I spent the past few days trying to figure out what I thought was a very simple thing. First some background…the main circuit for the robot is primarily to be powered by battery. Of course there will be times that the battery has to be recharged. For that I’m using the MCP73863 which has worked before (actually, I’ve only used the MCP73631). The real point I want to discuss is as follows. Each module has its own double li-poly battery pack. All the battery packs connect in parallel effectively pooling their energy. I do this in case one of the modules churns through more current than others. Now, in order to recharge all those batteries, one connects the robot to an AC adapter at a single point, and the power is distributed along the power bus throughout the whole robot. The power coming from the adapter is a few volts higher than the li-poly batteries, and I need to make sure that it doesn’t directly feed into the batteries. Why? Because it has to go through the recharging IC in order to properly regulate the recharging (I actually don’t know what would happen if it did connect directly). There’s a simple solution for such situations which is to use a diode to ensure that the electricity can only flow in one direction. The problem is that using a diode, even a Schottky incurs a voltage drop of at least .4V which in my case may be too much, as I need at least 6.5-7 Volts to run the circuit and my batteries will have a max voltage of 8.4V, and run down to about 7V. So Dropping .4V means that my circuit will stop operating that much sooner, and lose at least 30% of possible run time. This is what it looks like…
In looking for alternatives the best I could find were mentions of mosfets in similar applications, usually for reverse battery protection. Based on that information and through the following helpful conversation on a forum I came up with the following plan.
I’ve decided to go with a Schottky Diode instead of the Mosfet circuit for a few reasons. It’s simpler, requiring just one part. Because of that, it’s also cheaper. It works, ie it does what I want. It doesn’t have the side effects that I thought. The servos I am using are rated to work down to 6.5 v, but when I tested them they actually work well down below 5V, and still funtion all the way down to 3V! This is good for the Schottky because losing .4V is no longer a big concern. Lastly, the mosfet solution only kind of works, and not as well as the Schottky. Using the comparator to switch the mosfet on/off works fine, but there appears to be a few millivolts leaking through (150mV) though at worst it trickles through the mosfets 100k + resistance so any effect on the battery would be negligible I think. Worse though, when the battery power is run through the mosfet with Drain attached to ground (this is opposite the normal switching use) then there is a voltage drop of almost a Volt. This was unexpected, and I think is caused by the diode. I thought that once the switch closed the battery would bypass the diode and there would be no voltage drop, but that does not appear to be the case in my tests. At any rate, the simpler the solution the better, so back to my Schottky search. I’m still glad I went through this test both because it satisfies my curiosity, and also because I learned about comparators and mosfets.
Thanks to the Electro Tech forums for their help, and especially to hero999. I’ve never had such a useful exchange on a forum before. It can be a very helpful resource, more so sometimes than a web site, because the forum can respond to you.
UPDATE: K7elp60 just pointed out the LTC4412. It looks like an IC for power ORing (switching between sources). Thanks for that. I’m going to proceed with the Schottky for now, but if it doesn’t work then I’ll look at that part more.