Self powered High voltage indicator

I saw that, no biggie, make it a 10W.

Another non-issue, any kind of zener starts conducting 10s uA at around 0.8 x Vz, so his threshold is around 20V. What's the problem?

Yeah- a 60:1 range on current there.

All those issues are easily fixed...the circuit is a good start.

OK, fix it.

My thing would need fairly low leakage, too.

Actually, 200 volt zeners would be OK. That would take 15 LND150s, each at 200 volts, 300 mW each, lots of small parts. The zeners could probably stand the 5KV transient.

As in, today...

The more the big guys go simulator-specific or encrypted, the more business I can build making models that work everywhere >:-} ...Jim Thompson

.JPG

at

at 200

the 5KV

No need to take unnecessary chances. With some 100 pF caps along the divider chain (after the sacrificial series resistors) would tame that issue right nicely and only add a few pF to the probe tip capacitance. Perhaps use a 470 pF for the last (bottom) step.

Just a thought. I think V3 is a better design.

?-)

Jim,

I have not the same results (waveforms) with LTSpice. May be i missed somet hing or more likely i did not understand how it works ...

Habib.

Technically, about 6.8V

A 1N4148* has the knee at around 115V, stick two in series :)

(*) This one I have here

On 10/03/14 22.18, John Larkin wrote: ...

How about using a (variable) zener diode amplifier:

Variable Zener Diode:

/Glenn

If I were to do this, I would bend the rules a bit and put a rectifier on the front end with an appropriate current limiting resistor, charge a cap and then feed that into a Xenon flash tube to build a relaxation oscillator. What have I missed?

Or instead of a Xenon flash tube, use 50 NE-2s in series, each with a small cap across it. Or use a fluorescent lamp with no filament connection, but use an indicator lamp that's a better impedance match to the source.

It needs to work from 50 volts to 3KV, and also needs to furnish a little DC to power some other circuits.

It doesn't work.

As in, today...

The more the big guys go simulator-specific or encrypted, the more business I can build making models that work everywhere >:-} ...Jim Thompson

John,

Ok i'm suspecting it does not work. I spent an hour trying to understand Jim's ideas with his schematic but no way ... Please Jim explain me how is supposed your idea working.

Best regards, Habib.

Of course it works. John Larkin just can't resist being an asshole.

The simulation shows that it works...

Q1 & Q2 form a discrete equivalent of a UJT, which trips once C1 charges up to VZ + 1*Vbe (of Q2, and ignoring Vbe of Q6, since current is _very_ low).

When it trips it provides base current to Q4 pulling the main current draw thru the LED. Q3 & R4 serve as a current limiter so that LED current is held at ~2mA, until the current in the "UJT" drops below a few uA, then the "UJT" resets.

Controlling holding current proved the difficult part, given your wide voltage range, and the variation of the current thru the 1.5Meg dropping resistor.

Thus Q6 & R7 were added, which divert any excess current away from the "UJT", basically maintaining maximum capacitor charging current at ~20uA.

Complicated to understand for young guys (and assholes :-), but straightforward for those of us who have been designing working circuits for more than half a century.

If you still doubt that it works, enter it into LTspice and try it (be sure to enforce a small maximum time step... it _is_ an oscillator).

I have no problem designing at the transistor level... after all, I design chips with THOUSANDS of devices for a living. Peruse my website...

To view my 18 patents and the hundreds of chip designs I've done. ...Jim Thompson

It doesn't work because there is insufficient power available through the resistor to usefully light the LED. The max current you can get from 50 volts and 1.5M is 33 uA. The max power possible is about 400 uW, and that needs a better circuit to exploit. These are fundamental blunders that no hairball of parts can fix.

The oscillator is flakey too.

And it doesn't provide the aux DC output that is required.

Ha! Got you. You're so FUNDAMENTALLY IGNORANT you can't discern how it works... even though I just walked you through it with the explanation above.

I guess you're "intuition" can't cope with charging a capacitor, then dumping it into an LED ??

Larkin, you're such a pathetic ignoramus.

The OP called for an "indicator", and, in fact, said a one point, "blinking" was OK. You apparently, dumb as you are, couldn't conjure a circuit that directly blinked, rather than making an "aux DC" to run the blinker. ...Jim Thompson

That's dumb- the 1uF capacitor is supplying the LED current, not the 1.5M.

An UJT ... ok,

ok,

enforce a small maximum time step, e.g. 0.001ms or so ? I'll give it a try tomorrow it's too late now :-)

Yes of course i have seeing them. Impressive !

Best Regards, Habib.