Re: Light LED with 4 nanoamps
klugesmith, Mon Jun 16 2014, 05:40PM

By amperage, do you mean the short-circuit current? If it's that small, then the pile must have an internal resistance of 150 gigaohms. How do you measure the voltage?

The flashing light is trivial, and probably more luminous on average, if you use a neon glow lamp instead of an LED. Jut put an NE2 or equivalent (e.g. from disposable camera flash unit) in parallel with the capacitor, and you will have a relaxation oscillator.

Re: Light LED with 4 nanoamps
Zamboni, Mon Jun 16 2014, 06:29PM

By amperage, do you mean the short-circuit current? If it's that small, then the pile must have an internal resistance of 150 gigaohms. How do you measure the voltage?

***I used a Picoammeter to get the amperage. I use a Singer Electrostatic Meter to measure the voltage. I believe that your figures are correct, a dry pile has tremendous internal resistance.

The flashing light is trivial, and probably more luminous on average, if you use a neon glow lamp instead of an LED. Jut put an NE2 or equivalent (e.g. from disposable camera flash unit) in parallel with the capacitor, and you will have a relaxation oscillator.

*****I had pondered the use of a neon bulb. I was looking for something that would last for decades (or longer). As I understand it, a neon bulb will last for between 8 and 15 years. Perhaps even longer in "blink" mode".I was hoping for something that would last for at least 50 years.

My dry piles have already lasted for 5 years. There are dry piles like the Oxford Electric Bell that have lasted for over 150 years I hope that mine will last that long...or, dare I dream, longer. I suppose that I will be long gone before they fail to function.

Re: Light LED with 4 nanoamps
klugesmith, Mon Jun 16 2014, 07:40PM

This is fun stuff! Please tell us more.

I bet the unpowered shelf life of neon glow lamps is more than 100 years, just like incandescent lamps.
Nixie-tube clock enthusiasts could confirm that most 50-year-old nixies work, though some small fraction have gone bad with age. There's much less to go wrong in a 2-terminal lamp. Who knows how long non-hermetic encapsulated LEDs will last in benign environments? They seem OK in my 1970's handheld calculators and digital watch.

Neon specified operating lifetimes are similar to those of LED's.
A typical part here: ,
run at 0.3 mA for a brightness of 1.40 mcd (omnidirectional),
should be good for 50,000 hours.
That amounts to 15 ampere-hours of charge passed through the lamp.
How many AH of consumable material (per cell) are in your dry pile battery?

-Rich

[edit] I just learned a new name for the proposed oscillator circuit:
What kind of capacitor are you thinking of? Leakage current in capacitor or lamp could be a deal breaker.

Re: Light LED with 4 nanoamps
klugesmith, Mon Jun 16 2014, 08:49PM

Couple more things.

1. As mentioned before, your dry pile battery's anodes or cathodes must be chemically exhausted after some number of amp-hours or milliamp-hours.
If 4 mAh (that's a wild-ass guess) then at 4 nA it would last for a million hours (114 years).
Surprising that any battery could be short-circuited for years and have any juice left, but I guess that's what galvanic corrosion is all about.

2. Back to LED's. A few years ago I reported being able to see the light from an LED at steady-state current of about 12 nanoamps.

Re: Light LED with 4 nanoamps
Zamboni, Tue Jun 17 2014, 01:22AM

This is fun stuff! Please tell us more.

*****I make a (very expensive, lots of work to make) Device that uses a Zamboni battery to power a horizontal pendulum. The batteries that I make usually have more than 1500 volts per battery. Ocasionally, I mess up and the resultant (defective) battery only has around 500 volts. I was looking for something interesting to do with my "factory seconds". I have been working at making Zamboni Dry Piles for around 8 years. To my knowledge, I am the only peson in the world making Devices of this type. They look like this:


I bet the unpowered shelf life of neon glow lamps is more than 100 years, just like incandescent lamps.
Nixie-tube clock enthusiasts could confirm that most 50-year-old nixies work, though some small fraction have gone bad with age. There's much less to go wrong in a 2-terminal lamp. Who knows how long non-hermetic encapsulated LEDs will last in benign environments? They seem OK in my 1970's handheld calculators and digital watch.

*****I am old enough, so you would think that I would have thought of Nixie tubes. I think you are right.

Neon specified operating lifetimes are similar to those of LED's.
A typical part here: ,
run at 0.3 mA for a brightness of 1.40 mcd (omnidirectional),
should be good for 50,000 hours.
That amounts to 15 ampere-hours of charge passed through the lamp.
How many AH of consumable material are in your dry pile battery?

*****I went to Radio Shack and got a neon bulb that turns on at 90 volts. I hooked it up and it did not light ;-(
I am now going to try using a capacitor in parallel with the neon bulb. I don't know if the neon bulb (in an unlit state) draws off any current at all. I suppose that will be the next parameter I investigate.


What kind of capacitor are you thinking of? Leakage current in capacitor or lamp could be a deal breaker.

*****I am using a Mouser part number 871-B32652A2332J Polypropylene Boxed Capacitor (3300Pf) I understand that poly caps have very low leakage. I know the battery will charge the cap because I (foolishly) picked up the cap...and got shocked.


Thank you very much for you insightful suggestion. I don't know all that much about electronics and your encyclopedic understanding is very helpful.

Paul

Re: Light LED with 4 nanoamps
klugesmith, Tue Jun 17 2014, 02:44AM

Thank you, Paul. I admire your Zamboni pile creations. We both have plenty to learn about neon lamps in nanoampere circuits.

Answers might come sooner by experiments than by searching the Internet.
The latter just led me to this article about the dark effect

I wonder if your lamp is glowing at steady 4 nA, but too dim to see except with dark-adapted eyes, or only detectable with instruments?
If the canonical oscillator actually runs, then with 3300 pF of capacitance the discharge energy is only microjoules per pulse. Might be visible in the dark. Let's see what happens.

Re: Light LED with 4 nanoamps
Newton Brawn, Tue Jun 17 2014, 02:48AM

Is it possible to make a converter HV to LV and feed a LED ?

Re: Light LED with 4 nanoamps
Ash Small, Tue Jun 17 2014, 10:53AM


Parallel connection of the cells, rather than series connection?

Re: Light LED with 4 nanoamps
Antonio, Tue Jun 17 2014, 01:49PM

Interesting research. I don't see how to use conventional electronic switches (mosfets probably) to make a DC/DC converter at this low power level. Even if a suitable design could be found, powered only by the HV supply, the leakage of the closed switches would be probably greater than the available current. The neon lamp oscillator probably works, but with switching at 90 V most of the power is wasted. It's probably possible to use a neon lamp as switch, driving another lamp (maybe a LED) through a step-down transformer. Note that neon lamps have a small leakage current, through the glass envelope and through the gas inside, that is frequently doped with a small amount of radioactive material.

Re: Light LED with 4 nanoamps
Zamboni, Wed Jun 18 2014, 08:09PM

Thank you all so much for your help.
Rich, you were right. Only experimentation will solve this issue. My end result was that I can't get a neon to light. I was able to get an LED to light...though so dimly that you need almost total darkness to see it.
I will update if I find a way.

Paul

Re: Light LED with 4 nanoamps
klugesmith, Thu Jun 19 2014, 04:35AM

Paul,
You said your picoammeter indicates about 4 nA when connected directly to battery.
When the leads are reversed, do you observe similar magnitude with opposite sign?
This is analogous to checking the offset voltage of a sensitive voltmeter by reversing the leads.

What does the meter show when the neon lamp, or the LED, is in series?

What does it show when the 3.3 nF capacitor is in series? That ought to start out looking like the short circuit case. Then current should decrease as the capacitor charges, until we reach equilibrium between battery resistance and capacitor leakage. Time constant approx 150 Gohm * 3.3 nF = 495 seconds (no need to use a calculator or even pencil and paper for that one).
You could also monitor the capacitor voltage with your electrostatic voltmeter.

I've never personally worked with such low currents, but know that leakage over the surface of insulators (esp. if there's any ionic contamination) can be a problem.

Re: Light LED with 4 nanoamps
ZakWolf, Thu Jun 19 2014, 06:56AM

darlington pair???

Re: Light LED with 4 nanoamps
Zamboni, Thu Jun 19 2014, 02:34PM

Rich,
Thank you for your helpful suggestions. I had a couple of questions...

You said your picoammeter indicates about 4 nA when connected directly to battery.
When the leads are reversed, do you observe similar magnitude with opposite sign?
This is analogous to checking the offset voltage of a sensitive voltmeter by reversing the leads.

*****I was under the (perhaps mistaken) impression that you never hook up a battery with no load onto an ammeter. When I do my amp testing, it is always through a load. In my case, the load is my horizontal pendulum. You are suggesting that I hook up the battery directly to the pico-ammeter, with no load ?
I have switched the leads (by accident) and did get the negative, it was slightly different, but extremely close.

What does the meter show when the neon lamp, or the LED, is in series?

*****I have never tried that. I will give it a try and let you know. I found out that a neon bulb has about a megaohm of resistance, unfortunately, that is like a direct short to this type of battery. The LED would be a very interesting experiment. I will try it.

What does it show when the 3.3 nF capacitor is in series? That ought to start out looking like the short circuit case. Then current should decrease as the capacitor charges, until we reach equilibrium between battery resistance and capacitor leakage. Time constant approx 150 Gohm * 3.3 nF = 495 seconds (no need to use a calculator or even pencil and paper for that one).
You could also monitor the capacitor voltage with your electrostatic voltmeter.

*****I will try that also. Your figure of 495 seconds tracks very nicely with my experience. By the way...I would have needed paper, a calculator and the Internet to figure that out

I've never personally worked with such low currents, but know that leakage over the surface of insulators (esp. if there's any ionic contamination) can be a problem.

*****I can tell you that there are many interesting and very stressful problems with currents this low. Most people would have had the good sense to abandon this project years ago. I am just too stubborn (or maybe too stupid) to give up. Now that I have succeeded, the joy of watching my Devices operate makes it all seem worth it.

Paul

Re: Light LED with 4 nanoamps
Newton Brawn, Fri Jun 20 2014, 03:47AM

Hi,

If the battery can charge the capacitor with 1500V, the energy storaged in the 0.0033 uF capacitor could be 0.5 . 0.0033 . 1500^2 . 0.000001 = 3.7mJ.

This energy may be switched by the pendulun, modified to act as a normaly open contact, that moves by the electrostatic force... Is the pendulun acting as a switch send the energy to a converter that provides low voltage to a neon or LED.

Such energy may fed a converter that provides low voltage to the neon Assuming that a led or neon consumption is in the range of of 0.005 watts, the lamp can glow 0.0037/0.005 = 0.7 seconds..

By construction, I believe that the pendulun has insulation very high, much better than solid state switches, neon lamp or LED...

It looks like a dream, please forgive-me.

EDIT "0.0033 uF"

Re: Light LED with 4 nanoamps
Steve Conner, Fri Jun 20 2014, 10:19AM

Buck converter with pendulum switch? Nice idea :)

Re: Light LED with 4 nanoamps
Newton Brawn, Fri Jun 20 2014, 09:14PM

Hi Steve,

The idea is simple circuit where the battery charge the capacitor, and when the cap is full charge the eletrostatic force moves the pendulum (as a electroscope foil) that makes contact with the other terminal, discharging the cap to the converter.

The converter then start switching and feed the bulb for one second or less.

Someone may advise what is the minimal current that a NE2 or LED provides a visible flash ?

Regards

Re: Light LED with 4 nanoamps
Antonio, Mon Jun 23 2014, 02:31AM

Some calculations to see if it's viable to light a LED using a DC/DC converter using the pendulum as switch:
Let's assume that a LED can produce a visible flash with 1 mA, a flash lasting for 0.1 s, and the voltage drop in the LED as 1.7 V. This requires the energy E=1.7*1e-3*0.1=170e-6 joules (170 uJ).
A capacitor storing this energy at 450 V shall have the capacitance C=170e-6/(0.5*450^2)=1.7e-9 farads (1.7 nF).
A pile producing 4 nA charges 1.7 nF to 450 V in t=1.7e-9*450/4e-9=191 seconds.
Too long time.
Reducing the time of the flash to 0.01 s reduces the energy to 17 uJ and the capacitor to 170 pF. This capacitor can be charged in 19.1 seconds. Less bad, but the flash would be not very visible.
The converter could be a buck circuit, using an inductor and a low leakage diode, powering directly the LED.

Re: Light LED with 4 nanoamps
Bored Chemist, Mon Jun 23 2014, 07:46PM

The easiest thing to look at would be to connect an LED (or several) in series with the neon lamp of a neon oscillator.
Compared to the non conducting state of the bulb, they will just look like a conductor. but, when the neon strikes they will carry the current so they should light up. The current spike will be quite large, but very brief so it might not damage the LED- in principle, a small resistor could limit the current to a mA or so. ( A megohm is a small resistance in this sort of circuit)
I strongly suspect that the neon will last longer than the LED in this configuration, but it would be interesting to find out if it works.

Another approach, albeit one that doesn't address the real question of what to do with "dud" batteries would be to include two LEDs between the capacitor plates and the moving contact so the swinging pendulum switches the current through the LED directly.

In any event, those a really cool items.

Re: Light LED with 4 nanoamps
Newton Brawn, Fri Jul 04 2014, 11:08PM

Zamboni

See this thread




The coupling of the source and the lamp is made with a air gap. As the source has a capacitance, the discharge is made when there is enough energy in the source. and the energy storaged in the cap cam turn the lamp on.

Maybe you can try this arrangment to get some flash in the neon or led

Re: Light LED with 4 nanoamps
Newton Brawn, Sat Jul 05 2014, 01:40AM

Zamboni

Here some news, and RESULTS:

I just get a 1200pF polyester capacitor and charge it with 340volts using a diode and 220V line.

Then I disconected the capacitor from the diode and line,

And connected the cap with a NE2 neon bulb....

The result: I got a really visible red flash from the lamp !

The conlusion is that the Zamboni battery cam be connected with a 1000pF cap and a pendulum in such way that (when the cap is charged) the mobile blade of pendulum make a contact with one pole of the bulb, transfering the energy storaged in the cap to the bulb.
After the contact done, the cap voltage is reduced and the blade returns to initial position and starts to be charged again.


Here the proposed arrangement:
B= batery that provides high voltage,
C= capacitor that is carged by the battery and discharged throgh the pendulun to the bulb
NE2 = neon bulb 70V, 1/8W
P = pendulum, that work as a electroscope, opening the blade when charged

Note: the pendulum cam be assembled in the horizontal position, using the vertical shaft. for more sensibility.



]zzaamm_model_1.pdf[/file]

Regards

Newton

EDIT - fixing the typing

Re: Light LED with 4 nanoamps
Zamboni, Mon Jul 07 2014, 03:45PM


The conlusion is that the Zamboni battery cam be connected with a 1000pF cap and a pendulum in such way that (when the cap is charged) the mobile blade of pendulum make a contact with one pole of the bulb, transfering the energy storaged in the cap to the bulb.
After the contact done, the cap voltage is reduced and the blade returns to initial position and starts to be charged again.

*****I believe your proposed circuit will work. I am trying to engineer a set of conductive leaves that will allow your circuit to work. Thanks for your idea.

Paul

Re: Light LED with 4 nanoamps
Newton Brawn, Tue Jul 08 2014, 10:23PM

Hi Zamboni,

1-Refining my experiments, this time I have charged the 0.001uF capacitor with 170V and discharged through the neon bulb.
THE RESULT: I got a really visible red flash from the lamp !

2- The calculed time for 3 nanoamper charge a 0.001uF to 500V is about 167 seconds, (considering no leaking through the surfaces or insulation creepage) That means, you can get a light flash at every 2.8 minute.


3- The electrostatic switch may take some energy from the battery in order to move itself. The blade mass, self capacitance, friction will increase the one minute charge time.
A more efficient switch with the blade performoing a push-pull arrangement may be more suitabe.

Let us know your experiments

Regards

EDIT " 167 seconds"

Re: Light LED with 4 nanoamps
Antonio, Thu Jul 10 2014, 12:54AM

You can obtain more efficiency by discharging the capacitor through a transformer, reducing the high voltage in the capacitor to the level required by the neon lamp. This also allows the use of a smaller capacitor, using better the available energy. Try first the direct version. Leakage in the capacitor may be a problem. Use a high-voltage capacitor (or make one) and keep it dry and clean. The mechanical assembly uses very little energy, just the amount required to charge its capacitance just before the discharge, probably negligible in comparison with the energy required to charge a capacitor of even a few tens of pF.

Re: Light LED with 4 nanoamps
Newton Brawn, Tue Jul 15 2014, 07:50PM

Hi:
Iam trying to calculate a small transformer to couple the capacitor with the lamp.
The primary voltage could be 500V , secondary 83V, providing voltage to a NE2 neon bulb (or 2.5V secondary feeding a LED)
So far, I do not have problems. The transformer ratio could be 6:1 or 200:1 in the case using a LED.
Now lets see the primary turns:
Input data:
Core = Ferrite , 16 mm diameter, area 0.0002m2 , max induction 0.3T.
Primary wave shape = sine half wave pulse.
Pulse lenght = 0.005s,
Bmax max = Epk x 2 x t / pi x N x A. >> N= Epk x 2 x t/ pi x Bmax x A
N= 500x2x0.005 / 3.14x0.3x0.0002 = 26526 turns.
It is too much turns. My limits are 600 turns, 32awg or thicker wire.
Did I make a mistake ? Or any wrong approach ?
One way to reduce such primary turns, could be a reduce the pulse length, that could set to 0.00005 seconds. Them the primary will be 265 turns, very easy to wind in a confortable way. BUT now a low power converter is required to shop the 500V to 20kHz.
Any suggestion for such converter?
Regards

Re: Light LED with 4 nanoamps
Dr. Slack, Tue Jul 15 2014, 10:35PM

Can I suggest that instead of a complicated transformer, you go for something simpler, with one winding. This is basically an inverting buck converter. Your low current pile charges the capacitor, stores energy in it. Eventually, the switch closes, connecting the capacitor to the inductor. Current builds and voltage drops as the energy stored as voltage in the cap is transferred to energy stored in the inductor as current. The top terminal of the inductor sinks current from the capacitor.

Eventually, after 1/4 cycle of the resonant period of L1 and C1, the voltage will reach zero as the current peaks. Now the inductor will continue sucking current out of its non-grounded terminal. Without the lamp+diode there, the voltage would ring negative to the inverse of the previous +ve voltage (less losses) as it recharged the capacitor -ve. However, when it gets to -2v (LED) or -100v (neon) (round figures, don't quibble) it will suck charge out of the lamp instead of the capacitor. As the voltage across the inductor now is clamped to a low value, the current will decay only slowly, and the inductor will deliver its energy in a long low voltage pulse to the lamp.

This ought to simplify your calculations of the inductor. The core has to store the total energy of the cap without saturating. Once you've done that, the pulse length can be whatever it comes out to be, as long as it's not silly figures for current. The peak current from the cap / into the lamp is of course given by the energy that the cap dumps into the inductor, 0.5CV^2 = 0.2LI^2. The arrangement dumps essentially all of the capacitor energy into the lamp, and the choice of inductor and cap allow you to choose the peak current and pulse duration more or less freely, so there is no advantage in using a transformer per se.

Re: Light LED with 4 nanoamps
Tony Matt, Thu Jul 17 2014, 11:41PM

Hi Dr Slack,
Now it looks easy, nice converter.
Do you mind say the amount of henryes that such inductor has to have?

I have experimented charge a 0.001Uf with 170V and dischage it in a 6mm LED.
I got a very small flash, difficult to see.

Another experiment I replicate the Newton experiment I got good and visible flash on a neon
C=0.001Uf, V=170V, neon 10mm lenght.

Let me know the inductor volts seconds per amper

Cheers

EDIT : Please read 0.001uF in the place of 0.001Uf

Re: Light LED with 4 nanoamps
Dr. Slack, Fri Jul 18 2014, 07:50AM

Do you mind say the amount of henryes that such inductor has to have?

It's more complicated than that.

Decide what energy you want to dump into the lamp in a single shot.

Add a bit for losses, this is the energy that the inductor will need to store. If you are using an ungapped low permeability core, this will define the minimum core volume. If you are using a gapped high mu core, it will define the minimum volume of the air gap. If you are using an ungapped high mu core, then read up about energy storage in inductors.

Add a bit more for losses, this is the energy the capacitor will have to store. Choose a voltage level and capacitance jointly to give you this energy.

Now decide what peak current you want to send into the lamp. Finally the inductance so that 0.5*Ipeak^2*L gives your chosen energy.

Re: Light LED with 4 nanoamps
Tony Matt, Sat Jul 19 2014, 02:02AM

Hi Dr SlacK

Please correct me if I am dreaming or writhing
something crazi...

available capacitor = 0.001uF, charged with 500volts

available energy: E= 0.5 x V x C^2

E= 0.5 x 0.001 x 500^2 = 125 microjoules

Adding 30% energy extra = 156 microjoules

using a tv flyback ferrite hi mu, 16mm diameter.>> core area = 0.0002m^2

how I can get the air gap volume based on the 156 microjoule ?

Re: Light LED with 4 nanoamps
Dr. Slack, Sat Jul 19 2014, 01:50PM

how I can get the air gap volume based on the 156 microjoule ?

try this link, for magnetic energy stored per unit volume.



You will need to choose a Bmax. For ferrite, 0.4T may be pushing it, 0.2T might be more conservative.

However, if you are starting with a flyback core, that's designed to work in this energy storage mode, so unless it is actually gapped, it's likely to be low u to start with. However, do the sums assuming high u, extra air will merely mean you need a few more turns to get your inductance, and will reduce the likelyhood of saturation.

Re: Light LED with 4 nanoamps
Ash Small, Sat Jul 19 2014, 02:18PM

If you only have one pulse a minute or so, this equates to a frequency of ~1/60 Hz, so you don't need ferrite, do you?

Surely you can use a high permeability core, as used in 50-60Hz transformers, etc, as the frequency is so low?

(The reason these can't be used at high frequency is due to heating, but at these low frequencies this won't be an issue)

You don't need to use gapped ferrite here, any old laminated iron core will work.

Someone please correct me if I'm mistaken.

Re: Light LED with 4 nanoamps
Newton Brawn, Sat Jul 19 2014, 03:43PM

Tony

I have some old notes that says the density of energy in a air gap is about " 400000 joules x induction squared per cubic meter of air gap"

E/v = 398000 x B^2 [J/m^3]

E = Energia, joules [ J ]
v = air gap volume, cubic meter [m^3]
B = induction weber /m^2 = tesla [T]

has somebody a better figure? or a comment??

Re: Light LED with 4 nanoamps
Tony Matt, Sat Jul 19 2014, 04:16PM

The inductor is using a tv flyback core with air gap.

" If you are using a gapped high mu core, it will define the minimum volume of the air gap."

Yes, using a high mu core GAPPED, and assuming 0.3T as the Bmax in the core and air gap
and based on Dr Slack and Newton information,

the air gap volume for storage of the 156 microjoule could be:


J/m^3 = 398000 x B^2 >>> m^3 = J / (398000 x B^2)

m^3 = 0.000156 / (398000 x 0.3 x 0.3) = 4.4 x10^-9 ...m^3

but the air gap area = 0.0002m^2 = 2 x 10^-4

min. air gap lenght, d:

d= 4.4 X 10^-9 / (2 x 10^-4) = 2.2 x 10^-5 = 0.000022m = 0.022mm

air gap relutance:

R= l/(u x a) = 0.000022/ (4x pi x 0.0000001 X 0.0002) = 87500 amper turn per weber

or

Al = 11.4 uH/turn (600turns result in 4.11H)


EDIT: AL=11.4uH/turn^2

Re: Light LED with 4 nanoamps
Newton Brawn, Sun Jul 20 2014, 02:33AM

Hi Ash,

Well, the pulse lenght is much less tha 1 minute,

The repetition rate is about 167 second, ( The calculed time for 3 nanoamper charge a 0.001uF with 500V is about 167 seconds, say 2.8 minute ).

The available energy in a 0.001 uF capacitor charged with 500V could be ;

E= 0.5 x C x V^2 = 0.000125J

Applying this energy in a 2.5V 2mA LED. the LED time on will be:

0.000125/(2.5 x 0.002) = 0.025 seconds

The inductance could be :

V = Ldi/dt >>> L= Vdt/di = 2.5 X 0.025 / 0.002 = 31H

Is a too large inductance, the number of turns is above my limiite....

Re: Light LED with 4 nanoamps
Dr. Slack, Sun Jul 20 2014, 09:38AM

I'm not sure you've divided the right thing by th eright thing there Newton.

If we take 125uJ in the cap, and work out what inductance is needed to store the same energy at 2mA in an inductor, then it's 62H. This is a lot. Unless you want to be up to your elbows in very fine wire, it sounds like rethinking the LED current, or the cap energy, would be in order.

2mA might be the rated contionuous current for the LED, but what's the pulse current? Can it take 20mA for a few mS once every 3 minutes? That would reduce the inductance to a more manageable 620mH. 50mA? Now you'd be talking 100mH.

Iron or ferrite? 1nF * 62H resonates at 640H, so you might just get away with iron, with a lot of losses, unless you can find aircraft grade 400Hz cores. Using 620mH puts you up at 6.4kHz, defintely ferrite territory.


Now were you to design a transformer to do it, then either ...

a) it's a flyback, and you'd need still to store the 125uJ in the core. Using a lower primary inductance would draw a higher current from the cap, which is all fine. But what about the secoindary? If you want to deliver 2mA max into the LED, you are back with a current step-down, or voltage step-up transformer, so an up to your elbows in fine wire secondary.

b) it's a flux-coupled conventional transformer, with a large step-down in voltage. Either you'd need signaificant leakage inductance to control the LED current, or a wasteful series resistor. With a uni-polar input, the core must still be volt.seconds rated to tolerate the rise in flux over the length of the pulse.

Re: Light LED with 4 nanoamps
Ash Small, Sun Jul 20 2014, 09:50AM

Inductors of several H are readily available from the power supply circuits of 'vacuum tube era' equipment. They are reasonably compact and don't require being 'up to your elbows' in fine wire.

I have one here rated for 463V. Looks like it's ~10H, but I've not measured it yet.

Re: Light LED with 4 nanoamps
Dr. Slack, Sun Jul 20 2014, 05:32PM


50Hz iron is it?

Re: Light LED with 4 nanoamps
Ash Small, Sun Jul 20 2014, 05:48PM


Yes, but isn't this circuit operating at ~1/60Hz, or so?

EDIT: And if you dump a capacitor into an inductor, does it matter what the core is made from? Won't the discharge rate be entirely dependant on the inductance? (assuming the ESR of the capacitor doesn't dominate)

Resistance losses in the copper obviously have to be considered, but these 'iron core' inductors will have less losses than a ferrite cored inductor, which will require more, presumably thinner, copper wire for the same inductance, won't it?

Re: Light LED with 4 nanoamps
Dr. Slack, Mon Jul 21 2014, 07:03AM

If the 1nF cap and 60H inductor ring up one quarter cycle in 400uS, what frequency do you think the eddy current losses in the iron are interested in? the 640Hz fundamental frequency, or the once per 3 minutes that it happens? You're not worried about it getting hot, you're worried about the efficiency as energy out to LED divided by energy fromm the capactiro, or at least I am.

Re: Light LED with 4 nanoamps
Ash Small, Mon Jul 21 2014, 10:19AM


Ok, I think I get it, Neil. I'm not familiar with calculating eddy current losses, but I think I can see that at 450V there won't be a lot of resistance to the small amount of current we're considering here.

If I=V/R, then eddy current losses are entirely dependant on V and frequency, for any particular inductor?

I'm not sure I completely follow this, as the voltage is only applied for a relatively short period, every so often, and energy stored in a capacitor is a function of 1/2C and V^2, if I remember correctly, although I think I can see that, for any voltage, eddy current will be a constant, per unit of time. (EDIT: although V will be divided by the number of turns)

I can see that reducing eddy current losses will greatly improve efficiency, although using ferrite will presumably increase copper (ohmic) losses.

Would a powdered iron toriodal core be an inprovement over a gapped ferrite core? You presumably wouldn't need the same amount of copper, for example, due to the much higher permeability of the powdered iron.

Re: Light LED with 4 nanoamps
Dr. Slack, Mon Jul 21 2014, 11:24AM

Just think of it as inductor Q, energy stored divided by energy dissipated per cycle. The eddy currents in the iron put an effective resistance in series with the inductor.

Re: Light LED with 4 nanoamps
Ash Small, Mon Jul 21 2014, 07:55PM


If I had an easy way to charge a 1nF capacitor to 450V, I'd connect an led to the inductor and cap and try it, but I don't think I have an easy way to get 450V.

EDIT: If I get a chance I'll rig up a mains powered doubler, that should get to something around 460-480, maybe? (I won't rush into anything, though, not where 240ish mains is concerned. I'd need at least a fuse and a resistor between the mains and the doubler.)

Re: Light LED with 4 nanoamps
Tony Matt, Thu Jul 24 2014, 03:05AM

Hi Ash !

Your ideia of " easy way to charge a capacitor " is not bad at all....

1- supose that the bonanza battery carge 500V to the 0.001uF cap. The energy stored inside the cap will be 0.000125J.

2 - Then the pendulun moves and connect the 0.001uF in parallel with another cap. (40uF)

3- The voltage at the 40uF cap will be:

E = 0.5C.V^2 >>> V^2 = 2E/C >>>> V = (2E/C)^0.5

V = (2 X 0.000125/0.000040 )^0.5 = 2.5V

As the LED is conneted to the 40uF cap,, the cap will supply 2,5 volt to the LED.

Any suggestion or comments ?


EDIT : I the above calculation I have used the principle of " conservation of energy " to convert 500V to 2.5V, that is not the wright way.

The correct way is using the principle of carge conservation as:

1- Charge in the 0.001 cap with 500V:
Q = C x V = 0.001 X 500 x 10^-6 = 0.5 x 10-6 coulomb

2 - A capacitor with a charge of 0.5 x 10-6 coulombs and a potential of 2,5v shall have :
C = Q/V = 0.5 x 10-6 / 2.5 = 0.2 x 10-6 farads or 200nanofarads (nF).

So , please my apologies for my mistake.

The problem of two capacitors may be found in the internet.

Cheers

Re: Light LED with 4 nanoamps
Dr. Slack, Thu Jul 24 2014, 07:39AM

Hi Tony,

What makes you think all the energy in the 1nF cap will end up in the 40uF one? Do the sums again for voltage on the second cap, but this time use a non-zero approximation for the residual resistance of the interconnect, the switch, and the capacitor ESRs. If you don't fancy the algebra or the Laplace stuff, then just run it through a simulator.

Suggestion, try 1 ohm. Then try 10 ohm. Any difference in the final voltage on the 40uF cap?

Now do the same again, and assume a non-zero residual series inductance in the wiring and the switch. Is there any difference in the final voltage for the two resistance cases? Think about the Q of the residual series inductance and resistance.

Now try again with an enhanced series inductance.

Is a pattern emerging?

Re: Light LED with 4 nanoamps
Tony Matt, Thu Jul 24 2014, 07:16PM

Hi Dr Slack !

Of course my suggestion is just a concept.

We need to work, develope, test, and probably we may achieve a solution.

So far what I have for sure is:

1 - a good and visible flash in a NE2 bulb during the discharge of 0.001uF charged with 340V

2 - a poor but visible flash in a LED bulb during the discharge of 0.001uF charged with 340V


The transfering charge of a cap to another allways lose some energy. Losses in the ESR, pendulun contacts and some losses transformed in electromagnetic waves too.

I do not have a simulator, and the Laplace transformations are rusted... It is more than 45 year that I have play with such stuff.

Maybe some one can simulate such energy conversion, and guive us one idea about the losses, the role of a inductance or inductor to be added in such schema.

Re: Light LED with 4 nanoamps
Dr. Slack, Fri Jul 25 2014, 07:40AM

I'm not sure I'd flatter it by calling it a concept. Concept implies an idea for which the physics will work, but maybe the engineering hasn't been nailed, or the economics. The physics won't work. Connect a 400v 1nF cap to an uncharged 40uF one, and they both end up at around 10mV, regardless of the value of resistor used to connect them. Have a think about which conservation laws are used to arrive at that result. I love conservation laws BTW, as long as you pick the right one, it allows you to cut through all the mathematical cr@p and arrive at a limiting result with certainty.

Re: Light LED with 4 nanoamps
Antonio, Fri Jul 25 2014, 10:11PM

I made some experiments with this circuit, shown in the LTspice simulation below. To charge the 1 nF capacitor C1 I used an electrostatic machine (applying current in parallel, not voltage in series as in the simulation). My machine has on it the high-voltage diode D1, with about 20 V drop, that completes a buck converter (with some loss). The switch simulates a spark gap, that I set to breakdown at 450 V. The LED pulses visibly in the experimental setup.
The inductor L1 and the capacitor C2 are not critical. They control the maximum current in the LED, producing a short pulse of high current if small or a wide pulse of low current if big. The result is practically the same, as the total energy is what matters. A short pulse works better because the LED is more efficient with high current, and the spark gap operates for just a few tens of us anyway. The actual waveform seen in an oscilloscope is practically identical to the simulation.


A video, using 1 nF and 10 nF charged to ~500 V.

Re: Light LED with 4 nanoamps
Tony Matt, Sun Jul 27 2014, 04:39AM

Hi Antonio !

Amazing progress in the Zamboni project !

Now we have a stable spark gap that is very simple contruction, avoiding all problems of a mechanical/electrical construction, and a inductor that has a low inductance, simple and easy constuction.

If you do not mind, please inform the diameter of the coil, the number of turns and wire gage.

Cherrs

Re: Light LED with 4 nanoamps
Antonio, Mon Jul 28 2014, 01:33AM

I used a section of PVC tube with 8.8 cm of diameter. 430 uH requires 53 turns of #32 wire (My inductor is actually two in series, with the same effect). I obtained a significant improvement by adding another diode (1N4007) from ground to the node between the gap and the coil, making a true buck converter. I had to increase C2 to 680 nF to reduce the LED current to a less dangerous level. Note that my generator generates a few uA of current. 0.5 uA produces one flash per second.

Re: Light LED with 4 nanoamps
Newton Brawn, Tue Jul 29 2014, 03:27AM

Hi Antonio,

So nice get some light of so small energy...

And the air spark gap solved the switching problem with precision. The old technology of spark gap exeeded the actual semiconductor devices, at least in this project.

Maybe Antonio may try to check the result using a inductance of 1 henry in the place of 430uH. For me it is easy to wind at home a inductor of 0.5 to1.0 henry.

I suspect that 1 henry will allows the LED bright for few miliseconds, providing a better visual sensation .

Regards.

Re: Light LED with 4 nanoamps
Antonio, Wed Jul 30 2014, 12:35AM

I will see if I have a suitable inductor. A possible problem with a large inductor is that the spark must be active for longer time, with smaller current, during the time taken by the discharge of the 1 nF capacitor through the inductor. The time is proportional to the square root of the inductance, and the current to the inverse of the square root, approximately. This may cause an unstable spark, wasting energy. In the simulation, with 430 uH the spark lasts for 1.2 us, reaching 720 mA. With 1 H it must last for 50 us, reaching just 16 mA.

Re: Light LED with 4 nanoamps
Tony Matt, Wed Jul 30 2014, 03:15AM

Hi !

I undestand that the spark gap requires a good current to hold the conduction ( holdind current). If the current is low, the spark turns off. A experiment with Antonio gap may help find out this issue.

If the spark gap can not deal with so small current, the Zamboni pendulum has to providing the switching. The spark gap time on also may be require to be bigger, in this case the pendulum ( with some additional inertia) will increase the on time.

Other option is a manual switching.

Based on the actual dimensions of the 430uH inductor, a 0.5H inductor with same dimensions will require 34 x more turns, or 1700 turns. A old transformador used in the 5Y3 Vacuun Tube ( 350 - 0 - 350V ) without its stell core may have similar inductance.

Or a just wind a 600 turns coil with a ferrit core from a tv flyback .

Anyway, the results achieved so far are amazing !

Cheers

Re: Light LED with 4 nanoamps
Antonio, Sun Aug 03 2014, 12:09AM

I made some tests with larger inductors, with 30 mH, 0.45 H, and 5.5 H, all with ferrite cores and small. The LEDs continue to light up approximately in the same way (the energy of the discharges does not change) but the spark becomes unstable, stretching approximately as the simple simulation predicts, but splitting into a series of sparks, clearly separated with the two largest inductors. The gap voltage appears to be reduced by the large inductors, as the flash rate increases, but there is so much interference from the sparks that it's difficult to make precise measurements. I am not sure if my inductors were not damaged by the 500 V discharges. The result with the original 430 uH inductor, that supports high voltage, was more consistent with the simulation, with quite clean waveforms.
Another video, with the 430 uH inductor and the improved buck converter, showing the capacitive divider used to observe the spark gap breakdown voltage:

Re: Light LED with 4 nanoamps
Newton Brawn, Mon Aug 04 2014, 03:13AM

Hi Antonio,

Thank you so much for the simulation showing the results with several inductors and the clarification for my doubts.
I also agree that tha energy in the end of converter is to be maintained despite the inductor size.

Well, I hope Zamboni can show us his aparatus flashing the LED or the neon bulbs ...

Re: Light LED with 4 nanoamps
Newton Brawn, Sun Apr 12 2015, 04:23AM

Hi Zamboni !

We would like hear from you if you have done some work with the pendulum and lights !

Regards

Newton

Re: Light LED with 4 nanoamps
Zamboni, Sun Apr 12 2015, 11:58PM

I tried to replicate the designs that were suggested, but I was unable to make them work. Some of them required parts that I did not have.

I did have marginal success with taking the 4 nanoamp pulse that is delivered by the Swinger, and channeling it through an LED. I used a larger LED and had a better result. Part of the difficulty is it is just a 4 nanoamp "pulse", there is never a constant voltage.

I should have been clearer about the pulsed, rather than constant, nature of the voltage.

Some of the ideas that were suggested depended on a switch that would contact 2 poles at the same time. That never happens with my device.

I have enclosed 3 pictures that show the only success that I have had. The one called "overview", shows the entire Device. The one called CloseUp shows the clip leads hooked up to the LED.

As the Swinger rotates, the 4 nanoamps of current is delivered to Ground, but only as a pulse. not continuous voltage. I clipped in the LED between the ground terminal, and the batteries negative pole. That does give a barely discernible pulse each time the Swinger makes contact and delivers it's few hundred million electrons (I think that is accurate for 4 nanoamps) to the ground terminal.

These batteries are not able to connect in any way that creates a "short". About the only thing that can be hooked up directly to the battery is a capacitor. Anything else creates a direct short and that is a bad thing. I have hooked the Pos and Neg to a capacitor and it will charge it up to 1700 volts with no harm at all.

I tried hooking up a capacitor between the LED and the negative terminal. It did charge the cap, but there was no way to "switch" the cap on to discharge into the LED.

The very creative ideas that were suggested were wonderful, but I could not incorporate them into the Device. I bet any of them would have worked if I could have devised a way to make them "fit".

Frankly, my understanding of some of the ideas was so incomplete that I was not able to incorporate them. The people on this site are so well versed in electronics that a layman, like myself, can have a hard time following the train of thought.

I tried to get a picture of the tiny flash from the LED, but the camera would not show it. It is just too dim.

I wonder if having a bank of 10 large LED's and have the pulse fire all of them at once would work? I may have to try that. That may create a large enough flash to be picked up by the camera. If I do, I will certainly post a picture.

I really appreciated all the great suggestions.

Paul

Re: Light LED with 4 nanoamps
mister_rf, Mon Apr 13 2015, 03:21PM

Nice work. May I suggest you to use some high-efficiency LEDs for your experiments?

I have built a test circuit, controllable current source, connected to some various high efficient LEDs and cranked the dial down to observe the minimum threshold I could see the LEDs light.
It was as low as 10 nanoamps recorded by the camera, but in practice, in a complete dark room the light is quite visible with the naked eye down to 5 nanoamps.
I have been home-testing the following LEDs:

WURTH ELEKTRONIK 151053GS03000 green, 10.000 mcd, 522nm
KINGBRIGHT L-2523QBC-D X-BRIGHT blue, 2.300 mcd , 465nm
AVAGO TECHNOLOGIES HLMP-EG1A-Z10DD red 21.000 mcd 630nm
OPTOSUPPLY OSW54L5111P white 75.000 mcd
CREE C503B-RAN-CY0B0AA1 red 23.500 mcd, 630nm

Re: Light LED with 4 nanoamps
Zamboni, Mon Apr 13 2015, 06:26PM

Wow! Thank you for the suggestion. I will buy a few of the models that you mentioned and see which one works the best.

Thank you for all the effort.

I will post the results of my experiments.

Paul

Re: Light LED with 4 nanoamps
Ash Small, Mon Apr 13 2015, 08:07PM

A string of LED's across the capacitor will only 'fire' when the forward voltage of the string of diodes is reached (I think)

From memory, the forward voltage of most LED's is around 3 volts. A string of, say ten LED's will 'fire' when the capacitor voltage reaches ~30 volts. By choosing a capacitor of the 'correct' value, you should be able to vary the coulombs stored in the capacitor. Youi may require a resistor/inductor to limit the current to whatever gives a visible flash, as the capacitor discharges through the resistor/inductor and LED string (I think, maybe someone can confirm this, and fill in any gaps. I'm no expert in this field, but I think this basic idea, with tweaking, should work. Not sure how often it will fire, though, that depends on current produced and values of components used.)

Can anyone see any flaws in my reasoning?

EDIT: You quite possibly won't need a resistor/inductor to limit current, as most LED's are good for more than 20mA of continuous current, and can take pulse much higher for very short periods.

EDIT: Leakage 'could' be an issue, I'm not sure.

Re: Light LED with 4 nanoamps
Sigurthr, Mon Apr 13 2015, 11:31PM

You'll get incredible leakage over a range below the Vf. It (any diode) isn't a perfect switch unfortunately.

Re: Light LED with 4 nanoamps
mister_rf, Tue Apr 14 2015, 12:26AM

Yes, maybe that’s not so simple. Any LED is in fact a diode, and the ideal diode law gives an expression for the current through a diode as a function of voltage. But in practice at such very low currents, even that law it’s not accurate any more.


For example, in my previous LEDs test, there’s a string of 5 LEDs in total.
For a very large current (5mA) the total voltage drop is around 12Vdc. If we reduce the current up to 100nA, the voltage drops to 7Vdc, and for 10nA up to 5Vdc. I have measured less than 0.05nA for a total voltage drop of 3.5V.
This means you can’t charge the capacitor if the strings of diodes are permanently connected in parallel to the capacitor.

Re: Light LED with 4 nanoamps
klugesmith, Tue Apr 14 2015, 07:58AM

Nice work there.
There was a 4hv thread about ultradim LED's a few years ago:
in which an old white LED with unknown credentials was visible at 12.5 nA.

You make me want to go back and try some newer LED's. The ones most efficient at 10's or 100's of mA might not be the most efficient at 10 nA (see Figure 3 in the report below). It's really a contest of luminance, not luminous intensity or flux. Nice trick, running multiple specimens in series so we know their currents are identical.

Re: Light LED with 4 nanoamps
Ash Small, Tue Apr 14 2015, 01:29PM


So maybe if we can put a capacitor in the circuit before the switch, when it discharges it will provide more current?

Obviously, the capacitor will need to have virtually no losses.

As long as the capacitor charges to a higher voltage than the diode string, you'll get a 'flash'?

Obviously the 'flashes' won't be as frequent as without the capacitor.

Re: Light LED with 4 nanoamps
Zamboni, Sun Apr 19 2015, 12:34PM

I am having a hard time finding the OPTOSUPPLY OSW54L5111P white 75.000 mcd.
Any idea where I could get a very small order of 5-10 units?

Something on Ebay would be really handy.

Paul

Re: Light LED with 4 nanoamps
mister_rf, Sun Apr 19 2015, 04:06PM

I have bought those LEDs from here:

Re: Light LED with 4 nanoamps
Zamboni, Sun Apr 19 2015, 04:59PM

I have ordered them today. I will let you know what happens once they get here.

Hopefully, I will post a picture of the "flash".

Paul

Re: Light LED with 4 nanoamps
Newton Brawn, Tue Apr 21 2015, 04:03AM

Hi Zamboni,

If you set the moving pendulum in a " neutral " position, the pendulum can start the oscillation movement by it self ?

Re: Light LED with 4 nanoamps
Tony Matt, Tue Apr 21 2015, 04:42AM

Zamboni,

You wrote:
"My thought was to try to use the battery to charge a capacitor, and have the capacitor light the LED intermittently. I could not figure out a way to charge a cap, and have the cap "switch on" to light the LED, and then automatically go back to charging the capacitor."

Your concept is ok.

The way to charge the cap could be keeping the cap connected with the battery all time. The primary cap will be 1nF.

The device to "switch" the capacitor current to the leds will be the spark gap, as you can see in the videos above. IN YOUR CIRCUIT, the spark gap switching action has better performance than other modern switching devices.

And to " amplify " the current from capacitor to the led you need a buck converter. The buck converter is the 1N4007 diode, 430uH coil, and a 0.5 mF cap circuity, as shown in the Antonio's videos

I hope you can deal with the machine arrangement to incorporate the leds

Cheers

Tony

Re: Light LED with 4 nanoamps
Zamboni, Tue Apr 21 2015, 08:57PM

If I understand your idea properly, you are suggesting using a spark gap as a switch.

With that in mind, do you think this would work?:
The Swinging pendulum will deliver the 400 volt (at 4 Nana-amp) “load” from the positive terminal to the negative terminal. The “load” would then flow through the high voltage diode to the capacitor. That should return the terminal voltage to zero. For the Device to function properly, the negative terminal must remain at or near zero (the voltage can’t “build up” on the negative terminal). Once the voltage reaches the capacitor, it will charge until it reaches a voltage that is sufficient to trigger the spark gap. When the gap is bridged, the LED panel (I will probably need to switch to a neon bulb array) should flash.

Is that what you had in mind?

I have never tried to use a high voltage diode with the Device. I know that an LED will function, (even though the “flash” is almost invisible). I don’t know if the “high voltage” diode will even work at such low amperages. The leakage of the diode worries me.

Do you think it will work?

Paul

Re: Light LED with 4 nanoamps
Antonio, Wed Apr 22 2015, 12:55AM

I don't think that the idea can work in this way. But your schematic shows something that was forgotten in the previous discussions. The swinging pendulum is not short-circuiting the battery, but just taking a small charge from one terminal of the battery and, after swinging, returning the charge to the other and taking a small opposite charge. A possible solution is then to increase the charge that goes to the pendulum by adding a capacitor connected to it (by a thin wire that can be its support), and make the charge pass through a "floating buck converter", as in the picture. Another capacitor is connected across the battery, that has the negative terminal grounded. The LED will flash when the pendulum touches the positive terminal.

Re: Light LED with 4 nanoamps
Zamboni, Wed Apr 22 2015, 02:12AM



Yes. If stopped in the "middle" of the cycle, the Swinger will be attracted to one pole and start the cycle.

Paul

Re: Light LED with 4 nanoamps
Tony Matt, Wed Apr 22 2015, 03:02AM

Hi Zamboni,

In this project I think the best approach shall be separate the pendulum device from the light flash. Once we " dominate " and solve the flashing light issues them we may return our attention to the pendulum and integrate both into devices the battery.


Answering your questions:
1- Yes the spark gap is the best switch for discharge 500V from the cap / battery into the coil and diode. Note that the cap is connected in parallel with the battery all time.

2 - I try to understand your schematic and I am not sure it works. I will try post the schema showing what could work.

The diode could be a 1N4007, as tested and approved by Antonio's video. Two 1n4007 connected in series can handle 700V or more. during the switching the diode may comutate 2-10mA or more (for very short time).

Re: Light LED with 4 nanoamps
Tony Matt, Wed Apr 22 2015, 04:39AM

Here the schema more readable...

Re: Light LED with 4 nanoamps
Zamboni, Wed Apr 22 2015, 12:53PM


Antonio,
Your idea is quite intriguing. The only problem is that in my Device, nothing can touch the Swinger. The Swinger is held under a sapphire V jewel by a magnet. Any additional friction will stop it dead. I had thought about holding the needle directly on the magnet, but, because the Device rotates about 14 million times a year, the needle will "drill" into the magnet very quickly and cause the Device to stop. I had to add a sapphire bearing to reduce the friction enough to keep it going.

Thank you for you ideas. Matt has used your calculations in his circuit.

Paul

Re: Light LED with 4 nanoamps
Zamboni, Wed Apr 22 2015, 12:57PM


Tony,
I will round up the parts and see if I can get it to work without the Device. I will let you know as soon as I get it built.

Paul

Re: Light LED with 4 nanoamps
Zamboni, Thu Apr 23 2015, 09:06PM


The idea to “dominate” the method, and then try to get it working on my Device is a good idea. To that end, I connected a 3.3NF 2Kv capacitor to both the battery and my “Plexiglas and screw” spark Gap. The small neon bulb was connected through the spark gap.
The thought was that the battery would fill the capacitor, and then the capacitor would discharge through the spark gap into the neon bulb.

I figured that I would add the additional circuitry, once I verified that the concept was sound.

It worked. About every minute, I got a very visible flash of light. I, foolishly, used one of my “good” batteries (1700 volts) to test the theory. Turns out that even though the theory worked like a charm, the patient died.

I will explain, using my less than perfect knowledge of these batteries. As noted by another member, my dry pile batteries have about a 300 Giga-ohm internal resistance. The 3.3 NF capacitor has a substantially lower (though still very high) resistance. The end result was that, from the batteries point of view, the capacitor is a direct short. These batteries are extremely sensitive to direct shorts. I only had the battery connected for about 30 minutes. During that time, my 1700 volt battery Device drained down to a 1500 volt Device. Bummer…

I should have thought of this, I am painfully well aware of the sensitivity to “short circuiting” the battery.

I have researched “Glass capacitors”, that have 100 Giga-ohm resistance, but even they would appear as a short circuit. My conclusion is that I can’t ever have a capacitor directly hooked to the battery.

Has anyone ever of a capacitor that has greater than 400 Giga-Ohms of resistance? It would be nice to have a Tera-ohm of resistance.

The idea of using the “transferred” 4 Nano-amp charge that the Swinger moves from terminal to terminal is still viable as a concept.
That concept requires a way of keeping the transferred charge in the capacitor. The negative terminal must remain at or very near zero. I am guessing that the leakage back through the diode would be large.

Without about a Tera-ohm of resistance, there can be no direct connection to the battery. Direct hook up to the battery will cause failure…quickly.

The only thing that I can safely connect to the battery, is an Electrostatic voltmeter. The one that I have has a 1 X 10^15 internal resistance. I have left batteries hooked to the meter for weeks, with no damage whatsoever.

I have enclosed pictures of the Electrostatic voltmeter, the setup that I used and the spark gap that I made.

Paul

Re: Light LED with 4 nanoamps
Andy, Thu Apr 23 2015, 09:31PM

Hi Zamboni
Can you recharge or remove salt production, by running for a short time ac voltage source throught it, not 100% sure it will need a high ac source thought?Maybe have two piles, when one fires the otherone fires reverse porilty into the tempary low first pile as a ac source, and vice verser.
You could try instead of paper, titanium dioxide or caliucium copper titanate, so the pile capance matchs closer to the cap your trying to charge.

Re: Light LED with 4 nanoamps
Zamboni, Thu Apr 23 2015, 11:25PM


The LED's arrived today. I hooked them up and they work the same as the other LED's. There was no discernible difference.
Still far too dim.

Good idea though.

Paul

Re: Light LED with 4 nanoamps
Zamboni, Thu Apr 23 2015, 11:31PM


I actually tried Titanium dioxide several years back, the output was lower than what I am using now.

It may be possible to recharge them, but I think I will cut my losses and be happy with 1500 volts.

Paul

Re: Light LED with 4 nanoamps
Antonio, Thu Apr 23 2015, 11:39PM

Normal commercial capacitors really don't have a very high parallel resistance, specially due to surface leakage. When I use them in experiments the first thing to do is to clean their exterior carefully, apply some wax, and dry everything with a hair dryer. You can make a Leyden jar capacitor using a HDPE plastic pot, that has very low leakage. Can these batteries can be recharged?

Re: Light LED with 4 nanoamps
Andy, Thu Apr 23 2015, 11:54PM

By recharge, I mean force the protective oxide layer off the metal, if its Al metal and some others, not so much in the "secdonary battery" type way.

Re: Light LED with 4 nanoamps
Ash Small, Fri Apr 24 2015, 01:17AM

If I'm undestanding this correctly, 3,3nF is far too big a capacitor.

Try a capacitor in the picofarad range. Also, the 'buck converter' charging a larger cap to a smaller voltage may also be worth considering.

I'm probably stating the obvious by suggesting a 'larger battery', that goes against the original purpose of this thread, but paralleling a few Zamboni piles would probably also sort it.

I do think it's worth trying a smaller capacitor, though, and if leakage is a problem with 'commercially available' caps, look at Leyden jars, or other 'home made' capacitors which are designed for 'ultra low leakage'. You certainly have the engineering ability and understanding, but I do see that it's not the leakage that makes it appear to be a 'dead short'.

Re: Light LED with 4 nanoamps
Zamboni, Fri Apr 24 2015, 01:43AM


Antonio,
I don't actually know if/how they could be recharged. My gut feeling is probably not, but I really don't know for sure.

If I understand you correctly, are you saying that the internal resistance of a Leyden jar capacitor would be greater than 300 Giga-ohms?

Paul

Re: Light LED with 4 nanoamps
Zamboni, Fri Apr 24 2015, 01:46AM


I really don't know how to go about this, or if it is possible at all for this type of dry pile.
Paul

Re: Light LED with 4 nanoamps
Zamboni, Fri Apr 24 2015, 01:49AM


I will look into the Leyden Jar capacitors, do you believe that they will have an internal resistance greater than 300 Giga-ohms?

Paul

Re: Light LED with 4 nanoamps
Andy, Fri Apr 24 2015, 02:15AM

Type of saying the oppsite from Ash Small, but look for a comecial cap with leakage current in the nA, if mA then series them to low the leakage current, try 100-1pF/nF cermic cap with ten in series, and a low diletic constant glue and submerge the caps.
Firbeglass resin at about 5cm thick around cermic caps with ten in series, just check the leakage current in the datasheet as thats what is continue drawing power, after X amount of power is drawen the the Al metal will havea full protive oxide layer, and after you then use the charge on that metal plate, it will be difficult to recharge to keep the current supply constant, so starts at 100% then drops to 10% and will stay there untill you remove the layer, thinck Al electlight caps they have leakage current so it will still make power the pile, but at lower levels.
Reverse polarity by a high voltage in anti parrellel like back voltage or a 1700volt pile connected plus to plus, will force holes in the oxide layer, and then allow charge to build up at 100% untill its full oxidized again.

Re: Light LED with 4 nanoamps
Antonio, Fri Apr 24 2015, 03:10PM

Note that the battery may have been discharged not because of the leakage of the capacitor, but because the battery was repeatedly charging it and lighting the neon lamp. Of course this consumes much more energy than the required to just swing the pendulum. Leakage control at Gohm levels is not easy. Glass is useless as insulator, and even some plastics leak a lot. Surface humidity is the main problem, and if everything is dry plastics as Teflon, HDPE, acrylic and even PVC are excellent insulators.

Re: Light LED with 4 nanoamps
Ash Small, Fri Apr 24 2015, 06:17PM

According to Wikipedia (table shown on this page: )

Glass is 10x10^10 to 10x10^14 per metre, so not really good enough here, maybe if it's several mm thick.

Fused quartz is 7.5x10^17, which should be plenty good enough.

PET is 10x10^20 ohms per metre, and Teflon is apparently even higher.

All figures are at 20 centigrade.

There may be others. Fused quartz looks interesting., but as Antonio points out, some of the plastics may have advantages.

They will either need to be in dry air, a vacuum, or an insulating gas, most of which, from memory, are pretty toxic. Your machines already contain dry air, don't they?

Is it the leakage that's the main problem, or is it just the current drain charging a large(ish) capacitor that kills the pile?

EDIT: Any capacitor looks lke a 'dead short' until it's charged, then it looks like an 'open circuit' (apart from leakage).

Re: Light LED with 4 nanoamps
Zamboni, Fri Apr 24 2015, 09:17PM


You and Antonio both made the same point. You may be correct. I tend to defer to P.H.D's in matters of theory. Antonio's (and your) point seems likely. It may indeed be that there is simply a higher draw draining the battery.

whatever the reason, it makes the capacitor system untenable. I will have to look at ways of using the 4 nanoamp "pulse" from each swing of the pendulum.

My Devices use "regular" air. Nothing special.

I have not yet given up, I am still looking at a couple of ideas and will hopefully bring one to fruition.

I have taken 5 of the high efficiency LED's and strung them together. I placed them behind the negative terminal. When I did that, the 4 nano-amp pulse lit up all 5. They lit one after another in a moving pulse. Still quite dimly, but I am going to string more and see if human "persistence of vision" can be utilized to make it appear brighter than it really is. Just an idea, but it seems reasonable. I was thinking of a 5X5 LED grid...something like that.

If I could figure out how to make a diode function as a truly "one way" door, with no leakage, that would work quite well.

As always, thank you all for your ideas. It has been helpful.

Paul

Re: Light LED with 4 nanoamps
Ash Small, Sat Apr 25 2015, 01:19AM


If the capacitor is small enough, current drain will be much less.

3.3nF is obviously far too big.

If you reduce capacitance by a factor of a thousand, say a few picofarads current drain may well be acceptable.

Maybe reducing by a factor of one hundred would be sufficient?...say a few tens of picofarads, or maybe somewhere in between these values?

Adding a 'very small' amount of capacitance, and using your string of high efficiency LED's should give a brighter flash.

I think the issue here is that 3.3nF is far too large for your pile to charge.

I'm actually working on a project that uses tens of picofarads and a spark gap myself, but mine is designed for much higher voltages.

I'm building my own caps, and hope to post some 'very early' test results soon. My initial setup will be very crude, and will still need a lot of development after I start firing the gap, I just want a 'base point' from which I can develop the idea.

If the capacitor is small enough it should work, and it should be brighter than no capacitor at all.

Edit: I've just had another idea. You can do away with the spark gap if you use enough LED's in series. I'm not sure what the exact 'forward voltage drop' of your high efficiency diodes are, but the averave LED is around 3-3.3V. If you string enough of them together in series, say 150, the forward voltage drop of the string will be around 450V (you can adjust this to the exact voltage you require). Every time the capacitor charges to whatever total Vfd you choose to set, the string of diodes will 'fire'. Spark gaps are lossy, you don't actually need one

Maybe someone else can confirm this idea, or can find a flaw in it?

EDIT: LED's can sustain much higher current for a very brief time, generally 100mA for a diode rated for 20mA. If current becomes too high at these voltages you may need to add an inductor, but at the moment I'm assuming that won't be necssary.

EDIT: I can think of a few tricks you could use to use a larger capacitor, which doesn't fully discharge, so, due to the exponential drop in current draw as it charges, current drain on the pile would be less, but it's nearly 3am here now, and I need to give it a bit more thought before posting.

Re: Light LED with 4 nanoamps
Andy, Sat Apr 25 2015, 01:22AM

Ash about the leakage current or drain on the battery, they are the samething but negtively corlated, leakage current is small at first, drain is high, then leakage is high and drain is low.

Is a voltolic pile a battery ,a capactor, or a diode with the part that has the large effect on its outcome, a electrliclight capactor at a rating of 400volt can only withstand 5 volt wired wrong, but at the factory when the set the porilty they can supply 10mA to a 10000uF cap to set is oxide layer, with the rest of the current being supplyed by the reaction.

Just some thoughts ;)

If its a battery, a non-polar substance with high charge carrier density, if its a capactor a polar substance with the miniunm possable charge carriers.

Re: Light LED with 4 nanoamps
Ash Small, Sat Apr 25 2015, 01:49AM

Andy, with a well made capacitor leakage current s negligible. The leakage current of most electrolytics is awful in comparison, but I don't think it's the leakage current that's the problem. An uncharged cap 'looks like' a dead short to a DC supply, at first, anyway. When it starts charging the current slows exponentially, to the point where it never actually fully charges to the potential of the DC supply.

Re: Light LED with 4 nanoamps
Andy, Sat Apr 25 2015, 02:01AM

True that but with 4nA the 1700volt would drop to 1volt with a cap that large making the short draw 3nA, as long as the device can continue supply 4nA the size of the cap shouldnt be a problem.

Does the above make any sense?, I dont think it is a dc source as we know it, so to speak.

The easyest way to make this work is have a 70volt neon keep the cap the same, but make the pile 20-30kV, Q=CV2 and Ctot = 1/(C1+C2+C3...)

Re: Light LED with 4 nanoamps
Zamboni, Sat Apr 25 2015, 02:14AM


Your point is well taken. I will try a few pico-farads and see what happens. I think I will hook up my picoammeter while it is charging so I can see what kind of current draw is happening as the cap charges.

My batteries have a 4 nanoamp draw while running my Device, I have never measured the short circuit draw...mainly because short circuits are so harmful to my batteries.

This time I will hook up a "factory seconds" battery and watch the draw. Once I an confident that it is OK, then I can hook up a bigger battery.

I will let you know what happens...

Paul

Re: Light LED with 4 nanoamps
Ash Small, Sat Apr 25 2015, 02:18AM

Ok. As I said, if current draw is still to high I do have a couple of other ideas, but I'd like to see the results of your experiment first

EDIT: I think I may have found a flaw with my 150 LED string idea, but I'm sure there is a way around it. Using a shorter string and spark gap is one way, but there are probably other solutions.

First thing is to experiment with some smaller caps, then maybe experiment with caps that just partially discharge, but I'm jumping ahead here.

I think the 150 LED string idea only works with eg rectified AC, I did a Spice simulation of it a couple of years back. I think the diode string may prevent the cap from completely discharging, though.

Maybe I should do another simulation, but not tonight.

Re: Light LED with 4 nanoamps
Andy, Sat Apr 25 2015, 02:28AM

Zamboni, to help understand the current draw, teate the pile as a low Farad cap being charged by another battery, at 20kv max, it will charge upto 1700volt in a half second 0.01RC and then taper off, so the current draw max from the thing is based on max voltage(ie the voltage it reach with a dc short, removed then one second later), then resistance.

To save the expermint just look at ,capactor discharge time constant, match the voltage source to 1700v the capance 1pF, and keep changing the resance value to the amps reach 4nA in one second, that will tell you the tc of the load aswell

You have 25000000000ohm for the pile, and a max contious charge curret of 6.8e-8 A for a 1nF load cap, add a 200k resistor, for a tc off 0.024sec to reach steady current of 3.4nA for one second, with approx 300volt reached

The sim acrees with the values.



[Moderator edit by Mads Barnkob, double posts merged]

Re: Light LED with 4 nanoamps
Tony Matt, Mon Apr 27 2015, 02:40AM

Zamboni,

Do you have in your junkbox any resistor in the range of 500000 megohm ?

And

What are the values of your high resistors ?

Re: Light LED with 4 nanoamps
Newton Brawn, Mon Apr 27 2015, 03:35AM

Hi Andy,

I have other numbers for the time to charge the 0.001cap with 400V.

1 - assume that the max current to be draw from battery could be 20 nA (ZAmbone, please confirm)
This 20nA is the current drained from the battery to start to charge the cap 0.001uF

2 - If the battery supply 1700V, the series resistance (battery internal resistence + outside resistor ) could be 1700/ 20x10e-9 = 85000megohm.

3 - them the time to charge charge the 0.001 cap with 400V may be:
dt = Cxdv/i = 0.001x10e-6 x400/20x10e-9 = 20seconds or more .( if the current is constant)

The value of 200k resistor not was considered , it is too low in comparison with other resistors.

Re: Light LED with 4 nanoamps
Zamboni, Mon Apr 27 2015, 11:08PM


Largest I have is 1 mega ohm.

Paul


I believe that you are asking about the maximum safe current that can be withdrawn?

I do not know the answer to that question. I know that the maximum I have used is 7 nano-amps. My Picoammeter is not working properly, so I don't know the "short circuit" amp draw.

You time estimate almost matches what I observed. It took the battery 59 seconds to charge the 3.3nf capacitor that I was using. I have ordered a 2 PF capacitor, it should be here late this week. I will probably test on Saturday.

Paul

[Moderator edit by Mads Barnkob, merged double post]

Re: Light LED with 4 nanoamps
Ash Small, Tue Apr 28 2015, 01:47AM

I came up with this, I think it can be 'tuned' to any current we specify, to a point. I think it's self explanatory. The capacitor only discharges partially when the gap fires, then the second cap, which is a smaller value than the first slowly discharges through the resistor. When the voltage across the spark gap reaches breakdown voltage, it fires again, and the process repeats, ad infinitum.





Not sure how feasible it is to build, though. You charge the first capacitor before connecting the pile. If it only discharges partly, due to charging a smaller capacitor, the current stays in the 'pointy bit' at the top of the exponential charging curve.

The trick is to select suitable (and realistic) component values. Not sure what the frequency would be either, but that could be tuned to 'how bright you want the LED to flash', at least in theory. Components will certainly be a challenge.

EDIT: I think the value of the second capacitor is approximately the same as whatever value capacitor charged to the breakdown voltage of the spark gap, when discharged through the LED, lights it sufficiently. The first capacitor may be ten to twenty times the value of the second one (It needs to be big enough so that, once it partially discharges, it only re-charges at a rate the pile can handle). It should be possible to find a suitable resistor.

Re: Light LED with 4 nanoamps
Newton Brawn, Thu Apr 30 2015, 04:02AM

Paul,,

very good.

If you spend 60 seconds to charged the a 3.3nF cap, you will spend 20seconds to charge charge a 1nF cap.

Now you have to work in the circuit that will discharge the cap into LED.

I may sugest you follow the Antonio experiments.
The schematic was posted Wed Apr 22 2015, 12:39AM
the spark gap will not drain all voltage from battery

Re: Light LED with 4 nanoamps
Zamboni, Fri May 01 2015, 01:16AM


If you think it will work, I was going to try a 10PF cap, and a 100PF cap. Do those values sound good? If not, what would you suggest?
What resistor value would be a good start?

Antonio's design, though very inventive, will not work with my Device. There is no way to place a wire on the needle of the swinger. Any friction (from the wire) will stop the device.


Paul

Re: Light LED with 4 nanoamps
Newton Brawn, Fri May 01 2015, 02:16AM

Paul,


All you need is in the schema posted by Tony on
Wed Apr 22 2015, 12:39AM.

" 430 uH requires 53 turns of #32AWG wire wound on a section of PVC tube with 8.8 cm of diameter."

This schematic is the summary of all work done by Antonio.

Later, after you have adjusted the gap, insert a 100Meghom between the battery and C1

4 diodes 1n4007 in series will be fine

Re: Light LED with 4 nanoamps
Ash Small, Fri May 01 2015, 10:14AM

Won't the buck converter design place extra strain on the pile, as far as current draw in concerned?

By boosting the voltage, it leaves a voltage deficit behind, which places an even greater current draw on the pile?

Re: Light LED with 4 nanoamps
Newton Brawn, Fri May 01 2015, 10:06PM

Ash,

Yes, you are right.
It will place some strain in the pile, BUT just the time for Paul to adjust the spark gap. THEM he will place a 100 megohm between the battery and C1.
The gap will disrupt the voltage before the complete battery drain.

The converter is supposed to buck the voltage.

Re: Light LED with 4 nanoamps
Zamboni, Sun May 03 2015, 07:02PM



I will order a 100 Mohm resistor and give it a try.

Paul

Re: Light LED with 4 nanoamps
Ash Small, Sun May 03 2015, 07:26PM

As you're aware, there are different quality resistors on the market.

Not sure what the best type would be here. Maybe a film type?

Re: Light LED with 4 nanoamps
Zamboni, Sun May 03 2015, 11:16PM


I was going to get this one. Do you think it would be OK?

Re: Light LED with 4 nanoamps
Ash Small, Mon May 04 2015, 10:10AM

I imagine it will be ok for testing purposes. Apparently some lose accuracy over time. Maybe a 'name brand' with a datasheet, etc. would be better for 'production models'. Others here probably know more than me about this.

Re: Light LED with 4 nanoamps
Zamboni, Sat May 09 2015, 12:41AM


I saw this from Mouser. Does it look OK to you? Even has a data sheet. If it's good, I will order tomorrow.

588-MOX-750231006JE

Re: Light LED with 4 nanoamps
johnf, Sat May 09 2015, 07:40AM

Try VR68 range resistors they are very cheap max value 68 meg ohms
element 14 RS etc carry them

Re: Light LED with 4 nanoamps
Tony Matt, Sun May 10 2015, 12:45AM

Why not try first without the resistor ?

If it not work without resistor, you do not need order the resistor ...

Re: Light LED with 4 nanoamps
Zamboni, Wed May 20 2015, 11:52PM


I got the parts from Mouser and hooked it up. It does not really work. There is a tiny flash, but it is even dimmer than hooking up a regular 10MM LED between the negative terminal and the battery.

As I looked at the setup, another problem became more pronounced. The battery really does not like being hooked up to the capacitor. I noticed that the direct short of the battery and the capacitor is not for a moment or two, but for nearly the entire time. The battery is charging the cap until the spark gap fires, then charging again. That is the same as having a dead short the entire time.
That won't work.

So far, the only thing that can be hooked up "dead short" fashion is an electrostatic volt meter. The Tera-ohm resistance prevents the battery from "feeling" bad.

The other suggestion that uses a capacitor suffers from the same problem.

I really appreciate all the assistance that the forum members have provided, but it looks like it can't be done without harming the battery.

I did notice one very interesting thing. I will probably start another thread that describes a small, but interesting thing that happens when I have a 10MM LED hooked up after the negative terminal and before the negative terminal of the battery. The swinger delivers the 4 nanoamp charge to the negative terminel, the charge lights the LED (very dimly), then the swinger returns to the positive terminal to gather more charge. What is interesting is that when the, now depleted, swinger touches the positive terminal, the LED blinks again. That surprised me. It seems very interesting.

Again,
Thank you for all your assistance,
Paul

Re: Light LED with 4 nanoamps
Ash Small, Thu May 21 2015, 01:57PM

You could still try limiting the current to a few nano-amps with resistors, but you'd need a very low leakage capacitor.

If I had suitable resistors I'd try this with a Leyden jar, for example.

EDIT: Some form of 'water resistor', or even 'salt water resistor' might be worth looking into

Re: Light LED with 4 nanoamps
Antonio, Thu May 21 2015, 10:14PM

The capacitor is there to accumulate energy to flash the LED more brightly. There is no way to light a LED without consuming energy, of course. Resistors would not help. They would consume energy and reduce the energy delivered to the LED. Please provide a drawing of the system where the LED blinks when the swinger touches both terminals. Try a neon lamp instead of the LED.

Re: Light LED with 4 nanoamps
Zamboni, Wed May 27 2015, 12:53AM


Antonio,
The setup where the light flashes each time the swinger touches one side or the other is the older setup. It is the one where the LED's are attached between the negative swinger terminal, and the negative battery terminal. I have enclosed the original sketch, The only difference is that there is no capacitor or spark gap.

To be very clear, the flash occurs as the swinger touches either the negative terminal, which would be expected, or the positive terminal, which seems very strange to me.

I have tried a neon. It works OK, but just like the LED, too faint to be of any use...except in the dark, where you can barely see it.

Paul

Re: Light LED with 4 nanoamps
Tony Matt, Fri Jun 05 2015, 11:21PM

Zamboni,
Tell us the polarity of the hi voltage diode and polarity of the leds.
Regards


Antonio :
How about using similar arrangement of Bennet Doubler ? Nicholson Doubler ?

Re: Light LED with 4 nanoamps
Antonio, Sat Jun 06 2015, 01:45AM

With this schematic, the LED may really light up when the pendulum touches both plates. When it touches the negative plate, the charge that comes from the pendulum goes through the LED. When it touches the positive plate, the charge that goes to it comes in part from the negative plate, passing through the LED again. The effect is possibly more interesting with two LEDs, one at each side of the battery.
I could try to make the same using an electrostatic machine, as the rotating doubler of a previous video.

Re: Light LED with 4 nanoamps
Zamboni, Sat Jun 06 2015, 02:06PM


Antonio,
Your explanation makes perfect sense. No big mystery after all...just lack of knowledge on my part.

Thank You,
Paul

Re: Light LED with 4 nanoamps
Antonio, Mon Jun 08 2015, 01:16AM

I tried a similar setup using that machine of the video as power supply. The pendulum works without problems, but I didn't see any light with LEDs or neon lamps at the pendulum touches. My pendulum was small and with a different geometry, however. It's also difficult to keep well regulated the voltage produced by the machine, and the neon lamps had a tendency to light dimly following the ripple in the voltage. The voltage was higher too, probably around 2 kV.

Re: Light LED with 4 nanoamps
Zamboni, Mon Jun 08 2015, 01:57AM


Antonio,
The flash is very faint. I used a 10MM LED for maximum visibility. I have found the clearest results at night. I turned the lights off. The flashes are very apparent in the darkness. The battery voltage a little over 1700.

I use a 4 inch x 1/2 inch .002 thick Tungsten Pendulum.

Paul

Re: Light LED with 4 nanoamps
Newton Brawn, Sat Aug 08 2020, 12:55AM

It is so nice see again this kind of experiments !