Colpitts dc/dc converter

What are the input & output voltages? We need 3V in and 15V out. We are currently using a 13001 driving a 15:200 transformer, as you suggested many years ago. But we can certainly make it smaller with transistors.

If you don't need the low rise/fall time requirement that Larkin has, then a Royer/Jensen converter is pretty simple too. It used to be used in all PC power supplies before PFC was a requirement.

Cheers

Klaus

You're just hacking around with various oscillators and failing to address noise suppression. One of your last posts knocked small conduction angle waveforms for bad load regulation, so just exactly wtf is this?

The way to keep noise down is to minimize di/dt and dv/dt everywhere in the circuit: avoid hard switching and shoot-through, and keep the frequency low. The Colpitts is pretty good that way, especially if the base current is kept low so it runs in a nice sinewave mode.

As long as you're sure it will start reliably. Backing off on the gain without using AGC can make that a bit iffy.

Cheers

Phil Hobbs

This circuit self-AGCs. You apply more base current than you need to oscillate, several times more. When the oscillation gets so big that the collector swings down to ground, the c-b junction conducts and pulls the base voltage down. At that point, the p-p collector swing will be almost exactly 2*Vcc, with a small flat spot on the negative swing and near zero TC. Done right, it's absolutely reliable.

What I hadn't noticed before is that if you apply a lot of base current, you can get a big flat on the negative swing, and a corresponding bigger positive peak swing, and get more than 2*Vcc p-p swing. Because I want low noise, I wouldn't operate in this big-flat mode.

It looks like I'll only need microamps of photodiode supply, so the PV things look promising. Maybe piggyback one or two of them on top of my regular negative supply.

Ah, okay limiting by saturation. (I didn't look at the .asc file.) Most of my Colpittses have been for PLLs, so I've never designed one that used saturation as its amplitude limiter because if you use cutoff instead, you get much better phase noise performance. (Not as good as with AGC, but inside a PLL it's often good enough.)

Cutoff limiting is more vulnerable to startup problems unless you're well into Class C, I think.

Cheers

Phil Hobbs

When I ran your posted circuit in LTSpice, Ie(Q1) peaked at 86mA on every cycle, with FWHM of 2.3usec in a cycle time of 27.2usec, which is pretty much what you expect with any class-C oscillator

This is an unnecessarily high di/dt - the Baxandall Class-D oscillator was specifically designed to avoid spike-driving the tank circuit, and works with much lower di/dt values.

It costs you an extra inductor, but if you are as concerned about minimising noise as you claim, it might be worth the extra space and expense. It won't be 100 times the volume and ten times the price, no matter how ineptly you implement it.

Doesn't matter... I need tens of uA for the photodiode.

Looks like a charge pump may be better. I've got a nice sim, using a slow rr opamp as an oscillator, running from +-12, pumping up to about -34. The opamp has a low slew rate, running at about 8 KHz, and looks pretty good. The photovoltaic things are fun, but can only make about 7 volts and 25 uA each, too marginal for comfort.

Version 4 SHEET 1 1436 680 WIRE 992 -192 944 -192 WIRE 1024 -192 992 -192 WIRE 1232 -192 1184 -192 WIRE 1264 -192 1232 -192 WIRE 944 -160 944 -192 WIRE 1184 -160 1184 -192 WIRE 944 -64 944 -80 WIRE 1184 -64 1184 -80 WIRE 128 48 -48 48 WIRE 176 48 128 48 WIRE 336 48 256 48 WIRE -48 64 -48 48 WIRE 224 112 192 112 WIRE 240 112 224 112 WIRE 192 144 192 112 WIRE -48 160 -48 128 WIRE 128 160 128 48 WIRE 160 160 128 160 WIRE 336 176 336 48 WIRE 336 176 224 176 WIRE 416 176 336 176 WIRE 560 176 496 176 WIRE 704 176 624 176 WIRE 784 176 704 176 WIRE 944 176 848 176 WIRE 1008 176 944 176 WIRE 1168 176 1088 176 WIRE 1280 176 1168 176 WIRE 1328 176 1280 176 WIRE 160 192 128 192 WIRE 1328 208 1328 176 WIRE 704 240 704 176 WIRE 944 240 944 176 WIRE 1168 240 1168 176 WIRE 192 256 192 208 WIRE 224 256 192 256 WIRE 240 256 224 256 WIRE 0 320 -48 320 WIRE 128 320 128 192 WIRE 128 320 80 320 WIRE 160 320 128 320 WIRE 336 320 336 176 WIRE 336 320 240 320 WIRE 416 320 336 320 WIRE 448 320 416 320 WIRE -48 352 -48 320 WIRE 944 352 944 304 WIRE 1168 352 1168 304 WIRE 1328 352 1328 288 WIRE 624 368 592 368 WIRE 704 368 704 304 WIRE 704 368 624 368 FLAG -48 160 0 FLAG -48 352 0 FLAG 224 112 +12 FLAG 224 256 -12 FLAG 1184 -64 0 FLAG 944 -64 0 FLAG 992 -192 +12 FLAG 1232 -192 -12 FLAG 1168 352 0 FLAG 1328 352 0 FLAG 624 368 -12 FLAG 944 352 0 FLAG 1280 176 OUT FLAG 416 320 OSC SYMBOL res 272 32 R90 WINDOW 0 -48 52 VBottom 2 WINDOW 3 -41 53 VTop 2 SYMATTR InstName R1 SYMATTR Value 50k SYMBOL res 256 304 R90 WINDOW 0 73 55 VBottom 2 WINDOW 3 84 56 VTop 2 SYMATTR InstName R2 SYMATTR Value 50k SYMBOL res 96 304 R90 WINDOW 0 73 57 VBottom 2 WINDOW 3 82 57 VTop 2 SYMATTR InstName R3 SYMATTR Value 50k SYMBOL cap -64 64 R0 WINDOW 0 59 28 Left 2 WINDOW 3 58 62 Left 2 SYMATTR InstName C1 SYMATTR Value 1n SYMBOL voltage 944 -176 R0 WINDOW 0 58 39 Left 2 WINDOW 3 59 72 Left 2 SYMATTR InstName V1 SYMATTR Value 12 SYMBOL voltage 1184 -176 R0 WINDOW 0 60 44 Left 2 WINDOW 3 56 76 Left 2 SYMATTR InstName V2 SYMATTR Value -12 SYMBOL res 512 160 R90 WINDOW 0 -56 58 VBottom 2 WINDOW 3 -44 59 VTop 2 SYMATTR InstName R4 SYMATTR Value 50 SYMBOL cap 624 160 R90 WINDOW 0 -55 31 VBottom 2 WINDOW 3 -41 32 VTop 2 SYMATTR InstName C2 SYMATTR Value 1µ SYMBOL schottky 688 240 R0 WINDOW 0 -73 14 Left 2 WINDOW 3 -127 51 Left 2 SYMATTR InstName D1 SYMATTR Value BAT46WJ SYMATTR Description Diode SYMATTR Type diode SYMBOL schottky 848 160 R90 WINDOW 0 -70 28 VBottom 2 WINDOW 3 -54 31 VTop 2 SYMATTR InstName D2 SYMATTR Value BAT46WJ SYMATTR Description Diode SYMATTR Type diode SYMBOL cap 1184 304 R180 WINDOW 0 77 59 Left 2 WINDOW 3 70 21 Left 2 SYMATTR InstName C3 SYMATTR Value 10µ SYMBOL res 1312 192 R0 WINDOW 0 -56 48 Left 2 WINDOW 3 -62 87 Left 2 SYMATTR InstName R5 SYMATTR Value 30K SYMBOL Opamps\\LT1884 192 112 R0 WINDOW 0 60 40 Left 2 WINDOW 3 36 87 Left 2 SYMATTR InstName U1 SYMBOL cap 960 304 R180 WINDOW 0 72 52 Left 2 WINDOW 3 67 16 Left 2 SYMATTR InstName C4 SYMATTR Value 10µ SYMBOL res 1104 160 R90 WINDOW 0 -56 58 VBottom 2 WINDOW 3 -44 59 VTop 2 SYMATTR InstName R6 SYMATTR Value 200 TEXT 324 -96 Left 2 !.tran .25 TEXT 144 -184 Left 2 ;PHOTODIODE NEGATIVE CHARGE PUMP TEXT 240 -144 Left 2 ;J LARKIN SEP 15 2013

A push-pull Cockcroft/Walton...

Replace inverters with slow OpAmps.

(Extra doo-dads on the schematic are to convert back and forth between analog and digital domains, so you can "see" voltages rather than ones and zeroes.) ...Jim Thompson

RL