Joule Thief

[Farmhand]

Yeah Pirate, it's important to remember that the current limit is the peak conduction current value not the average, so a higher
inductance will lower the peak current value and store more energy, hence the better efficiency, I've erred on the higher inductance
side and so at lower voltages this might be why i cannot get the full output voltage, maybe the higher inductance restricts the
current too much at lower input voltages and so maybe mine isn't restricting the current by way of the chips current limiter maybe
mine is limited by the higher inductance. Scoping the setup under different conditions will tell the story I think.

If we connect a super cap at zero volts to the output with about 1.5 volts input we should see the current limiting in action
because the super cap will look like a short circuit to the chip. Still the chip should take care of business and charge the cap.
I should try some things and take notes. Like try to charge a 25 farad capacitor to 3 volts from 0v with over 1.4 v input.
I should see the voltage wave form from the mosfet drain get chopped to restrict the input current.

I've fried 3 DMG1012 mosfets in my other (own built) circuit because of too much current due to too high of an input voltage and
too much load with no limiting in place.  Always things will get fried if we try enough stuff. 

..

I need to set up a 1402 circuit, with the only difference between it and the 1402 circuit as the inductor, if there is a value that can be used by both I might do that, maybe 33 uH. If not I'll try to get the recommended value even if I must adjust or make them.

For capacitors I just used a 10 uF electrolytic and a 100 uF electrolytic, if I put a smaller better quality cap in parallel with them
that would improve efficiency as well, Parallel capacitors on the output of boost converters is a good idea even identical parallel
caps.  That's why I like Tinsels method of making big traces on hand drawn (mostly) PCB's because I can solder more caps on or
indicator LED's ect. I don't think signal traces should be too big though or too small or too long. 

..

[MarkE]

Really?  I have read in many places that will only work down to 1 volt, including on their website and videos.  Are they just rounding up the starting voltage from the chip? (.8 volts) and not mentioning how low it will still run?

I will get one and test it for myself.  If what you say is correct, it will work better than what I am trying to build, and I will have a lot more than $6 in each board.

I must be getting old.

Thanks mark,

Bill

(Pasted from Sparkfun's site)

"Description: The NCP1402 is a 3.3V DC-DC converter. The breakout board will accept voltage inputs between 1 and 3 Volts (such as 1 or 2 AA batteries) and output a constant, low ripple 3.3V output capable of sourcing up to 200 mA."

I checked their data sheet but it is just a copy of the 1402's data sheet I already have.  I do not see where it says it will run below 1 volt.

ETA:  I just ordered one board and I will test to see how low it will go.  If this works, this is cheaper than I could get a blank board in 100 quantity from a supplier!  Plus, I won't have to solder those damn smd chips.  I will let you all know.

Bill

What happens is that the load current that the board can support and still regulate drops rapidly below 0.9V input.  This is due to limitations of the NCP1402 and the inductor.

The 3.3V output board regulates at 3.29V down to:

0.55V input with a 1K load, 3.3mA
0.42V input with a 2K load,  1.7mA
0.30V input with a 4.7K load, 0.70mA

[Farmhand]

What happens is that the load current that the board can support and still regulate drops rapidly below 0.9V input.  This is due to limitations of the NCP1402 and the inductor.

The 3.3V output board regulates at 3.29V down to:

0.55V input with a 1K load, 3.3mA
0.42V input with a 2K load,  1.7mA
0.30V input with a 4.7K load, 0.70mA

Cool info, yes I'm working with a load of three five mm bright white LED's which draws about 23 mA at 3 volts output. I can't
remember if I measured the input, I need to buy some good CSRs in different watt values and small R values.

I've got super caps in .22F x 5.5 volt - 1.5 F x 5.5 volt  - 25 F x 2.7 volt - and 140 F x 2.5 volt. My 1402 Parts are 5 volt output
parts and my 1400 parts are 3 volt parts.

How does the circuit deal with a 3 volt output with a 4 volt input ? go to Buck (down) converter ?

Thanks Mark, your a big help, I could make a list of all the guys that have given me real help on these forums, it's not many and
you all don;t get along or agree on everything, but have one common trait, lots of experience and learned knowledge as well
as the will to help out. Much appreciated !

..

[MarkE]

Cool info, yes I'm working with a load of three five mm bright white LED's which draws about 23 mA at 3 volts output. I can't
remember if I measured the input, I need to buy some good CSRs in different watt values and small R values.

I've got super caps in .22F x 5.5 volt - 1.5 F x 5.5 volt  - 25 F x 2.7 volt - and 140 F x 2.5 volt. My 1402 Parts are 5 volt output
parts and my 1400 parts are 3 volt parts.

How does the circuit deal with a 3 volt output with a 4 volt input ? go to Buck (down) converter ?

Thanks Mark, your a big help, I could make a list of all the guys that have given me real help on these forums, it's not many and
you all don;t get along or agree on everything, but have one common trait, lots of experience and learned knowledge as well
as the will to help out. Much appreciated !

..

The NCP1400 and NCP1402 are boost devices.  They cannot go down in voltage.  They are used with a flyback diode that passes the input voltage to the output via the filter inductor if the device is not switching.

There are much more expensive ICs out there that buck or boost as needed.

[Farmhand]

Pirate, reading the data sheet the "soft start" function should stay in operation when we connect a super cap at 0v to the output,
it should restrict the duty of the input until the output voltage reaches 1.5 volts then go into full power mode if it can. I'll try it with
a 1.5 Farad 5.5v cap, to try it with a 2.7 volt cap with the 3 volt output I'll need to use a diode in series with the output to the cap to
utilize the voltage drop across the diode to get a full voltage of less than 2.7 volts so I'll need at least 0.3 volts drop and 1 Amp
diode. FR302 might work.

I'll try to get a shot of the output voltage ripple with my single 100 uF electrolytic capacitor and the load current