New layout: adapted RG Keen's charge pump

[Valoosj]

As I need an 18V supply for some pedals and I don't feel like purchasing a new adapter, I decided to adapt a charge pump by RG Keen.



This thing charges up to about 25V and then it goes through a regulator which brings it back to 18V. I hope this gives a nice clean 18V without any issues. When using the 78L18 you should be able to get 1A out of it. At least, that's what Auke Haarsma told me 
I added the ICL7660 as an option for the IC since I killed my first MAX1044 after about 5 seconds in a different project.

I'm not 100% sure of the way I added C1 and C3, but I think they are correct.

Oh, this is where I got the original schematic from:
http://www.geofex.com/  I can't give a direct link, so do a search for 'voltage' and then in the results choose 'New page 1'.
There is some other stuff mentioned there, but I can't quite figure out the use of that transistor switching. Is it necessary when you always use this charge pump? 

[gigimarga]

It's possible to use this circuit to get 24V/100mA?

[snap]

This thing charges up to about 25V and then it goes through a regulator which brings it back to 18V. I hope this gives a nice clean 18V without any issues. When using the 78L18 you should be able to get 1A out of it. At least, that's what Auke Haarsma told me 

the L versions of  those regulators go to 100mA max!

[Valoosj]

This thing charges up to about 25V and then it goes through a regulator which brings it back to 18V. I hope this gives a nice clean 18V without any issues. When using the 78L18 you should be able to get 1A out of it. At least, that's what Auke Haarsma told me 

the L versions of  those regulators go to 100mA max!

Auke said that the 7818 gives 100mA max and the 78L18 gives 1A max. Maybe he's wrong 

Something else that I just noticed is that the diodes can't cope with 25V... 1N5817 Schottky-Diode 20V, 1A.
Maybe it should be all 1N4001 then.

[Valoosj]

Kurtlives just mentioned to me that I need some filtering, so I'll add a 330uF from 18V to ground and from 9V to ground.
I'll post my findings when I finish making this sucker. 

[zyxwyvu]

Auke said that the 7818 gives 100mA max and the 78L18 gives 1A max. Maybe he's wrong 

The 7818 can do 1A, the 78L18 can do 100mA.

Something else that I just noticed is that the diodes can't cope with 25V... 1N5817 Schottky-Diode 20V, 1A.
Maybe it should be all 1N4001 then.

You can also try the 1N5819, which is a higher-rated (40V vs. 20V) 1N5817.


If you're looking to get a significant amount of current, you should look into the LT1054. The MAX1044 is only rated for 10mA, while the LT1054 is a pin-compatible replacement that can do 100mA. The only difference is you do not nee to connect pins 1 and 8 (the LT1054 normally runs at a high enough frequency to avoid problems, and that pin has a different function on it).

[Auke Haarsma]

Auke said that the 7818 gives 100mA max and the 78L18 gives 1A max. Maybe he's wrong 

The 7818 can do 1A, the 78L18 can do 100mA.

This is true. I must have mixed them up Yorick. Apologies

[Valoosj]

Auke said that the 7818 gives 100mA max and the 78L18 gives 1A max. Maybe he's wrong 

The 7818 can do 1A, the 78L18 can do 100mA.

Something else that I just noticed is that the diodes can't cope with 25V... 1N5817 Schottky-Diode 20V, 1A.
Maybe it should be all 1N4001 then.

You can also try the 1N5819, which is a higher-rated (40V vs. 20V) 1N5817.


If you're looking to get a significant amount of current, you should look into the LT1054. The MAX1044 is only rated for 10mA, while the LT1054 is a pin-compatible replacement that can do 100mA. The only difference is you do not nee to connect pins 1 and 8 (the LT1054 normally runs at a high enough frequency to avoid problems, and that pin has a different function on it).

And what about the ICL 7660?

[Auke Haarsma]

did you check the datasheet?

[Mark Hammer]

Q:  We've had discussions about heterodyning of two clock signals.  Does increassing the value of the step-wise storage capacitors in this type of voltage multiplier circuit reduce the "clockiness" of the doubled/tripled supply voltage, or is it simply sufficient to make the final storage cap at the very end of the chain (where V+ will be taken from) bigger?

[earthtonesaudio]

Good question Mark!  I'd like to know as well... 

A few things I do know that you should keep in mind when powering circuits with voltage multipliers:

-Ripple increases with increased number of stages (that might be the "clockiness" thing).
-You don't get a free lunch, so as voltage goes up, current goes down (plus losses through diodes and caps).
-The output is high-impedance so it sags a lot under load (but your regulator might take care of that).

But then again, most stompbox circuits are low power so some of these considerations might not matter much, if at all.

For more power you can use the oscillator in that chip to drive a relatively simple switcher using a few other parts.