Cool joule ss/ssg - Another view

[Slider2732]

With reference to Brad's pretty amazing findings with the circuit here:

After a chat with Chet, I decided to start a topic here for discussion, because we may bring in the Dovada or other circuits that are kinda sorta quiet. Also, there tend to be some highly sceptical know it alls in some threads on the main board.

These are my opinions and my take on what is going on.
First thing - the LED isn't being lit by any drawn milliamps, it's being lit by the back collapse of the drive coil. The very high coil resistance is something always of note in such a circuit - the higher the resistance, the higher the inductance and capacitance - you can find a situation much like a tank circuit within a coil. Such a situation aids the ring down noticed in the scope shots.
Also relevant is the 5 wire RomeroUK special coil. Regardless of personal opinions about the Muller, that coil works. Also a bifilar pancake coil has relevance, for the same reasons.

The 20% On and 80% Off is a key player. The circuit is running at the speed that it can, which is (from memory) around 5.5kHz.
A looser or less wound coil would enable a faster running, but also a greater duty cycle. I'm not sure of that, but it's the way it usually works.
The small amount of input in the uA, is enough for many blocking oscillators to give a good LED output, it's how my Garden Lights work. By the time the coil has collapsed and rung down and at the speed with which it does so, we see a continuous brightness in the LED output. Anything above 25Hz would do for that, here we have thousands of times faster.
The run battery will very slowly deplete, the charge battery will gain - 4 times faster in this case than the depletion of the run battery because of the duty cycle and the way that the ring down extends throughout the off phase. The net result is a net increase being shown across both batteries.
Ni-CD don't run like lead/acid, which i've just noticed Brad addressed in the main forum topic.

I'd like to see a circuit diagram, but am going to try to work it out from the video.
Haven't built an SSG type for a long time, yet it's a 10 minute thing in some ways, the rest is waiting around for something or nothing to happen

[Slider2732]

The build coils are 2x coil from dancing flowers.
I know they have a high resistance but have to test them yet.
Edit - 485ohms, 422ohms
The batteries will be either 2x Ni-MH 200mAh or 2x AA 350mAh Ni-CD.
I would imagine the AA's that Brad is using are the 350mAh Ni-CD's ?!
Transistor is likely to be a 2N2222A, unless I can find a high hFE C1815 or S9014.

Oh, solar garden light circuits, such as those using a 330uH choke and 5252F 4 pin chip are great, but nowhere near as efficient as a bifilar 300 winds ferrite cored coil, C1815 'GR' transistor, 1meg resistor and '103' 0.01uF ceramic cap...just for reference

[TinMan]

Here is the schematic of my circuit.

Edit-this is an air core coil set-please disreguard the core depicted in the circuit.

[Slider2732]

Edit - Cheers 



Tentatively, first run.....

IT WORKS !


Starting at 1.008V on the run battery and 0.634V on the charge battery.
After 5 mins, the voltages were 1.014V and 0.639V.
Both appear to be charging.

Or I should say 'appeared'. I tidied the wiring and inadvertently shorted the drive battery  
It went down to 0.972V and the LED went out.
But, recovery etc should bring it up, will keep an eye on it...or indeed replace the battery.
Upside - I found that touching the 100ohm resistor with a finger will significantly brighten the LED.
Run battery is now at 0.995V and I might just leave it, to see if the LED starts to come on again.

Here's my first run vid ('Unlisted' vid due to kinda sloppy build so far),
Taken because, on connecting the last connection it started up by itself. The switch was already 'On' and the LED came on.

Full spec:
2x Dancing Flower coils approx 450ohms each
C3198 'GR' transistor
100ohm resistor
1N4007 diode
Superbright white ex garden light LED
2x 200mAh Ni-MH 1.2V rechargeable batteries - not used in last few days
1x On/Off switch for the LED and charge battery

[TinMan]

Here is the 20 hour run video.The run battery voltage has started to rise!after 18 hour's??!.I have never seen a run battery of any type in these system's rise above the starting voltage-but this one seems to be on the way up.We have .003 volts to go on the run battery,and we have hit the starting voltage.The charge battery voltage has exceeded that of the run battery start voltage.For something so simple,i am seeing things here that i havnt seen in any other pulse system-using nicads.
Now off to watch Marks video.
http://www.youtube.com/watch?v=1ZTjJHrcro8

[Slider2732]

Yes Brad...this circuit replication of mine has been charging BOTH too  

Fixed the none LED problem - disconnect the meter !
It was the positive lead to the meter that was stopping the LED from being on.
Lead position is important too. Put the drive coil two leads next to each other and LED brightness will suffer.
If the LED goes out inadvertently while moving leads around, you can touch the Positive end of the drive battery with a multimeter probe and it will come back on.

Oh, to add and clarify, these are Ni-Cd not Ni-MH as I thought they were.


30 min run:
Start of both 1.680V
At 30 mins 1.702V

Update at 1hr:
Combined - 1.723V
Drive battery - 1.062V
Charge battery - 0.661V

Update at 1 1/2hr:
Combined - 1.743V
Drive battery - 1.073V
Charge battery - 0.667V

40mV per hour

[TinMan]

To all
As i was makeing the last video(yet to be uploaded),i noticed that although the solar lights are all the same kind-some have NiMh batteries in them insted of the nicad's.The lite green and dark green seem to be the metal hydride's,and the yellow ones seem to be the nicad's(that i have found so far).So i have been chargeing metal hydride's and not nicad's-but there is a nicad in the larger setup now,being charged from a super cap.

[PhysicsProf]

  Very interesting -- and reminiscent of earlier circuits in some ways... 

   Can someone check the Frequency of the oscillation?  Is it 50 Hz (Oz) or 60 Hz (US)?

[Peterae]

I could be wrong but from a bit of experience with cells and colour i have found
Yellow to be nicad
Green to be NIMH
Pink LIPO4
Blue LI

I am not sure if this is a standard but probably is googleable  

[TinMan]

The frequency is 5.856kHz-no sign of my 50Hz AC in there anywhere.

[TinMan]

Hi Peterae
Your spot on with your color bands and type's

[Slider2732]

Indeed, the colours check out here. Good observation...these 200mAh Ni-Cd's in my runner are yellow.
Only deviation are some black and red Ni-Cd's at 350mAh in some older garden lights.

@Brad - So, in other words, we've swapped what we thought our circuits were charging ! lol
You were charging Ni-MH when I thought I was, I'm charging Ni-Cd when I thought they were Ni-MH !
Funny old world eh.

Updated figures, the circuit was left running while I slept.

02.20 Combined voltage 1.680V
04.20 Combined voltage 1.749V
12.20 Combined voltage 1.845V
Present drive battery 1.121V
Present charge battery 0.724V

The charge rate has slowed a bit, but it's still charging away merrily.
The LED may be slightly brighter.

Another will be built and this one will explore caps.

[PhysicsProf]

The frequency is 5.856kHz-no sign of my 50Hz AC in there anywhere.

Excellent!  thanks for this development and measurement Tinman!

Also Slider -- excited by this work.  The battery seems to be "getting a charge" out of it, too!   

[TinMan]

Indeed, the colours check out here. Good observation...these 200mAh Ni-Cd's in my runner are yellow.
Only deviation are some black and red Ni-Cd's at 350mAh in some older garden lights.

@Brad - So, in other words, we've swapped what we thought our circuits were charging ! lol
You were charging Ni-MH when I thought I was, I'm charging Ni-Cd when I thought they were Ni-MH !
Funny old world eh.

Updated figures, the circuit was left running while I slept.

02.20 Combined voltage 1.680V
04.20 Combined voltage 1.749V
12.20 Combined voltage 1.845V
Present drive battery 1.121V
Present charge battery 0.724V

The charge rate has slowed a bit, but it's still charging away merrily.
The LED may be slightly brighter.

Another will be built and this one will explore caps.

Still going up-cool
I have tried super caps on my larger setup,and the voltage on both caps went down.As soon as i replaced the charge cap with a battery,the battery started to rise in voltage.This is a clasic example as to why you just cant simply change batteries for cap's.A cap has the same(or lower)parallel resistance as a battery,but a much higher series resistance.As we have the two batteries in series,we cannot simply replace them with caps and think it will work just the same.So when you hear people say!just swap the batteries for cap's!to proove that your machine is OU-some time's that just wont work.We are charging a cap with a cap,and we will have that 50+% loss straight up.
I am now charging the battery with a super cap-we'll see how that one go's.

[Slider2732]

@PhysicsProf - It really is great, in terms of very well known tech with a very very subtle twist.
Particularly the use of 2 diodes end to end and one being an LED, as it's similar to the Dr. Stiffler SEC circuits.
Wonder if the pair can go to the Base and a cap go over the resistor, hmmm

Have built a second, cleaner circuit and this one is showing similar traits to the first...the lack of LED firing which is infuriating !
I can touch the Emitter and Collector and get a flash, but then the LED doesn't stay on. Disconnect the charge battery first, then do it, still nothing.
The first circuit does the same thing and moving the wires around or touching the battery contacts with a finger used to switch it on. It's staying off too. Should have just put it to one side running really.
Anyway, first circuit is still going, though slowed now with the charge battery not seeing much.
Drive battery - 1.168V
Charge battery - 0.735V
The charge battery is climbing slowly back to the 0.749V of 04.20am, before I messed it up...finicky things these if mucking about with them.
The drive battery is continuing to climb, which I can't work out for the life of me lol


Points noted on the caps, this second build retains the battery holders.
Biggest difference, the use of 2x 24V relays. 2.7k each for resistance.
Uses a 500ohm pot.
Transistor is a C1815 'GR' again.

[Itsu]

Hi,

here my cool joule replication.
I have to kickstart it with my FG, else it won't start.
B2 battery slowly decreasing, B1 battery slowly increasing.

Diode is 1n4001, rest similar as Tinman.

i will have it run overnight.

Video here: http://www.youtube.com/watch?v=OIpO2fmuWI8&feature=youtu.be

Regards Itsu

[Slider2732]

Function generator to kickstart ! neat idea, thanks Itsu, will try that here too.
By the way, you must have an impressively sized cat there...sounded like it brought a shelf of parts down or something !
My usual in those situations is "*%^*&% CAT !!!!", beep as the camera goes off, beep as the camera goes back on again, compose the self and "Ok, so here we have the circuit running..."


The first circuit here carries on running, with no LED lit.
Drive - 1.179V
Charge - 0.753V
The charge battery has now surpassed the original highest of 0.749V
So it's all going, but no light output.

The second circuit still isn't behaving, but now i've changed to a pair of 150ohm relay coils.
It has shown something which I would like confirmation of.
The charge battery was fully out of its holder. The LED therefore had no connection to cathode. I connected a 3rd coil, of only about 8ohms, by 1 wire from that cathode of the LED, the other end was disconnected. On tapping the Emitter and Collector of the transistor together with a piece of wire, the LED will flash dimly !

[Slider2732]

Circuit 2 is now running
The fix, was to ditch the idea of steel/ferrite cores and go back to dancing flower type coils.
There'd been a dancing Hawaii woman type solar toy on the windowsill for a while and a coil from a dancing flower was already on the bench. So, the young lady from the windowsill has donated her coil.
I connected them up and on the first short of the Emitter/Collector, it fired straight up.

Will let it run overnight.

[Itsu]

Slider,  she indeed is impressive (see picture), its a Maine coone, but small compared to the 3 mail maine coons we also have.
Here problem is that she is very nosy, fortunately, the box she dropped was closed, so i only had to pick it up.


The overnight result of my cool joule shows more decrease of the B2 battery (-176.3mV) then increase of the B1 battery (+85mV), so overall a negative result.

I will let it run some more days to see whats going to happen.

Regards Itsu

[TinMan]

Slider,  she indeed is impressive (see picture), its a Maine coone, but small compared to the 3 mail maine coons we also have.
Here problem is that she is very nosy, fortunately, the box she dropped was closed, so i only had to pick it up.


The overnight result of my cool joule shows more decrease of the B2 battery (-176.3mV) then increase of the B1 battery (+85mV), so overall a negative result.

I will let it run some more days to see whats going to happen.

Regards Itsu

Hi Itsu
Im not sure why your run battery(B2) is dropping so fast?.It may pay to check that your DMM's done draw any current when reading voltage.I have two that draw around 1 to 2 Ma when reading voltage.You can check each meter simply by using another DMM set to Ma in series with the DMM reading the voltage to see if it draws any current while reading the voltage.With these low power circuit's,it dosnt take much of a draw to throw the charge and draw rate out.

[Itsu]

Hi Tinman,

ok,  tried that, but no reading at all on my DDM in series with the voltmeter (B2), it shows 0.000 in the DC 2mA position.

Regards Itsu

[TinMan]

Hi Tinman,

ok,  tried that, but no reading at all on my DDM in series with the voltmeter (B2), it shows 0.000 in the DC 2mA position.

Regards Itsu

I just tried a 1n4001 diode insted of the 1n4004-my charge battery charge rate went right down?.If you have any 1n4004's,swap it out with the 1n4001 you are using.I believe the 4001 is only good for 40 volt's.I know we see only around 5 volt's on the scope with the circuit(LED to charge)hooked up,but disconect the charge battery and have a look at the voltage on your scope.

[Slider2732]

Runners and non runners circuits are equally important.

Also errors in describing circuits...which is what i've done.
I said "C1815 'GR' again" for the second build, in fact it's only a C1815 'GR' on the second.
The first is as originally described, using a C3198 'GR'.
Both are comparable and the 'GR' same hFE designation threw me, sorry about that.

This second circuit ran over night.
It lost 20mV.
Before the run, the diode was changed out from a 1N4148 (100V 200mA rated) because I wasn't sure that it was still ok, with the LED not coming on. It now has a 1N4001 type black/silver band but the number is 1R100ZC.
I'll change it to a 1N4148 and then to any others, to see if there are differences.
Also, my 500ohm pot is presently sat at 142ohms, so will be decreased to 100ohms.
And, remove the two small relays that still sit on the circuitboard.

First circuit trundles on though:
Run - 1.195V
Charge - 0.797V

[Itsu]

If you have any 1n4004's,swap it out with the 1n4001 you are using.

amazing, as i did not had a 1n4004 readily available, i tried several other diodes like a uf4007, a 1n4148, and even a germanium diode 1n60, but all showed decreasing voltage for B2 battery, some diodes worse then others.

Finally i located a 1n4004 on an old piece of pcb in my junkbox, and this directly showed an increasing voltage on B2!!

It could be the battery is recovering now, but none of the other diodes showed this.
So both batteries are increasing now, so lets see where this goes to.........
b2 start voltage when 1n4004 was inserted is 0.9823V, now after 15 minutes its 0.9872 and climbing.
B1 is steady on 1.064V

Regards Itsu

[Slider2732]

Reanalysing my first circuit - the diode reads as 'RGP 9'. I have hopes it's a 1N4004 equiv but haven't looked it up. Same size, colours.

The second circuit now has no relays on it, so will run while we're out for an hour or so.
The batt voltages will presumably sink, so i'll then change to a known 1N4004.
Then decrease the pot to 100ohms like my first one. Brad's being a higher rating of resistance is what i'm looking at there.

Good info about the diodes, thanks Itsu and let's hope yours is now a runner circuit