MPPT vs PWM - The Solar Controller Test!

[wildebus]

There are lots of different thoughts and opinions about the different flavours of Solar Controllers and the Pluses and Minuses of MPPT vs PWM models.

I thought I would actually test out a few controllers to see just what is what and how they really compare in reality.

I am not using actual solar panels for this testing as that will not provide a consistant power input to the controller under test. So instead I am using the following to provide the power:
1 x 9V 2A Transformer; 1 x 9V 2A Transformer; 1 x 9V 2.4A Transformer. I have wired these in series to give a Power Bank that supplies a 27V input with a current of around 2.1A - so a Wattage of 57W.
a regular 57W would not be a bad power level from say a 100W solar panel so represents a fair test in that respect. What this testing will not do is check one of the key features of the MPPT controller - the way it attempts to maximize the power in as the solar irradiation changes (the Maximum Power Point Tracking), but that is just for MPPT so is more for a test between different MPPT controllers so not a key factor for this test.

I have 5 different controllers I will be testing - 2 MPPT units and 3 PWM units. One of the MPPT units is a Victron Smart Solar and will be used as the benchmark as it is well known & respected and generally regarded as Best of Breed when it comes to Solar Controllers.

The Battery in the test is a slightly tired 75Ah Lead Acid Leisure Battery. I am commencing the test when the battery would be requiring a bulk charge and has the same (fairly minimal) load in it, so would be able to take for many hours the full 57W that the power bank could offer.

With the Power Bank proving a constant power (within the variances of transformers) and the Battery taking all it is offered, we can be pretty confident that any variances will be down to the Solar Controllers.

I have a Power Monitor on both the input and the output of the tested controller. This will tell me the Voltage, the Current and the Power in Watts both into the Controller and out to the Battery- and also the cumulative power (Wh) in and out over time


So Let's Get On It!

These are the Solar Controllers to be tested:

SC - Test Controllers by David, on Flickr


Test #1 - Venus MPPT 20A Controller (note the 'Venus' name has nothing to do with the Victron Venus)
This is the Controller. Physically very large. Well featured with temp probe for dynamic voltage adjustment once it gets to the Absorption Mode.
SC - Venus_MPPT_20A by David, on Flickr

The Test Results

SC - Venus_MPPT_20A - Test Results by David, on Flickr

Key Things to observe here:
The Input Monitor shows just what the Power Bank is providing - 27V and a current of 2.15A - 57.9W.
The Output shows the Battery level - 13.26V. And the key thing, which is a feature of the MPPT Controllers - the output current is raised above the input current (2.15A in, 4.08A out) so there is minimal power loss over the input - so with an input power of 57.9W, we get 54.1W out - an loss of around 7% (efficiency of 93%).


Test #2 - Venus PWM 30A Controller
This is the Controller.

SC - Venus_PWM_30A by David, on Flickr

The Test Results

SC - Venus_PWM_30A - Test Results by David, on Flickr
Now this is interesting
The MPPT Controller had an average power in of 51.3W/Hr over the test period. That is from a power input of 57W/Hr into the controller.
The PWM Controller in comparision had an average power in of 29.3W/Hr. That is just 57% percent of the power - a massive drop in performance.
What is the reason? PWM Controllers are unable to effectively convert excess voltage to extra current to maximise the power input. So when you look at the monitors you see the input voltage is cut to just above the battery voltage and the current in and out are pretty well the same.
If you didn't know otherwise, you might assume that the Solar Panel (the Power Bank in this case) is actually not putting out a higher voltage, but it definately is. And to confirm this, I disconnected the battery to get a truer picture of the input voltage ...
SC -PWM No Battery by David, on Flickr

So in essence, any voltage greater than around 1V more than the battery voltage is wasted.
A basic 12V Solar Panel has a Voltage of around 20V or so. as the battery voltage increases, more power is realises from the panel, but there is always a fair waste even when the battery is at is absorption level of around 14.4V.


I'll be carrying on the testing with the remaining controllers as further comparisions over the next few days (I need to leave the battery to drain off a bit overnight between each test), but the data so far to my mind actually suggests to me that an MPPT controller will be much better then generally thought with even the most basic panel.

Once I have completed the tests with the five controllers, I will repeat the test on an MPPT and a PWM controller, using a Power Bank input at 18V (so remove one of the transformers) to do a comparision at a level closer to the single 12V Solar Panel to see the difference.

 

[SquirrellCook]

Is it not that simple using a power supply to mimic a solar panel?
Please correct me if I'm wrong.
You don't measure a panels potential yield by multiplying the OCV by it's Short circuit current.
All the controllers I've looked at will adjust the charging load to collapse the voltage to a predetermined amount.
From my experience of mppt they come into their own when the solar yield is lower.

 

[wildebus]

Is it not that simple using a power supply to mimic a solar panel?
Please correct me if I'm wrong.
You don't measure a panels potential yield by multiplying the OCV by it's Short circuit current.
All the controllers I've looked at will adjust the charging load to collapse the voltage to a predetermined amount.
From my experience of mppt they come into their own when the solar yield is lower.

When you say the "controllers you've looked at", were they both MPPT AND PWM?

a Solar Panel is a power source, pure and simple. It is no different to any other power supply with the same potential voltage and current except the solar panel is very inconsistent in its power output. Using a PSU to test a Solar Controller is a perfectly valid thing to do.

The potential realised of a solar panel at any given time is simply the Voltage x Current (P=IV) at that time - it really is that simple. The battery monitors show the data coming it, they are not reading a label

Ref "All the controllers I've looked at will adjust the charging load to collapse the voltage to a predetermined amount" - Don't agree. Look at the MPPT numbers - that voltage in just about the same as the voltage the Power Supply provides. Where is it collapsed? What Controllers have you looked at and how have you determined that statement?

 

[wildebus]

Let's discuss the validity or otherwise of driving a Solar Controller by a PSU rather than a Solar Panel.

As I mentioned at the beginning, doing so will mean the Power Point Tracking part is not really exercised due to the consistent nature of the PSU over the fluctuating power generation of a PV panel due to light variances. But excepting those times, there is no real difference between the power sources apart from the consistancy of the PSU.
Have a look at a typical day using an MPPT Controller.

Once the sun has risen the voltage is fairly consistant at around the 70V level until the sun starts to fall. The current ramps up and then falls as expected as the day progresses. The dips will be due to cloud and shade (for example, the time around 13:00 always goes low for me as a tree blocks the sun. For information, the voltage and current above are being generated by 4 100W panels in series, each with a Vmp of 18V and a Voc of 21.6V, so the voltage results are entirely consistant with that and a PSU of around 72V would be a good alternative for testing.

How does that compare to a PSU feeding an Controller? Well pretty well the same in fact. The voltage goes up to pretty well what can be provided (27V in this case) and is ner enough flat lining as there is no effect from shading and light intensity as the day progresses.

So when it comes to how well a Solar Controller is able to convert power at the kind of Voltage Levels that a Solar Panel Array can provide, using a Power Supply to mimic a Panel of a given output (in the test setup I have, it is essentially mimicing a 60W panel operating in full sunlight) provides the basis for consistant comparisions.
It is not the full story as, as already explained, it is not testing how well a controller will be dealing with dusk and dawn harvesting, but that is a small part of the day as you can see from the first chart.

So after that interlude, back to the testing.

 

[wildebus]

Test #3 - "Newest version MPPT 20A 12/24V Auto solar regulator charge controller PWM BR"

The description is not that clear as it mentions both MPPT and PWM. Personally I think that is so when someone does a search for "MPPT Solar Controller" on eBay, this item comes up and it is not actually an MPPT at all but just a PWM unit (the price of under £6 would support that).

This unit will get tested once the battery dropped some charge, but in the meantime, I connected it up and it does indeed give the hallmark appearance of a PWM

SC - MPPT-PWM - 20A - PowerUp by David, on Flickr

Now actually for under £6, it looks pretty decent as PWM controllers go in terms of appearance and build and could be a reasonable buy, but from my initial impression, do not be fooled into thinking it is an MPPT unit.

Performance results to follow .....

 

[SquirrellCook]

Please don't think for one moment Dave that I'm trying to belittle you.
My mastery of our language is poor at best, and I've been told I can offend unintentionally.
As I've said before I admire what you do and your dedication to it.
It just reminded me of a UPS test demonstrated to me many years ago with a power supply.

I've noticed that you don't consider poorer solar delivery time as important, but If you have an small solar array or not enough room for a big array to offset your consumption. The suboptimal conditions are important.
I guess for those that like to go away during the winter and remain off grid it's important too.

 

[wildebus]

Please don't think for one moment Dave that I'm trying to belittle you.
My mastery of our language is poor at best, and I've been told I can offend unintentionally.
As I've said before I admire what you do and your dedication to it.
It just reminded me of a UPS test demonstrated to me many years ago with a power supply.

I've noticed that you don't consider poorer solar delivery time as important, but If you have an small solar array or not enough room for a big array to offset your consumption. The suboptimal conditions are important.
I guess for those that like to go away during the winter and remain off grid it's important too.

I am not saying a controller that will optimize performance in poorer conditions is not important. quite the reverse. But when it comes to eeking out the first bit and last bit of solar, that amount of energy is very small compared to the overall amount harvested in a day, be it winter or summer. A controller that say throws 10% of available power away across a day is far worse then a controller that is not optimized for dusk and dawn.
However, it is just not something I am testing here as it is not possible to do unless I had a bank of panels, a bank of batteries and could run all the controllers simulateously as that is the only way - without a set regulated power supply - to do a comparative test unless you were somewhere like the Arizona desert or the like where the conditions from day to day were pretty well identical for a decent period. In the UK? and in October? forget that.


What I am testing - and to be honest I think it is pretty obvious - is how well a controller takes power in, with power parameters similar to a solar panel, and puts that power into the battery. I made it very clear that the test would not be testing the Maximum Power Point Tracking Feature and how well that works. If someone wants to see a test of how good a controller is at dawn at pulling in power, this is not the test to look at.

I still don't understand your comment "All the controllers I've looked at will adjust the charging load to collapse the voltage to a predetermined amount" unless that was for PWM Controllers only - where the 'pre-determined' voltage seems to be Battery Voltage +0.2V or so. I have not seen that behaviour on ANY MPPT controllers.

 

[SquirrellCook]

No Dave, I'm not singing the praises of the PWM devices. Quite the opposite.
Some of the claimed MPPT ones I had purchased proved to be PWM in their behaviour.
Since buying the Victron MPPT I'm more that happy with it. It does what I would have expected and more.

 

[RAW]

Please don't think for one moment Dave that I'm trying to belittle you.
My mastery of our language is poor at best, and I've been told I can offend unintentionally.
As I've said before I admire what you do and your dedication to it.
It just reminded me of a UPS test demonstrated to me many years ago with a power supply.

I've noticed that you don't consider poorer solar delivery time as important, but If you have an small solar array or not enough room for a big array to offset your consumption. The suboptimal conditions are important.
I guess for those that like to go away during the winter and remain off grid it's important too.

Not wanting to rock the boat at all but @wildebus has put a lot of time and effort into all this and writing the results up for us all.
My only comment on how it's done would be to suggest maybe a video of the tests as well as being useful. Saying that I am not great at making videos or like being in the Frame unlike @Geeky Philip so I choose not to make Video's of what I do either !! Perhaps that will change in the future.

As a baseline for the test @wildebus has chosen to take Power from a steady source into the Controllers, this makes perfect sense as it then provides a baseline for the tests. To do the tests under natural light would require a great deal of time, energy and money to set up arrays all working at the same time in the same location and outputting power to 5 batteries all the same type. A big ASK and also it would not work as well as a steady input for testing as even a slight variance in the position of a panel will result in a different power output.
Anyway enough said.

Great work @wildebus and thanks very much, I will stick with Victron I think and will potentially have another solar install looming if I commit to purchase the Van I am seeing tomorrow and sell Sylvia.

 

[wildebus]

Ref Videos, In preparation of doing the major Electrical update on the T5 I had in (and posted a thread about) I did actually get a little digital video camera before hand specially for this and thought I would do some videos on the work .... But I decided, like yourself I think, making videos was not for me. Hats off to those who do and I am an avid viewer of lots of them, but ME making them? nah!

(as it happens, I just recording 2 minutes of a battery monitor which I will be posting shortly. Exciting huh?)

 

[wildebus]

Test #3 - "CMP-2420 - Newest version MPPT 20A 12/24V Auto solar regulator charge controller PWM BR"

Test #3 - "Newest version MPPT 20A 12/24V Auto solar regulator charge controller PWM BR"

The description is not that clear as it mentions both MPPT and PWM. Personally I think that is so when someone does a search for "MPPT Solar Controller" on eBay, this item comes up and it is not actually an MPPT at all but just a PWM unit (the price of under £6 would support that).

This unit will get tested once the battery dropped some charge, but in the meantime, I connected it up and it does indeed give the hallmark appearance of a PWM

SC - MPPT-PWM - 20A - PowerUp by David, on Flickr

Now actually for under £6, it looks pretty decent as PWM controllers go in terms of appearance and build and could be a reasonable buy, but from my initial impression, do not be fooled into thinking it is an MPPT unit.

Performance results to follow .....

To conclude Test #3 ....
This is the controller

SC -CMP-2420_MPPT_PWM_10A by David, on Flickr

And the Test Results

SC - CMP-2020 MPPT_PWM_20A - Test Results by David, on Flickr

IMO, this confirms that the controller is a PWM unit and performs just the same as any other PWM unit

 

[wildebus]

Test #4 - 'Generic' PWM 10A Controller

This is the controller used for Test #4

SC - Generic_PWM_10A by David, on Flickr

I am calling it 'Generic' as it is a unit that is sold by countless vendors on eBay, Amazon and other outlets and is very common. The units can be different colours (most are the blue like shown above, but can be in Black as well as occasionally a few others). I don't believe the actual internal electronics vary.

And these are the Test Results

SC - Generic_PWM_10A - Test Results by David, on Flickr

I have one more PWM Controller to go (it is under test right now) and will then be finishing off with the Control Controller - the Victron MPPT Controller.
I am finishing off with the Victron as I can validate (or otherwise!) that the power output from my little frankenstein Power Supply is the same at the end as it was at the commencement of this bank of testing.


Looking at the results so far, the 3 PWM controllers have an average of 29.3W/Hr, 29.6W/Hr and 28.3W/Hr.
Very close results and close enough within test environment variences to say they are basically the same. I am confident enough to be able to predict the 4th PWM Controller will be putting around 29W/Hr into the battery. We shall see if that is the case at about 6PM

 

[RAW]

To conclude Test #3 ....
This is the controller

SC -CMP-2420_MPPT_PWM_10A by David, on Flickr

AND SOLD OUT !

 

[wildebus]

AND SOLD OUT !
View attachment 1765

Well, If you want one, I have one for a tenner (including postage) . Otherwise it will just go back to China.
As long as you accept it is NOT an MPPT (sadly I bet many people are buying this because it says it is and the price is low enough that they will just write it off as a "oh well" buy when they find it is not) and you want an PWM controller it seems a perfectly decent budget product.

 

[Pugwash69]

I'm finding this very interesting. I fitted a cheap (£17.35) charge controller in my van when I started and suspect it's barely worth that much money. I'm probably going to go for a Victron.

BTW did you see the YT video by Will Prowse? He had a few of the same models you show here.

This is the Anself CMG-2420 I want to replace.

 

[wildebus]

I am quite surprised that he is seeing no significant difference between PWM and MPPT in terms of power input.
I will be reducing the input to 18V and comparing an PWM and MPPT once each controller is tested.
And finally use a 2nd battery, albeit a small 36Ah one and run a parallel test - MPPT vs PWM, each with a 100W Solar panel side by side. I can use the Victron monitoring info to view incoming and outgoing voltage and current and use the two monitors on the PWM so keeping a real-time view on both. The difference in battery size should not matter for a relatively short test period with low current (I am not expecting much from a 100W panel at the end of October!)

EDIT: It would be good to run a full day test encompassing dawn and dusk, so given one of the batteries is just 36Ah, I think I will use 30W Panels instead of 100W. They will operate just the same as the 100W in terms of voltage but obviously at much lower current so a full day should not be close to fully charging even a small battery at this time of year

 

[trevskoda]

Thanks for the tests ,may now fit my new unit,no point having it sit in a box,mind you i am getting a good charge from the 6 buck unit.

 

[wildebus]

Test #5 - Venus PWM 20A Controller (again, note the 'Venus' name has nothing to do with the Victron Venus)

This is the Controller

SC - Venus_PWM_20A by David, on Flickr

And the test results
SC - Venus_PWM_20A - Test Results by David, on Flickr

So was my prediction right?

Looking at the results so far, the 3 PWM controllers have an average of 29.3W/Hr, 29.6W/Hr and 28.3W/Hr.
Very close results and close enough within test environment variences to say they are basically the same. I am confident enough to be able to predict the 4th PWM Controller will be putting around 29W/Hr into the battery. We shall see if that is the case at about 6PM

The 4th PWM worked out at 28.8W/Hr


My conclusions with the PWM Controllers ....
Some may have some extra features such as temperature compensation (such as the 'Venus' units)
Most have a digital display but some don't.
Units with digital displays vary in the info shown - some have power, some have current
Some Units have more programmable settings than others to allow a little fine tuning

But end of the day, all the PWM units have as close to identical performance is doesn't matter.

 

[wildebus]

Thanks for the tests ,may now fit my new unit,no point having it sit in a box,mind you i am getting a good charge from the 6 buck unit.View attachment 1773

Be interesting to see what difference you see (you could also change from Parallel to Series on the panel? that will eek out a touch more start and end of the day and give fractionally less cable loss)
The Dual Battery bit is interesting - next week I am getting in a prototype standalone starter battery trickle charger to test, which will do a similar job I expect? excepting it will be good for solar and EHU. I am assuming the starter battery trickle on the Votronic is just active with solar?

 

[Lord Lucan]

Fantastic - excellent testing and detailed in-depth explanation.
I'm up for a mppt.

Can I buy the Victron after testing is finished?

I also need a panel and battery