Author Topic: Need help understanding how an MPPT charge controller works  (Read 4804 times)

0 Members and 1 Guest are viewing this topic.

Atokatim

  • Jr. Member
  • **
  • Posts: 99
Need help understanding how an MPPT charge controller works
« on: August 17, 2014, 10:25:29 PM »
I am interested in building my own MPPT charge controller, but need quite a bit of help understanding each part of an MPPT charge controller.

The first part is solar panels in series creates higher than battery voltages.  From what I have been reading about MPPT chargers is there is a DC to DC converter that steps down this voltage to a certain voltage. 

Does the output voltage of the DC to DC converter ever change unless the input voltage falls below the output voltage? 
What is the difference between using PWM to lower voltage to a predefined state and using a DC to DC converter? 
 - With PWM, you can change the "On time" of the pulse (Period) which in turn increase / decrease amperage, then you can change the distance between pulses (Duty Cycle) to adjust voltage correct?

The second part is the charging the batteries.  I am not going to go into this until I understand the first part.  Most of the time, if I ask 5 questions, only 1 will get answered and then the rest get lost in all the posts and then never answered....

All help is greatly appreciated!

joestue

  • Hero Member
  • *****
  • Posts: 1759
  • Country: 00
Re: Need help understanding how an MPPT charge controller works
« Reply #1 on: August 17, 2014, 10:53:34 PM »
pwm chargers have no ability to store energy in the charger itself. they instead dump the energy stored in the parasitic inductances of the circuitry into zenars or other snubbers.

basically the pwm charger is setup to charge the battery, at the highest current possible, which would be 100% duty cycle. i can't comment further on them because i don't have one, i don't know why you would ever need one, except to limit the power flow coming from the solar cells.. so that you don't have to have a dump load.

mppt chargers are designed to store energy and perform a true voltage conversion to achieve the most amount of power out of the solar panel.
for example, suppose you have a 12v battery, and an 18 volt solar panel.
highest current out of the solar cell is 1.1 amps at 12 volts with a pwm charger, (it would be running at 100% duty cycle.) you get 13.2 watts.
now suppose the highest power out of the solar panel is 16 volts at 1 amp, dumped into a mppt charger you get 12 volts at 1.33 amps into the battery.

that's the difference. the maximum practical increase in power with an mppt charger is about 30% and most of that is weather related. meaning, ideally, you wouldn't need one.
My wife says I'm not just a different colored rubik's cube, i am a rubik's knot in a cage.

Ungrounded Lightning Rod

  • SuperHero Member
  • ******
  • Posts: 2865
Re: Need help understanding how an MPPT charge controller works
« Reply #2 on: August 18, 2014, 03:14:31 AM »
I am interested in building my own MPPT charge controller, but need quite a bit of help understanding each part of an MPPT charge controller.

The first part is solar panels in series creates higher than battery voltages.  From what I have been reading about MPPT chargers is there is a DC to DC converter that steps down this voltage to a certain voltage. 

Does the output voltage of the DC to DC converter ever change unless the input voltage falls below the output voltage? 
What is the difference between using PWM to lower voltage to a predefined state and using a DC to DC converter? 
 - With PWM, you can change the "On time" of the pulse (Period) which in turn increase / decrease amperage, then you can change the distance between pulses (Duty Cycle) to adjust voltage correct?

The second part is the charging the batteries.  I am not going to go into this until I understand the first part.  Most of the time, if I ask 5 questions, only 1 will get answered and then the rest get lost in all the posts and then never answered....

All help is greatly appreciated!

A max power point charge controller works by using a DC-to-DC converter to accept power from the solar panels (or other power source) at a higher voltage than the output load and trades away extra voltage for increased current in the down conversion.  This lets you run the power source at a voltage where it produces substantially more power than it would if its output voltage were held to the charging voltage of your battery pack.  The extra power from running the power source this way is far more than the losses from the DC-to-DC converter, so you get a substantial improvement.  (In some cases a controller is designed so the power source voltage is several times the battery voltage:  As long as you're down converting anyhow, why not reduce your wiring losses with the same mechanism?  B-) )

A solar panel's open-circuit voltage is the number of cells in series times the bandgap (which decreases with higher temperature).   Its short circuit output current is proportional to the light intensity (photons create free electrons).   It output voltage is the open circuit voltage minus resistive losses from current through the wiring - mainly the metalization on the cells (which increases slightly with temperature), and also minus the voltage drop of any series protective diode, as long as the current doesn't approach the short-circuit current.  (If the load tries to pull more, the voltage will drop percipitously until the short circuit current is not exceeded.)

Panels are designed with enough cells that the open circuit is substantially higher than the load battery they're expected to drive.  This is so they will still drive essentially full current into the batteries through the wiring resistance, even if the bandgap voltage is degraded by high panel temperature.  That means there's usually considerable additional power available for an MPPT controller to harvest by running the panel output at a higher voltage.

The output voltage of an MPPT controller is controlled by the charging voltage of the batteries.  There are no options (except for things like limiting the charging rate to prevent battery damage, cutting off charging when they're full, or occasionally continuing to charge them past full, to drive an equalizing charge into them.)

MPPT controllers for permanent magnet wind generators have a complex job to do:  A wind generator's generated voltage (equivalent to a panel's open circuit voltage) varies with RPM.  The current drawn puts a torque load on the shaft, which lowers the RPM, but gradually, so the response is not immediate.  Meanwhile the torque from the blades varies with wind speed, RPM, and yaw direction, and the relationship is not simple, because it's' due to the complex aerodynamics of the blades.  The alternators also have a maximum current above which they will burn out - but must be kept under load (and slowed down into stall by deliberate overload) so they don't overspeed and tear themselves up.

A typical MPPT controller for a windmill cycles the output load voltage slightly, slower than the mill's reaction time, sees which side of the cycle (higher or lower than setpoint voltage) produces more power, and walks the load voltage in that direction.  But it does it within the constraints of:  Limit the battery voltage / amount of charging, limit the current from the mill by overloading it to slow it down (yes that momentarily RAISES the current).

For a solar panel array as the generation, the problem is simpler, because the response is immediate.  (Also:  The panels can be safely shut down by shorting them.  This dissipates the power they would have generated as heat in the panels - but that's what happens if you leave them sitting out in the sun unconnected, too, so it doesn't heat them any hotter and thus damage them.)

Unfortunately, though the solution for doing an MPPT controler for solar is simpler than wind, I don't know exactly what it IS.  B-b

Flux

  • Super Hero Member Plus
  • *******
  • Posts: 6275
Re: Need help understanding how an MPPT charge controller works
« Reply #3 on: August 18, 2014, 10:50:51 AM »
Everything is pretty much covered I can't add much.

For direct charging with a panel across the battery it is very much a compromise. The panel voltage has to be high enough to deal with the worst case with panel hot and battery fully charged. On a cold day with low battery the panel voltage will be way above optimum. The dc converter of the mppt controller adjusts the effective charging voltage for optimum.

The converter ratio will change to give the optimum (mppt definition, maximum power point). It will change mostly with battery voltage but will also change with panel temperature and light level. It can dither to find the optimum point or it can measure the parameters and calculate the optimum ratio both schemes are used.

As long as the panel has excess volts it will sort things out. Panels in series provide a higher voltage and as mentioned you can virtually eliminate the wiring loss by keeping voltage high and current low.

For this to work it needs to be a true energy store dc converter which converts high voltage low current to high current low voltage. A PWM scheme is useless as it only regulates current. PWM charge controllers are purely used as dump controllers to divert excess battery current above what the battery can handle.

This is the basics, I can't go into the detailed circuitry to track the optimum point, there are a lot of methods.

Flux

OperaHouse

  • Hero Member
  • *****
  • Posts: 1308
  • Country: us
Re: Need help understanding how an MPPT charge controller works
« Reply #4 on: August 19, 2014, 12:45:52 PM »
My UNO fridge program that runs everything controls a couple of power point converters.  I call these junk box converters because they are made from old UPS and PC power supplies.  They are designed to use 490Hz switching speeds to make everything simple.  An opto isolator provides isolation and level shifting to drive the FET.  The inductor is just a transformer from a UPS.  A little insulation was cut back so the two 12V windings could be placed in parallel for very low resistance.  This is the first 12V prototype that shows just how crude  it can be made.  Another one is for 36V panel operation and operates at 52V.  This converter is off 95% of the time and the power from those panels goes to heat water at the panels power point.

I keep saying power point because the controller does not track.  It really doesn't need to.  Power point mostly varies with temperature.  During the day it doesn't buy you much.  OK, it hummmmmmmmms a little.  Only because it doesn't know the words.  My wife can't hear it and it is a pretty good diagnostic. The 36V version uses a BIG inverter transformer that must have been from 1,000W version and it has a little core loss.  It only works about an hour a day at low light levels except when it is cloudy and raining. 
« Last Edit: August 19, 2014, 12:51:56 PM by OperaHouse »

Atokatim

  • Jr. Member
  • **
  • Posts: 99
Re: Need help understanding how an MPPT charge controller works
« Reply #5 on: August 20, 2014, 12:29:55 AM »
Thanks for the input everyone!  I am learning about how to design my own MPPT controller with a processor I am quite familiar with.  Problem I have run into is amperage and charging.  I have always been into low voltages 12V and under and less than 500ma.  Now that I am getting into the 20+ amps range with the solar panels I have purchased, I am worried something will go wrong and be catastrophic.

What I am attempting to do is create a circuit that uses a Buck Converter that is controlled via PWM from the processor.  A DC to DC converter already made just seems way too expensive for me unless I am just looking in the wrong places.  Since I would be using a Buck Converter, would this be more efficient than direct PWM to the batteries?  Does anyone have a schematic on a buck converter design (including part numbers of FET's and drivers)?  I am having a hard time grasping how to drive a mosfet at 20khz to 50khz.  I can drive one at 1khz but it is audible and probably not good for electronics around it.

joestue

  • Hero Member
  • *****
  • Posts: 1759
  • Country: 00
Re: Need help understanding how an MPPT charge controller works
« Reply #6 on: August 20, 2014, 02:03:27 AM »
At really slow frequencies like 10 to 50khz it doesn't matter what you use for a mosfet, almost all your losses should be conduction losses at those frequencies. for example you should be able to send 10 amps at 50% duty cycle through an ancient, irfz46 fet soldered to a 2 oz circuit board, without a heatsink (that's about 2 watts of conduction losses)
Another common fet is the 50n06 (60 volts, .022 ohms)
If you can find some broken inverters they are full of ^those and other similar parts.

If you want 500Khz then you'll have to use latest generation stuff.

This is a half bridge driver: http://uk.farnell.com/international-rectifier/ir2011pbf/driver-half-bridge-200v-8dip/dp/1080612
Mouser is cheap. so is Arrow,  both cheaper than digikey.

A part like this http://www.mouser.com/ds/2/427/si1539dl-106033.pdf can be used to provide sufficient amps to drive the fets quickly. Also popular is to use npn/pnp transistors to do the same thing. discrete transistors are also appropriate.
Do a google image search for mosfet driver and you'll find dozens of different ideas.

To successfully operate a buck converter, you have to be able to charge up the boot strap capacitor to power the highside level shifted fet.
The best way to do that is build a proper synchronous buck converter, the low side switch turning on is what enables current to flow through the boot strap diode. some half bridge drivers will turn off the low side switch once current starts flowing in the wrong direction (buck converters are boost converters working backwards) I don't recommend you try and do this because it changes the control loop.
I am a fan of fixed synchronous operation.


As a general note, make everything as low inductance as possible.
The loop formed by the input capacitor (yes, you do need one) and the two mosfets is the most critical. don't connect them with wires 6 inches long.
Also important is the distance between the gate driver and the fets. one inch of wire is enough. don't try and breadboard the gate driver and connect it to the fets a foot away. 2 inches is enough for problems to start to become visible.
« Last Edit: August 20, 2014, 02:11:58 AM by joestue »
My wife says I'm not just a different colored rubik's cube, i am a rubik's knot in a cage.