trigger LED

OK, this would work, with a few more parts,

and maybe a little math.

Is on the fast schottky diodes is maybe 100 nA or so, so 100K stretch per stage is possible, maybe 10K to be conservative. Even 1K would work in three stages.

A stretch ratio of 1e7 might just be possible in one stage. Discharge at 100 mA, charge at 10 nA.

-+- | .-. | | R1 '-' | V ~~> --- | ||--' ||-->|--'|--+ | ---

:-)

Cheers, James

That's my trick! The Is of the diode is the discharge. Cgs is 20 pF or so, maybe some Miller added, so we need a diode that leaks maybe 1 nA reverse. Not a schottky.

A 123 type one-shot has two inputs. I could clock B, and connect Q to A, so it locks its own triggers out while it's active. That might help.

(Note proper use of its and it's )

(Not to be confused with Is )

Huh, no bleed resistor? leakage from the Schottky? (or cmos gate?)

George H.

not at all, or not visibly?

Here's my favorite fault-indicator LED circuit. Its input is a /dis line, accepting wire-or'd outputs from multiple o/c comparators and gates. The /dis signal makes a BSS84 MOSFET charge the cap. It's discharged by base current of an NPN pulling current through an LED at some higher V.

. LED Fault Indicator . 40ms stretch . . 3.3V 5,12V etc . | 3.3V | . 10k | A . | S K LED . +- G pMOS | . | D | . | | C NPN . /dis +---- B . | E . 0.1 | . | 680 . Gnd | . Gnd .

Low-barrier schottkies have predictable reverse leakage in the 100 nA range, and a CMOS gate has pA of input current, so it works.

Here's an RF detector. Works fine.

Beta-graded transistors, like BCX70, help things like this. The old rule "never design around beta" doesn't always apply.

It's just a visual trigger indicator, so it's not fussy.

I think that's a winner -- triggers the '123 doesn't see, can't confuse it.

Cheers, James

Hey, I tried to look at your cute circuit but that 4mA LED poked my eye out! .-)

(I get complaints about LEDs being too bright at 2mA these days.)

Cheers, James Arthur

We ran the first Cree blue LEDs at 50 mA. As they got better, customers started complaining.

Somethin like

The red blip is a burst of 7 5ns wide pulses, the blue spike is a pulse through the LED connected to the HIGHLY _ILLEGAL_ 555's discharge transistor pin.

Kind of a lot of parts to light an LED.

Sometimes there's an advantage to using a few extra parts to convert a problem you don't know how to solve into one you do; detecting missing pulses in a low-frequency train of them is easier than detecting isolated pulses in the vast expanse of time when maybe nothing happens from like an "information theory" perspective, or something.

Logic is flawed and it takes 2xtpd to latch a discrete gate latch like that.

?????????????????????????

Thanks, I guess there's no reason not to use a diode both ways.

As long as you can live with the temp dependence.

George H.