Using a cheap MPPT BQ24650 module as solar trickle charge siphon

[cybermaus]

For a UK narrowboat installation we are adding some solar. In summer for some living, and in winter to keep the batteries topped up.
In summer, we hope to get 30Amps peak out of it, in winter hopefully 1 amp to counter for self discharge, strictly no standby users allowed (other then the bilge pomp, which hopefully sits idle at a measured avg 10mA)

Of course we also have a engine starter battery, and a bow thruster batter. It would be nice of those could also stay topped up.
Without having to add extra solar.

After some mucking about with some resistors and diodes, I am now using a cheap MPP(T) module to do this:
Its not really a MPPT, more a fixed point MPP. I set it like this:

- Tuned down input MPP from its 18V setting to 13.2V
- Tuned the output voltage from 14.4V (lithium) to 13.8V (lead)

As a result, when the input battery is less then 13.2V (in other words, not actively charging) the MPP turns off. Measured Drain is about 10uAmp
So no risk of draining the domestic batteries, and only siphoning energy when solar is actually active and batteries are nearly full.

Also, if for some reason the starter is nearly empty, or even already stone-cold dead, the device will limit max 2Amp, and nicely stay at 45 Celcius
And in any case, it will stop if domestic gets below 13.2V, so a dead starter cannot also kill of the domestic batteries.

Test setup works quite well. Or at least, seems to work well in testing. More actual results on this next year.
As stated, I had it sit at full 2 amp for several hours, and also I had it fully fill up the starter battery, seeing the charge current slowly drop to nothing, without overvolting the battery.
And I tested that indeed activity and even standby drain stops when it gets below 13.2V. So in the lab, it works quite well.

https://www.aliexpress.com/item/1005006061650369.html


Edit: some more testing showed that if you short the input while the output is still connected to a battery, there is a backfeeding short. Good thing the module was fused on the PCB, it survived.
They do show a diode on the eBay, which may be a good idea. However, I would put one (a Schottky) on the input side, not the output side.

[spostma]

Thank you for investigating this modification.
Could you indicate what you changed exactly to achieve this?
This could help others customizing these modules for other CC/CV applications.

[cybermaus]

Modified? Used as is!
Like I said, I just tuned the voltages. As visible on the link/photo, there are two adjustable potmeters (multi-turn)

So it is not so much a mod, as a creative use case.

Some more detail though:
- Temporary connect 1K resistor and voltmeter to output.
- Connect finely tuneable power supply from 10V to 20V to input.

- Turn MPP_t pot (the one toward the middle) so that the switch-on point is at 13.2V
- After your input turn-on-point is tuned, turn the power supply to well above that point, for example 16V
- Turn V_out pot (the one on the edge) so that the output voltage is 13.8V
Or whatever voltages you desire.


As stated, it may be good to add a Schottky diode on the input side, to avoid backflow. Input side as output side is where it measures charge voltages, and we do not want to mess that up too much.
As mine survived thanks to its 10 Amp PCB fuse, and with my battery-2-battery setup I have little voltage levels to play with, I wonder if I should simply keep relying on the fuse.