+9/-9 VDC supply...?

[petemoore]

  I'll try to synopsize...
  I read that opamps don't care 'where' they are biased, because of the DC blocking caps...as long as they have 'room' to swing the signal...it doesn't matter.
  I tried to replace two batteries with 1 "R divided" 18v PS, to get +9v/-9v [DC]...I used a two 22k bias string between V+ and V-.
  I can't figure why it doesn't work...nothing's blown...it works fine with batteries but CHoWZem....
________________________________________________________________
  I have a CA Reverb, works great with two batteries...one - and one + from each clip are connected and Reference ground [IIUC], then I have +9 and -9 connections...for power supply.
  Starting with a wall wart measureing ~18VDC, and used a 22k string for voltage divider.
  This gives me -9/+9 when I measure with the DMM's clip connected at the 22k 1/2v point
  I clipped the two commons together [right at the battery clips], and connected the new power supply [wall wart instead of batteries], but I'm finding out 'ya can't just do that and have it work...'
  Kinda ofa nube I guess with the power supply dept., I thought I could just whip up a split supply and apply it...
  I decided to test out my reverse polarity protections also, just in case I got the reverse.
  With the power supply connecting in what I think is the 'rightway, I get a mild power up click. When reversing the ends of the 22k's connections to the + and - connections, no click.
  I checked the power supply @ ~9v+/-, when connected to the circuit, double checked that the two battery clip lugs were indeed connected [gnd.] by the circuit board, tried both polarities and NOGO...basically I just twisted a two 22k string together, twisted the WW leads to the string ends, where the 22k's meet got wrapped around the battery clip lug @ Gnd, then I had what was measuring as +9/-9 where I could stick it to the open lugs of the '1/2 connected by offsetting' battery clips.
  I would like to press this Reverb Tank Unit into service this weekend, but batterie consumption is much too high to be reliable or practical.

[petemoore]

  Bumpity Bump

[R.G.]

You have a high resistance ground. Connect a 10uF to 47uF cap in parallel with each of your splitter resistors. I think it will work better.

You do have the "ground" connected to signal ground directly, yes?

[petemoore]

  Here's the schematic:
  http://www.generalguitargadgets.com/diagrams/stage_center_reverb_sc.gif
  The PS Switching looks to be done at the jacks, but ground looks to be the center of Voltage, or between V+ and V-, where these two connect from the batteries.
  I'm working 'outside the box' using batteries powers up the effect.
  I simply tried to mimic what the batteries do [V-, Gnd., V+], using a WW [measuring ~18vdc], into a divider string of 22k/22k.
  I've installed the PS Filtering on the CA Reverb perfboard build, large caps oriented and placed between V- and Ground, V+ and Ground, even V- to V+ IIRC, noise isn't an issue at all.
  I've yet to get the reverb effect to operate using 1 WW into divider...though the measurements I've taken show this as almost exactly the same readings I get when batteries are connected [V-, Gnd/Ov, V+] or  9+ / Gnd /9- [DC readings], the effect isn't functioning on the power supply.
  I'm just connecting to the two battery clips...
  I suspect there is a reason I'm not aware of or something I'm missing with the grounding scheme.

[moosapotamus]

Huh... I didn't think you could get negative voltage out of a resistive divider.

If you put in 18VDC instead of 9VDC, shouldn't you get +18V, +9V and 0V (gnd)? If you want +9V, 0V and -9V, I thought you needed (pos and neg) regulators, or something like the MAX1044. But, I guess I'm mistaken.

Here's a bit that you may have seen before, might be helpful...
http://sound.westhost.com/project43.htm

... live and learn.
~ Charlie

[petemoore]

  Adapting that divider schematic to what I tried would be just reverse the V.R. and Gnd...IOW ground in the middle, where the two 22k's connect.
  >>>Or exactly what is shown in Figure 1 via supplied link.   
  from there V+ and V_ were measured with the DMM, just like the batteries provided AFAIKTell, however the circuit seems more discerning than I in this matter.

[petemoore]

At this point it's all suggestions welcomed...anything except batteries
  Really what I'm asking is:
  for a PS Setup as in the GGG schematic of the CA Reverb Unit...
  " A resistive divider doesn't do a bipolar supply? " 

[soundcollage]

http://www.geofex.com/circuits/+9_to_-9.htm

[petemoore]

http://www.geofex.com/circuits/+9_to_-9.htm

Figure two looks like it might be exactly what I could use easily with the CA Reverb?
  I figured it might work a little bit even with the current differences between V+ and Ground or Ground / V-
  Having not tried the power supply system and applied it to the Reverb Driver circuit...I don't know...I've built stuff before when I GUessed it looks good
  A bit more complex than stringing a couple resistors together, and would require a small board...Do you think this is compatible for use with the CA Reverb?

[R.G.]

What is ground? It's a place we elect to call "zero volts".

Opamps don't care what voltage you bias them to as long as it is acceptably between the most positive and most negative, and has a low AC impedance.

It's that last that I was referring to when I said "parallel in some caps". If the ground has a high AC impedance to both power supplies, it does not act like we expect ground to act.

Huh... I didn't think you could get negative voltage out of a resistive divider.

Sure you can. Just ground the middle. Bingo, positive and negative voltages. The problem is that the "ground" in the middle cannot eat much current without moving around. If you parallel it with caps, the caps "short" the + and - sides to the ground for AC, and things start looking OK. It's the same thing as grounding the positive side of a battery - what you have left is -9V. If you ground the negative side, what you have left is +9V. The battery doesn't care. If you split it with resistors, you get +/- 4.5V. But to get the opamps you connect to it to "respect" it as ground, the ground point has to be able to source and sink any currents and stay reasonably at the same DC voltage. The same issue comes up with a resistor split Vbias, and for the same reasons. I go into this in my article on selecting resistors for Vbias at GEO.

" A resistive divider doesn't do a bipolar supply? "

It does. The two resistors have to be low enough that any DC currents into/out of the junction do not move the "ground" voltage too far towards either the positive or negative side, and there needs to be a BFC (Big Freaking Capacitor) in parallel with the resistors to swamp any AC currents.

And go read http://geofex.com/circuits/Biasnet.htm

[soundcollage]

No reason to stereo jack switch if you are using a power supply. That's to save battery life.

[Unbeliever]

Don't miss the bit at the end of RG's page:
"The LM386 power amplifier chip comes in an 8 pin DIP package, is quite cheap, and self biases to half the power supply."

I use LM386s all the time to 'solve' the 1/2 V bias 'problem'.

[petemoore]

  Ok...that's it then.
  And makes perfect sense.
  I'll work something up and test it.
  I thought I'd get a pop at least, as often as I've used voltage dividers, I see now how ground has become 'formed' between the two resistors and could get 'pulled on' and how that is less stable than 'Ground' ground.

[petemoore]

  I strapped a couple BFC's on the 22k's...
  Perhaps the 22k's are too large
  Or the 470uf caps I used from V- to 'gnd' and Gnd to V+ were too small.
  Whatever the case is, I've bought myself 1 second of lush reverbosity...after power up I get circuit function for exactly a second...lol.
  It has at least improved that much so far, and I can work with it for a while and see if I can't get the PS to be fine with it.

[petemoore]

  I'll have to try bigger caps or smaller divider resistors...
  or see if I fry an Lm386 by supplying 18.? Volts to it..
  AFAICsee the 386 data sheets say 'wide supply range 4-12v or...5-18v...[said something else about 4v, but I don't know what it meant.
  Supply Voltage 15v
  Maximum Voltage 22v
  So max AFAIKTell is 15v or ...22v...or something like that...can't really tell.

[Nasse]

Hope this post of mine does not cause any confusion but anyway... it is another way to skin the cat, simple old rectifier trick that makes + and - supply with center 0 possible if you have some single winding suitable transformer lying around in the junk box. Much similar idea better circuit and nice pcb can be found at generalguitargadgets if I remeber. It works for small load like opamp or few needing bipolar supply (yes, you can not feed a power amp with this).

I was in a real hurry when I did this, so it looks ugly. And the pic quality is low, I think some of them like the hand drawn vero layout is just and just readable. No drilling or cutting of vero tracks needed... The wall wart is some small power 12 V a.c.

http://img.photobucket.com/albums/v411/Nasse/unchecked.jpg
http://img.photobucket.com/albums/v411/Nasse/psu3.jpg
http://img.photobucket.com/albums/v411/Nasse/psu2.jpg
http://img.photobucket.com/albums/v411/Nasse/psu1.jpg

[petemoore]

  I went for a time with 2x 5k6 and a couple RBFC's...3300uf
 That won 15 seconds of fine Reverb.
  :icon_idea:Then...I went with ~ -19vdc / +15vdc, [using Two Wallwarts] and that works but it distorts a bit when the input is boosted...
 Another day doing some fine learning about PS's I'd say.
 I'm considering using a 5v or 6v Zener to bring the 19v down closer to being balanced with the other power supply.
 I thought of trying the 15v WW and dividing with an LM386, but then I'd be at too low a supply voltage. Then there's the WW measureing 18.3v, which is .3v more than the LM386 is rated for...which is a good bit less than <10% over tolerance...
 LOL...
 How close is the voltage power supplies of the same make and model put out...lol.
 The reverb itself only took twice as much messing with and I still haven't really figured out the PS...but I'm having fun trying to figure it out, and the sloppy bias makes a sorta cool distortion distorter..

[petemoore]

  Since it can be done with two adapters, one for + one for -...
  I'm wondering if a dual adapter is available...separate outputs for each supply, fairly closely voltage matched.
  That looks pretty slick Nasse, what AC voltage are you starting with. If it had a schematic and an AC adapter of usable voltage I might try it, it's a bit hard for me to make out what connections you have going on there.

[R.G.]

I went for a time with 2x 5k6 and a couple RBFC's...3300uf
  That won 15 seconds of fine Reverb.

If that is the case, what is happening is that you are pulling a significant amount of DC current out of the "ground". A resistive divider simply can't do that unless it is letting over 10x the DC current flow through it.

Given that, here are your options:
- An active ground splitter: Something like the LM386 chip, but perhaps one of its bigger brothers. National makes a whole series of these, and the LM386 is the little guy. The LM384 is an 8 pin minidip version and will happily exist on 18Vdc and supply up to +/-1A of "ground" current. It self-centers, like the LM386.
- Stacked regulators: you can put two DC regulators in series across the 18V. This will provide the necessary DC currents. You lose the voltage across the regulators, so low drop out regulators are necessary for your marginal 18V raw supply.
- A dual-half-wave rectifier like Nasse is proposing. These work well if your opamps are not sensitive to power supplies t hat rise asymmetrically. Some opamps go gaga if one side of the power rises before the other. Some don't mind at all.
- A true dual DC power supply, of whatever ilk

[moosapotamus]

http://www.geofex.com/circuits/+9_to_-9.htm

Figure two looks like it might be exactly what I could use easily with the CA Reverb?
  I figured it might work a little bit even with the current differences between V+ and Ground or Ground / V-
  Having not tried the power supply system and applied it to the Reverb Driver circuit...I don't know...I've built stuff before when I GUessed it looks good
  A bit more complex than stringing a couple resistors together, and would require a small board...Do you think this is compatible for use with the CA Reverb?

If you are not intending to ever use batteries, go with Fig 1. I'm going to guess that this would work just fine with the CA Reverb. I've used the MAX1044 that way number of times with good results. Works great in my Maestro FSH-1A, for example... and you only need a 9VDC wall wart.

~ Charlie