I have been playing around with the soft latch circuit that enables PITS to be turned on / off with a single non-latching switch. I had problems getting the original circuit to work correctly on the breadboard.
Whilst wiring up this alternative circuit, I found out that I had a silly moment and identified the termination of the NPN transistor incorrectly and had swapped over the collector and emitter. In the photo below, I have crossed out the incorrect terminal markings and written in the correct ones.
With the circuit rebuilt with the transistors wired in correctly, the original circuit now functions correctly and will turn both on and off. So I decided to do a comparison between the two circuits, the logical placed seemed to be to start with a comparison of the BOMs. As can be seen in the table below, the original circuit uses 10 components and the alternative 11. However, the alternative circuit can be built slightly cheaper than the original circuit.
For both circuits, I measured operating current in three different states and the operating times for different options of each of the circuits.
It was noted that the alternative circuit drew slightly less current than the original circuit, the alternative circuit also dropped to the minimal current within a second of turning off, where as the original circuit took over 10 seconds for the current to drop from around 5.1uA to the 0.02uA.
The alternative circuit originally took 7 seconds for the circuit voltage to drop down to zero, this time was reduced down to 2 seconds by reducing the 4.7uF capacitor down to 1uF. Although to be honest, for my application, I don't think the slow voltage turn off is any issue at all.
One problem I did have with the alternative circuit, that I have not resolved yet, was that it would not shutdown when the input voltage was less than 3.8V, which could be a problem for me given that I intend to power the circuit from an 18650 cell.
I also looked at modified circuits for the original circuit after reading through the comments on the EEVBlog.
The first modification was to add a resistor between the capacitor and the collector of the the second transistor to reduce the current surge into the transistor base when the switch was closed.
The second modification was to add a 1nF capacitor to the transistor bases from ground. This was done to prevent the circuit from automatically turning on when power is first applied. I personally didn't have an issue with this, but I did also notice that the negative spike during start up was removed when this capacitor was fitted.
Due to the issue surrounding the low battery voltage on the alternative circuit, I will stick to the original soft latch circuit. This leaves me with the PCB to modify to correct the original mistake and add in the two extra components.