Lasersaber strikes again. A joule thief king ?

[Vortex1]

No goals are set for the design, but I can supply working circuits if you like, first just give me the design constraints i.e. what do you want the circuit to do?

Regards,
No idle talker

[conradelektro]

No goals are set for the design, but I can supply working circuits if you like, first just give me the design constraints i.e. what do you want the circuit to do?

1) The MAX931 multivibrator delivers a nice square wave whose frequency and duty cycle can be defined in a certain range, lets say 100 Hz to 5 KHz and duty cycles from 1 % to 50 %.

2) In addition the circuit can source up to 40 mA continuously from its output, lets say 20 mA would be enough to drive the gate or base of a very high current MOSFET.

3) The square wave from the MAX931 circuit can have a peak to peak Voltage of 2.5 V to 11 V (depending on the supply Voltage).

Now do that with less than 5 µA at 3 Volt or less than 14 µA at 11 Volt.

(One or two µA could still be saved by carefully selecting resistor values, at 11 Volt even more. The MAX931 circuit should be able to do it at around 3 µA at all Voltages from 3 V to 11 V if designed carefully. I think this is hard to beat with the same number of components.)

Attached please find a circuit diagram which you could redesign if you have time and ideas.

Greetings, Conrad

[Vortex1]

Greetings Conrad

Agreed, your original circuit is hard to beat if driving a single LED and your source voltage is limited to 2.5 Volts.

If you like nice clean rectangular pulses this is definitely a good way to go. You could substitute a cheap high gain low power op-amp for the Max931, a comparator is not absolutely required.

If you desire to drive many series LED's from a low voltage such as 2.5 volts, then the second circuit with the addition of the boost inductor is a good way to fly.

If 2.5 volts is not a constraint and you wish to drive many series LED's and only use the Max931 without the boost inductor, then you are limited to how many you can drive by the input voltage.

Since you can control duty cycle and frequency, there is no need to worry about inductor saturation, so no pre-emptive switch-off of the inductor is required. The second half of your circuit will work well "as is" provided the inductance is properly sized for the expected duty cycle and frequency.

I'm still uncertain what is the design goal? To run an LED for the longest time with barely visible light output while drawing the minimum amount of current from a given source voltage?

Unknowns are: peak led current required and duty cycle, type of LED required.

You could also use a simple relaxation oscillator based upon a unijunction transistor or PUT or even a simple two transistor multivibrator if the voltage is allowed to be up to 11 Volts.

I'll try to find some circuits in my archives and post them.

Kind Regards, Vortex1

[conradelektro]

I'm still uncertain what is the design goal? To run an LED for the longest time with barely visible light output while drawing the minimum amount of current from a given source voltage?

Unknowns are: peak led current required and duty cycle, type of LED required.

@Vortex1: thank you for taking the time to analyse the circuit.

Well, what are the requirements? Let's say it like a layman:

Drive a white LED at a "nice brightness" with as little power as possible. It would be desirable to do it with a single 1.5 V AAA battery. (This excludes the MAX931, so I decided to use two AAA batteries, the MAX931 still works with 2 Volt, so the circuit would work with two fresh batteries in series at 3 V and run down to 2 Volt eventually depleting the two batteries.)

"nice brightness": the LED is visibly shining in day light but far from nominal brightness. In darkness one would see the LED well but it would not really give useful light. (So, it is a pretty useless exercise.)

LaserSaber seems to be able to do it with 9 Volt and less than 10 µA (two red LEDs). I would like to beat that. Visibly blinking the LED is not allowed (that is why I used 120 Hz).

The trick obviously is to hit the white LED with very short pulses at more than 3 Volt, e.g. with 10 Volt spikes, because the LED would look brighter than with 3 V pulses.


Possible useful application: one uses these new 220V (or 110V) LED lamps which give bright light at 6 Watt or 10 Watt. It seems to be possible to drive these LED lamps with short 220 V to 400 V pulses at a useful brightness (e.g. like a flash light) at very low Wattages. I could quite nicely light such LED lamps with 0.1 to 0.5 Watt (with a 3 Volt supply Voltage and a Joule Thief type circuit). This would be a battery driven emergency light which allows to wander through the house in case of a power outage.

I attach circuit and photos from a project I did some years ago. I want to do a similar project but with the MAX931 instead of the microprocessor and a LED lamp instead of a gutted CFL.

The microprocessor needed 50µA to 100µA at 2.5 Volt.  I used a fly back transformer salvaged from a black and white TV set. I could connect the miroprocessor to my PC and with a program I could flexibly change frequency and duty cycle. It was quite difficult to find efficient settings for various power demands.

Greetings, Conrad

[Vortex1]

Hello Conrad

Thank you for explaining in more detail what your desired specifications are and also for showing your fine work. I will be sketching up some circuits for you, but first I want to test them myself on the bench, so that will take some time.

I want to try a programmable unijunction transistor driving a NPN transistor. Thats only two active components and three or four resistors and a capacitor.

The other circuit will be a two transistor multivibrator, four resistors and one capacitor.

I like to squeeze as much out of a few common low cost discrete parts as I can. My circuits will be better than the ones attached from the web. The PUT circuit will drive a transistor with it's cathode inplace of LED. The LED will go in the collector.

Other variations of these circuits can be found on the web, but the frequency of operation is too low, they will have to be tweaked for higher pulse rate and low power operation.

Kind regards
Vortex1