Self Running Micro TPU, with closed loop.

[4Tesla]

There is no way to have OU with this circuit without applying an outside input, which is why I mentioned it a couple pages back that a pulser motor was needed to maintain both charging and recycling charge. I'm almost done both the pulser/flywheel and the circuit. should be interesting. This circuit is unique in that it uses least amount of parts to get the job done (efficiency) and configurable. The pulser/flywheel should produce that extra time increase by maintaining the charge duration in any case.

P.S. does anyone know of an electronic circuit equivalent of a mechanical flywheel? is there even one? I was thinking of capacitor/coil combinations, but it does not exactly show both mass, centrifugal force and angular momentum ? or do I have the wrong idea here?

I believe the idea of the TPU is to use the Earth's magnetic field as the outside input.

[sanmankl]

@All,

After playing around with multiple coil configurations (torroid size, wire size, wire type, no. of turns) my best resulst is about 28m 52s with 2 x 5mm white and 3 x 3mm Blue LEDs hooked up in parallel. Single 3mm Blue LED gives me approx 28 minutes as well.

I won't bore all with the details for what I really want to put it to all is

                    "EMDevices has his circuit going to 2+ hours with only single blue 3mm LED"

and none of us (or that I know of) has match his performance. I want to replicate his setup as close as possible, get to as close the run time as possible (maybe 80% of his run time? 1.5 hours?). EM run time and circuit is my base line for now.

I see a lot of discussions going about improving circuits to get better times, etc but to me, I would not want to do that simply because I have not achieved what EM got for his. So, I'm sticking to the original circuit/plan.

Questions for EM are

1. What wire is he using?
    From I can see from his picture, it's an insulated one. Is it multi-stand or single core? Copper, alloy?

2. What about the torroid core?
    EM says that he got his from a recycle shop and it's about 1.5" diameter.
    But based on his photo and I measured 11 plugpoints on his board, it's more  like .75"?
    From what I can see, it's black? Maybe it's not a pure ferrite core? His coils of 30 turns/15 turns/7 turns do not wound all the way round the core.
    Does this matter? Also, the 15 turns x 2 for collector wire in series. Again, would this matter?

3. NPN transistor.
    I have tried, 2N2222A, 2N3904, BC237, BC548, BC337. All similar results. I have even tried FET 2N7000 with even poorer result than the NPNs.

4. Cap.
    1000uF (normal? low impedence?)

5. EM says he charge his capacitor up with 5V.
    Did he get better run times with 9V, 12V?

Given the hugh quality variance with electronics components, I would be happy if I can get past the magical 1 hour run time. Once I have that, I'll move on to incorporate some of the ideas posted here. To me, EM has proven that his circuit is sound to go 2+ hours. That is my replication goal. After all, the term replication means "to duplicate", right?

Cheers, sanmankl (cp)

[jcims]

@ All
A reminder that no one here claims that this is OU.  The goal is to learn, experiment, have fun, and see if it can be pushed that way, and then scaled up.  I am sorry I do not have a working scope at this time.  It is frustrating for me as well. 

Hi Bruce,

If you replace the condenser in a cheap audio microphone with a CdS cell, set it in proximity to the LEDs and record it in Audacity or some other audio recording software, you'll get a slightly more objective measurement of light output, rate, etc.  Shouldn't cost much more than $2.  If you have an old mouse around, you can rip the phototransistor out of the internals to get a similar effect.

(Of course the other option would be to wire the output directly to the microphone input, with something to regulate the voltage if it's going to exceed 5vdc by too large a margin)

[Bruce_TPU]

@samnakl

Please remember that EM stated that his LED was some super efficient hybrid type, and that is why it ran so long.  He used a regular LED and he said it ran for "a little bit" but never gave a time.

I want to stay away from increasing the efficiency of the LED or increasing the Main 1000 uf cap size and learn how to make this thing rock using "other" methods, such as coils, types of wire, frequency close to shumann's, etc.  This is the base line I have set for my self.  Just blinking an LED for a length of time does me no good, unless the coil is directly involved.

Your test with the iron wire core was "inspiring".  can you also test 4Tesla's circuit, to bring the circuit down to around 7.3 Hz if you have a scope.  Also a current reading would be awesome.  Thanks!

@ Mr. Ramos
A quote from Scorpio about his super cap.  "Yes, is it really 0.22 Farad! = 220000uF/5.5V."

It ran so long because it was 220,000uf.  With my configuration I could get about 110 hours run time with that.   , with 8 LED's.

@ 4Tesla
The diameter of my toroid is 2".  Also, please post a revised schematic, since you have added more resistors and a second pot.  Thank you! 

Holiday Cheers,
Bruce

[quantum1024]

I understand that the original TPU was to use the earths magnetic field as an outside power source, which is what got my attention, however it's simple enough to check and see if it is an influencing factor, shield the toroid in a small metal can and check the run time.

Duty cycle (on/off periods of load and circuit-frequency) plays a large role in the efficiency of the circuit. The trade off is less light to more light.

You could also use Litz wire which will increase your run times/efficiency. Super caps are an excellent way to optimize as well. 

Increasing the voltages would increase the efficiency and decrease losses, the trade-off is down conversion.

I have seen leaky capacitors have electrolytic reactions (acts like a battery due to chemical reactions, crystal formations between plates), just take an electrolytic cap and short it, then wait a minute and look at the voltage, this i would say is a large contributing factor in the different results.

# 4telsa - current check with meter will be out to lunch in ac circuit. must be dc to do a proper check. in this circuit you are dealing with oscillations, hence ac. no use. I high five your paw and raise you a paw.

I agree that that circuit should be kept holy electronic if possible, I'm just looking at all options and keeping an open mind.