555 timer voltage doubler.

Why reinven the wheel!

That is the site that prompted me to ask.

The circuit is reliable but the voltage will drop fast when you load it (as stated in the text). I doubt it to be usefull for a chain of LEDS. You'd better use the power source as is and put the LEDs (or shorter chain of LEDs) in parallel.

petrus bitbyter

I plugged the circuit into LTSpice. The result for a 12v supply had the output voltage at 20v, and with a single LED down to 18v.

I know that SPICE has a margin for error, but for the cost involved, it is easier for me to get a 24+v power supply.

If you want increased voltage at a decent amount of power, use the 555 in a self boosting SMPS configuration.

Feed the 555's Vcc via an inductor & forward biased diode (cathode to Vcc pin - inductor to supply I/P) then drive a MOSFET with the 555's O/P and connect the drain to the junction of the diode/inductor. You will need voltage sensing and control of the M/S ratio as even with the minimum 4.5 Vcc, the boosted voltage will easily exceed the 555's Vcc max. A very simple way of regulation is burst mode, sense the boosted voltage and switch the reset pin when the voltage exceeds a set value - use a large electrolytic on Vcc.

1N4148s are a lot better than 1N400x, which are SLOW. Im too lazy to sim/calc the switching frequency, but with a 10nF cap and a 15k (ish) timing resistor it'll be around 10kHz or so.

make sure the caps are good low ESR caps, eg Rubycon ZLH series. If you got it from DSE, its probably crap.

and for a bit more grunt, place a complementary emitter follower between the 555 and the first cap. NPN + PNP, bases tied together connects to

555 with 100R resistor. Emitters tied together connects to cap. NPN collector to +12V, PNP collector to 0V. BC327 + BC337 is a good start, bigger transistors = more grunt (technical term)

Cheers Terry

Complementary emitter followers are great for beefing up the drive current, but don't forget that each B/E junction costs you 0.7V off the drive amplitude - certainly significant where the Vcc is 5V and the peak O/P may be a bit less impressive even with 12V Vcc.

The bipolar 555 can source/sink 200mA which is more than plenty for driving a diode capacitor charge pump.

and a PITA at 3.3V Vcc

depends entirely on the load. OP wants to drive LEDs.....

but C = 220uF, dV = 12V so dT = C*dV/I = 220uF*12V/0.2A = 13ms

OK the 200mA isnt a fixed figure, but what the aforementioned calc shows is that the current limited output effectively gives a lower bound on the output impedance & hence regulation.

NB: 12V/0.2A = 60R, so Tau = 13ms.

This suggests that there is a meaningful upper bound on the amount of capacitance you can drive - making it larger increases Tau, and once Tau gets up near the switching period, the load regulation gets worse. Im not going to analyse it, but its pretty obvious Zout will have a minimum for some value of C, getting larger with smaller C, and asymptotically approaching Zmax = (impedance of 555 + ESR of cap) for larger C.

high ESR caps and pitifully slow diodes add to the total output impedance.

Toss in an emitter follower (FZT651/FZT751 and now you get 5A), the effective slew rate goes up (dt = 0.5ms into 220uF), the effective Z555 drops (to about 2R4), Zout drops.....

Cheers Terry

If the OP wants significant power output then my earlier suggestion might be useful, make a boost converter by driving a MOSFET to switch the current through an inductor, the data sheet suggests the 555 guaranteed down to 4.5V but most chips will go down to about 3V - this is a bit low for fully efficient gate drive so it may be convenient to supply the 555 from the boosted rail, the flyback voltage can easily boost the voltage to more than the recommended Vcc max so I'd suggest using a zener and transistor to shunt the reset pin as a coarse burst mode regulation - the CV pin can be used to vary the duty cycle for more precise regulation.

thats an understatement. for voltages that low, you need to use a FET spec'd for operation at Vgs = 2.7V. "ordinary" FETs have Vth = 3-4V or so, and are seriously unimpressive when Vgs isnt much larger than Vth. I recently had to replace a 2N7002 with an NDS355AN for this very reason (Vcc = +3.3V).

A little bipolar is probably a better choice for low Vcc - it'll be happy down to 0.8V. Just bang a 1nF speed-up cap across the base resistor. even the humble BC547 works pretty fast when you do that - I have had built ~ 100,000 not gates using BC847 & BC857s (SMT versions of

547 & 557) with 10k Rb and 1nF speedup caps - slew + propagation delay < 70ns. The not gates were used to invert the primary SMPS FET gate drive to then drive the synchronous rectifiers, in a 36-72V:2.8V 50W DC-DC converter (94% efficient, NO heatsinking, inside an IP56 plastic box) used to drive a 12x12 RGB pixel array, forming one tile of an LED video screen. 4x3 tiles arranged into a 48x36 pixel panel (840mm x 630mm), buy as many panels as you want (US$5k/m^2) :)

besides, FETs and bipolars are basically the same from a driving perspective; the only difference is that bipolars have a mechanism to automagically suck out base charge, so you have to keep sticking it back in there. From a turn-on perspective though, its just about injecting charge fast enough.

so it may be convenient to supply the 555 from the

the OP certainly could. Of course the post was re. a 555 SC doubler, which is why I talked about it. An even simpler SC doubler would use a CD40106. Put all 6 gates in parallel, then set the compound gate up us a relaxation oscillator - 10k from OP to IP, 10n from IP to 0V (values wild-assed guess). OP drives a diode pump a-la the 555 version (of course using 1N4148s and low ESR caps). output impedance can be improved by soldering another Cd40106 directly on top of the first. Repeat until stack gets unwieldy... :)

This works even better when Vcc < 7V, as you can use a 74HC14 (down to about 2V!), which is a LOT gruntier than a CD40106, 6 in parallel pumps out more than 0.5A! add a sizeable NPN/PNP complementary emitter follower, and you can poke thru quite a bit of power.....

and if you are going to make a smps, make it properly. by the time you piss around enough to make a crude smps controller from a 555, you might as well buy a UC3842, and get some decent functionality. for a novice, peak current mode control should be mandatory, it'll save quite a lot of dead parts.

Cheers Terry

In my experience with using a 555 in a self boosting configuration, even a low gate drive will invoke enough drain current to get things going, from that point onwards the Vcc rises rapidly. The circuit I built used a MOSFET liberated from the SMPSU of a scrap monitor - I can't remember the type but I think it needed about 6 - 8V for full conduction (the circuit was powered by a 4.8V NiCd pack), the first prototype didn't have any regulation and the Vcc shot up to about 30V, the spec sheet for the Hitachi 555 I used said Vcc max 18V so its a miracle it survived. The boosted Vcc was brought under control by tying the reset pin to Vcc via a 4k7 resistor and shunting it with a control transistor who's base was fed from the boosted Vcc via a zener and current limiting resistor. This crude regulation was adequate for my application so I left it at that, but if the OP is driving LEDs there might be a noticeable fluctuation in the brightness, so a little extra circuitry driving the CV pin might be needed to provide continuous smooth MS ratio control. Since LED brightness is directly related to current, I'd suggest using an OP-AMP to compare the voltage across a current sensing resistor with a reference voltage and directly control the CV pin.

>