Improvements on the "DicksonChargePump"

[steveschnepp]

I wondered if we could simplify the Dickson charge pump to only use 1 clock without the need to invert it.
And it seems it works indeed, as you can connect intermediate capacitors to ground instead of the inverted phase.

It does charges much slower, but it works, and will save a pin on the Arduino.

Access to an interactive simulation.

[tardate]

neat idea @steveschnepp - essentially a "half-wave rectifier" if I'm following properly. Nice compromise - twice the cycles to charge but save a pin.

[steveschnepp]

I'm searching a way to have 0/5/12V from 1 pin. something like

pin	rst
0V	0V
5V	5V
50kHz	12V

[tardate]

@steveschnepp the closest I think I could get is using 2 pins (similar scheme for the 8051 programmer). PS: added a new project for the "half-wave" approach, see https://github.com/tardate/LittleArduinoProjects/tree/master/Electronics101/Power/DicksonChargePump/HalfWave

[steveschnepp]

@tardate, don't stop a fool from doing impossible things! And huge kudos for the HalfWave one as a new project.

Here is the circuit that does exactly this, with only 1 pin, guess my intuition was right.

Notes:

• The input is a little bit changed, to be able to be able to emulate nicely "0V/5V/50kHz", but in real life, you would only use one GPIO.
• I cheated to keep the number of parts low, but it remains within the datasheet specs for the RESET line.
• I used a 50k resistor to emulate the load, as the AT89C2051 datasheet says that the pull-down resistor of the RESET pin is between 50k and 300k.
• I used simple to source numbers for every other parts. Only the 20nF capacitor isn't, but nothing that 2 of 10nF can't do.

Access to an interactive simulation.

RESET == LOW

At 0V, we have only 60 mV, but it should be low enough to be seen as LOW

RESET == VCC

At 5V, we have only 4.3V, but it should be close enough to Vcc to be seen as HIGH

RESET == HIGH VOLTAGE

At almost 12V, thanks to the Zener diode.

[tardate]

isn't that still 2 pins @steveschnepp ? (one for CLK, and one for switching +5V?)

[steveschnepp]

No @tardate, the 2 pins with 5V is VCC rail, which doesn't count like a pin.

And the 1K/10K resistor array with 5/0 is just for the ease of the simulation. In reality it is the CLK pin that will be 0, 5 or CLK, without any resistors. I added those to ensure no short circuits when playing around.

As the PCM pin that you used for CLK can also be set at a solid 0V or 5V DC 😉

I'll craft some updates to the 8051 programmer if I find some time.

[tardate]

@steveschnepp ok got it - that's very clever;-)

I'll have to throw that on a breadboard!

[steveschnepp]

My goal is to have a vusb atmega328 driving the programming. Thus not requiring a full blown Uno board.

My targets for programming are selfish (i only have those) :

• atmega328, with High Voltage erase
• at89cx051
• at24cXXX EEPROM

Ideally, leveraging a slightly patched avrdude on the host side.

Having the patch upstreamed into avrdude, and the perfboard/pcb mercilessly copied by chinese resellers are my ultimate goals 😉

As, you can always measure your success in how many chinese copycats you have on ebay/aliexpress/...

[steveschnepp]

Also, using a Schottky is only useful for the one that is between the 2K2 and the 10K resistors.
As the 12V doesn't need to be raised beyond, but it does raise the 4.3V to 4.5V.

That said, I don't think it will make a huge difference, as the datasheet says "> 0.7 Vcc (3.5V)" is enough for a HIGH.