Is an MPPT controller just basically a buck/boost converter with extras?

Thread Starter

KevinHowJones

Joined Jun 19, 2022
30
It seems to me that a buck/boost converter would provide the optimal power available to the load. The current to the load will always be a function the target voltage (eg 13.2V) and the load impedance. Whether the PV is above or below the target voltage, the converter will always be trying to supply the needed current at the output by drawing the whatever current from the PV to attempt to maintain Pout = Pin (limited of course by the maximum current available from the PV.)

So if you are strictly/only interested in maintaining the maximum power point, would not a buck/boost converter work for that?
 

Papabravo

Joined Feb 24, 2006
21,159
It is more likely that a transformer is involved, in which case you have a related topology which is similar to a buck converter. I'm not aware of how a boost topology would apply. It seems like the PV output will be greater than the battery voltage in most cases. Since there are multiple topologies that could be used more details would be helpful in answering your question.

EDIT: The "extras" you mentioned would include sensors for light intensity and temperature. These are not normally part of standard DC-DC conversion schemes.
 
Last edited:

ronsimpson

Joined Oct 7, 2019
2,989
Maximum Power Point Tracking.
There is usually a buck, buck/boost, or boost converter involved.
A buck/boost, or any other type of PWM, has a job of holding the output voltage constant.
Exception: a LED driver has the job of holding the current constant.

Let's say you have a 12W light bulb, 12V x 1A. If the PV can produce more than 12W, the PWM will take 12W from the panel and make 12V for the bulb. (not max power at all) If the PV can produce only 11.9W then the PWM will suck the PV voltage down to 0V and give up. (not good)

Maybe a better example is a battery charger set to charge at 1A and 12V. The same thing will happen. But if you are running MPPT then the controller will drop back the power or increase the power to get the most. If the PV can only make 6W then the PWM will try 12W and see the PV voltage collapse, then try 10W and 8W and 6W and 5W then it will stay at 6W. This will only charge the batter at 0.5A because that is the most power it can get off the PV.

There is a lot of smarts in a MPPT controller.
 

crutschow

Joined Mar 14, 2008
34,285
It seems to me that a buck/boost converter would provide the optimal power available to the load.
Not by itself.
The maximum power to the load (typically a battery) is determined by the characteristics of the power source (a solar panel), not the converter output load, and (from the maximum power point theorem), is where the apparent load (converter input) impedance matches the source impedance.
That point varies with the illumination on the panel, so a MMPT (Maximum Power Point Tracker) varies the current drawn from the source slightly and continually up and down around the point where the maximum power is extracted from the panel (the MPP point).
That point is not at the maximum current from the source, but is generally where the voltage is about 1/2 the open-circuit voltage (load impedance matching the source impedance).
So the MPPT controller measures the power from the panel and adjusts the converter (Buck, Boost, or Buck-Boost depending upon the voltage difference between the source and load) to operate at that point.
 

Ian0

Joined Aug 7, 2020
9,671
The panel voltage at which maximum power occurs varies very little with light-level, but it does vary with temperature; so you can make a crude (but quite effective) MPPT by regulating the panel voltage and compensating for temperature.
Not as good as the clever algorithms, but much better than no MPPT at all.
 

Papabravo

Joined Feb 24, 2006
21,159
As has already been pointed out, a dropping PV panel voltage makes it really hard to boost the voltage and provide significant amounts of current. You are on the wrong side of the power conversion process. There has to be a point where you can't win that battle.
 

Ian0

Joined Aug 7, 2020
9,671
This is the curve of power vs. cell voltage.
Screenshot at 2022-06-19 18-56-49.pngThe problem for the feedback system is that it has to find the peak of the curve. Should it increase or decrease the cell voltage to find it?
You can see that if the voltage were fixed at about 33V then it wouldn't be far wrong. But then you have to look at the graph of peak power vs. voltage at different temperatures.
 

bassbindevil

Joined Jan 23, 2014
824
There's a circuit floating around that uses a 78S40. The obvious drawback is that battery common is not the same as solar panel common. OTOH, some PWM charge controllers also work like that.
http://web.archive.org/web/20091124055655/http://www.freewebs.com/acselectronics/buildppt.html

A revised version of the schematic as a pdf:
https://www.fieldlines.com/index.php?topic=135301.0
https://www.otherpower.com/images/scimages/2006/solar_mppt.pdf

https://www.aaroncake.net/forum/topic.asp?TOPIC_ID=8849
this provided the link to the original page on Archive.org, but also mentions a more modern TI chip for battery charging with MPPT:
https://www.ti.com/product/BQ24650
Boards based on this chip are available from "overseas" starting at around $20, which isn't bad since Digikey would charge me almost as much for one chip plus shipping if they had stock, which they don't.

If you just want a basic battery charger, the 78S40 looks buildable from junkbox parts. But for off-grid life where you need to budget your energy consumption, something smarter that measures and logs energy production would be better.
 
Last edited:

Ian0

Joined Aug 7, 2020
9,671
There's a circuit floating around that uses a 78S40. The obvious drawback is that battery common is not the same as solar panel common. OTOH, some PWM charge controllers also work like that.
http://web.archive.org/web/20091124055655/http://www.freewebs.com/acselectronics/buildppt.html
It's half right - keeping the panel voltage fixed keeps it very close to maximum power point, provided that the panel temperature doesn't vary.
The peak power voltage varies at -0.3%/°C - and the panel can vary from -10°C on a winter morning, to more than 60°C in full sun on a summer afternoon. So the voltage varies much much more with temperature than it does with light level.
Without any temperature compensation, it will be hopeless.
[edit] Also, you can’t make a DC inductor on a ferrite toroidal.
 
Last edited:

Ian0

Joined Aug 7, 2020
9,671
This is probably the simplest functional circuit.
It is a hysteretic regulator, with the MCP1401’s Schmitt trigger making it oscillate. Temperature compensation is done by the transistor, which has the same tempco as the panel.
However, it is an “open loop” system, so won’t work quite as well as systems with feedback.3DEC7C79-4BB2-4B0B-BA25-7A56D6D182CF.jpeg
I haven’t shown how the 12V supply for the gate driver is derived from the panel voltage, as it will depend on the panel voltage.
 
Last edited:

Ian0

Joined Aug 7, 2020
9,671
So is this performing an MPPT function?
Yes.
if you are trying to extract the very last milliwatt from the solarpanel, then you could probably achieve better with a more complex circuit and/or a microprocessor,but bear in mind that a more complex circuit takes more energy to run!
There are a few things that make it less than perfect:
1) the transistor would have to be touching the solar panel for it to track the temperature perfectly.
2) maximum power voltage isn’t absolutely invariant with light level.
3) it has a preset that needs to be adjusted.
 

Thread Starter

KevinHowJones

Joined Jun 19, 2022
30
I assume the two diodes on the left are the PV, and the horizontal lines on the right are the battery.

What is the MCP1401 driving? I don't recognize that symbol.
 

Ian0

Joined Aug 7, 2020
9,671
Also, do you have an idea of values for the inductor and the zener?
I assume the two diodes on the left are the PV, and the horizontal lines on the right are the battery.

What is the MCP1401 driving? I don't recognize that symbol.
I drew it quickly with a pencil- I think I have a SPICE simulation somewhere.
The zener is in fact a schottky (it does look zener-ish)
The MCP1401 drives a MOSFET.
The inductor value (and that of the capacitor across the panel) depend on the voltage and current of the panel.
The output isn't shown - I don't think you said it charged a battery, but that joins to the output capacitor.
 
Top