controlling or predicting differential

[stevez]

I would like to design a voltage regulator that essentially duplicates the function of the old electro-mechanical regulators but I do not understand how I can predict or control the differential between the steps. I would like to drive two relays with the output - primarily because the users would like it that way (they understand "zero charge" or "half charge rate").

The charge states are "no charge", "half charge" and "full charge" as it describes the amount of charge being applied by the generator.

I would like "no charge" at 7.4 volts, "half charge" at 7.2 volts" and "full charge" at 7.0 volts and less. The problem: as the battery charges it crosses the 7.0 volt point (or 7.2 or 7.4) and if there is not enough differential or hysterisis then the system will be unstable and might just toggle or oscillate back and forth.

I would like to know how I can design something with a predictable (and preferably adjustable) differential in order to avoid this instability. Example: if the voltage is well below 7.0 volts the system will be at full charge and as voltage rises past 7.0 volts the system will switch to half charge. In some cases the voltage may drop ever so slightly on this changeover and I don't want it to go back to full charge - I'd rather it be 6.9 volts. I know there is differential or hysterisis in most systems but I'd like to be able to predict what it is and adjust it somewhat if possible. Does anyone know how to do this??

This seems like a perfect fit for a microcontroller but I have not progressed enough yet with my learning to actually design something. I'd also like to understand how to control or predict the differential for other applications.

Thanks.

 

[Roff]

Use 3 sections of an LM339 quad comparator for your comparators. Connect the inputs of the unused comparator section to ground. The outputs are open-collector, and can drive up to 10ma each, so you might be able to drive a small relay or LED. Otherwise, add transistor buffers to the outputs as needed.
Now comes the good part: Use a CD4053 for threshold voltage selection. The CD4053 is a triple SPDT analog switch. See the attached schematic for typical application. Be sure you pay attention to how you hook up the voltages. If you get them reversed, you won't have hysteresis - you'll have an oscillator.
Hopefully, you can figure out how to design the threshold voltage and hysteresis circuits. If you need help with that, post your request.
Obviously, you'll need 2 of the relay drive circuits and the 3rd one will drive an LED or some other indicator.

 

[stevez]

Ron - thanks. Looks like a good place to start. I could use a simple regulator and resistors to derive the appropriate reference voltages - and pots to "fine tune." I'd have to scale all of the voltages back because battery voltage can dip way down but that's not a big problem. Only worry would be as switch transistions between the two reference voltage sources but one way to find out is build it and see. Actually, the first thing to do is build one and see just how much hysterisis exists "naturally" then go from there but the method you've described allows the differential to be increased significantly.