the collector that would stop current flow from the LED in that direction.The transistor would also be open when and if the current could flow in that direction.I think you will find that all ss system's would not be receiving a charge from the LED's,as the current flow is in the wrong direction-and also because the collector is open when the current from an LED array could flow.Even when the C-E junction is close-the positive charge would be flowing straight to the negative of the run battery-as the emitter on the npn is ground(or negative).So i cannot see any way that any number of LED's could be adding to the system?.
If you have a chance, please post a schematic, in any form you wish.
I'm sure all SS systems would not be prone to charging from the LED in the circuit but the typical blocking oscillator is prone to that function even when a diode is in series with the LED, especially then, if that diode is a typical silicon glass bodied switching type and that diode is polarized the same as the LED.
The diode and LED act like two solar cells in series.
The output polarity of an LED in solar cell mode is the same as in LED mode. Where you apply negative to light the LED is where the output of the LED is negative when shining light on the LED. So, it is quite possible to charge a battery with one component, the LED. The same simple circuit with a diode added to it to make sure current only flows one direction will apply an even higher voltage when charging the battery under a light source.
Below is the schematic for the circuit.As you can see,it is just the ss ssg circuit with an extra LED on the charge output-along with high ohm coil's.Also note that there is no way for any current from any number of LED's to pass back to either of the batteries.
Edit-this is an air core system,so please pay no attention to the core depicted in the schematic.
Hi have just found out that even though all these solar lights are the same kind,some have nicad's and some have Ni MH batteries in them.I havnt worked with metal hydride batteries befor,so im not aware of how there chemical makeup reacts with pulse charging-but we will find out shortly.
Thanks for the schematic, Tinman.
Very interesting -- and reminiscent of earlier circuits in some ways...
Can you check the Frequency of the oscillation? Is it 50 Hz (Oz)?
Or is there a 50 Hz "carrier" along with higher-frequencies?
Thanks for the posted circuit.
I'll back out to leave you to your excellent work. Just keep in-mind that an LED as a source or sink has the same polarity.
WW
Hi WW
I dont know yet if there is anything in this,but i am doing a rundown test on the first charged battery now-along side a battery that was charged by conventional means.I do know that the LED has the same polarity when used as a source or sink-but the current flow as a source is opposite to that of it being used as a sink.The circuit is just the berdini ss ssg circuit,but with high ohm coil's,an LED and NiMh or nicad batteries.But the ability to remove what is the trigger coil away from the drive coil and still have it oscillate is the bit that has me stumped?.I can move it out of magnetic range-and it still runs-Mmm?.
Hi PhysicsProf
The frequency is 5.856kHz.There is no sign of the 50Hz of mains power within the signal.I also switched of the grid power to the shed via the isolation switch in the video,so as to eliminate that chance-but the battery still kept charging.
Tinman
That's what interests me as well. I've built a similar circuit in the past and I could get self oscillation under certain conditions, but the coils had to be in close proximity, however I've never tried it with those beefy looking audio transistors you have.
As far as the voltage climb, that's the old debate. One battery has a different voltage current slope then the other charged battery, so since the current draw is so small you won't see it change in a day or few days, but the other since its almost drained, will show a faster voltage change.
The old trap that "Bedazzled" fell into! lol
EM
Quote from: tinman on 2013.04.28, 13:34:08
Hi WW
...the ability to remove what is the trigger coil away from the drive coil and still have it oscillate is the bit that has me stumped?.I can move it out of magnetic range-and it still runs-Mmm?.
With any circuit involving long lead lengths and frequencies in and above the kHz range there is a strong possibility that unexpected capacitance is also involved.
A good test of this idea would be to rearrange the wiring so the base-emitter circuit leads are at right angles to the collector-emitter circuit or create more separation between wires running parallel to one another.
What's also interesting, there is no DC bias?
QuoteBut the ability to remove what is the trigger coil away from the drive coil and still have it oscillate is the bit that has me stumped?.I can move it out of magnetic range-and it still runs-Mmm?.
The single FET oscillator circuit I posted a while back would work with 4 feet or more separation....those em waves are not afraid to travel.
Have aether? we'll travel. :D
Quote from: EMdevices on 2013.04.29, 13:15:00
Have aether? we'll travel. :D
Not sure what we have just yet EM lol.The same run battery is still going,and second charge battery is in the unit.The first charged battery ran the garden light for 3h50m-so about 40 minutes short of a battery charged conventionally O0-keeping in mind that we only used .003 volts from the run battery to charge it ^-^.
The run battery has drooped .007 of a volt from start voltage two days ago-so not doing to bad,and running the LED aswell-although not very bright.
tinman,
that is impressive, but I must ask, how discharged was this first battery? (I'm assuming it was charged by the circuit from the run battery)?
QuoteThe first charged battery ran the garden light for 3h50m-so about 40 minutes short of a battery charged conventionally -keeping in mind that we only used .003 volts from the run battery to charge it .
If you want to get fanatical with this circuit, I suggest seting it up to measure the Power Input and Power Output, similarly to how Tesung does it, I would suggest a resistor for measuring the current, by observing the voltage drop across it, however, unlike what Tesung does, I suggest building an OP Amp to amplify this low voltage, maybe 100 times, since you have so little current flow from the run battery. I think a 10 ohm resistor would be fine as well. The Op Amp would have to be calibrated accurately, but its doable. As we all know, measuring the current flow is the most accurate way to determine power flow out of a battery, because voltage alone is not enough especially around mid-charge when it doesn't change much.
EM
@EM
The charge battery was discharged in one of the solar garden light's.It stops running when the battery voltage gets to around .74 volt's,so the battery was quite discharged when i started the charge cycle.
Here is day 3 update video.The larger coil circuit(lower resistance)didnt do aswell,but that may have been the tip2955 pnp transistor i used?.
http://www.youtube.com/watch?v=kzu4-rTQOok
TM
Off topic
I brought you something to Chew on over lunch.
http://www.ijee.ieefoundation.org/vol3/issue1/IJEE_13_v3n1.pdf
From here
http://www.overunity.com/13286/resonance-and-hho/msg359035/#new
Thx
Chet
@Chet
That is interesting,but very hard to believe-Mmm :o But i guess with the many claim's out there that turn out to be bogus,it isnt hard to think that way.This is why i post what i have,and the guy's here that are better at reading a scope,can tell me what im seeing-although i am learning fast on scope's(online course).
Now here is a little something that should have the best of them thinking.-Now how exactly is the trigger coil powering the base of the transistor?
http://www.youtube.com/watch?v=z7DlD8MIEes
I posted the answer.
Quote
Very easy.
There is a magnetic storm in the environment. Here the shishing in the background?
That is a wide spectrum noise generation. Magnetic white noise.
Try putting the trigger coil in an aluminium tube.
Chet,
The paper you referenced in #27 above (http://www.ijee.ieefoundation.org/vol3/issue1/IJEE_13_v3n1.pdf) is
interesting. It would seem to confirm other
studies which have been done on the pulsed
electrolysis phenomenon where increased
efficiency of hydrogen and oxygen production
has been documented.
It also would seem to confirm what Bob Boyce
had discovered regarding nanosecond pulses
at certain critical frequencies.
There are questions regarding the simplified
schematic diagram in the paper and how the
pulses are actually produced but if the process
can be duplicated experimentally it would be
a major breakthrough.
The circuit does bear a similarity to the Stan
Meyer Voltage Intensifier.
We shall see...
TinMan,
Your oscillator circuit which makes use of
the 3055 Power Transistor seems to defy
logic and common sense, but there must
be an explanation somewhere in there.
Have you discovered some unknown
characteristic of this transistor which is
not normally used in low voltage/low power
circuits?
Is the transistor type critical at all or can
others be used to attain the same result?
The DC current gain of a 2N3055 lies between 20 and 70 at a current of 4 Amps. You have way less than 4 Amps so the current gain can be as high as 125 at room temperature. This is certainly more than enough gain for oscillator criteria.
The DC resistance of your coil is 178 Ohms so with 1.2 volts your max load can be no more than 6 mA. This is well off the chart for a 2N3055 gain curve.
Your tickler coil in this case could be called a "loose coupler" (old radio jargon) and will work just fine as a drive for the 2N3055 as long as long as you don't significantly increase collector current.
Your aluminum tube is not really a shield as magnetic pulsations can easily pass through or enter the ends. It needs to be completely enclosed in a ferromagnetic high mu shield to disrupt operation.
A system with positive feedback and gain a bit greater than "1" will satisfy oscillation criteria under the correct circuit conditions. You have way more than a gain of "1" so the continued oscillation with increased distance between coils does not surprise me.
I was able to do this with a simple FET oscillator.
From your scope waveform it looks like you are not hard switching the transistor, rather just barely tickling the base.
Quote from: ION
Your tickler coil in this case could be called a "loose coupler" (old radio jargon) and will work just fine as a drive for the 2N3055 as long as long as you don't significantly increase collector current.
What is so puzzling about how the circuit
operates is that once oscillation is initiated,
the "tickler coil" seems unnecessary insofar
as feedback is concerned.
Sweeping a magnet past the tickler is sufficient
to initiate transistor turn-on and startup the
oscillation, but one must wonder what the
actual feedback mechanism is which sustains
the oscillation. Is it internal to the transistor
such as the Collector/Base capacitance or is
it something else?
Perhaps TinMan will substitute a small signal
transistor such as the 2N2222 in place of the
3055 to make comparisons.
Dumped:
I agree with what you say. Under certain conditions the Miller effect capacitance (C-B) can form an oscillator with the B-E capacitance and shunt inductance. This is quite the possibility. Also stray capacitance external to the device can contribute.
Your idea is good too replace the 2N3055 with a 2N2222 or equivalent. 6mA is well within the collector current rating.
tinman,
Have you tried disconnecting everything going to the base?
If it still oscillates I have your answer C.C
This type of oscillator does oscillate because of the NPN or PNP transistor junctions.
For a NPN oscillator the NP layer between collector and base will charge the trigger
coil each time the pulse in the power coil goes negative. For a PNP transistor based
oscillator the positive pulse in the power coil will charge the trigger coil. There is no
need to have a magnetic coupling between the trigger coil and the power coil to get
this type of oscillator to oscillate. If you want a self start oscillator then put a capacitor
in series with the trigger coil and a bias resistor to get the base over the trigger
voltage threshold.
GL.
Oh!
GL is right. I didn't realize your oscillator wasn't self-starting and should have known from your schematic.
Had it been, I suspected your polarities were reversed for the transistor in use and you built yourself an NRO (negative resistance oscillator).
Quote from: ION on 2013.05.01, 18:57:42
Your aluminum tube is not really a shield as magnetic pulsations can easily pass through or enter the ends. It needs to be completely enclosed in a ferromagnetic high mu shield to disrupt operation.
From your scope waveform it looks like you are not hard switching the transistor, rather just barely tickling the base.
Hi ION
The tube is steel,and has two large magnets inside it. I can not see any small weak magnetic field effecting the trigger coil in this situation.
I watched the latest tinman video, and I agree that the oscillator feedback is predominantly through the transistor's own internal capacitance, between collector and base. No magnetic coupling is needed between the coils.
EM
PS
Here's a circuit simulation I just put together in LTspice. I had no problem starting the oscillations with a quick pulse. Note: the coils are not magnetically coupled.
Quote from: EMdevices on 2013.05.02, 02:30:17
I watched the latest tinman video, and I agree that the oscillator feedback is predominantly through the transistor's own internal capacitance, between collector and base. No magnetic coupling is needed between the coils.
EM
PS
Here's a circuit simulation I just put together in LTspice. I had no problem starting the oscillations with a quick pulse. Note: the coils are not magnetically coupled.
>>"oscillator feedback is predominantly through the transistor's own internal capacitance, between collector and base."
No, not because of capacitance, but because of the DIODE layer between the collector and base.
GL.
Quote from: Groundloop on 2013.05.02, 09:01:28
>>"oscillator feedback is predominantly through the transistor's own internal capacitance, between collector and base."
No, not because of capacitance, but because of the DIODE layer between the collector and base.
GL.
With a diode layer between B&C,would it not act like a capacitor?.
Your comment has given me an idea GL-diode layer between C&B ;)
Quote from: tinman on 2013.05.02, 09:13:43
With a diode layer between B&C,would it not act like a capacitor?.
Your comment has given me an idea GL-diode layer between C&B ;)
Tinman,
Yes, a diode also have capacitance and this can be used in HF applications to get a signal path or
to tune a LC tank circuit by varying the voltage over the diode. In you circuit it is the PN diode
path between collector and base that is the main reason for charging your trigger coil so that
the oscillation can happen.
GL.
Wavewatcher,
you were right, it is indeed an NRO.
I replaced the coil on the base, and just give it an initial pulse to the base to start it, then it is electrically shorted to ground. I've played with NRO's before if you remember, so it should of jumped out at me, but that coil on the base confused me.
edit: I haven't played with LTspice in a while, and just discovered I'm being an idiot. If I simulate for a longer time it becomes apparent the waveform decays, so the gain is < 1 :-[
EM
@Tinman
I have a close but not perfect replication of your circuit and it works as described. This is day 2 and the LED is still bright. The charge battery started at .5v and is not 1.2v approximately. The other battery started at 1.55v and is now about 1.44v. I used 2 different setups. The first setup uses inductors and the second is a quad coil using 28 guage mag wire. It is an air core but each winding is only about 21 ohms.
Quote from: stprue on 2013.05.04, 11:52:39
@Tinman
I have a close but not perfect replication of your circuit and it works as described. This is day 2 and the LED is still bright. The charge battery started at .5v and is not 1.2v approximately. The other battery started at 1.55v and is now about 1.44v. I used 2 different setups. The first setup uses inductors and the second is a quad coil using 28 guage mag wire. It is an air core but each winding is only about 21 ohms.
Nice going stprue.
I have another that is looped back onto the run battery(no charge battery),and the voltage has been climbing steady for 3 days now(gone up .6 of a volt)I believe that Slider done this some time ago,but haven't heard back from him yet on his results.
The original has been running 6 days now,and i have just swapped the run and charge batteries around-as the charge battery hit 1.34 volt's.Now we have to get the run battery back to it's original starting voltage of 1.228 volt's.It was already at 1.223 volt's after just 40 minutes running,so it should be done by tomorrow.I will then run that battery in one of the solar light's,and see if it will do the 4.5 hours run time like a well charged battery dose.
That sounds pretty cool. Do you have a new diagram of your new set up?
Quote from: ramset on 2013.05.01, 02:58:53
TM
Off topic
I brought you something to Chew on over lunch.
http://www.ijee.ieefoundation.org/vol3/issue1/IJEE_13_v3n1.pdf
From here
http://www.overunity.com/13286/resonance-and-hho/msg359035/#new
Thx
Chet
The pulsed method draws about 0.57 watts, while the traditional DC electrolysis draws about 18 watts at 12 V, for the same amount of gas production. (0.58 mL/s)
It should be no surprise DC electrolysis is very inefficient, most of the current just heats up the water. On the other hand, these guys hit it with a short 200 nano second pulse, and the cell rings (apparently at 100 MHz) and thus its incredibly efficient it seems, but I don't know how it compares with other efficient cells, because there's other techniques out there.
Do we have any chemists here? I would like to know how much power is theoretically required for generating 0.58mL/s flow rates at STP conditions, giver a certain concentration of electrolytes.
EM
On Tinman's cool-Joule, Lidmotor just posted a great little vid showing how to turn this into a radio transmitter:
http://www.youtube.com/watch?v=8pLFXxKGnKw&feature=em-uploademail
Quote from: PhysicsProf on 2013.05.04, 21:44:01
On Tinman's cool-Joule, Lidmotor just posted a great little vid showing how to turn this into a radio transmitter:
http://www.youtube.com/watch?v=8pLFXxKGnKw&feature=em-uploademail
Lidmotor finally "discovers" AM modulation! :D
EM
Hi Tinman,
i toke another shot at your cool joule, now using 2 inductors (150uH) instead of the coils.
I also used a breadboard and some excess wiring to increase the stray capacitance/inductance in the circuit.
Transistor is a 2n2222a metal cased.
But again the same behaviour was noticed, increase of the charging battery voltage and (more) decrease of the run battery :-(
I will let it run overnight to see if things change for the better.
Video here: http://www.youtube.com/watch?v=Jl8O0_en8gA&feature=youtu.be
Regards Itsu
Quote from: Itsu on 2013.05.04, 23:46:24
Hi Tinman,
i toke another shot at your cool joule, now using 2 inductors (150uH) instead of the coils.
I also used a breadboard and some excess wiring to increase the stray capacitance/inductance in the circuit.
Transistor is a 2n2222a metal cased.
But again the same behaviour was noticed, increase of the charging battery voltage and (more) decrease of the run battery :-(
I will let it run overnight to see if things change for the better.
Video here: http://www.youtube.com/watch?v=Jl8O0_en8gA&feature=youtu.be
Regards Itsu
It would be best to use 2 batteries of the same charge level to start. If the run battery is at its highest resting level, it is in a range of not decreasing in voltage level as much as a battery that is much lower when current is taken or given from them.
A very 'similar' thing happens with caps, where the run cap is full and the charge cap is empty in a situation such as this. But with a different map of how it goes down. ;)
The cap to cap experiment will look like good results when getting toward where the caps are becoming equal in voltage because they are both higher than half of the initial run cap voltage. It gives you the feeling that you've got it! You will think that you have filled the charge cap more than half way and the run cap still has more than half to pump the charge cap to full and have some left seemingly. But the end result is that the run cap, still declining is now limited to emf and trying to pump up a more than half full cap that is above the run cap voltage.
And even then, using 2 batteries of the same starting charge, the batteries should be switched over every so 'often'. Then after the testing period is over, leave the batteries to rest and check them to see if the voltage has gone up or down for both batteries.
This will eliminate non linear conditions that seem like a good thing. ;)
Mags
tinman,
On the next battery swap, please measure the voltage between either the emitter or base and the collector of that large metal can transistor you are using, with no batteries installed.
I get ~300mV on one of my 2N3055's unless I apply some heat to it then the voltage goes up a bit.
Good approach, Mags:
QuoteAnd even then, using 2 batteries of the same starting charge, the batteries should be switched over every so 'often'. Then after the testing period is over, leave the batteries to rest and check them to see if the voltage has gone up or down for both batteries.
This will eliminate non linear conditions that seem like a good thing. Wink
Mags
O0 This would be impressive, if the charge battery rose significantly higher than the SAME starting value of the two batteries, while the run batt dropped less...
Update
My replication is on day 3. Light level looks the same. Run battery is at about 1.392 and the charge battery is at about 1.321. It will be interesting to see what happens when both batteries have equal charge.
Quote from: WaveWatcher on 2013.05.05, 03:13:18
tinman,
On the next battery swap, please measure the voltage between either the emitter or base and the collector of that large metal can transistor you are using, with no batteries installed.
I get ~300mV on one of my 2N3055's unless I apply some heat to it then the voltage goes up a bit.
That's interesting, where is it coming from? Are you measuring in circuit or disconnected?
Tinman
Is it possible for you to measure the inductance of the coils? and also find the first frequency where the current goes to zero, so we can calculate the parallel capacitance of the coil? I hope you have a frequency generator. I'm trying to accurately model the circuit, and I have a sim that works better now, but I want to refine it.
Thanks
EM
WW,
I just measured an LED in sunlight, and I get close to 1 volt! :o
I did not know they work in reverse, thanks for mentioning that.
But there is a trick, since they don't have much capacitance or current, the mere fact of connecting the voltmeter probes reduces the voltage to 0.01 or less, but with a capacitor in parallel, when I connect the probes I see 1 volt and it starts to drop rapidly (cap is not electrolytic, and my digital voltmeter is not the best)
So here's a fun challenge for everyone: build a self flashing LED with no batteries, just a capacitor to accumulate the charge and then dump it back through it in a quick higher current pulse. O0
EM
PS. I tried one of those clear white LEDs, in sunlight coming through my window, and I get over 2 volts on my oscilloscope. No capacitor. O0
Quote from: EMdevices on 2013.05.05, 14:54:06
WW,
I just measured an LED in sunlight, and I get close to 1 volt! :o
I did not know they work in reverse, thanks for mentioning that.
But there is a trick, since they don't have much capacitance or current, the mere fact of connecting the voltmeter probes reduces the voltage to 0.01 or less, but with a capacitor in parallel, when I connect the probes I see 1 volt and it starts to drop rapidly (cap is not electrolytic, and my digital voltmeter is not the best)
So here's a fun challenge for everyone: build a self flashing LED with no batteries, just a capacitor to accumulate the charge and then dump it back through it in a quick higher current pulse. O0
EM
PS. I tried one of those clear white LEDs, in sunlight coming through my window, and I get over 2 volts on my oscilloscope. No capacitor. O0
LEDs can't handle much load unless you series/parallel them. Then, you can power almost anything if you have the inkling to die of solder smoke.
The UV and IR ones work just as well but not so well with visible light ;)
The 2N3055..... Any PN junction is a Seebeck generator when the heat energy flow is the correct direction. There was someone selling emergency power supplies to top your wood burning stove or kerosene lamp. (survivalist kook). You could charge your car battery with it or power a radio and small lamp. It was made with arrays of modified 2N3055's.
Your challenge.... I wasn't able to find a source but all but the flashing part can be found in perpetual grave markers. No battery or solar cell but I'm sure it had a whopping super-cap.
EM
I wonder what the efficiency of the LED is when receiving sunlight, if it's over 17% people have been buying the wrong devices to go on the roof ;D
I wonder how a 3watt or even a 20W, LED performs and most come with a nice Lens on top albeit a narrow angle
ok, after 24 hours of my last video, we now see the same behaviour as reported by tinman and slider.
Run battery decrease of 62mV
charge batt increase of 145mV
But as ION already mentioned in the other threat:
QuoteBig error to compare battery voltages at two different points on the charge / discharge curve and think that it is a meaningful comparison of energy used / recovered.
But its fun to see it happen, video here: http://www.youtube.com/watch?v=wlC1z_i8_Lk&feature=youtu.be
Regards Itsu
@ ww
the 2N3055 is a NPN junction.
????????????????????????????
think........PNP probably can also
@EM
this task is possible with enough 2N3055's. Take the cap of and you have an NPN that can utilize solar and heat.
http://www.youtube.com/watch?v=2bUc8RqSFXs
Quote from: EMdevices on 2013.05.05, 14:13:19
That's interesting, where is it coming from? Are you measuring in circuit or disconnected?
Tinman
Is it possible for you to measure the inductance of the coils? and also find the first frequency where the current goes to zero, so we can calculate the parallel capacitance of the coil? I hope you have a frequency generator. I'm trying to accurately model the circuit, and I have a sim that works better now, but I want to refine it.
Thanks
EM
Hi EM
I just orded a new scope,benchtop power supply and SG on friday-all atten gear.The new lab(inside the warm house) is also under construction,so as soon as im up and running i will get those figures for you.
It has come apparent that half blind measurements are no longer going to cut the cake.The one thing i still need to get is a benchtop DMM that has atleast a 5 decimal point reading-but that will take another month or two of saving lol.
Quote from: tinman on 2013.05.05, 22:14:57
Hi EM
I just orded a new scope,benchtop power supply and SG on friday-all atten gear.The new lab(inside the warm house) is also under construction,so as soon as im up and running i will get those figures for you.
It has come apparent that half blind measurements are no longer going to cut the cake.The one thing i still need to get is a benchtop DMM that has atleast a 5 decimal point reading-but that will take another month or two of saving lol.
Good -- better measuring instruments are always welcomed!
In particular, what SG model did you get? I need a new SG myself... Thx, Tinman.
Quote from: stprue on 2013.05.05, 20:52:44
think........PNP probably can also
think......PNPNPNPNPNPNPNPNPNPNPNPNPNPNPNPN........................ Then, all you need to worry about is where you connect the heat & cold sinks O0
Quote
this task is possible with enough 2N3055's. Take the cap of and you have an NPN that can utilize solar and heat.
They aren't sensitive to visible light but are sensitive to IR even more when the body caps are removed. More current may be obtained by jumpering the base and emitter together. There are a couple of u-tubers who think they work like a visible light solar cell but are incorrect.
Solar power flux is about 1000 watts/m^2, on a sunny day.
For a tiny 3x3 mm LED area, the available power is about 10 mili watts.
However, the semiconductor efficiency is low, lets say 10%, so I expect about 1 milli watt the most.
I'm assembling my tiny blocking oscillator circuit to give this a try. we need to engineer these circuits for micro amp operation, because there is so little power coming in through the LED that a poor choice of capacitor, that has high leakage will prevent the voltage from building up to operational voltages ( 0.7 v or greater)
EM
Only IR rings a bell. This would be better because they would work at night as well. Thanks for the info WW.
Be aware. Not all of these transistors work the same way. I imagine that sensitivity to IR could be a problem in some circuits. It was probably an unadvertised 'feature' for the cheaper parts.
Day 5 and my replication is still going strong. LED is bright. Voltage in both batteries is somewhere around 1.33v. Hopefully today I will see if they stay level with each other.
@Tinman
Any updates?
Quote from: stprue on 2013.05.07, 14:11:46
Day 5 and my replication is still going strong. LED is bright. Voltage in both batteries is somewhere around 1.33v. Hopefully today I will see if they stay level with each other.
@Tinman
Any updates?
Just trying a few different circuit mod's at the moment,so will let you know how they go.The original one is still running away charging the original run battery from a battery that the original run battery itself charged.
@Tinman,
In the original circuit, did both the run and charge batteries level out around the same voltage?
Quote from: stprue on 2013.05.08, 11:33:15
@Tinman,
In the original circuit, did both the run and charge batteries level out around the same voltage?
no-in all my circuits the charge battery rose higher than the run battery starting voltage
So are you occasional switching the run and charge batteries or just leaving them in their respective spots? IF thats the case did your run battery level off in voltage or is it still slowly draining?
The run battery will drain down eventually,as there is no return been given to the run battery.This is where you have to see if the charge battery received more charge than the run battery lost.I have now set up a resistor drain board.Here i place the charge battery on the resistive load,and document voltage drop over time.I can then plot a graph with these reading's,and compair them with a battery that was charged in a conventional way to the same starting voltage.
That seems like it is a good test. I look forward to your results.
Day 20 still going. I have modified a few things at this point to see if I can recover more energy. I replaced the 1N4004 with a filter through capacitor and the 150ohm resistor with a 5K pot. I think it is adjusted to around 1.5K, which makes the LED more dim. I have also added a small solar cell from a calculator to the run battery to see if this can charge the battery up beyond what it uses on a daily basis. We will see.
Just had to come back and revisit the cool joule now that i have a 2 channel digital scope-->and glad i did.
For L1 & L2,i am useing a couple of coils from solenoids out of an old washing machine(the ones on the valves that control water flow). Both the steel sleeves have been removed so as they are air core coils/inductors. I have also removed the diode that was between the collector and LED. The new schematic is as below. I have also taken a scope shot(below)-Blue trace across emitter/base,and yellow trace across emitter/collector. I have switched to a TIP35C transistor as well. What i find interesting is-first,the transistor seems to be switching on with only 480mV on the leading pulse. But the trailing pulse has me a bit confused. How do we get the second pulse on L2 when the base voltage is still negative?
As before,i am getting a combined battery voltage rise that is now .7 volts higher than when i started. Although no longer fooled by rising voltages,it is still an interesting effect,and even more so when useing NI-CDs
I _think_ that what you are seeing is the ringing of one of the coils when the transistor is off, going through two peaks then being "struck" again by the transistor when the base voltage reaches sufficient level. I don't have a matched pair of coils with sufficient inductance and resistance to check with my own equipment; I suppose I should wind a pair just to be sure, but that's what it looks like to me from your scopeshot.
I also don't have a TIP35C in my box. Does it do the same kind of thing with a 2n3055 or equivalent type?
Quote from: TinselKoala on 2015.03.24, 03:54:42
I _think_ that what you are seeing is the ringing of one of the coils when the transistor is off, going through two peaks then being "struck" again by the transistor when the base voltage reaches sufficient level. I don't have a matched pair of coils with sufficient inductance and resistance to check with my own equipment; I suppose I should wind a pair just to be sure, but that's what it looks like to me from your scopeshot.
I also don't have a TIP35C in my box. Does it do the same kind of thing with a 2n3055 or equivalent type?
With the 2n3055,the second peak is about 1/2 a division lower on the scope. Other than that,it is much the same.
Quote from: TinMan on 2015.03.26, 03:28:31
With the 2n3055,the second peak is about 1/2 a division lower on the scope. Other than that,it is much the same.
That makes sense. Just by inspection, it looks like that double-peak ringing is at about 40 kHz (one full cycle in 25 microseconds). So you could try a non-powered test with batteries disconnected to see if the coil-transistor resonance is in fact near 40 kHz: Loop a couple turns of wire around the L1 coil and connect that to your signal generator through a 50 ohm resistor. Hook the scope up across the coil-transistor connections. Sweep the FG's output frequency and look for the maximum voltage that you get on the scope. If it peaks at the same 40 kHz, then you've identified the source of the double-peaks as the resonant ringing of the coil-transistor combo.
I think, maybe.
I did this test on my little version with the two 1mH toroids and BC337-25 transistor and it worked perfectly to confirm that the oscillations are due to the coil-transistor combo ringing. At a much higher frequency than your version, of course!
This has to be one of my all time fave circuits and am glad you are re-testing Tinman.
Also, that TK is here and moving things along.
Recently, an original build Cool Joule ran for 3 months continually on 2x 100mAh yellow solar garden light Ni-Cd AAA's !
It was put on a table and practically forgotten about. When it stopped, a quick short across the transistor and off it would run again for several days.
Circuit used 2x 'dancing flower' 500ohm coils side by side and a superbright white LED. Transistor being a (iirc) C1815. Lumens were low, yet from knowledge of the terrible quality of the garden light batteries, it lasted well past regular discharge traits.
My new scope hasn't arrived yet, but it may well be worth trying to hook up my wobbly buttoned erratic scope to follow along.
>:(
Quote from: Slider2732 on 2015.03.27, 14:12:24
Recently, an original build Cool Joule ran for 3 months continually on 2x 100mAh yellow solar garden light Ni-Cd AAA's !
It was put on a table and practically forgotten about. When it stopped, a quick short across the transistor and off it would run again for several days.
Circuit used 2x 'dancing flower' 500ohm coils side by side and a superbright white LED. Transistor being a (iirc) C1815. Lumens were low, yet from knowledge of the terrible quality of the garden light batteries, it lasted well past regular discharge traits.
My new scope hasn't arrived yet, but it may well be worth trying to hook up my wobbly buttoned erratic scope to follow along.
Very intriguing, Mark!
Ps -- I checked and they say your scope has shipped, so its on its way!
Hi Slider,
i put something together yesterday and toke some scope shots and made measurements, perhaps they come in handy when your new scope arrives.
Using a MJE3055 transistor, a big (10mm) blue led and some transformer windings as aircoils (191 Ohm/106mH and 185 ohm/93mH) and 2x 1.2V 750mAH NiCd batteries
Its running for 24 hours now, need to tickle the base to get it started
Yellow is the voltage across the collector to ground (emitter)
Blue is the voltage across the base to ground
purple is the current to/from the lower nicd battery (2mA/Div)
green is the current through the total battery stack (1ma/Div)
So current is very low and hard to register on the scope.
Video here: https://www.youtube.com/watch?v=iCaiNJ0Zd2Q&feature=youtu.be
Regards Itsu
Scope delivery note.
I double checked on that "already Shipped" and apparently that was from the west coast USA to their
east coast shipping terminal ,where they will then be shipped to fill orders.
long story short....they'll be shipping to customers by 4/1.
Thanks for the scope updates :)
This circuit will be ideal as a tester for getting used to it.
Thanks for your results Itsu, always well shot useful videos. Btw, the batts appear to be 250mA, rather than 750ma ?
Say, i'm just thinking about something. The whole scopes container goes from a California port to the East coast, then they put a few on a truck for deliveries westward and presumably a couple of them end up back in California ? LOL
???
Slider, i just double checked, they are 700maH:
Regards Itsu
Haha, so we were both a little out. Guess who read the charge rate on your video by mistake ;)
Should last well at that rating if fully charged and allow for good accurate discharge graphing.
:)
Well, just checking in with my version. I'm using the extra diode, a BC337-25 transistor, a blue LED and 2 toroids of 1mH each, about 30 - 32 turns of #33 wire on each one. I still haven't been able to duplicate tinman's double-peaks but as you can see the collector and base waveforms are pretty similar to Itsu's, but at a higher frequency due to the lower inductance.
As I said above, I've also confirmed that this frequency is the natural ring frequency of the tank formed by L1 and the transistor (with no batteries connected). I looped a couple of turns thru the L1 toroid, connected that to the FG thru a 50 ohm resistor, connected scope across coil-transistor connections, and swept the FG until I got maximum voltage on the scope, then read frequency on the Philips counter.
Ok, i could try that, but first want to leave it running to see how long it goes.
I use the primary windings of some 220V/12V transformers now, but i started with the both secondaries (around 1 Ohm each) and the frequency then was around 300KHz and
pulling much (relative) more current, so yes, the used L's heavily determine the pulse frequency.
Also no double pulse seen here like in tinmans screenshot.
Regards Itsu