Strange behavior of a simple LME49710 instrument (guitar) buffer circuit

Sam Blaza wrote: "I built a simple buffer circuit using a LME49710. I chose the device for its high input impedance and also because it did not require offset nulling. It sounds good but I encounter a strange behavior when the circuit is disconnected from and then reconnected to the power source (single supply, regulated 9VDC). The circuit does not turn fully on (very low output, lower than unity). It would require for the circuit ground to be disconnected from the power supply ground to "reset" the circuit and enable it to fully power up. Could this be because the op amp is designed for dual-supply and I am using it with a single supply? Is there something I could do to make it work with a single supply? I could post a schematic if that would help."

Sam - I was able to duplicate this issue exactly in what may be a similar circuit to yours. My test circuit is a +6dB non-inverting guitar input designed to use a TL072 BiFET input.

To avoid loading guitar pickups a high value bias resistor, typically 1M, is used to bias the non-inverting input to mid-supply typically around 4.5V. The LME49720 (in your case a '49710 single) is a bipolar input op amp. The LME49720 has a low 10 nA typical (72 nA max)  bias current requirement which is orders of magnitude better than a NE5532 which is 1 uA. For the LME49720 it would appear that a 1M bias resistor would be fine given that at the maximum bias current of 72 nA across a 1M Rbias only 72 mV of error would occur. This may not be the case however.

On power-up, with a 1M Rbias, my guitar input would latch-up using an LME49720 but works properly using a TL072 or MC33178. For it to start-up properly Rbias had to be reduced to ~100K. I suspect, but do not know for sure, that at start-up the bias current requirements for an LME49710 may be significantly more than the 72 nA maximum. The back-to-back input protection diodes from the non-inverting to inverting input may also play a role during start-up.

The LME49710 or 49720 can be run on single supplies using rail-splitters and AC-coupling without issue but it appears that the DC impedance for input bias current should be lower than what the bias current specifications alone suggest. My recommendation would be to use a FET input part such as the TL071 for the high-impedance input. A 100K Rbias load is simply too low for most guitar pickups and the high impedance needed may not provide enough bias current for this particular bipolar input op amp.