MPPT solar controller and LiFePO4 battery for backpacking

[karrak]

If you had any clue how Buck Converters work you would know that. Below is the Buck Converter in its simplest form. If the FET looses it reference signal which is what happens if the battery is disconnected, the FET shuts off and the Input is directly connected to the output through L1 and a forward biased diode. .

Wrong yet again, either you are very confused, or you want to confuse everyone else or is it both? The circuit diagram you have shown if for a boost converter that boosts the voltage. Hint, look at the webpage address of the image you used. The buck power supply does the opposite.

With reference to the circuit diagram below, if we replace the Rectifier & Reservoir Capacitor with a Solar Panel and the Load with a Battery we end up with the basic circuit for an MPPT controller. As you can see the Switching Transistor is between the Solar Panel and the Battery. If the Transistor is switched off the Solar Panel will be disconnected from the battery. A PWM controller is basically the same except that it doesn't have the Flywheel Circuit. For those that want to find out more about buck switch mode power supplies see http://www.learnabout-electronics.org/PSU/psu31.php



Simon

Off grid 24V system, 6x190W Solar Panels, 32x90ah Winston LiFeYPO4 batteries installed April 2013
BMS - Homemade Battery logger https://github.com/simat/BatteryMonitor
Latronics 4kW Inverter, homemade MPPT controller

[Sunking]

Wrong yet again, either you are very confused, or you want to confuse everyone else or is it both? The circuit diagram you have shown if for a boost converter that boosts the voltage. Hint, look at the webpage address of the image you used. The buck power supply does the opposite.

I am typing this very slowly Simon so you can understand and follow along. Yes I clicked the wrong converter. The Transistor in your circuit is constantly switched on/off to some degree. Pulse width is modulated That means there is always voltage on the output. Without a load (battery) the voltage will go to source voltage because there is NOTHING for the Inductor and Capacitor to discharge the energy into. The CAPACITOR charges to the Input Supply Voltage and holds it. A first year electronics student would know that. You failed the test.

Like I said disconnect your battery and see what happens. Now get lost. You are only here to start trouble.

[Dave C]

Simon and Sunking,

Could the problem be as simple as the controller supplying charging voltage that's too LOW for the battery? The controller's specs show a CV voltage of 14.2V, and the battery's specs show a charging voltage of 14.6V. Is it likely that the BMS would disconnect the battery because of a 0.4V difference?

[jflorey2]

I am typing this very slowly Simon so you can understand and follow along. Yes I clicked the wrong converter. The Transistor in your circuit is constantly switched on/off to some degree. Pulse width is modulated That means there is always voltage on the output. Without a load (battery) the voltage will go to source voltage because there is NOTHING for the Inductor and Capacitor to discharge the energy into.

Incorrect. The schematic shows a standard nonisolated buck converter. (There's also a bridge rectifier on the input to convert AC to DC; it is not relevant to this discussion.) The controller modulates the duty cycle of the switching transisitor to maintain a certain voltage at the output. If there is no load, the controller will go to zero duty cycle, and the output will not go to source voltage. That's why switchmode power supplies can regulate their output voltages from zero to maximum load.

This is an important concept to understand; it is why a well designed MPPT controller will never exceed its set output voltage even if there is no load. That is not true of PWM controllers.

[Dave C]

Guys, any thoughts on my last post?

[inetdog]

Guys, any thoughts on my last post?

The problem is NOT that the CC output voltage is too low. All that would do is stop charging the battery when it reached the preset voltage, and the BMS would not mind that at all.

The BMS appears to be designed to disconnect the battery completely when the battery terminal voltage gets too high OR when the battery voltage reaches the safety limit and the BMS cannot shunt enough current to keep charge current from flowing into the battery. It will keep it disconnected as long as there is too high a voltage present on its input terminals. But it should not prevent you from connecting a load to the battery at that point and using it.
Forget any other tests until you have connected a LOAD to the battery and then measured the battery output voltage (with the Genasun disconnected.) At that point, with the LOAD still in place connect the Genasun and see what voltage it produces.

[Dave C]

The problem is NOT that the CC output voltage is too low. All that would do is stop charging the battery when it reached the preset voltage, and the BMS would not mind that at all.

The BMS appears to be designed to disconnect the battery completely when the battery terminal voltage gets too high OR when the battery voltage reaches the safety limit and the BMS cannot shunt enough current to keep charge current from flowing into the battery. It will keep it disconnected as long as there is too high a voltage present on its input terminals. But it should not prevent you from connecting a load to the battery at that point and using it.
Forget any other tests until you have connected a LOAD to the battery and then measured the battery output voltage (with the Genasun disconnected.) At that point, with the LOAD still in place connect the Genasun and see what voltage it produces.

Thank you for your input. I don't currently have a GV-5 on hand to test with. I plan to order one, probably a unit custom programmed to match the battery's charge parameters.

[Sunking]

Simon and Sunking,

Could the problem be as simple as the controller supplying charging voltage that's too LOW for the battery? The controller's specs show a CV voltage of 14.2V, and the battery's specs show a charging voltage of 14.6V. Is it likely that the BMS would disconnect the battery because of a 0.4V difference?

No the Genasun is configured to Float a LFP battery just below 100% SOC. 14.4 is the voltage for 100% or 3.6 volts per cell.

Your battery with two ports has a HVC on the input, and LVD on the output. Both input and output are the point electrically across the battery, but use a relay contacts to open/close the two paths. Both the Controller and battery are stand alone units that do not communicate with each other. That is a major problem with Lithium batteries and their BMS. To operate effectively both the charger and BMS must be able to communicate to be effective.

Your battery appears to disconnect the Charge source when the battery is at or near full charge. For a Stand Alone product is OK. Again as I said a long time ago not ideal for Solar Applications.In a Solar Application you want to Float the Battery so when it is fully charged and still usable sun light, you use solar power instead of battery power. If you manage to recharge by say noon, and disconnect the controller, you run off battery until th enext morning when the cycle starts again. Does that work? Yeah it does but a piss poor design concept IMO as you are not utilizing all the power from the panels, and causing excessive wear on your battery that need not be.

You battery judging from the connectors type is intended for Ham Radio operator market to charge it from a bench power supply. When the battery charges up, it disconnects from the charge source. Then use it on a mobile radio. Why do I think that? Because I am KF5LJW. I have built a many DC power supplies in 40 years as a ham operator.

73's

SK

[Sunking]

Incorrect. The schematic shows a standard nonisolated buck converter.

Semi correct Jeff. If there is No Load (battery), there is no place for the capacitor to discharge. Read any MPPT controller instructions. Thou shall connect battery first before connecting the panel. Thou shall disconnect panel before connecting battery. There is a damn good reason the instruct you to do that. Solar MPPT controllers are not designed as Switch Mode Power Supplies. They must have a load to operate into. To make them behave like a SMPS would require more circuitry and expense for something that is not required or needed.

Take any MPPT controller you want and try it. Disconnect the battery with bright Sunshine on the panels. One of three things is going to happen:

1. Voltage out will rise to Vin.
2. Let the magic smoke out and never work again.
3. Spills all the ones and zeros out going stupid.

What do you have to loose to prove me wrong? I can tell you without a doubt a Genasun will rise to Vin or about Voc of the panel. It clearly states that in their documentation. Most others will let the magic smoke out, or go stupid. Try it if you got the balls to do it.

[Dave C]

Your battery appears to disconnect the Charge source when the battery is at or near full charge.

I don't think that's what happens. The mfr. said that when it reaches full charge, it backs down current to a few mA, and expects CV voltage.

[PNjunction]

For my backpacking solar setup, I use a PowerFilm R28 rollable 28W panel and a Bioenno Power 4.5 Ah LiFePO4 battery with internal balancing and BMS.

Good panel and I recommend Powerfilm highly. Strangely, their marketing material showing a nominal 15v or so, almost prevented me from buying them! I suspected a nominal rating, or something that won't scare unknowledgeable consumers who don't know diddly about solar. I was right.

Take your panel into the sun, and sure enough, the open circuit voltage is anywhere from 18-22v like most "nominal 12v" panels. Even in high-heat conditions laying directly on concrete. Anyway, I've covered that elsewhere. Powerfilm is spendy, but totally worth it if you need to go that direction.

Here's my problem: I currently have a compatible controller from Bioenno Power. However, it's PWM, not MPPT, so it doesn't maximize solar panel performance, plus it requires a button press to display status.

Understand that BOTH mppt and pwm use the pwm technique to control the last CV voltage stage. The mppt part is more beneficial when the battery is heavily discharged. Yes, it is more efficient, BUT consider that LFP's super low internal resistance, AND slightly higher operating voltage, puts you into a better part of the "mpp curve" with a simple pwm controller to start out with! So the question is are you not achieving full charge with your existing pwm controller, or are you spending a lot of money drinking the marketing kool-aid?

Understand that I fully endorse mppt, and the fact that it IS more efficient, but if you don't take your application needs into account, you could be just wasting money on technical improvements you may not take advantage of often enough to justify it. Kind of like putting a Chrysler Hemi into Toyota Corolla. Anyway, getting off track here...

I tried the Genasun GV5, which was small, light, and had great features, but with one fatal drawback: it passed panel open circuit voltage (about 18V if I recall correctly) to the battery terminals when the battery was fully charged. I think this was because the GV5's BMS conflicted with the battery's BMS.

It is operating properly. There is no conflict. When charging is finished according to the Bienno protection circuitry, it OPENS THE CIRCUIT to the battery. Thus, you are measuring the OCV from the Powerfilm panel.

Yes, it does seem Genasun does it backwards - most controllers won't pass any voltage if no battery is attached. It is in fact a major newbie mistake resulting in many demands for replacements or refunds. They "test" the controller with no battery attached, and don't realize that controllers MUST have a battery for proper operation. This is *critical* before making assumptions.

Of course once a battery IS attached, the Genasun charges with appropriate voltages, not just a simple ocv dump.

As for the error status of too high a voltage: this is a protection against you connecting it to say a 24v battery. In the case of no battery at all, this is NOT A NORMAL SETUP, and the error you receive is NOT NORMAL either. The fact that it puts up ANY error message to warn you that you have a major battery problem (missing!) is a feature, not a bug.

In other words, it cannot electrically detect the difference between a totally dead zero-voltage battery, and one that is completely missing! If they had an error of "low voltage", that could be just as misleading when it can't electrically tell the difference if there is one physically missing or in fact a totally fried 0-volt LFP attached.

Had you used an LFP battery with no internal circuitry, like a Shorai or Antigravity LFP motorsports LFP, you wouldn't have second-guessed Genasun!

Warning - don't go too gaga over Bienno-power batteries. That is, if you start thinking big (20ah or more), doing large parallel / series custom packs of Bienno may be waaaay more expensive than just 4 Winston/GBS/CALB cells in series...

Understand that most responsible manufacturers will put out a "CYA" liability specification for a bms / protection circuit. Those customers that don't understand LFP, or in fact batteries at all, may seek litigation, refund-scams and the like when their improperly engineered diy system goes poof. The bms requirement helps cover that.

[Dave C]

Understand that BOTH mppt and pwm use the pwm technique to control the last CV voltage stage. The mppt part is more beneficial when the battery is heavily discharged. Yes, it is more efficient, BUT consider that LFP's super low internal resistance, AND slightly higher operating voltage, puts you into a better part of the "mpp curve" with a simple pwm controller to start out with! So the question is are you not achieving full charge with your existing pwm controller, or are you spending a lot of money drinking the marketing kool-aid?

Understand that I fully endorse mppt, and the fact that it IS more efficient, but if you don't take your application needs into account, you could be just wasting money on technical improvements you may not take advantage of often enough to justify it. Kind of like putting a Chrysler Hemi into Toyota Corolla. Anyway, getting off track here...

Wow, thanks. Having zero knowledge of solar, this is exactly the kind of advice I've been hoping for.

It is operating properly. There is no conflict. When charging is finished according to the Bienno protection circuitry, it OPENS THE CIRCUIT to the battery. Thus, you are measuring the OCV from the Powerfilm panel.

This is the piece of information that was missing. Had Bioenno told me this back in September, it would have saved me many hours of testing, phone calls, emails, a return of the GV-5 and posts in this forum. Actually I deduced this last night. Originally, they told me that the battery backs down to a few mV when fully charged, so I assumed that it didn't disconnect in the course of normal operation. But in the course of conversing yesterday by email, it came out that after shifting to CV, when the battery voltage hits 14.6V, the charge current goes to 0 A.

Yes, it does seem Genasun does it backwards - most controllers won't pass any voltage if no battery is attached. It is in fact a major newbie mistake resulting in many demands for replacements or refunds. They "test" the controller with no battery attached, and don't realize that controllers MUST have a battery for proper operation. This is *critical* before making assumptions.

From my experience with Genasun's tech support and customer service back in September, I would say that they needed to improve their communication with newbies. Actually, in all of my communication with both manufacturers, I didn't have all of the data I needed to make an intelligent decision until just now. Kevin at Bioenno has always gone the extra mile in terms of his availability--he's even made unsolicited follow-up calls. And I have a much better impression of Genasun's customer service in recent conversations.

Of course once a battery IS attached, the Genasun charges with appropriate voltages, not just a simple ocv dump.

As for the error status of too high a voltage: this is a protection against you connecting it to say a 24v battery. In the case of no battery at all, this is NOT A NORMAL SETUP, and the error you receive is NOT NORMAL either. The fact that it puts up ANY error message to warn you that you have a major battery problem (missing!) is a feature, not a bug.

In other words, it cannot electrically detect the difference between a totally dead zero-voltage battery, and one that is completely missing! If they had an error of "low voltage", that could be just as misleading when it can't electrically tell the difference if there is one physically missing or in fact a totally fried 0-volt LFP attached.

Had you used an LFP battery with no internal circuitry, like a Shorai or Antigravity LFP motorsports LFP, you wouldn't have second-guessed Genasun!

All helpful info. Thanks.

Warning - don't go too gaga over Bienno-power batteries. That is, if you start thinking big (20ah or more), doing large parallel / series custom packs of Bienno may be waaaay more expensive than just 4 Winston/GBS/CALB cells in series...

The main selling point to me of Bioenno was that their batteries do internal balancing. To me, this made using LFP in my application possible. I had read that LFP batteries degrade quickly if they're not balanced, and the required input current for the balancing chargers I found was too high in a single-panel solar situation. So internal balancing did it for me.

Understand that most responsible manufacturers will put out a "CYA" liability specification for a bms / protection circuit. Those customers that don't understand LFP, or in fact batteries at all, may seek litigation, refund-scams and the like when their improperly engineered diy system goes poof. The bms requirement helps cover that.

I wouldn't be inclined to take on a diy LFP project with no solar experience and limited electronics experience.

So, it sounds like I would be fine with the 14.2V GV-5 and the Bioenno battery. Except I have one more question: When I asked Kevin what is the max voltage his BMS can handle without frying, he answered 15V. Well, the BMS was hit with much more than that every time I connected it to the GV-5. So, I wonder if the BMS might be damaged. It doesn't seem so. The battery still works fine with the Bioenno controller. I wonder if a steady diet of seeing 18 - 20V will damage it over time.

I like the GV-5 because it's small, light and marine-grade (good for backpacking), and it displays status in a way that makes sense to me (not requiring a button push). If I'm certain that I can ignore the error status and high voltage when the battery is fully charged, I'll order another GV-5.

[jflorey2]

Semi correct Jeff. If there is No Load (battery), there is no place for the capacitor to discharge.

Nor is there any place for it to charge. Again - you do not need a load to guarantee a buck converter will not go overvoltage.

Read any MPPT controller instructions. Thou shall connect battery first before connecting the panel. Thou shall disconnect panel before connecting battery. There is a damn good reason the instruct you to do that.

There is indeed a damn good reason to do that. On MPPT controllers, the housekeeping supply (the power supply that provides power to run the oscillator, gate drive, feedback amp, processor etc) is powered from the SECONDARY or load side. This is because the power from the PRIMARY or line side goes away sometimes (i.e. at night) whereas the power on the secondary side (the battery) never goes away while the system is commissioned. On an AC/DC SMPS, the housekeeping supply is powered from the PRIMARY side. That's because on an AC/DC, there is no power on the secondary side.

So if you apply power to the primary side of an SMPS converter, several things might happen:
1) Smoke (unlikely)
2) Nothing. No power to the gate drive, so no switching, so no output.
3) Ugly startup. If there is enough leakage through the system (due to for example a floating gate) then the output voltage will start creeping up. Once the output voltage is enough to start up the housekeeping supply, the device will go into regulation.

[Sunking]

So if you apply power to the primary side of an SMPS converter, several things might happen:
1) Smoke (unlikely)
2) Nothing. No power to the gate drive, so no switching, so no output.
3) Ugly startup. If there is enough leakage through the system (due to for example a floating gate) then the output voltage will start creeping up. Once the output voltage is enough to start up the housekeeping supply, the device will go into regulation.

All true Jeff except a Solar MPPT Controller is not SMPS. They are Buck Mode Current Sources. Open th eload on any current source and you know what damn well happens. If you don't, then you are not who you say you are. .Every MPPT controller has large electrolytic capacitors on the output. Those caps must have a place to discharge. True if you disconnect the battery, the gate driver pulse width goes to minimum, (not 0% not thus shutting off the switching transistor completely), The output caps cannot discharge without a load or battery connected. For Petes Sake try it rather than preaching theory on SMPS when a controller is not a SMPS. I have tried it on many on many models, and have schematics of a few MPPT controllers. .

This is all pointless as we already know for fact the Genasun goes to Voc when the battery is removed.

[Sunking]

Dave don't feel bad, even lots of engineers do not fully understand Lithium Batteries. They are completly different than any other battery chemistry. One of the huge problems is the technology is still in development and there is no Industry standards. The biggest miss so far is what is a BMS? Ask 100 people and you get 100 different answers. There is NO definition. Every manufactures has a proprietary solution.

A BMS can be as simple as a volt meter reading pack voltage. It can be a cell voltage monitor that read individual cell voltagges. They can read cell temps, , they can discharge cells, they can disconnect chargers, they can disconnect the load, they can balance cells either by bleeding them or taking power from one cell and transfer it to another cell. They can be passive or active. They can be centralized or distributive. They can be analog or digital. They can be anything you want or not want.

There is only one market sector where standards are being pushed, and that market is the Electric Vehicles using CAN BUS, an ASE proprietary communication protocol like Internet Protocol is a standard communications protocol. As of now using lithium batteries in solar systems is not feasible, thus not a lot of things out there. There are a few companies that have integrated lithium products, but each one is proprietary and will not work with other manufactures.

Your own battery is a great example. Most integrated lithium batteries look and behave like any standard Pb battery. They only have one connection point or a positive and negative terminal. Inside they do have a form of a BMS, but very limited using Cell Balance Boards, and some may have a LVD. The only requirement is the charge source has to be a CC/CV at a specific voltage. For 12 volt batteries using LiFeP04 is 14.2 to 14.4 volts. Your battery has more inside the box and keeping is somewhat secretive. It is using two ports, one for charge, and one for load. The charge side has some form of a High Voltage Cut-Off circuit. Perhaps HVC might be the wrong term, but the effect is it disconnects the charger source when the batter is charged up. Solar Charge controllers do not like that as you have discovered. It is incompatible with most charge controllers. Your battery appears to be made to use a DC Power Supply that must supply 14.6 volts and limited to 2 amps maximum output. A Solar Charge Controller is not a DC Power Supply.

What PNJunction and I have been alluding too is if you know how Lithium batteries operate, what the Do’s and Don’t are, a person does not need much in the way of Automated BMS. You can use a Protein based BMS aka common sense and strategy. Joe consumer cannot do that. Lithium batteries are dangerous if abused and capable of self-destruction and worse. However for Joe Consumer and equipment manufactures POV an Automated BMS must be used to make Lithium batteries safe. However to do that you have to give up cash and some battery service life. Unfortunately there is no standards to say how it is to be done and compatible with any manufacture like there is with all other battery types.