DIckson charge pump?

Rayman

I was looking into how to generate the +/- 10 VDC I need for some opamps...

pedward had mentioned how you can use the MAX232 type chips to borrow some of their voltage.
But, was wondering how they do it...

I think they are using Dickson charge pumps. It's just a series of diodes and capacitors and two out of phase square wave sources.

Anyway, I think with just 2 Prop pins, we can make + and - 10 VDC using a pair of 4-stage Dickson charge pumps.

I think it would be fun to try, even if I don't wind up using it...

Anybody tried this yet with a Propeller? (or think this won't work?)

jmg

The trickier part of the charge pumps, is not generating the square Wave, but the synchronous rectifiers.
Series Schottky's are the simplest passive, but they lose some voltage. With a perfect analog-switch transfer, you still only add Vcc per cap, and you need a pair of SPDT switches, per cap.
So 3v3 to -3v3 and +6v6 is the first direct step, (or -6V6 if you go to +6v6 first) and you can add Caps and more SPDT switches....

A MAX3232 is ~54c, so is not a bad choice, but it is only spec'd to give +/- 5.4V at 1.8mA - above that, you'll likely need an inductor based converter ?

Heater.

I'm sure it will work and have been meaning to try it. I'd be very interested in your results.
There are many voltage doubler/booster circuits around the net a lot using CMOS gates running at 15v to get up to 100 or more volts. Often using gates in parallel to get more current drive into the doubler. Afriend of mine uses such a circuit to get a 100v to drive a tube in a guitar effects amp.

Coming from a Physics background these things look like Cockroft-Walton voltage doublers. Never heard of Dixon before. No idea what the differences may be.

Toby Seckshund

I did have a pump running off of an AVR, to try and get the -Ve for a sinple '232 interface. Even with the extra help of a +5V rail and using Schottky diodes the results were not that good. With the lower 3.3V the diode losses would be much worse, As Heater. said the +15V start is a good way to get around those pesky half volt drops. I think that I had to sort out a diode that had only 0.2V drop to ger it to work.

Where is Germaniun when you need it?

Duane C. Johnson

Take a look at this for a description of the Dickson charge pump.

Charge_Pump_Circuit_Design/45xch01.pdf

Basically the difference between Cockcroft-Walton and Dickson is the CW has the capacitors in series and D has them in parallel. OK, this is a big deal if generating 800_000 volts, not so much for our low voltage applications.

Also, CW can be driven with a single ended driving source and D must be driven differentially. D also needs about twice as many capacitors. Well ya, cuz its differential. But this also means twice the current output.

I experimented with the D circuit using 2N7000 MOSFETs instead of diodes which worked nicely. I was generating 4 times the supply voltage and it regulated better than the diode version.

I haven't tried this with a Prop yet, did with a PIC, but using a single pin to drive the CW multiplier and with a high resistance feedback voltage divider into a second input pin one can actively regulate the output voltage. Should work well.

Duane J

tonyp12

first doubles the input voltage, then inverts the doubled voltage.The input voltage ranges from +1.5V to +6.0V.
http://www.mouser.com/ProductDetail/Maxim-Integrated-Products/MAX865EUA+/?qs=1THa7WoU59EQP5N64Fw0o0PMchOxUj8HvodbJmj6rP8%3d

The MAX680/MAX681 are monolithic, CMOS, dual charge-pump voltage converters that provide ±10V outputs from a +5V input voltage.
http://www.mouser.com/Semiconductors/Power-Management-ICs/Charge-Pumps/_/N-41dteZscv7?P=1z0vmchZ1yzt4fqZ1yzt2xnZ1yzrrcwZ1z0vmc3&Ns=Pricing%7c0