Hi all,

This is just a quick question really as I tried looking this up but I can't find much info. I have an STU95N2LH5 p-type mosfet in the post to try to make a solar pulse charge circuit to be used on lots of projects. What I am trying to achieve is a 0-19v 10w solar cell will charge a capacitor and when the voltage reaches 18v a zener diode then connects a discharge route via the mosfet then through a regular schottkey diode (for lowest forward voltage drop) to the positive terminal of a 14.8v li-Ion battery pack. This is in opposition to an MPPT setup that from what I can tell does very much the same thing but uses an inductor and a transistor to convert to pulses for the battery. What I'm concerned with is efficiency really, it needs to slurp every bit of juice and charge the battery. Is it realistic to use such a simple setup. I've seen some circuits over at BEAM that are called solar poppers using 2N3904's and 2N3906's but they didnt seem to work right. Then I was reading about the electrical isolation of mosfets and figured perhaps the gate was leaking somehow on the bipolar 2N390X's. I made a post a while back about a similar system but it simply had the diode dump to the battery from the cell direct. In this I'm hoping the lower resistance of the capacitor charging up and dumping to the battery is better than just direct but I may need to be corrected. I've noticed people use capacitors to help batteries but I don't understand if maybe just add extra capacitors to the direct output without the trigger or vice versa is better. Experience welcome, thank you.

 

Do you have a schematic?

 

 

You will not get a very good power usage efficiency with this scheme. Additionally, you may cause the battery to fail.

A solar cell is basically a current source, so it puts out the same voltage up to it's max current, then it tends to stay at about that current until the voltage drops off to zero. When you charge the cap it will basically be a constant current charge with a linear voltage rise from the battery voltage to the 18V limit you set. Note since the voltage isn't as large as it could be you are tossing power away.

Then you dump this 18V charged cap direct to your 14.8V battery. SOMEWHERE there will be some resistance, and that resistance is all that limits the battery charging current (while it too burns off power) as there is a 3.2 volt difference here. So you are again wasting power and dumping uncontrolled slugs of current into the battery.

If you want every last drop of power then you'll have to go with a MPPT scheme. With MPPT the solar cell's output is constantly measured to find the point where the largest current and voltage (actually: power!) may be drawn, then converted to the proper voltage for the battery at the highest current possible. They are basically switching power supplies that convert the cell's maximum output to what the battery needs.

 

The way I intend for it to run is controlled to a fashion and if I'm correct is refereed to as continuous and discontinuous mode when applied to dc dc converter operation theory. In continuous mode the gate is open due to high solar power input being above the trigger voltage and the battery just charges as it would normally. In discontinuous mode it pulses at the frequency depending on how long it takes the source capacitor to reach trigger voltage. The voltage of 18v was selected but now that I check perhaps 16.8v would be better considering I'm using Li-Ion not NiCd like in my last project. I didn't include on the picture but I was going to include a battery overcharge protection circuit too but I have not decided if to use an IC or perhaps even integrate into the projects microcontroller.

Reading about MPPT I still don't feel 100% comfortable. When you attach a battery direct to a solar cell with a diode the solar cell will always slowly creep up the battery voltage till it gets charged and floats high. When you use MPPT you just have the cell wired to an inductor that pulses at whatever speed provides the highest combination of volts x amps. Reading some pages on arduino MPPT there is a constant sampling that must take place constantly changing the power point and testing to see change.

Now I read on some forums that an MPPT tracker is exactly the same as a dc dc converter. I have a nice dc dc converter that was made from two LM modules LM2596S and LM2577S. The converter takes 2.6v -30v and then outputs much the same selectable on the output. Could I just use this and then just pick the perfect battery charge voltage and get MPPT boost during low power and amps boost due to excess voltage being reduced during high power? Or am I missing something?

 

Could I just use this and then just pick the perfect battery charge voltage and get MPPT boost during low power and amps boost due to excess voltage being reduced during high power?

Quite likely, but the devil is in the details. I'd probably go straight to experiments if it was me, taking great care at each step not to get a bad surprise. In other words, use your DMM to measure everything and don't make assumptions. Use current limiting resistors (or light bulbs) until you're more confident about how things are working. Blocking diode on the battery is a good idea, too.

 

Yes blocking diode for sure and I was going to adjust the dc-dc and add the forward voltage difference of the diode so it balances out. But to clarify if I get another dc-dc converter that is say 95%+ efficient (unlike this one which is 86%-90% depending) it should be as good if not better than one of those expensive home MPPT units excluding the amperage capability. I'm mainly just concerned with getting my method and application spot on. So by applying this technique and using a good quality dc-dc on all my solar projects I can say I'm using MPPT with confidence because that is what the dc-dc is doing?

 

I've seen some circuits over at BEAM that are called solar poppers using 2N3904's and 2N3906's but they didnt seem to work right.

They do not use those as the voltage trigger, they usually use a 1381 Voltage trigger to fire the 3904's and 3906's..... but these voltage triggers do not go as high as you need, I believe the highest voltage for these triggers is only 4.9 volts with the 1381U.

but I am sure that if the voltage is fed through the 1381U through some resistors set up as a voltage divider, may just work for what you need (maybe ).....

 

Yes I also looked into these triggers before and came to same conclusion. Do you not think that a reversed zener diode and a mosfet is a good substitute? Compared to the resistor divider network and an out of spec voltage trigger ic? To compare would we perhaps have to find out if the resistor network or the reverse leakage of the diode is worse?

 

...So by applying this technique and using a good quality dc-dc on all my solar projects I can say I'm using MPPT with confidence because that is what the dc-dc is doing?

Nope. You need to monitor both the charging current and the solar cell voltage while controlling the charge current to achieve maximum power usage.

 

What is the best way to control current then? Il add it after the dc-dc before the battery.

 

You may be able to use an LM317 set up as a current controller....https://www.national.com/an/LB/LB-35.pdf

 

Thanks il look into that

 

Ok my mosfet arrived today but it does not work how I thought it would in the circuit I drew on first page.
http://www.hobbytronics.co.uk/datasheets/ST95N2LH5.pdf

I have tested the mosfet and it passes voltage like a switch almost when I bias the gate but the gate stays on till I pull it to ground again. This would mean the gate would only trigger once then stay on till the sun goes down regardless of voltage drops. I'm not sure how to integrate a pull to ground if the voltage drops below a threshold. All I can think of is it forces me to use my Arduino as I know I can pull pins to ground using DigitalWrite function. Problem really though is the Arduino wastes 5v@20mA just for that one function and negates the purpose of building this simple trigger circuit even if I used say an AtTiny85 that uses 3v@10mA. Also on the mosfet i noticed that the voltage being passed is 1v less. With regards to the zener I can't get it to function correctly as intended in circuit. The zener appears to leak enough juice to bias the mosfet before the trigger voltage and leaks much the same way as the BJT's I bought it replace. Is it unreasonable to use a zener to trigger a mosfet in this way?

 

this is how I have my mosfet connected in my battery charge monitor circuit, you could try something similar...

 

As shown, use a pulldown resistor with the highest resistance you can live with. The resistor and the tiny capacitance of the gate are an RC tank. Big R means slow decay and thus slow pulldown of the gate voltage. Little R allows very fast switching but of course adds a load to whatever is driving the gate. Find a good compromise for your application. Be sure you avoid long times at less than fully on, if the MOSFET is passing much current. It'll overheat.