MPPT Charge Controller for LifePO4 Batteries

[T1 Terry]

It's one of those things that can be hard to test. A simple switch between MPPT and PWM controllers and reading the amps going into the battery will only work if the solar panels remain in parallel and the STC Vmp is between 16v and 18v. this does not suit an MPPT controller, they prefer a higher solar input voltage Simply taking the voltage and amps at one side of an MPPT controller and comparing it to the voltage and amps the other side will give a distorted picture of what is really available from the solar panels. The MPPT controller will hold the solar input voltage at a point that suits its algorithms best, not what the max current flow could have been if it was directly connected to the battery via a PWM controller.

T1 Terry

[noelex 77]

Originally Posted by T1 Terry

Simply taking the voltage and amps at one side of an MPPT controller and comparing it to the voltage and amps the other side will give a distorted picture of what is really available from the solar panels. The MPPT controller will hold the solar input voltage at a point that suits its algorithms best, not what the max current flow could have been if it was directly connected to the battery via a PWM controller.

Using the display that is on some MPPT controllers listing the current in and out is a very poor way of estimating the advantage of the MPPT conversion. The displays exaggerate the difference. Which is deceptive.
If you measure the true values you get a much better indication. As you point out as you reduce the input voltage the output current will rise slightly and this effect should be allowed for if you want an accurate estimation of the gain.
However, the current only goes up very slightly at the reduced voltage. This can be seen from the following graph

[T1 Terry]

Interesting graph and highlights what I was saying about the reduced Vmp as the cell temp climbs. at 25 deg c the panel in the graph has an approx Vmp of 17.5v, at 75 deg C the Vmp is down around 11v? To charge a battery with a terminal voltage of 13.5v you need a panel voltage of 14.5v to 15v, depending on the size conductor and cable length. looking at the chart, the 8 amps output is now only 4 amps at 15v, an MPPT controller can't fix that problem, but it will use some of that 4 amps to power itself, the PWM controller does as well, but only a fraction of what the MPPT controller does.
You can tell just how inefficient MPPT controllers are compared to PWM controllers, check out the size of the heat sink and the fact there are no fans required for a PWM controller.

T1 Terry

[goboatingnow]

Originally Posted by T1 Terry

It's one of those things that can be hard to test. A simple switch between MPPT and PWM controllers and reading the amps going into the battery will only work if the solar panels remain in parallel and the STC Vmp is between 16v and 18v. this does not suit an MPPT controller, they prefer a higher solar input voltage Simply taking the voltage and amps at one side of an MPPT controller and comparing it to the voltage and amps the other side will give a distorted picture of what is really available from the solar panels. The MPPT controller will hold the solar input voltage at a point that suits its algorithms best, not what the max current flow could have been if it was directly connected to the battery via a PWM controller.

T1 Terry

Sorry terry. Your post shows a fundamental misunderstanding of Vmp and mppt.

In PVs it all about max power transfer not mxa current transfer

Dave

[T1 Terry]

Look at the chart Dave, even at 50 deg C the Vmp is around 15v, how can you get more power out by using a piece of equipment that eats some of the power you are trying to harvest?
If you measure the amps going into the battery using either controller the gain or loss will be obvious, the battery voltage will be common to both controllers. This is the bit the average punter wants to know, how many more Ah will the extra $$ get me towards recharging my battery.
By changing the figures to watts makes a much more impressive difference, 10w to 15w looks impressive, but at 13v it's 0.77 amps against 1.15 amps, that's .38 of an amp or 0.38Ah if the system can maintain that additional input for an hr, and that is the bottom line, the average punter just wants to replace the Ah they used out of their battery to get it full again, but it would be hard to sell a $700 plus controller with a sales line that you will get 1/2 Ah more by spending all that extra money eh.

T1 Terry

[goboatingnow]

I not going to argue the cost issue terry that's really a different issue.

The fact is once the panel can deliver with its operating point at Vmp it will always outperform pwm .

In absorption phase if the panel cannot supply Wmp then whether pwm or mppt is better is more a function of the charge control algorithm. But it's not a function of pwm or mppt

In Li where virtually max current can be applied right up the cutoff point , especially if the cutoff point is below the upper knee, the. Mppt will always deliver better charging efficiencies. Of course individual products may have poor implementations

The overhead of mppt in itself is very small. The power conversion losses are the main issue.

So yes you are right when looking at the pounds and shillings argument. In that case its a function of costs of ownership versus performance. I don't disagree there.

The technology advantage remains and it explains the host of mppt chip sets appearing at the moment.

Dace

[noelex 77]

Originally Posted by T1 Terry

[COLOR=black][FONT=Verdana]Interesting graph and highlights what I was saying about the reduced Vmp as the cell temp climbs. at 25 deg c the panel in the graph has an approx Vmp of 17.5v, at 75 deg C the Vmp is down around 11v?

I make the Vmp about 11.9v, but if your Vmp reaches this level you are in serious trouble. The charge current will be very low with any controller and a parallel connection.

The "series" boys will be jumping up and down and pointing out that with panels connected in series and a MPPT controller you can regain nearly all this lost power. Perhaps doubling the power. Which is true, but fortunately cell temps this high and Vmp's this low are not typical.

[s/v Jedi]

Originally Posted by T1 Terry

If you are getting 20% to 25% more because you changed controllers has little to do with it being an MPPt controller and far more likely to be a case of either crap wiring or the previous PWM regulator was crap.
If you can get more than a 5% improvement using MPPT compared to PWM in bulk mode, you have either a seriously badly match system or crap wiring.

You're hitting a wall. My boat does not have any crap wiring nor PWM controllers. My measurements were done with a Wh counter (not an amp meter that can't be used to measure energy transfer) and it compares my Outback MPPT controller to a direct connection without any controller. In your view, a direct connection is the ideal one because it has no heat sink and it is exactly like what a PWM controller does during the bulk phase, or can hope to do.

I can advice reading up on Amps and Volts vs Wh to determine power transfer. Then reading about impedance matching should complete the picture. Denying this simple mechanism and connecting a solar array to a battery is like attaching a 600 Ohm antenna to your VHF. Yes some power will be transferred, but it's pathetic. The only reason that it works somewhat with solar panels is that the manufacturers designed them to even work somewhat when directly connected. They did an amazing job, but it is far from ideal.

Here's the proof:

MPPT vs. Shunting Controller - YouTube

[noelex 77]

Originally Posted by s/v Jedi

Here's the proof:

It might have been a good demenstration if they had actually used a solar panel

The assumed Vmp of 17.4v is unrealistic, or at least untypical. It also assumes the the solar panel current is identical at 17.4v and 14.2v which is false.

(The on/off shunting regulator is never a good choice at least go for a PWM regulator, but the drawbacks of this sort of regulator did not influence the test)

[s/v Jedi]

Originally Posted by noelex 77

It might have been a good demenstration if they had actually used a solar panel

The assumed Vmp of 17.4v is unrealistic, or at least untypical. It also assumes the the solar panel current is identical at 17.4v and 14.2v which is false.

(The on/off shunting regulator is never a good choice at least go for a PWM regulator, but the drawbacks of this sort of regulator did not influence the test)

why use a solar panel to measure a controller on the bench? It would be highly unusual to do so and results would be flawed when even the slightest change in light would occur.

The gold standard for a 12V solar panel is the 140W Kyocera, which has a Vmp of 17.7V... are you debating the 0.3V difference with 17.4V ?

I don't have the graph at hand at the moment but the current at 17.4V and 14.2V is very similar. Imp is 7.91A and Isc is 8.68A (which is all the way down at 0V).

No, I believe MaineSail did a pretty good test there

[T1 Terry]

Originally Posted by s/v Jedi

why use a solar panel to measure a controller on the bench? It would be highly unusual to do so and results would be flawed when even the slightest change in light would occur.

The gold standard for a 12V solar panel is the 140W Kyocera, which has a Vmp of 17.7V... are you debating the 0.3V difference with 17.4V ?

I don't have the graph at hand at the moment but the current at 17.4V and 14.2V is very similar. Imp is 7.91A and Isc is 8.68A (which is all the way down at 0V).

No, I believe MaineSail did a pretty good test there

It is obvious you don't understand much about solar panels. Unless you place the solar panel on ice the STC figure will never be a real world figure so using STC figures for theoretical test results will always be flawed.
A solar panel in the sun gets hot, at the point within the panel where the action is happening is not exposed to any cooling at all, it has a sheet of glass on top that is being heated by the sun and a sheet of reflective material on the back to maximise the cell efficiency, it gets real hot in there. The solar panel works best and gives it's max output when perfectly aligned with the sun, this is also where the max heat is transferred to the panel, how on earth could the panel temp remain at 25 deg C?

Next issue is how you measure input and output from an MPPT controller. Have you ever connected a CRO (cathode ray oscilloscope) to the input side of an MPPT controller? If you understand what you see on the screen you would understand why attempting to use digital equipment to read such a mix of frequencies is extremely inaccurate without a massive amount of filtering to reduce the signals to the ones the digital equipment can handle without giving false readings. Try putting an analogue volt meter and amp meter on the same connection points and do the math, this will at least reduce the error factor of the digital equipment due to the type of loads within there 2 analogue instruments, but the figures will still not match, nor will the figures be as high as they were without the analogue meters attached, hopefully then you will realise your test methods are flawed and any results are meaningless. But no doubt you will have it set in your mind the digital readings were correct and the analogue meters are causing interference and false readings

The bottom line is, if you want to charge your batteries using solar panels in the real world, you need to do the testing in the real world, not a lab bench or in the theory books, but in the real world using instrumentation that will accurately show the true values either side of the controller or more accurately, at the battery terminals and the entry point of the controller. Only then can you determine the percentage, if any, one controller type has over the other.
The simple one is by reading the Ah into the battery at the end of the day from each controller type, alternate controllers on alternate days, readings over a 1 mth, 2 mth and 12 mth period if you really want the full story, average out the reading and determine what the real difference is between one controller and the other.

T1 Terry

[senormechanico]

Good digital meters have averaging circuitry.
My Fluke Scopemeter from the early 90's has it and it's even user adjustable.

[dennisail]

Originally Posted by s/v Jedi

Here's the proof:

That proof is a joke. Proof would require some solar panels. As terry says, VMP will be way lower in the real world due to heat.

[noelex 77]

Originally Posted by s/v Jedi

The gold standard for a 12V solar panel is the 140W Kyocera, which has a Vmp of 17.7V... are you debating the 0.3V difference with 17.4V ?

You are quoting STC figures. This is a common area of misunderstanding. STC figures are a consistent standard that is fortunately used by all solar panel manufacturers, but these numbers are designed to reflect how the solar panel will perform in the real world. The data is actually collected with brief flashes of light to avoid the panel temperature rising.
Real world cell temperatures are much higher and this has a significant impact on Vmp as the temp coefficients are around 0.1v for each 1c temperature rise.

Typical cell temperatures around high output times are 15-25c over STC temperatures (unless you cruise Arctic waters when they will be lower, or if you have poor ventilation under the panels when they will be higher). Cell temperatures can be easily measured if you want to use your actual results.
Some solar manufacturers actually report these results.

NOCT or normal operating cell temperature conditions. These are measured with an air temperature of 20c.

The NOCT Vmp for the Kyocera panels is:

Vmp=16v

We can of course calculate this as well using the temperature coefficients (assuming a 20 degree rise) 17.7 - (20x0.08)= 16.1v

Added to this we need to allow for some voltage drop in the wiring. If we work on conservative 3% voltage drop we need to deduct a further 0.5v from these numbers.

17.4v is not an appropriate choice for typical 12v solar panel Vmp at the controller.

[noelex 77]

Originally Posted by s/v Jedi

I don't have the graph at hand at the moment but the current at 17.4V and 14.2V is very similar. Imp is 7.91A and Isc is 8.68A (which is all the way down at 0V).

Yes the I/V curve is quite steep so the effect is not large. A quick calculation shows if the bench supply had modelled the typical solar panel I/v curve the current gain improvement from the the MPPT controller would have been reduced by about 13%.