Instead of charging with a plugpack, I am using a 12 volt, 350ma watt solar panel (it needs to be this small due to physical constraints). To set the rate of charge of the batteries I can change the value of RProg from 10k for 130ma charge rate, to 1.1k for 1000ma. I would like to charge at the highest rate possible however this is dependent on the amount of sunlight.
If I use a resistor to give me a high charge current, the MCP73213 will try to draw too much current from the solar panel in low light periods (morning, afternoon and cloudy days) forcing the solar panel voltage down, thereby shutting down the MCP73213.
If I use a resistor that gives me a low charge current, I will be able to charge in a wider range of light conditions but I will miss out on fully utilizing the full charging potential of the solar panel around midday.
I would like to put a transistor (PN100?) in place of the RProg resistor to regulate the current, with the base somehow connecting to the output of the solar panel. This way, if the voltage from the solar panel started dropping off (say below 10 volts) due to excessive demand by the charger, the transistor could ease off the charging current. If the sunlight strengthened, The transistor would up the charge current to make the most of the solar panel potential.
However, I have searched for material on what values of resistors (and configuration of such) would be required to connect the transistor up in such a fashion but can't pull it together.
I am thinking the Rprog resistor is part of a 1) voltage divider or part of a 2) current follower circuit. In either case a Rprog could be replaced by a transistor but, you will need to find current through Rprog and the voltage across Rprog at both minimum and maximum current settings. Once this is known you can begin to size your transistor components.
If the Rprog resistor is part of a voltage divider. As the Rprog resistor gets smaller "voltage across Rprog gets lower" the current to the battery goes up. If you look at the internal diagram you will see they are using op-amps to compare voltage of Rprog against a reference of 1.21 volts. You will have to use a voltmeter to test voltage levels across Rprog. Once those levels are known you may be able to create a variable current input "from your solar panel" to Rprog to regulate the charge current.
The Rprog resistor is part a current follower circuit and as Rprog resistor gets smaller, the current level increases through Rprog and also to the battery as a response from the chip.
I would like to put a transistor (PN100?) in place of the RProg resistor to regulate the current, with the base somehow connecting to the output of the solar panel. This way, if the voltage from the solar panel started dropping off (say below 10 volts) due to excessive demand by the charger, the transistor could ease off the charging current. If the sunlight strengthened, The transistor would up the charge current to make the most of the solar panel potential.
The major snag I see with this is the "law" of the regulation. That means how the regulator responds to the stimulus (solar voltage size). It is this law that you will want to linearise, and that can be tricky with desecrate components, normally it involves op amps and feedback diodes. Often it is better to use a micro controller and a look up table to do the control.
However, until you know the relationship between input voltage and charge you can't begin to linearise it.
Therefore I would recommend that you do a series of measurements with an input voltage to your transistor increasing in small steps and measuring the charge rate. I would also put a potential divider in front of the transistor to give you more voltage range between fully off and fully on.
Finally did some testing as requested by cylcegadget. I ran the charge current at the upper and lower limits at which I would charge the LiPos given the solar panel that I have.
Test 1 Test 2
Charge current thru LiPos 105 ma 400 ma
Voltage across RProg resistor 6.3 volts 1.00 volts
Current through RProg resistor 0.07 ma 0.34 ma
RProg value 12.3 Kohm 2.925 Kohm
Given the results above, would a transistor setup still be the way to go?
Your data you have is definitely going to be helpful. Reading Grumpy_Mike's post you may have some complications but, I think you can try a transistor in place of the RProg resistor to start testing.
The transistor will connect as: PROG pin --- transistor collector transistor emitter ---GND Solar panel + ----- base resistor ---- transistor base.
The pn100 transistor is hFE "gain" of 100. You need up to 0.34ma from collector to emitter to get max charge rate. 0.34A base to emitter /100hFE = .0034ma base current required.
Assuming your solar panel produces 12v max "you should check this" (12v - .7v diode drop) /0.0034ma base current = 3324 resistor from solar panel to base of the transistor.
You should probably start with a resistor a little bigger than the estimated 3.3k to make your tests safer.
Keep in mind, this is a theory based on the data given. In a real life application it may not be so simple. You will have to test it and see what the results are.
I was thinking along the same lines but wasn't sure.
I had though I would have to put a voltage divider across the solar panel terminals with the base resistor (of transistor) feeding off the 'mid point' of the two resistors.
You don't think that this would be of any use?
Also, without getting too esoteric, would another transistor be better suited. Should I be looking at a lower gain?
I chose the PN100 because they are common and cheap (and I have a few).
LiPo charging is a two fold regulation, first is constant current based on the C rating of the cells in question with a terminal voltage contraint imposed over the top of that of Typically 4.4V per cell. Input source current and/or voltage limitations may put you below these values.
If you use a voltage divider, you can use a smaller resistor on the base of your transistor to get the same base current. It would take some thinking as to what resistors to choose because it would create a series/parallel circuit. It might be of good use but, I am not sure of the advantages/disadvantages.
As far as your transistor, if you had less hFE "gain" it would make your circuit easier to tune and maybe if noise becomes a problem, it would reduce that issue also. I can say that with higher gain, you will waste less current through the base to emitter so, you will be more efficient.
If Grumpy_Mike has any suggestions about transistors or voltage dividers, I would follow his advice as he probably has the most experience.