Zener turn-on question

CD4000 logic is, as Jim says, like molasses at 5V. In addition only 30uA of zener current are required to turn on the NPN and unless it is a very tiny zener its leakage current will make that "sloppy".

If you never need more than 5V you'd be better off with 74HC.

Depends on the expected load.

A lower DCR inductor :-)

And you might want to reduce the bottom 22k (would be nice to write in designators).

Ok, I'm trying to recall the 74HC equivalent of the 4093..74HC132 yeah?

Was hoping for around 5 mA.

Can do

Hey, that schematic wasn't MY handiwork (though my penmanship probably isn't that much better)

Seems I can't go too low, though, I tried a significantly (physically) larger 680 uH L with a DC resistance of only a couple ohms and it doesn't appear to want to start up with that load.

Yup. Though I usually do those with simple Schmitt inverters, not NAND. Then the transistor tugs across the cap.

That isn't going to fly with 3V -> 5V and CD series logic.

I am surprised the CD4093 starts with anything at 3V.

It would also be good to reduce the 100k to less than 50k and increase the 100pF if you can stomach the slight increase in idle power.

Does buffered vs. unbuffered 4000 series make a difference? Currently on the breadboard is the "B" variant.

Heh, into only the IC itself and a 100nF bypass cap as a load the circuit as it is now starts up and oscillates all the way down to ~1.7 volts

Normally you'd want the buffered one. Non-buf is for crystal oscillators and when using CD-logic for analog purposes.

Without an inductor? I wouldn't do that, it could fry the load and the IC.

No, no, that's with the L connected and the bench voltage connected to the inductor as shown in the diagram.

I'll draw a schematic of exactly how I have things connected for clarity/specific advice as soon as I get a chance and double-check all the component values, I understand it's hard to know exactly what's going on when what I've built is not exactly the same as the original schematic

Good idea. A scanned hand sketch is fine and it doesn't have to be pretty. Even beer or coffee stains are ok on this newsgroup :-)

Maybe use a bandgap instead of a zener. Low voltage zeners are really soft.

At low power, you can also do a charge pump.

TLV431? I don't think I can get away with a bipolar TL431, IIRC they need at least a mA of supply current to make them "snappy."

The bipolar type works great as a constant voltage source for common cathode 7 segment LED displays in the color of red, driven from 5 volts. Stick one in the cathode lead, '431 reference pin to cathode, done. Brightness is just about right for most lighting conditions, no current-limiting resistors required.

How about an LM385? It's my standard low power reference. 30uA max for a guaranteed 'snap' as you call it.

Looks good, readily available, pretty cheap; since it's two terminal and only comes in 1.2 and 2.5 I'd likely need to use two in a variation of the circuit I posted. Maybe a 2.5 in the position the Zener is currently in, and a 1.2 hung off the transistor emitter. Plus ~0.5-something for the Vbe at low current. If I had to pick I'd rather be closer to 4 than closer to 6 volts on the output

Ok, this is how things are lashed up currently:

These are scope shots, first set at 1.7 V input from the bench supply, all pics with 10x probes, 50uS/horizontal div, 0.1V/vertical div, DC coupling

At junction of inductor and 4093:

Junction of C1 and R1:

Vdd pin:

This is at ~2.5 volts input:

Still using the old Kik. Can't seem to bring myself to retire it...

Label on the schematic should be "4093B" not 4096B obviously

Lose the inductor. Two 4093 gates. Take the voltage from the input output node between them. Take that and the output as your output. You've just created a voltage double. The output is not ground referred but to negative at the output - input node.

piglet posted that.

Something I didn't understand: If the oscillator's output is "high" when the inductor becomes fully discharged, the synchronous rectifier then begins

So, for efficiency, the duty cycle has to be >>> Maybe use a bandgap instead of a zener. Low voltage zeners are really

LM385-ADJ.

Long ago (1980's?), there was a cool Design Idea in one of the magazines that showed a cap multiplier made from CMOS-gates-as-synch-rectifiers. No diodes, IIRC.

I can't find it. Too bad--it might come in handy.

I recently designed a CMOS Marx generator for low voltages--3V lightning!

Cheers, James Arthur

Analog switches are cool for Marxist power. Connect any number of caps in parallel, then in series.

I was thinking I could get a lot of doorknob or hockey puck caps, charge them one at a time to 30KV, and physically stack them. Maybe drop them into a plastic tube.

I came up with another configuration that Phil unkindly called the Groucho Marx Generator.

At least he didn't call it the Harpo Marx Generator. :)