Hi, I have designed a charge controller that charges a single cell, 3.7V battery via a solar panel. I need an analogue solution that detects when the battery is charged and switches it from being connected to the charge controller to another circuit (the load). When the battery is discharged, say down to 3.2V, it switches back to the charge controller again to be re-charged. I think this may be possible with the use of a zener diode and resistor network but am unsure how to implement this. Would this work?

 

There are two states: Load on, Load off.
The transition from Off to On is at one voltage (3.7V).
The transition from On to Off is at a different voltage (3.2V).

You need a "window comparator" or a "single comparator with hysteresis" driving a switch.

 

Do you have a supply voltage from which to operate the comparator that is higher than the battery voltage being sensed?

If so, what? How well regulated?

 

Your strategy prevents direct powering of the load with the panel, which is the most efficient arrangement. Why not just use the traditional arrangement?: The panel connects the battery via a blocking diode, and the battery directly powers the load.

 

Your strategy prevents direct powering of the load with the panel, which is the most efficient arrangement. Why not just use the traditional arrangement?: The panel connects the battery via a blocking diode, and the battery directly powers the load.

This offers no protection for completely discharging the battery into the load if the sun doesn't shine...

That might be ok, or not?

 

Do you have a supply voltage from which to operate the comparator that is higher than the battery voltage being sensed?

If so, what? How well regulated?

The only other supply voltage is the 9V solar panel and this is bound to fluctuate.

Your strategy prevents direct powering of the load with the panel, which is the most efficient arrangement. Why not just use the traditional arrangement?: The panel connects the battery via a blocking diode, and the battery directly powers the load.

The load is not able to be powered directly from the solar panel due to requiring a much larger power output.

 

This offers no protection for completely discharging the battery into the load if the sun doesn't shine...

Sorry, I was taking for granted that there is still a low-voltage cutout in there somewhere.

 

The only other supply voltage is the 9V solar panel and this is bound to fluctuate.

So you need a voltage reference against which to compare the battery voltage.

The load is not able to be powered directly from the solar panel due to requiring a much larger power output.

Where do you want to switch the load? High side or Low side?

Can you suffer a VCEsat drop across the switch or does it have to be a FET with a lower on resistance?

Here is something similar I did recently for my wife's greenhouse. It is for a 12V SLA battery. In this case, it uses an unregulated wall-wart as a source, but a solar panel would work as well. It turns on a 5A pump when the battery voltage reaches ~14.4V, and runs the pump motor until the battery discharges to ~12.5V. The wall-wart can only supply ~1A, so the pump duty cycle is about 20%.

It has all of the elements of what you need, but the values would have to be rescaled for your voltages.

 

Sweet. So R6 gives the hysteresis?

 

Sweet. So R6 gives the hysteresis?

The voltage divider Rx/(R6+Rx), where Rx is the effective resistance of R1//(R2+R9), sets the amount of hysteresis. I arrived at the appropriate value for R6 by repeated trials using LTSpice.

 

So you need a voltage reference against which to compare the battery voltage.

The charge controller circuit does have a 6V regulated voltage I think should work. Would I be able to use the window comparator as shown on page 12 of this document and include hysteresis as shown on page 7?: http://ww1.microchip.com/downloads/en/DeviceDoc/41215C.pdf