1.)
Use one battery and a "rail splitter device".
The "rail splitter" could be:
a.)
A resistor divider bypassed by caps.
b.)
A special device like the TLE2624
c.)
A power buffer with resistor divider at the input
d.)
Same like c.) but made of a power opamp with gain set = 1
e.)
???

2.)
Use two batteries in series. The junction of pluspole / minuspole
in the middle forms the ground. The left-over pluspole is +ve
the left-over minuspole is -ve. This is a real dual supply.
Make sure to use a DPST to switch both -ve and +ve in a real
circuit.

3.)
Use a wall wart supply with a power transformer that has its secondaries
center tapped.
The center tap is the ground, the outer transformer lugs
go to the rectifier.

This typically happens when you accidentally leave a headphone amp on for a few days. If you left the headphones plugged into it, you're likely to find them fried when you come back.

Now, to be fair, alkalines become useless below about 0.9 or 0.8V per cell. You wouldn't allow a several-cell battery to get below 1V while listening -- the amp would be sounding bad well before it hit that point. For instance, a 9V battery becomes useless by about 5V, so it would take serious neglect for it to reach the critical ~1V area. But, what happens if you fall asleep with headphones playing and only a few hours of battery life left? Or any number of other ways for this to happen.

Virtual ground circuits fix this problem because they always split the battery pack evenly. I suppose bad things could still happen with a very low battery supply, but at least with a vground circuit you're trying to avoid the problem.

dual batteries using the low impedance ground from between the batteries as system ground

active splitter like the tle2624 in parallel, output connected to low impedance battery ground ( thru a low ohm resistor to limit tle current draw to <15mA ) tle will load the higher V battery to pull up its virtual ground output and the extra drain will discharge the high battery to balance the +/- batteries V

tle should be able to balance battery V as along as the amp circuit doesn't have too unequal +/- ps load - all op amp circuits will be ok

if your operating scenario includes letting the circuit totally discharge the batteries, maybe a low voltage cutoff circuit is a better idea, most batteries last more cycles without deep discharges

I am planning to build a cmoy altoids tin amp for my Iriver to use with my new Allesandro MS1's and im debating if I should go with the powersupply that is shown in the schematic or build myself a buf643 buffered powersupply. Basically what im trying to understand is, is a buffered ground basically designed to prevent DC offset when batteries get to the point where the offset goes so high that it burns the headphones out?

If I dont let amp die on me, (I never fall asleep with headphones on and I wouldnt use it when the quality is bad) do I still need to build a buffered virtual ground power supply? Would it be smart to build a buf643 powersupply eitherway? its a little extra money but is it really worth it if none of the above happens.

[Edited by slibus on 05-31-2005 at 09:24 PM.]

A buffered ground gives you an amount of safety that the
dual-battery-center-connected ps cannot deliver.

If you are very sure that no-one ever "abuses" your amp
and you have a proper battery treatment then you will like
the connected battery scenario.

Additionally you could build a simple voltage monitor consisting of two LEDS
with each series resistor high enough to let the LEDs glow dim at
full charge. Modern low current leds won´t consume much power
and their characteristics show a "sharp enough" cutoff of the
forward voltage. If the leds are both off it´s time to change the
batteries.
Or use a cheap voltage meter with a double pole switch (maybe center off)
to check the voltage frequently. This looks cool. But it probably
won´t fit into the Altoids tin.

I've just finished a 3 channel version of Sijosae's design for the 74HCU04 and it sounds as good as the A47 but using only 2 AA alkaline batteries. Certainly more usable in my opinion as a portable since AA batteries last longer and are easily available.

Do you think I should change the 270UF caps to 470UF in my altoid amp as well?

slibus:
I found bass response is a bit better with bigger caps. My CMOY
works with a rail splitter 1000uF and 4.7kOhm.
If you want something overkill (and good sounding):
Reduce the rail splitter resistor to each 1kOhm and enlarge the
cap to 1000uF. Put a 0.1uF film cap between each opamp supply pin
and ground as close as possible (!) to the chip.
I´m sure you will be surprised !
(Ok. I know it´s hard to put bigger caps and those extra parts into the Altoids box.)

[Edited by heady on 06-02-2005 at 03:45 AM.]

Question: Where is battery life greater, 2 batteries and resistor divider or 1 battery and a buffered ground?

Also, how long would these playtimes be?

---

Edit: I have .1uF caps on the input(big beasts, had to relocate them to fit), they arents as close as they could be to the inputs but they work. Also I have two caps 220 a peice, so far they work exceptionaly. Ill have to think about dropping in larger ones. I found a guy online that sent me a pair of op227 on a brown dog dip adapter, so im gonna give those a go today and see whats up. But so far I am very very satisifed and the amp works great with my Alessando MS1's.

[Edited by slibus on 06-10-2005 at 07:31 PM.]