Sepic Charger

If you just run a SEPIC into a capacitive load, then it will continually extract power from the input, and that power will have to go somewhere -- "where" being the load. Keep stuffing energy into a cap, and its voltage will have to continue to go up.

I suspect that if you were to put a battery on the output, you'd find that there's always at least some current going into it.

I don't know what component values you're dealing with, but you may want to check what frequency the thing runs at with the original software.

This isn't something that I'd want to just do just by blindly diddling with software. At least simulate it in SPICE.

On a sunny day (Sun, 30 Nov 2014 17:40:55 -0800 (PST)) it happened captoro wrote in :

I see it is using a PIC micro, I use this everyday, also for battery charging:

i added a 500ohm resistor and it works great. i also though of completetly stopping the pwm on the transistor. therefore no output. then back on when the voltage drops below 12v.. any thoughtsnon that?? actually woupd not eve n need to play much with duty cycle. just turn on and off the transistor..

i added a 500ohm resistor and it works great. i also though of completetly stopping the pwm on the transistor. therefore no output. then back on when the voltage drops below 12v.. any thoughtsnon that?? actually woupd not eve n need to play much with duty cycle. just turn on and off the transistor..

i added a 500ohm resistor and it works great. i also though of completetly stopping the pwm on the transistor. therefore no output. then back on when the voltage drops below 12v.. any thoughtsnon that?? actually woupd not eve n need to play much with duty cycle. just turn on and off the transistor..

i added a 500ohm resistor and it works great. i also though of completetly stopping the pwm on the transistor. therefore no output. then back on when the voltage drops below 12v.. any thoughtsnon that?? actually woupd not eve n need to play much with duty cycle. just turn on and off the transistor..

You need to address your stuttering problem... your post appeared on S.E.D FOUR TIMES! ...Jim Thompson

The devil is in the details...all the way back to the original concept.

Your choices might vary considerably given the insolation at the latitude and weather and time of day you deploy your system. Under conditions where you solar panel puts out only 6V, can you get enough power out of it to be useful? Does that outweigh the complexity or additional losses at high insolation?

Operated under such low light conditions, I'd suggest you need a higher voltage panel. A buck converter is far simpler and may give you better performance even with a nominally 12V panel. Depending on your battery technology, you may be talking more about 14.x volts instead of 12V output. Sometimes, the losses incurred by the effort to work at very low voltage outweigh the energy gains.

Ignoring overcharge, undercharge, max/min limits for everything for the moment...add that back to the software later...

For battery charging, all you care about is current. For a buck converter, the duty factor is the ratio of output voltage to input voltage. You can measure both, or just estimate it. Initialize the duty factor to that number and measure the current. Now, increase or decrease the duty factor. If the current goes up, do it again. If the current goes down, go the other way with the duty factor. You dither back and forth at the peak power that you can get, at that instant, from the system. Hall effect current sensors are cheap and low loss.

If your solar array has high enough voltage at the desired insolation target, this method is hard to beat.

We also built a sun-position tracker using the same concept, but the mechanics were considered too unreliable for deployment on a mountain-top inaccessible for most of the year.

tly stopping the pwm on the transistor. therefore no output. then back on w hen the voltage drops below 12v.. any thoughtsnon that?? actually woupd not even need to play much with duty cycle. just turn on and off the transisto r..

Hi Mike,

I am doing tests with my power benchtop power supply, so current supply is no problem. I also want to adapt the same projet to a automobile battery , where the voltage will vary between 9v to 13v. Besides halting the pulsati ng of the boost converter, the voltage keeps increasing in the output cap (no load) no matter what duty cycle I use.

Ken

Is there some reason why you wouldn't put enough solar cells in series so you only needed a buck rather than a sepic or buck/boost?

On a sunny day (Tue, 02 Dec 2014 23:04:25 -0800) it happened miso wrote in :

I dunno, I have a 12V panel that gives almost 20V open circuit. When output drops below 12V there is no power anyway (no light). So you only need to step down the voltage if you need constant 12V or top battery. Current limit is a must in my view, especially to charge some batteries. Those regulators are about 6$ with free shipping on ebay. I have several. I also have some boost converters to then step it up for laptop charging and monitor (19V, 14V). The 14V case may need the buck-boost as a full battery maye be more, but it will probably work (tolerances).

My missive was intended to suggest that, given the disclosed parameters, I think you'd be better served by an alternative approach.

Within the region of interest, a solar cell supplies current. A battery being charged consumes current.

Your test setup supplies voltage and the battery is a voltage. Your test model does not fit your application and will not give you useful insight.

For example, you cannot supply a voltage to a battery. It will explode.

But, when the elements are properly sized, your battery can handle all the current the solar cell can supply...at least for a short time.

If you model your system as a voltage supply and a voltage battery, you must add a bunch of stuff to keep it from exploding. If you start with a current model, much of that is taken care of by the self-limiting inherent in the solar cell.

Back to your problem... A regulated voltage power supply is a linear feedback control system. Ignore, for the moment, that the process is nearly 100% digital. The output is a liner VOLTAGE and the feedback is a linear VOLTAGE.

The feedback loop MUST remain closed at all times. Typical power supplies are one-way devices. They supply current, but cannot sink it. If the output voltage exceeds the setpoint, the loop becomes disconnected and the output is out of control.

The schematic discloses a small fraction of the "real" circuit. There are parasitic elements inside the devices and in the interconnect system. If any of those parasitic elements causes current to be dumped into the output cap, the voltage will rise.

That's why VOLTAGE supplies need a minimum load current to guarantee stable operation. Sometimes, that's included inside the power supply box. Sometimes it is an external requirement.

If your power supply works correctly with a load, you're getting what you paid for...but maybe you bought the wrong supply for your application.

So, Think CURRENT. Determine how much voltage is required at the input to supply charge to your battery at it's maximum voltage. Call that Vin.

Decide how low the light/insolation can get and still charge your battery. Select a solar cell that has output voltage > Vin at that level of insolation and has adequate size to provide the energy you require under worst case weather conditions. If you want maximum conversion efficiency, build a MPPT buck converter. If you want more efficiency, build a sun tracker. Pick a battery size with sufficient capacity to outlast the worst case cloudy period. Double everything. Calculate the cost. Faint. Wake up and get a new hobby ;-)

Solar power is fun to play with, but it rarely makes economic sense if you have ANY other alternatives. There are some interesting seeming exceptions if you think about conservation and function instead of power production. With low power electronics, it becomes feasible to make stuff like solar-powered school-zone radar speed displays. The cost of the solar array, while high, is CHEAP compared to the cost of digging up the street to access a utility pole 100 feet away. It's "no alternative" on a smaller scale.

tetly stopping the pwm on the transistor. therefore no output. then back on when the voltage drops below 12v.. any thoughtsnon that?? actually woupd n ot even need to play much with duty cycle. just turn on and off the transis tor..

is no problem. I also want to adapt the same projet to a automobile batte ry , where the voltage will vary between 9v to 13v. Besides halting the pul sating of the boost converter, the voltage keeps increasing in the output c ap (no load) no matter what duty cycle I use.

HI, I think you answered my question when you mention that a small load is needed to keep it stable. Most euqation for buck boost do not take into con sideration that load resistor. I think this is a very important part. Also I was trying to figure out the internal working of DC-DC chip from Linear, but the datasheet does not say enough, but I think there is a hidden load r esistance. thank for your help

ken