low current inverter

Oh, good grief. You are indeed in the Slowman class of pendants.

Crap!

s/pendants/pedants/

And you are following JF's lead in avoiding on-topic content.

And most likely the last step you would ever take.

I know you have a fetish for mens asses but please leave mine alone.

Jamie

Good Lord, John. Have you been reading what you've written lately? Grow up!

I like to discuss electronics. Too many old hens here want to divert technical discussions to personalities.

So what *is* your problem with it? Exactly? What are the "errors"?

I find it admirably minimalistic.

If had done it, I would likely have had an input C, an output diode feeding another C, then a resistor feeding the LED. That is how one normally designs things; the Schmitt oscillator usually has an input C, a diode pump has the output diode+cap, a led has a resistor in series with it. That is the naive approach, but they are all useless parts.

Only issue I see is a possibly high current draw due to the transition current.

Charming

--- You like to play at discussing electronics so you can use that as a vehicle to stay in the limelight.

---

--- So calling people old hens instead of foregoing that and, instead, discussing technical matters isn't hoisting yourself with your own petard?

OK after all this I had to try it.

(Oh I had a 74HC14, run from +3V.. but no blue led's... I put two red in series.. and finally it's on a piece of white protoboard!) (note w = the frequency in radians /sec.) What to pick for C? Zc=1/(w*C), but w~1/RC, so Zc~R! (that's weird) So I stuck in a 1 k ohm and 0.1uF (that seems big for the C...) The thing worked but the oscillation frequency was up in the mega Hertz.

I mucked about a bit, but I finally had to add a cap to ground at the input of the inverter... that made everything under control.

George H.

That works on average, though the peaks will always be limited by the gate output. Which isn't horrible because a ~74HC at 3V will be in the 50-100 ohms range, perfect for direct driving LEDs.

Some years ago, I breadboarded a crude Z80 microcontrollery sort of project. A pair of 74LS273 latches drove an 8x8 LED matrix (8 digits of seven segment-plus-dot displays) with no current limiting resistors. The typical pull-up of 'LS is weak enough (~20mA S/C) not to really care. Neither the LEDs nor latches ever got hot or burned out, even under worst case conditions (whether all outputs active -- anodes high / cathodes low, or just one segment).

If I did that again, but with 'HC, I think I might add current limiting resistors after all. Then again, I'd probably do it at 3.3V as well, where the drive would be perfect for a red LED array without resistors.

Tim

To get the frequency down, use a big feedback resistor on the schmitt. Like a megohm, maybe. 1K is almost a short, so the gate will run fast, basically prop delay limited. That will waste power.

The charge pump cap should be smallish. Think in terms of, well, pumping charge.

I = C*V*F

where I is the average LED current and V will be a couple of volts maybe. Solve for C.

The 5-stage ring oscillator would be fun too.

No. This circuit will pump current into an LED whose forward drop is over 3 volts. Or pump a well-defined current into a blue LED whose forward drop is close to 3 volts... too close to make a series-resistor current limiting scheme practical.

Once you have a frequency, the average LED current is determined by the capacitor value, not by the resistor. It's a charge pump.

The parts only work "at cross purposes" if you don't understand it and pick silly values.

It's a very simple circuit. It's amazing how many useful, or just fun, circuits can be put together from just a few parts.

We used to use 74xx gates to drive (old, inefficient) red LEDs directly in the pullup direction.

The first blue LEDs I got were Cree SiC, which looked OK at 50 mA. A good driver was +5 volts, 39 ohms, and a couple of 7438 sections pulling down. As the LEDs got better, my customers complained of being blinded by the LEDs on our VME modules. Now we run about 2 mA.

You just don't get it, it negates (as you would say) its purpose.

Jamie