Joule Thief

[gyulasun]

I looked for a datasheet on the D1960 transistor on the web but couldn't find one (D1961 and up),

Edit:  In case of confusion for beginners, the D1960 listed above refers to the 2SD1960 transistor, on the Fuji circuit board it is marked as D1960.

Hi Paul,

I managed to find some data on the 2SD1960 transistor, see this link: http://www.datasheetarchive.com/2SD1960-datasheet.html  see on the right hand side vertical column  and it here is a good replacement type from KEC http://www.keccorp.com  their type is KTD1146, data sheet is here, page 28: http://www.datasheetarchive.com/pdf-datasheets/Datasheets-115/DSAP001244.pdf

rgds,  Gyula

EDIT sorry I gave a different link, the correct one is: http://www.datasheetarchive.com/pdf/Datasheet-021/DSA00362797.pdf 

Another good possible replacement is a SANYO 2SD1145, see data sheet here: http://www.datasheetarchive.com/pdf-datasheets/Datasheets-30/DSA-580790.pdf 

[Koen1]

@Koen1,
Germanium transisters I think are the way to go, if one turns on below the usual .6volt silicon ones, then it stands to reason we can get better usage out of it. Theres also something about the humble LED, theres a lot more this device can do that hasnt been discovered as yet. Can it be used as a transister itself? just a throw away line, could be wrong.

LOL Jim, are you bugging my appartment or something? I just told a friend how I thought we could do so
I have a few of these green blinker LEDs, they blink on and off by themselves, and as far as I've been able to
figure out they also have a little resistance buit in... I am not sure but I think there is a sort of Zener type
setup in these LEDs that only allows the thing to light up when a current of sufficient voltage and amperage
passes the resistor and the zener, and after that it has a little off time.
But I am not sure, could be less Zener and more transistor as in that it regulates its own passing current...
Anyway, the effect on throughput would seem to be similar: only pulses are passed at regular intervals.
So yeah, maybe we can use such blinking LEDs to produce the input pulses eh?

@all: On the subject of the comparison of the JT to Tesla's toroidal transformers and the TPU,
there are a few things I'd like to point out. First of all, yes indeed, the generator was a DC generator
with the slide contact commutators placed at such positions on the generator shaft that the output
came in the form of AC that could easily be produced in two out of phase feeds. Simply placing
the slide contacts on opposing sides of the shaft would give 180 degrees out of phase signals,
placing them at 90 degrees on the shaft can give either type of 90 degree offset AC signal.
Telsa very often used his generator and commutator design to produce the input pulses or AC
which he then used to power his transformers and Tesla coils. So yes, in the toroidal transformer
design as posted a few pages back, the generator is not just a dynamo, it is also an AC generator
that produces two out of phase signals. We should be able to build a transistor-based setup
that produces two such AC signals that are out of phase by 90 or 180 degrees and reproduce
this "rotating" magnetic field trick of Teslas.

Second, in the post I just mentioned, it was stated that the input signals are 180 degrees out
of phase. I am not sure about this, I think it may be 90 degrees instead. It seems to me that
at 180 degrees out of phase, the magnetic "poles" generated by the input coils do not clearly
"rotate" in a single direction, rather if you look at the displacement of these "poles", it would
seem that the displacement in the ferrite toroid clockwise is equal to the displacement counter
clockwise, and this does not really constiture a rotation... If we were to feed the input signals
at 90 degrees, the "poles" do "rotate" in a clear direction and the displacement should have
a clear clockwise or counterclockwise direction.
But I must admit I haven't discussed this with anyone else yet so perhaps I am just way off?
Thoughts on this are very welcome!
I have read in a few of Tesla's designs that are sometimes named in TPU discussions that
the signals had a 90 degree offset...

Third, the spark gaps on the secondaries with the one secondary coil connected backwards,
now that's just weird.
The spark gaps themselves are not, the backward coil is.
Spark gaps were clearly used by Tesla in just about all of his transformer designs and he was
also clear on why that was preferable: a spark discharge through the air is actually a large number
of discharges back and forth, with a bias to the positive side. The charge does not jump the gap
in one discharge in the direction of neg to pos, but rather it jumps the gap via the ionised air
conduit, then most of the charge jumps back, and forth again, etcetera, untill most of the charge
has moved to the opposite end of the spark gap and disappeared into that side of the circuit.
According to Tesla's own words, this produced a high frequency oscillatory signal with every
pulse discharge over the spark gap. In the classic Tesla "coil" transformer, he used this oscillation
as input to the transformer primary, which of course caused a great many pulses of upward
transformed voltage in the secondary.
We could of course try to replace the spark gaps with Zener diodes of specific voltage...
That would effectively simulate one crucial quality of the spark gaps: to only allow discharge
when the voltage is high enough, so that the capacitor (or other circuit parts) can accumulate
a higher voltage before allowing the input to flow on through the circuit.
But that would not produce the same typical bursts of high frequency oscillations as the spark
gap would. Instead, it would theoretically produce one clean pulse of charge.
However, I am again not entirely certain of this, since I recall Bedini and some others showing
scope shots of pulsed DC from a transistor-based supply, and those shots seemed to show
a high frequency signal superimposed upon the DC pulses, if the scope had a very very high
resolution. I am still not sure if that is the case with all tranformer-produced DC pulses,
or if that depends on the type and quality of the transistors.
Can any of our older and wiser friends here say something about this?
Thanks!

Best regards,
Koen

[gadgetmall]

Well I'm on my 6th Fuji mod and still no go, one thing I've noticed is that not all boards are made the same, out of 6 only 2 were made the same.  Also, out of 6 only 1 had a D1960 transistor.  Some have LEDs instead of NE2s, others have the 4th transformer pin soldered while others don't, some have all 4 pins soldered on the trigger transformer while others only have 2 soldered, and on and on....

I think the best thing would be to get the components off the original Fuji cct boards and onto a separate cct board and start from point A to point Z until we can nail it down to a perfect setup and avoid all these discrepancies between the original circuit boards.

I looked for a datasheet on the D1960 transistor on the web but couldn't find one (D1961 and up), it's also not listed in Jeanna's "Transistor specs for transistors database. Information on Transistor prefixes of 2N 2SA 2SB 2SC 2SJ 2SK 3SK and more.pdf" so if it is a requirement to making it work we need to find an equivalent substitute.  I found a site where someone listed using the BC 337 and/or D1960 transistors but without the specs of the D1960 it's hard to tell if they are the same.

At any rate, the closest I got to an output higher than 1.5V while adjusting the 20K pot was not enough to light a CFL but was giving really strange outputs on the DMM, I would get an overload at the 20 VDC scale but only 1V at the 500VDC scale!  If I set the DMM at the 500 VAC scale I could adjust the pot to get as high as 64 VAC.  I just replaced the battery in the DMM and same results so it's not the battery and my other DMM confirmed the same results....

Edit:  In case of confusion for beginners, the D1960 listed above refers to the 2SD1960 transistor, on the Fuji circuit board it is marked as D1960.

Regards,
Paul

Hi Goat . I have modded all three  the aa with neon the aa with led and the aaa fugi . they are all modden the same way . replace the 220-ohm  resistor and short the diode. ..  Never had a problem and i got 22 of them all different types  and they all light up different Modded CFLS / I have no idea unless the cfls in Canada are using something different in the gasses  ? all mine light up all CFLS  . 9  13 24 and 45 watts .. all of them are spiral ... The ones in the usa have all the nasty stuff in them .. mercury Vapor . also you cant read the voltage unless you put a diode in series with your meter and your meter is 2000 volts or higher . has to be on the ac scale with a diode and only reads one way . if it Ols your meter dont go high enuff . when they are modded correctly i get a reading of 1900 and something on the 2000 volt scale .

Gadget

[yaz]

Tried looking up the 2SD1960 and it's substituted by a Matsuhita 2SD965, which is then cross referenced to a ECG 11.
Hope that helps.

[innovation_station]

Hi Goat . I have modded all three  the aa with neon the aa with led and the aaa fugi . they are all modden the same way . replace the 220-ohm  resistor and short the diode. ..  Never had a problem and i got 22 of them all different types  and they all light up different Modded CFLS / I have no idea unless the cfls in canada are using something different in the gasses  ? all mine light up all CFLS  . 9  13 24 and 45 watts .. all of them are spiral

really?

i have a few fugies... i will find them and take a pic

let me know if you modded the one i have ....

thanks gadget...

ist!

just finally got the mk2 6 divider mounted ..  i may cut back on the recharge batteries ...  so they charge higher and quicker ... might just have 2 of them insted of 4

i was thinking of adding a tiny jt as a indicator light 1 for each input  battery ...  and mounting the switch and indicator light jt on the other side of the supply battery ...