Getting a bipolar power supply from simple 9V dc supply

[Balthazar]

Hi

In this thread, where I ask about how to make a low pass resonant filter...

http://www.diystompboxes.com/smfforum/index.php?topic=100679.0

...I was recommended to make a state-variable filter. I thought that sounded cool, so I searched a bit, and figured TI UAF42 might be a good chip for the purpose, since it has almost everything readily implemented. Anyway, it requires a positive and a negative power supply, in addition to ground. I was thinking of driving it with a simple 9V supply, like a battery. I tried doing this the naive way in SPICE, and it worked, like this:

VBattery VsPlus VsMinus DC 9V
RVs1 VsPlus  0 4k
RVs2 0 VsMinus 4k

That is, I make a voltage divider with two equal resistors, and connect the middle to ground. This worked, but transient analysis showed that it was very unstable. I suppose that might be because the load is not stable. Or something. Anyway, I did my search and found that virtually everyone answering the question of how to make a bipolar +/- supply from a simple dc supply wanted to see the schematic, so I'll do that here. Also, almost everyone recommended using a virtual ground, and get some chip or complex stuff involving opamps to do it. But I though, in a stomp box you do have a "real" ground, so wouldn't it be simpler to use that? So I found one schematic doing it my way here:

http://www.generalguitargadgets.com/pdf/ggg_eq6b_sc.pdf

They do the same as me, only with 10k resistors, and they add one 10uF capacitor in parallel with each of the resistors to stabilise the voltage. Plus a diode, maybe to protect the circuit if someone connect it the wrong way? I tried it in SPICE, it worked and was much more stable. So, what do you think, should I make a virtual ground, or is the solution from generalguitargadgets above good enough for my purpose? I said I'd post a schematic, here it is:



I left the power supply bit open. I don't know much about electronics, so I think there's a lot of errors here. I think perhaps the R5 should come after C1 and surely some other issues, but that doesn't matter yet. You see what I intend to do. So, the only question is really how you recommend I split the voltage to get a bipolar +/- supply.

[R.G.]

http://geofex.com/circuits/circuit_sweepings.htm
+9V and -9V from one battery.

It wouldn't be a bad idea to read all of geofex.

[Balthazar]

Thanks! That was a great idea. The chip is actually spec'd to require at least +/- 6V, but I was encouraged by the many people who had made it work with similar chips at +/- 4.5V. With +/- 9V I have nothing to worry about. Thanks for your recommendation for reading too.

[Seljer]

If you're just going to use batteries and aren't planning on adding a standard DC jack, the simplest solution is to just put 2 batteries in there 

[Balthazar]

I know..... I just want it to be as cool as possible. Cool means: 1) Have a standard DC jack. 2) Don't require two batteries.....

[armdnrdy]

Look to a charge pump. LT1054CN

If you can't figure it out from the data sheet, I'll help you.

[Balthazar]

Thanks. It seems to be the same thing as R.G. recommended. The article he linked to recommends a MAX1044, which is about 1/2 the price of the model you recommend. Do you think I should opt for the higher priced one instead?

[armdnrdy]

I've never used the MAX1044. I use the LT1054CN because it can accept a higher voltage and is rated for a higher output current.
Like the MAX1044 the LT1054CN also has a frequency adjustment feature to keep any audible whine in check.

The MAX1044 has a reputation of failing as well. Let Google be your friend to verify.

For a few dollars more the LT1054 makes more sense!

[Keppy]

The MAX1044 will burn up if you apply more than 10v to it. An unregulated 9v adapter commonly puts out 12v or more unloaded, which would fry the chip. Even a regulated adapter in many cases (Boss PSA for example) is 9.6v, not 9.0v. Even many batteries start above 9.0v, putting you at risk. LT1054 can take up to 15v supply and has a higher output current rating. It's a drop-in replacement for the 1044, so it will work in any layouts that you find. The MAX1044 is fine, you just have to be careful to stay within its limits, while the LT1054 has more wiggle room for pedal purposes.

[Balthazar]

Thanks. I had already googled it.... Looks like LT1054 is the way to go.

[armdnrdy]

The MAX1044 will burn up if you apply more than 10v to it. An unregulated 9v adapter commonly puts out 12v or more unloaded, which would fry the chip. Even a regulated adapter in many cases (Boss PSA for example) is 9.6v, not 9.0v. Even many batteries start above 9.0v, putting you at risk. LT1054 can take up to 15v supply and has a higher output current rating. It's a drop-in replacement for the 1044, so it will work in any layouts that you find. The MAX1044 is fine, you just have to be careful to stay within its limits, while the LT1054 has more wiggle room for pedal purposes.

Yes, Yes, and Yes!
Since we can't really can't control which power supply is going to be used with the effects we build for others, it is a good idea to use a charge pump who's input rating has a bit more "breathing room" than the average wall wart.

[armdnrdy]

Thanks. I had already googled it.... Looks like LT1054 is the way to go.

Good choice!

There is a company Snarling Dogs who made a Wah/ring modulator. The main problem with this unit when it came back to Snarling Dogs was the MAX1044!

[R.G.]

The funny thing about the sensitivity of the MAX1044 to overvoltages and lack of current capability is that I didn't notice it for a long time. I've had a pedal with the 1044 in it in volume production for six or seven years. We've had the normal random failures in that part, but not larger than the failures in things like opamps. I was so paranoid about switching noise getting into the circuit that apparently my decoupling dropped the incoming voltage enough to keep them under the danger voltage level, even with an incoming voltage bumping right into the 10V death limit.

In formal mathematical circles, that design process is know as the Method of Offsetting Errors.

[Tony Forestiere]

In formal mathematical circles, that design process is know as the Method of Offsetting Errors.

icon_lol

I believe you have referred to this design principle as "making it not matter" or some-such.
This is what I use for the CA stuff when I want to use a single battery: http://geofex.com/circuits/Biasnet.htm. (although I usually use two batteries anyway)

[armdnrdy]

In formal mathematical circles, that design process is know as the Method of Offsetting Errors.

That is the design process that I use 50% of the time but I don't know it!

I gave the wrong part number for the charge pump. Well...it wasn't wrong per se...The Linear Technology part will work just fine but is pricier than the Texas Instruments LT1054CP.

A little shopping goes a long way. Here's where I purchase mine from. (For well under a buck!)
http://components.arrow.com/part/detail/938008S7271929N7713

[Balthazar]

I also was alittle worried about the 10vdc rating, and I'm a newbie, so I'll stick with the safer option LT1054. I live in Norway, and I don't find the CP version in the webshop I usually buy electronics. Maybe there are other webshops in my country. Farnell in Norway refuse to ship to individuals. If I can get under a buck, it may be that I'd just order all of it from the US. The CN version is about 10 bucks in the webshop I usually buy things, about the same as the price I'd pay for shipping it from the US, with standard mail.

[Madkatb]

I recently posted about the same subject. http://www.diystompboxes.com/smfforum/index.php?topic=100375.0
I purchased 5 on fleabay with free shipping for just over $3 each. I just checked and the free shipping is good to Norway too. Delivery was just under three weeks.

[Balthazar]

Thanks. Did you mean ebay? I found some other great deals on ebay too, and many ship for free to Norway or very cheaply. This way I can get the component costs down to where they're no concern. Only component quality is a concern then.

[Wounded Paw]

I just tried a transient analysis of the LT1054 as an inverter in LTspice with a 9V supply and output voltage never reaches -9V.  It only gets to about -8.6V which is within the specs for voltage loss in the datasheets.  My question is how would you get a 'perfect' +/-9V supply or even a +/-8.6V.  Would you have to then regulate the +9v supply after the LT1054 to make it exactly +8.6V.  It's probably not a huge concern, just curious.

[armdnrdy]

You are correct!

The negative voltage is always a bit less with a charge pump.

But what happens when you use two batteries to create a bipolar voltage as many older effects did? There was no mention in the user manuals of battery voltage matching. 

And unless the circuits were engineered by NASA scientists.....I believe that the battery drain wouldn't be exactly what you would call even.

I've built effects with the LT1054CP and compared the performance by pulling the charge pump out of it's socket and inserting two 9 volt battery clips tied together in a bipolar voltage fashion....and have experienced no difference.