Reliable way to get +15vdc from 9vdc supply

[LyleCaldwell]

I've searched and searched, and due to the nature of the forum, posts seem to be about problems people encounter, but there's usually no follow-through.  Mark and RG post suggestions and the original poster never says whether the suggestions work.

So, I know great ways to get 15vdc from a 9v or up AC supply, but I want to know if there's a bullet proof way to get it from a 9vdc supply.  I'm looking at RG's charge pump design on geofex, and from searching here, when people encounter noise, RG has suggested filter caps and such but I haven't seen posts where people said "yes that worked thanks."

So, does it work?  If I have whine, will the proper filter caps plus connecting pin 1 on the IC let me eliminate it?  I don't want to make or order PCBs for a supply that ultimately won't work.

Of course, I'm inclined to believe that anything RG says works does, in fact, work.  But I know that any design might not work in every application, and that some circuits are more problematic with a charge pump.  The circuit I'm making uses two opamps (TL072 and 4558) and a few LEDs, but no LFO or freq sweeps.

[LyleCaldwell]

Will this work?  I'm concerned about putting the -9vdc to ground.  Will the 7815 "see" 18v coming in based on the swing between +/- 9v?

[Processaurus]

You may be aware, but the MAX1044 can only supply 10mA reliably, I've burned one out before.  I don't know how much your regulator would consume, and if even if its less than that, if there would be any current left over for the pedal.  If your circuits draw is less than 10mA, how about getting rid of the whole regulator thing, and just put a few rectifier diodes in series with the power running into the charge pump?  If you drop the supply voltage to the charge pump down to about 7.5v with the diodes, it doubles that and you have ~15v.  You could use an LED too, probably, and have a free power indicator.

There are other charge pumps with greater current capabilities, I think there was a thread by Nelson recently about them.

Whatcha making?

[R.G.]

There are a few things that you need to understand about charge pump multipliers.

First, the circuit actually makes twice the voltage you feed it, less a diode drop. In actual operation, with 9.0Vdc at its input, a 1044 gives you about 17Vdc when unloaded. That's just barely enough to run a 7815 and get regulated 15V. If you power it from a battery and the battery is less than 9V, then you get less than 17V out, and the 15V regulator won't regulate, but will just subtract about 2V from whatever voltage it gets. So if the "17V" sags to 16V, the "15v" output drops to 14.

Next, the thing can't supply an unlimited amount of power. The output of the 1044 charge pump looks like there is a roughly 80 to 100 ohm resistor in series with it. So if your circuit pulls 10ma, the voltage out of the charge pump can sag by up to 1V. There are solutions to that, like using more 1044's in parallel, but it makes things more complicated.

Next, the circuit idea you have is good, but the circuit has a mistake that won't let it work. You've shown the circuit as a voltage inverter, making -9V from +9V. The -9V output is on pin 5. But you have both the -9V output and pin 3 (chip ground pin) labelled "Gnd", and if you hook those together, it shorts the output of the chip. So you have to figure out what you are going to connect to ground in the circuit you're powering. It will be either the 15V regulator ground or the pin3 ground, but it can't be both.

If I have whine, will the proper filter caps plus connecting pin 1 on the IC let me eliminate it?

Whine in converters comes from either poor filtering or poor grounding. First of all, using a Max1044 mostly eliminates that, as with pin 1 and 8 connected, the frequency of the oscillator is bumped above audio. Your circuit may in fact be whining, but you won't hear it. Second, you must never, ever let the pulsed currents drawn by the converter share a ground or power supply wire with your circuit. If you do, the resistance of the shared wire can cause whine (if the whine is audible, which it's not for the 1044). Wired properly and with the correction noted above, it should not whine.

I don't want to make or order PCBs for a supply that ultimately won't work.

That's a very good approach. There's only one way to ensure that - get your circuit working as you like it **before** you lay out or order PCBs. No PCB should ever be the first trial of a circuit like this. Get it running on the breadboard first. Then your PCB has a chance.

The circuit I'm making uses two opamps (TL072 and 4558) and a few LEDs, but no LFO or freq sweeps.

The "few LEDs" bother me. An LED is often 5-10ma all by itself. The 1044 approach will probably be fine for the opamps, but a few LEDs could push it ove where it can supply enough current.

[LyleCaldwell]

Thanks.  I guess I'll just have to either spec an 18vdc supply (because it's easier to get than a 15v and the PP2 can supply 18v) or spec a 9vac supply and use my tested voltage doubler circuit.  I know this will draw more than 10ma.

[theman]

I haven't tested this, but if it's in an app note from TI, I think it should work!

http://focus.ti.com/lit/ds/symlink/lt1054.pdf

Look at Figure 22: Positive Voltage Doubler.

The LT1054 can handle 100 mA.

[Paul Perry (Frostwave)]

Here's a problem I've seen with 9v input voltage doublers: some doublers have a MAXIMUM input of 10V. Now, maybe there won't be a problem with a 9v battery (except the other way around, when the 9 droops to 6.5!).
But, if you are runing plugpacks into the input, who knows what is going to happen?? There are plenty of nominally 9v plugpacks that when lightly loaded might give you 11 or more volts.

[nelson]

I haven't tested this, but if it's in an app note from TI, I think it should work!

http://focus.ti.com/lit/ds/symlink/lt1054.pdf

Look at Figure 22: Positive Voltage Doubler.

The LT1054 can handle 100 mA.

http://www.elixant.com/~stompbox/smfforum/index.php?topic=41005.0

I have a few LT1054, I have never tested them though.

I drew up the PCB based on the app note, but the reply in that thread indicates that the voltage drops.

Any solutions?

[LyleCaldwell]

My voltage doubler is based on the Rane power supply, with 9vac doubled, then rectified, then regulated.  No dropping issues, no noise, but it requires  AC.