Efficiency comparison question between PWM and MPPT and a simple relay LifePO4

[bigfoot22]

deleted

[Seekonk]

Personal Disclosure: I was around when the 555 came out and the initial ones had problems. I've never gotten over that and consider any 555 design a joke. There are far better chips which are decades old which will give a much better design.

A constant voltage supply is all you need. Chances are you will never get near top charging current with your source. An occasional manual top balance is all that is needed.  MPPT indicates solar panels.  What do you have.  I modified a standard 120V to 12V LED switching power supply to work at 60V for my array.  These can be set up to maintain a minimum panel voltage as well as lower voltage buck converters for near MPPY performance.

[Geoff-AU]

Just how efficient is simply connecting the solar panel directly to the battery then interrupting it once the voltage goes above a threshold?

Direct connect from a 12V panel (max power at 17V) to a 12.8V LiFePO4 costs you around 25-30% efficiency, but if ultimate reliability is your goal then take the money for your fancy MPPT charger, buy 25% bigger panels instead, and don't worry about it.  Charge termination is the only thing to consider, for intermittent use I am relying on the BMS in my pack and the wide safety margin of LiFePO4 chemistry.  But I would not do that for continual use, I only solar charge when camping.  For continual use I'd probably stick a buck converter (or PWM charger) on it.  MPPT is only worth it if the difference between panel voltage and battery voltage is large.

[Seekonk]

Blah blah blah blah.  Those arguments don't make sense to me. Frankly, solar people don't make sense to me either. Everyone on solar forums is doing things wrong.  I have a camp with refrigeration, hot water, dishwasher, and large capacity clothes washer with all cycles using hot water.  And I have to keep myself alive at night with a machine. I do this reliably with only a car battery and electronics that cost next to nothing.  I was hit by lightning which knocked everything out. I'm running around trying to get the basics back working and my wife asks me in the middle of all this to make her parmesan popcorn. So, I took time out for this simple request. Everything was back working in two hours.  That sounds simple to me.

All you need is a simple circuit to keep the panels at a constant voltage.  A little improvement is adding a few extra components to provide temperature tracking. That gives you 95% of the performance of a full MPPT.   this circuit fakes any switching power supply or buck converter into thinking the output voltage is too high and it reduces the output current to what the panels can provide efficiently. In essence making the circuit a Linear Current Booster.

[gf]

Only problem with that is that I cannot afford the inefficiencies of a direct connect charge controller AND I cannot afford the possible unreliability of a MPPT/PWM charge controller mainly because I'm "going bush". This is for a possible installation on a tiny home where roof space is minimal and battery space isn't that great either but the demands on the system are highish. LED Lights, a 12/24v fridge, small LCD TV, possible occasional microwave oven usage (a luxury for when the batteries are fully charged). The greatest consumer of current would be a gaming laptop which I don't think is even possible to run for more than a few hours per day. And the most I think in solar panels I can get onto the roof would be about the 300-400w range. Maybe higher.

Why do you put so much emphasis on the reliability of the charge controller, but not consider the reliability of the inverter or the battery as well? Low-power charge controllers are not expensive. If you want redundancy you could use e.g. a separate charge controller for each panel, and additionally keep a spare if replacement becomes necessary.

[Siwastaja]

All you need is a simple circuit to keep the panels at a constant voltage.  A little improvement is adding a few extra components to provide temperature tracking. That gives you 95% of the performance of a full MPPT.

This is true of course, but then again designing "full" MPPT does not, in my opinion, add any cost or complexity, but that could be just me, because I would have no problem doing it; I would use a microcontroller anyway and can design reliable electronics around it.

The benefit of your temperature-compensated constant voltage switch mode converter is, one can do it with analog components if they are more "fluent" in analog electronics, and as you say, it's really good enough. So I'd recommend that too if MPPT seems like a struggle.

But mostly it's the same. Same power stage (usually buck, can be boost of course, or any topology), with a way to regulate input voltage.

[Faringdon]

Your relays will be breaking and opening often, and will wear out.
You should use MPPT...because it draws current out of the panel without overheating the panel with the current, which is a waste of energy....(MPPT reduces the current to the right level)
Eg if the panel has a 12v o/c voltage...then typically you need to draw current out of it, at that exact  rate, which keeps the panel voltage at 12v minus about  10% (ie 10.8v)
If you dont do this, then you waste your precious solar energy.

[Siwastaja]

I'm going to continue to research the MPPT route.

This is the route I would go:
* pick a microcontroller. If unfamiliar with micros, do some led blinkers, hello worlds etc. first.
* wire a topology such as synchronous buck with bootstrap fet gate driver, current sense resistor and current sense amplifier
* make the microcontroller Analog Comparator (etc.) generate interrupts when you discharge a capacitor on the current sense resistor
* make the microcontroller generate PWM signals
* combine the two, make it stop the PWM on overcurrent signal
* now you have a DC/DC which does not blow up on every experiment - apply Vin
* make it output fixed duty cycle, which is Vout/Vmp from the solar panel datasheet (compensate to the expected temperature)
* now it's already 1000x better than the 555
* if still interested, go for MPPT:
* wire input voltage measurement (voltage divider + capacitor)
* add a control loop (PI loop works fine) to adjust PWM duty to regulate that voltage.
* replace fixed duty with this regulation loop.
* now it's 2000x better than the 555.
* wire that current sense not only to analog comparator, but ADC as well.
* calculate power (voltage * current)
* add perturb&observe algorithm: make changes to the voltage setpoint, recalculate power to see if it gets better or worse

[Geoff-AU]

Only problem with that is that I cannot afford the inefficiencies of a direct connect charge controller AND I cannot afford the possible unreliability of a MPPT/PWM charge controller

So..  you can't hack simple and you can't hack complex?

OK good luck.