Yes an old analog.
I also dug out this ancient RS module, 2 AA cells to 9v. Chip is BP51C. Only gives 9.09v out with 1.2k load with 48mA input at 2.5v, so 58% efficient. However, I want to power LT1115 so a resistor divider for the virtual earth 0v rail isn't going to work. Hence needing a true + - supply.
I could scope this RS module out and use a capacitor and diodes to provide a negative rail. Possibly.

 

The module is from a Texas Instruments calculator. The chip was available as a part. I can't download the pdf on my mobile, so a screenshot.
With one cell, the positive output is not regulated anyway. Going to try taking a capacitor output from SW/feedback pin of chip and try two diodes to get a negative rail. As in the discrete circuit.
Does the chip mask give any clues to the discrete circuit issues?

 

Yes an old analog.
I also dug out this ancient RS module, 2 AA cells to 9v. Chip is BP51C. View attachment 348873Only gives 9.09v out with 1.2k load with 48mA input at 2.5v, so 58% efficient. However, I want to power LT1115 so a resistor divider for the virtual earth 0v rail isn't going to work. Hence needing a true + - supply.
I could scope this RS module out and use a capacitor and diodes to provide a negative rail. Possibly.

My results on trying a feed from the coil on the Texas module.
The pulse width driving waveform has a frequency of about 10kHz. Used two 10uF back to back to centre of diodes, as in original posted discrete circuit. With 100uF smoothing capacitor, ripple with no load is 2.5mV. With 1.2k load, ripple is 80mV. Voltage maintained at about 9v. Module positive output also has 1.2k load.
I am going to need some more smoothing for use with a high gain circuit based on LT1115.
Positive and negative fixed 5v linear regulators?


I am still at a bit of a loss as to why, or even how, the discrete circuit is supposed to oscillate. What sets the frequency, coil self-capacitance?