555 timer boost converter (and buck converter) switching power regulator

[Someone]

With the demise of the 555 timer competition this may as well be revealed here in full detail, I have never seen anyone else present a regulated supply based on the 555 chip but its very capable of generating useful supplies including those often annoying several hundred volt sources for vacuum/gas tube technology that spawn endless threads on this forum. I have run these discontinuous current configurations over thousands of hours at loads up to several watts, with suitable FETs and diodes they can produce 200V and more in a single step and 500V+ when using the doubler. Let loose they can produce flashover or direct failure of capacitors, diodes, and FETs, the voltages and currents involved are dangerous and they are not suitable for beginners to experiment with. Exact components are left as an exercise for the reader.



A stable reference Vcv is supplied by the forward biased diode Dr (approximately 0.6V) where the current through Dr is from the internal resistor divider from VCC to GND inside the 555 chip. These internal resistors also provide a second reference for the TRIG input comparator of Vcv/2. Lower cost can be achieved by using a resistor instead of a diode, at the expense of poorer line regulation.

At startup the feedback voltage supplied to TRIG will be below the setpoint Vcv/2 and trigger the FET to conduct. The switching frequency is dominated by the current building up in the inductor L1, and the current flowing through Rq reaching the reference threshold Vcv. In discontinuous conduction mode this will cause the the regulator to have a constant on time proportional to L/(VCC*Rq), low MHz switching rates are possible but constraints are typically minimised at several hundred kHz. Off time is a combination of Ri, Df and Ci but the Ri*Ci product impacts on the peak current sensing.

Peak switching current is set by Vcv/Rq, but during startup Df is strongly forward biased to turn off the FET. Some 555 timers will operate satisfactorily with Df replaced by a resistor and rely on silicon quirks (biasing protection diodes etc) to turn off the FET, then this feedback resistor can be taken from the feedback node to the DIS pin or current measurement node to achieve the same feedback modulation at a lower cost. In any case the inductor, FET, and diodes will see much larger current pulses during startup.

With the FET turned off the L1-D1 combination will feed current into the output load, Cl and Rl. The operating voltage is set by the R1-R2 divider and is approximately = Vcv*(R1+R2)/(2*R2), C1 provides basic compensation for light loads.

A voltage doubler can be easily added to the output to provide a supply which can be switched off to 0V by the RST pin, or simply for achieving higher voltages. Component selection becomes important with the high voltage peaks on the FET drain and may require a FET with specified avalanche characteristics, similarly the switching diodes will require a peak current rating to withstand the startup inrushes.



Finally it is possible to build a regulated buck converter, but lacking current limiting it does not have the elegance or robustness of the boost design. If operated from a current limited supply it is however much safer to experiment with and learn about loop compensation.

[T3sl4co1l]

Heh, cute.

The lack of voltage feedback would seem to suggest using a TL431, which would make it trivial to add an optoisolator for isolated applications, too.

You'd need a startup (low current UVLO) circuit to do it from mains, though.  The usual way being, a UC3842 (which, operationally, isn't really much evolved from what you've got already), supplied from a large resistor, so that it charges up the aux supply, turns on, and runs itself from an aux winding.

Tim

[NiHaoMike]

For the buck variant, you won't get enough gate drive for proper operation.

[Someone]

For the buck variant, you won't get enough gate drive for proper operation.

Yes, it will operate with a large voltage drop across the FET if you run it all from the same supply. Thats fine for low loads, or you change to a separate bias and power supply voltages.

[Buriedcode]

Necroposting... but its relevant!

I designed something similar but with flyback instead of a stock boost, using the 555 as, well an SR latch really.  Peak current limiting using the same method of low side current sense resistor, and setting the 'control voltage' to 0.7V with a diode.  Just like that.  However, I really don't think the comprators on the 555 are ground sensing, so surely there is a limit, or range, of voltages one can use for CV.

Has anyone built the above schematics and verified?  I guess I could just knock up a breadboard 555 and test it for myself - the specs of the comparators aren't in the datasheets, which is reasonable because in almost every application of this ubiquitous chip, they only see 1/3-2/3 VCC.

[tantric]

hi, what do i need to change in the circuit to make it variable output voltage?

or what do i have to change to make it output around 400volts?