Electrical – How is this circuit working? This is Flashing LED oscillator, I guess.

Re

• I understand but I do not not how C1's current flows and comes from where and goes to where. But here, I think the voltage at "out" and Q3:C which is supply voltage, adds up.

The output if driving an LED would depend on Vsupply being < Vf_LED_On but > VfLEDOn/2. When Q3 is off C1 charges to vsupply via Vsupply-R3-C1-R7//R4//Q1_be.

When Q3 turns on the left side of C1 goes from 0 to Vsupply so the right side of C1 goes from Vsupply to ~= 2 x Vsupply *(as voltage across a capacitor cannot change instantaneously and left end was raised by Vsupply so right side must be too.

2 x Vsupply acrosss D1 turns it on if VF_D! < 2 x Vsupply.

This is a bad wayto drive D1 as it provides large chort current pulses. IF the circuit is driving D1 at about mean rated current then current peakswill be >> I rated. Moderm LEDs only allow a small margin - which this will exceed if LED is near mean rated current.

BUT cap C1 is charged vi R3 + R7 etc. If Vsupply = 1.5V then Icap charge = 1.5/27k = very very very little. .

_______________________________________.

Ocsillator action:

Q1 Q2 Q3 are effectively inverters.

Remove C2 temporarily

Force Q1b low -> Q1C high Q1C high -> Q2b high -> Q2 on. q2 on -> qQ2clow -> Q3 on -> Q3C high.

So the 3 Q's are high low high

And Force Q1b high -> Q1C low -> Q2B low -> Q2 off Q2 off-> Q2C high -> Q3 off -> on Q3C low

So the three Q's are low high low

This like 3 CMOS inverters in a row with output connected to input.
As their are 3 you get net inversion.
Connect out to in as here and it tries to "chase its tail" With pure R biasong it will settle dowm to some stable point/

BUT R5 + C2 connect across an inverting point - when one end is high it is driven by a low input and when low it is drive by a high input. This provides the negative feedaback to ensute ris system is unstable and will always chase its tail.

This is not a very "exciting" circuit. It id not obvious that it does anything overly well. But, it may.

Consider this circuit:

Try simulating that. Make L1say 1 mH to start. Can be say 100 uH or whatever.
Play with R1 & C.

Does it oscillate.
Note this is for 1 cell. Q2 grounds Q1's base - noyt good for high Vsupply

Components may be re-added as desired - this was a stripdown attempt AFAIR.

If R2 C1 time constant is long - say 0.5 second, you have a LED flasher. If R2 Ca time constant is short you have a boost converter and steady LED driver (it seems visually).

I "invented" this circuit in about 2000 and published it on-web in ?2001?.
I say "invent" as Q1 Q2 oscillator was used by Noah to power a deck light. I simply added L1 which was wholly obvious -and very successful. I would be completely non-surprised if this had been invented by 3,456 people, 10 years before I suggested it, BUT I've never seen any earlier versions. Several close version turned up inside 6 months of my publishing the original. No doubt a fluke - they didn't mention me :-).

The PUT is a nonlinear device that "switches on" when its anode terminal reaches a voltage determined by the gate terminal (approximately one diode drop above). It "switches off" when the anode current drops below a certain threshold known as the "valley current" (set in part by the source resistance of the gate terminal).

When your potentiometer gets down to 52K or so, it is capable of supplying enough current to keep the PUT turned on. At values higher than this, the current is low enough to allow the PUT to switch off.

Curiously, a "programmable unijunction transistor" as shown in this datasheet is really a 4-layer/3-junction device more akin to an SCR than an actual unijunction transistor.