The problem with dividing the supply voltage is the opamp headroom. Whether you use a LM1458 or TL072, the output voltage swing can be as bad as 3V less than the supply (but usually it's only 1-1.5V less). So with +/-4.5V supply you could be down to +/-1.5V headroom on the output .
Worse than that, FET opamps can flip the phase under certain conditions especially when going negative and will "create unwanted results".
Using a resistor divider will waste some battery life but you can live with that. But opamps like to be kept up in their sweet area and the divider will puts this right on the edge even with a fresh battery.
I'd suggest either finding a rail-to-rail opamp, using two batteries or using a voltage converter (eg: LTC1144 http://www.linear.com/pc/productDetail.jsp?navId=H0,C1,C1003,C1039,C1014,P1377) to maintain a full range.
Best of Luck,
Brian Dunn
Worse than that, FET opamps can flip the phase under certain conditions especially when going negative and will "create unwanted results".
Using a resistor divider will waste some battery life but you can live with that. But opamps like to be kept up in their sweet area and the divider will puts this right on the edge even with a fresh battery.
I'd suggest either finding a rail-to-rail opamp, using two batteries or using a voltage converter (eg: LTC1144 http://www.linear.com/pc/productDetail.jsp?navId=H0,C1,C1003,C1039,C1014,P1377) to maintain a full range.
Best of Luck,
Brian Dunn
--- In cgs_synth@yahoogroups.com, Dave Kendall <davekendall@...> wrote:
> I
> guess I could use a resistor divider with a 9V battery to, generate
> ±4.5V, with the centre tap to GND on the PCB, and something like a 47uF
> cap between the centre tap and 0V on the battery (which is now -4.5V on