OPAMP ASTABLE

Opamp triangle generators used to be ubiquitous in PWM motor drives.

Done right, you get proper PWM, without spurious PPM.

So don't do that. Instead, use a Cockroft-Walton string of capacitors and rectifiers; the current delivered is constant-ish during the triangle ramp, and falls off only at the turnaround times. You can save parts by using a voltage-tripler instead (which only delivers current on the up-ramp phase), and compensate somewhat by making a ramp generator instead of symmetric triangle wave. There has to be SOME filtering on the HV end, either way. I'd use a shunt regulator and limit resistor for such a small current requirement.

I don't think so. The diodes will conduct at the peak of the triangle, so there will be a current spike there.

I'm thinking I'll use a slow opamp as a square wave oscillator, driving a diode-cap multiplier. If we pick an opamp with a low slew rate, we'll have a trapezoid drive. A series resistor can soften up the charging spike.

te:

by

e a

You'd be better off with a Baxandall class-D oscillator. The rectifying dio des still turn on at the peak of the sine wave, but the dV/dt is slow there so there isn't much of current spike. Your photo-diode isn't drawing much current so you will do most of your charging right at the peak of the sine where dV/dt is very low.

Not so. The impedance of the Cockcroft-Walton rectifier is capacitive, so constant dV/dt on the input makes constant current output. This is NOT just a filter capacitor and diode, it does NOT have that particular current-at-the-peak character. Do the engineering right, and size the series capacitors according to the current you're drawing.

On Thu, 19 Sep 2013 16:07:11 -0700 (PDT), whit3rd wrote:

Can you make that work? Besides, I want a voltage source, not a current source. A C-W is, after all, generally used as a *voltage* multiplier.

This

Try this:

Version 4 SHEET 1 1136 680 WIRE 688 -704 384 -704 WIRE 848 -704 688 -704 WIRE 896 -704 848 -704 WIRE 928 -704 896 -704 WIRE 688 -640 688 -704 WIRE 848 -640 848 -704 WIRE 384 -560 384 -704 WIRE 688 -512 688 -576 WIRE 848 -512 848 -560 WIRE 240 -384 96 -384 WIRE 384 -384 384 -496 WIRE 384 -384 304 -384 WIRE 384 -320 384 -384 WIRE 384 -160 384 -256 WIRE 512 -160 384 -160 WIRE 640 -160 576 -160 WIRE 640 -96 640 -160 WIRE 384 -80 384 -160 WIRE 96 48 96 -384 WIRE 240 48 96 48 WIRE 384 48 384 -16 WIRE 384 48 304 48 WIRE 96 112 96 48 WIRE 384 112 384 48 WIRE 384 208 384 176 WIRE 96 240 96 192 WIRE 96 352 96 320 FLAG 96 352 0 FLAG 384 208 0 FLAG 688 -512 0 FLAG 848 -512 0 FLAG 896 -704 OUT FLAG 640 -96 0 SYMBOL voltage 96 224 R0 WINDOW 0 66 47 Left 2 WINDOW 3 51 100 Left 2 WINDOW 123 0 0 Left 2 WINDOW 39 0 0 Left 2 SYMATTR InstName V1 SYMATTR Value PULSE(-10 10 0 1m 1m 1u 2m 1000) SYMBOL schottky 400 176 R180 WINDOW 0 -84 52 Left 2 WINDOW 3 -116 13 Left 2 SYMATTR InstName D1 SYMATTR Value 1N5819 SYMATTR Description Diode SYMATTR Type diode SYMBOL cap 304 32 R90 WINDOW 0 74 32 VBottom 2 WINDOW 3 94 32 VTop 2 SYMATTR InstName C1 SYMATTR Value 10µ SYMBOL schottky 400 -16 R180 WINDOW 0 -84 52 Left 2 WINDOW 3 -116 13 Left 2 SYMATTR InstName D2 SYMATTR Value 1N5819 SYMATTR Description Diode SYMATTR Type diode SYMBOL cap 304 -400 R90 WINDOW 0 79 32 VBottom 2 WINDOW 3 87 31 VTop 2 SYMATTR InstName C2 SYMATTR Value 10µ SYMBOL schottky 400 -256 R180 WINDOW 0 -84 52 Left 2 WINDOW 3 -116 13 Left 2 SYMATTR InstName D3 SYMATTR Value 1N5819 SYMATTR Description Diode SYMATTR Type diode SYMBOL cap 704 -576 R180 WINDOW 0 -50 47 Left 2 WINDOW 3 -50 15 Left 2 SYMATTR InstName C3 SYMATTR Value 10µ SYMBOL res 832 -656 R0 WINDOW 0 64 42 Left 2 WINDOW 3 57 78 Left 2 SYMATTR InstName R1 SYMATTR Value 10K SYMBOL res 80 96 R0 WINDOW 0 62 45 Left 2 WINDOW 3 67 78 Left 2 SYMATTR InstName R2 SYMATTR Value 1 SYMBOL cap 576 -176 R90 WINDOW 0 79 32 VBottom 2 WINDOW 3 87 31 VTop 2 SYMATTR InstName C4 SYMATTR Value 10µ SYMBOL schottky 400 -496 R180 WINDOW 0 -84 52 Left 2 WINDOW 3 -116 13 Left 2 SYMATTR InstName D4 SYMATTR Value 1N5819 SYMATTR Description Diode SYMATTR Type diode TEXT -40 -512 Left 2 !.tran 50m TEXT -224 -656 Left 2 ;TRIANGLE-DRIVEN C-W VOLTAGE MULTIPLIER TEXT -72 -584 Left 2 ;JL SEP 19 2013

If you use the full-wave C-W circuit, you get continuous output current when you have a virtual short circuit on the output (i.e. a 30V Zener diode will be attached directly to the rectifier, no series resistor). It's wise not to run the rectifier near its peak output voltage, you'd want the C-W drive to deliver maybe 200V into a short circuit. No energy is wasted, the current limit is capacitance not resistance.

This works well when the two branches of the C-W have matched capacitors and diodes, but unmatched components will, of course, result in some ripple at the triangle-wave frequency. A voltage-multiplier (half wave version of C-W) with a sawtooth instead of symmetric triangle would be OK. The application was for only 500 uA, so 'high' recovery-phase currents aren't an issue. An inductor and catch diode on the output will keep your Zener biased.

I don't understand that. Got a Spice model or something?

There's one surprising exception to that: make your product a Windows-based USB peripheral. That way, if you ship something that causes the customer' s PC to bluescreen immediately after they take your device out of the box a nd plug it into a USB 3 port, they will actually blame Microsoft!

Eventually the customer will try plugging your new gadget into a USB 2 port , at which point it will work as advertised, and you won't hear anything ab out it until they have an unrelated question or need to place another order .

I never realized just how low peoples' regard is for Windows-based instrume ntation until I, um, heard about some guy who did this.

-- john

Of course the transformer drive has to come from the power supply rail. My goal is to make it small and slow, with no big di/dt or dv/dt events anywhere.

I can't see how my opamp astable can have any hang states. Each amp drives itself towards its linear operating point, and once both go linear, it has to oscillate.

There are

I don't like fixing not-understood bugs by fiddling. It's that thing about return customers.

talks about using inverters to make ring oscillators that are guaranteed to oscillate. It's old and obvious stuff, but clearly put.