Questions regarding battery type and what voltage moving forward

[georgia088]

I would suggest, especially with mosfets to start a thread on an electronics forum.

Beware of the body diode.

On PCBs. I use KiCAD. PCBWay or AllPCB, depending on which one has the cheaper "fast" shipping. Takes about a week. PCBs... £7. Shipping £35.

I agree surface mount is a lot less scary than you think it will be. Just get good magnification and VERY bright lightening. As bright as you can stand without getting a headache. You can pick up cheap microscopes for £30. I would recommend, considering going to one that at least has it's own screen etc, ideally an HDMI output. ~£120.

Hot air can be awesome, definitely worth it for a whole board with dozens of parts. You can also use the iron and a nice flat tip and for some chips it's faster than air. I especially found soldering an ESP32 wifi daughter card to a PCB was a pain in the butt with air. Far too much mass. The iron filled all those little castlated contacts perfectly in a quarter the time.

I also use a syringe and paste. I am however buying one of the pump handles for it as that stuff takes a LOT of force to get moving out the needle.

If the board gets beyond about 20-30 components I'll fork out the £10 for the stensil next time. Putting down the paste is nearly as fiddly as placing the components.

Heres a video I made when I was learning how to do surface mount ICs. The magic is in how the solar paste flows onto the chip legs itself, like magic!

https://www.youtube.com/watch?v=rgcupYOnvBo

I will get on an electronics forum to discuss the mosfet and other PCB component issues. I just dabble in it and obviously far from an expert. You seem to be quite knowledgeable when it comes to electronics, which forum do you suggest?

Also, thanks for the video! I may be giving smd soldering a try if I can convince the wife this hobby is worth a few new toys!

Thanks!

[venquessa]

That is not true. The charge controller profile should support the battery type being charged. Maximum charging voltages and end point for a charge cycle will vary considerably between lead acid and lithium type batteries.

With my RV the use of lithium-phosphate batteries cut my charge times from the solar panels by more than 30% compared to the original lead acid batteries. But this required a charge controller with a charge profile for this type of battery.

It's one of those things. The more layers you peel back the more contradictions emerge in the various levels of interpretation. Sort of like the dunning-kruger effect.

I can't argue as your approach errors by a factor on the side of caution, rather than caveliery (sp?).

I will go as far as to say... In the specific case of LiFePO4 cells in 4S~12V 8S~24V and 16S~48V configs. Their "ideal" charge characteristics and phases are very close, to similar lead acids. Also their tolerances beyond and around their "ideal" is actually wider than the lead acid they are replacing. They will stand up to MORE abuse than the LA they replaced. They are sold internationally as "drop in replacements" in everything from cars to golf buggies.

The reason it is still good to continue the strict "Don't mess with lithium, do it 100% right all the time" message and it is EXTREMELY important.

Not all lithium cells or packs are the same. "Lithium batteries" cover a fair vast range of battery types, cell types, charge voltages, tolerances and worst of all, differing severity of outcomes for breaching those tolerances.

Take for example a radio controlled helicopter battery pack. It might be an 4S pack, but that is not a 12V battery. It has a peak voltage of 16.8V. A 16.8V, maybe 5Ah, pack that will happily run a 500W motor for 10 minutes and happily provide short bursts of 2000W. So called "High C rating. LiPo". Often they are sold as having peak outputs of 30, 40, 50C and sustained output for 10-20C. From a 5 Amp hour pack the size of a portion of lasange! Not weighting much more either! However, these packs will burn your house down if you don't treat them with extreme respect. Raw pouch cells which swell and get warm in use! Change their pack voltage by 10% due to temperature. Have an over-voltage tolerance of 100mV is you are lucky and will start to get hot and then vent tongues of fire not much more than that 4.3V per cell. They have no BMS, no fuse, no case, no shell, no protection what-so ever, no interior balancing. Literally if you short one, or your charger goes on the fritz and shorts on.... you have literally seconds before it swells, goes thermal runaway and them into full venting fire meltdown.

There are at least a dozen different lithium battery types. The reason LifePo4 are popular is because they are far more laid back than the other chemistries. The cell in your mobile phone is of the middle ground type. Same chemistry as the high discharge RC pack, but with a very, very gentle and diligent BMS and power management system to keep it healthy. I believe Tesla also use that chemistry and it does require a very smart BMS to keep a pack that size healthy for 100k miles.

EDIT: The other aspect is "Life span". Tolerances and lifespan tend to end up in the same equation. That's undoubted. If you run the pack at 100% of it's capabilities, red line to red line at the red line every day, day in day out, it will still last longer than the lead acid would. It will not however last as long as the pack you baby sat and nursed, were gentle with and added your own "comfort" margin and "engineering grace" to. Just be careful, in many instances you can go too far and do worse by the battery by babying it too much. Keeping it constantly fully charged, like you would a lead acid, is not ideal.

Again, we need to look at what people are typically coming from and what they are going to. Lead acids with a cyclic life span to 80% capacity of 800-1000 with a 50% DoD. They cost $1 per Ah. LFP packs which, if kept below 14.40V and above 10V will do over 3000 cycles at 100% DoD. If you baby them, say only use 90% DoD that goes up to 4000 cycles to 80% capacity. That's over 10 years of daily cycles.
(*All figures IIRC but can be looked up on the datasheets for batteries and cells independantly).

[georgia088]

It's one of those things. The more layers you peel back the more contradictions emerge in the various levels of interpretation. Sort of like the dunning-kruger effect.

I can't argue as your approach errors by a factor on the side of caution, rather than caveliery (sp?).

I will go as far as to say... In the specific case of LiFePO4 cells in 4S~12V 8S~24V and 16S~48V configs. Their "ideal" charge characteristics and phases are very close, to similar lead acids. Also their tolerances beyond and around their "ideal" is actually wider than the lead acid they are replacing. They will stand up to MORE abuse than the LA they replaced. They are sold internationally as "drop in replacements" in everything from cars to golf buggies.

The reason it is still good to continue the strict "Don't mess with lithium, do it 100% right all the time" message and it is EXTREMELY important.

Not all lithium cells or packs are the same. "Lithium batteries" cover a fair vast range of battery types, cell types, charge voltages, tolerances and worst of all, differing severity of outcomes for breaching those tolerances.

Take for example a radio controlled helicopter battery pack. It might be an 4S pack, but that is not a 12V battery. It has a peak voltage of 16.8V. A 16.8V, maybe 5Ah, pack that will happily run a 500W motor for 10 minutes and happily provide short bursts of 2000W. So called "High C rating. LiPo". Often they are sold as having peak outputs of 30, 40, 50C and sustained output for 10-20C. From a 5 Amp hour pack the size of a portion of lasange! Not weighting much more either! However, these packs will burn your house down if you don't treat them with extreme respect. Raw pouch cells which swell and get warm in use! Change their pack voltage by 10% due to temperature. Have an over-voltage tolerance of 100mV is you are lucky and will start to get hot and then vent tongues of fire not much more than that 4.3V per cell. They have no BMS, no fuse, no case, no shell, no protection what-so ever, no interior balancing. Literally if you short one, or your charger goes on the fritz and shorts on.... you have literally seconds before it swells, goes thermal runaway and them into full venting fire meltdown.

There are at least a dozen different lithium battery types. The reason LifePo4 are popular is because they are far more laid back than the other chemistries. The cell in your mobile phone is of the middle ground type. Same chemistry as the high discharge RC pack, but with a very, very gentle and diligent BMS and power management system to keep it healthy. I believe Tesla also use that chemistry and it does require a very smart BMS to keep a pack that size healthy for 100k miles.

EDIT: The other aspect is "Life span". Tolerances and lifespan tend to end up in the same equation. That's undoubted. If you run the pack at 100% of it's capabilities, red line to red line at the red line every day, day in day out, it will still last longer than the lead acid would. It will not however last as long as the pack you baby sat and nursed, were gentle with and added your own "comfort" margin and "engineering grace" to. Just be careful, in many instances you can go too far and do worse by the battery by babying it too much. Keeping it constantly fully charged, like you would a lead acid, is not ideal.

Again, we need to look at what people are typically coming from and what they are going to. Lead acids with a cyclic life span to 80% capacity of 800-1000 with a 50% DoD. They cost $1 per Ah. LFP packs which, if kept below 14.40V and above 10V will do over 3000 cycles at 100% DoD. If you baby them, say only use 90% DoD that goes up to 4000 cycles to 80% capacity. That's over 10 years of daily cycles.
(*All figures IIRC but can be looked up on the datasheets for batteries and cells independantly).

Interesting info. I thought that most solar charges even the cheap ones had a “lithium profile”. Do they not? I am not sure that these “profiles” on the cheap ones actually do much or anything different than they do with the lead acid, but I don’t know. I did notice on your YouTube Chanel you are playing with a mppt charge controller. Did you design a mppt charge controller? Are you actually using it in practice or was it just experimental?

As to the 12v or 4s lithium pack I am contemplating getting for my dock, would you think it is a must to have a bms with interior cell balancing? I know you will not want to answer this for fear of persecution and liability, but if fused properly and placed in a case/area on a metal dock where fire risk was minimal, are they any more dangerous than the drone or helicopter batteries you described earlier laying under your bed in your house? Just curious as to your thoughts. I know safety is the utmost importance and some people will scold me for even asking the question, but most people wear seat belts in cars yet none where seat belts and helmets in cars. The latter would be safer….

Enjoying the conversation and learning a lot from everyone.

Thanks!

[georgia088]

My posts keep getting flagged as possible spam and must be approved by a moderator. Any idea why?

Thanks!

[SunEagle]

My posts keep getting flagged as possible spam and must be approved by a moderator. Any idea why?

Thanks!

It could be due to our antispam software or something you have posted or just about anything else the Forum software feels is bad. All I can say is try to be patient and a Moderator will review your post and Approve it.

[venquessa]

Interesting info. I thought that most solar charges even the cheap ones had a “lithium profile”. Do they not? I am not sure that these “profiles” on the cheap ones actually do much or anything different than they do with the lead acid, but I don’t know. I did notice on your YouTube Chanel you are playing with a mppt charge controller. Did you design a mppt charge controller? Are you actually using it in practice or was it just experimental?

Personally, I wouldn't have a charge controller that didn't expose all of the parameters and allow most or all of them to be user editable. However it is extremely important that you DO set it up. Buying a controller which by default supports lithium or has a "lithium" mode specifically designed for LFPs is a good start, but I would still want to verify what parameters that lithium mode uses. They aren't always to every ones tastes by default.

The only 2 I've owned have been EPEver. Cheap, without being stupidly so, those 40A inductor coils cost a lot on their own! Out of the box if you want "lithium" you have to buy the more expensive lithium model. The "normal" ones allow you to select the battery type from the LCD panel, but does not include Lithium. There is a "USER" battery type which you can select, but not configure on the LCD. This is quite common, the LCD provides basic stats and control, but the obligitory USB cable to the PC and you can define the battery profile. It's a total of about 25 settings.

Like most of them, they have many "phases", Bulk, Boost, Equalisation and Float, being the most common.

For lithium. I get rid of the equalisation charge (defaulting to 15.0V or 15.5V!), setting it's voltage to the same as "boost" and it's time to 0. Set the max charge voltage and the boost voltage to 14.40. Overvolt disconnect 15V. Set boost time to 30 mins (some prefer shorter). Float voltage I set to 13.6V. It picks up the loads nicely there in pass through. It will remain in float unless due to clouds or night fall the battery is pulled below the "boost recon" voltage setting, when it will switch back into bulk, then boost. You can also set the max charge current.

What you might get from a true lithium profile are things like active current limiting during "absorption" to allow top balancing BMS's to keep up. If you are really lucky (expensive) you'll get one that has "comms" and a battery from the same brand which can "talk back" to the MPPT and tell it what it wants.

As to the 12v or 4s lithium pack I am contemplating getting for my dock, would you think it is a must to have a bms with interior cell balancing? I know you will not want to answer this for fear of persecution and liability, but if fused properly and placed in a case/area on a metal dock where fire risk was minimal, are they any more dangerous than the drone or helicopter batteries you described earlier laying under your bed in your house? Just curious as to your thoughts. I know safety is the utmost importance and some people will scold me for even asking the question, but most people wear seat belts in cars yet none where seat belts and helmets in cars. The latter would be safer….

Without cell balancing, interior or exterior, the issue you will encounter is pack charge disconnections before the pack hits it's full charge voltage. Without a BMS, you will overvolt a cell eventually. There are techniques to put a set of cells into a state of balance and just let them stay there. I'm not convinced.

I ran my 4S pack without a BMS at all for a month. March. As the sun started to get stronger, days longer, the pack started to get fully charged. First time I baby sat it. Carefully watching. One cell got to 3.7V with a 200mV spread across the pack. Not a big problem. I used the battery mostly before the next sunny day. (It's ireland!), next time it got to fully charged one cell hit 4.0V while the charge controller was still seeing 14.34V and still pushing 15 amps. I had to intervene. For 2 weeks, if it was sunny, my options where to nurse the thing with 1 and 2 ohm loads, occasionally supporting the low cells with a carefully placed and calculated bench power supply (also pushing back against the MPPT with that), or if I had to go out to the shop, turn the panels off.

If I had a non-balancing BMS, the instant that first cell hit 3.70V the whole charge circuit would disconnect (diode blocking the MPPT). The cell would settle back slowly to 3.65V and the charge reconnect. Depending on the BMS it might do that all day long. Others have a back off tolerance. "1 minute disconnect. 3 triggers inside 5 minutes, sleep for 1 hour." type thing.

This is still possible WITH a balancer and you do what it this way. It's just that from 3.3-3.4V and 5-10mV min/max a balancer kicking in (say above 75% SoC) can go a long way to helping prevent early charge termination due to cell imbalance occuring in the first place.

Similar on the way down. A single cell hitting 2.5V will undervolt disconnect the whole pack. This one is a little more serious and not for the batteries sake, but the charge controller. If your MPPT is pushng 20A from 100V's of panels and the BMS diode blocks it, that current is not going to stop, it has momentum, particularly in regards to the massive inductor. The controller should protect itself, but almost all of them state they should never be left connected to the panels when the battery is disconnected! So, just like my cooking of a multimeter in conditions it should not have cooked in, the manual did say "Don't do that." so no warranty claim. If that's a £40 EPEver 10Amp unit, fine. If it's a big victron £450 job, probably not fine.

I'm going to be experimenting with a fall back power source for the MPPT, Something it can't charge, with diodes, that will provide it with 24V surrogate battery for this event or the event a fuse blows or a breaker pops, leaving the MPPT to fight for itself against the panel surge.

Dogs nuts, JK-BMS is £35ish. £70 for the big ones. Cheaper ones with passive balancing and over/undervolt and even bluetooth are extremely common. There is a reason for that. The same factories that make them for LFP pack manufacturers sell them out the back door on Ali*/BG/Feebay/amazon. The difficulty there is knowing what you need and hoping it's what you ordered.

On liabilities. I am not worried about stating opinons and reusing facts I've learnt, no worries discussing electrics or even software. I'm just a "bloke on the net". It would be a different story if someone commissioned or paid me to write said software.

Your point on seat belts noted. If people were serious about safety as 1st priority they would have 4 point harnesses and full dual steel safety cages, just like world rally cars. The number of them I've seen go rolling down a mountainside at 80mph and the two fellas get out and walk away!

A point I like to give me is that unlike mountains, risks can be moved and offset. Just like the 4 point belts and roll cages. The point at which needs outweigh comfort and convenience say it's a bit much for popping down to pick up the kids.

On fire safety. I first point out, mines in a plastic box, because I'm lazy, cheap and haven't found a suitable metal one. It's in a garage with bare block walls, no cieling and no furnishings. The metal box is only mean to contain the heat. With bulk lithium cells, ala Aliexpress in the blue wrap, are already contained in a metal case. They do tend to hold onto their guts, they have to. Electric forklift fires are expensive. The "run away heat failure" mode tends to do exactly as you would expect. Generate a ton of heat, quite a lot of swelling pressure (580kg/m2 or higher). If they do vent or breach it will be entirely hot gas, possibly hot enough to set fire to soft furnishings or flamibles like paper/maybe, though unlikely wood. The box MUST be vented. Consider where that vent comes out, keep combustables well away from all of it and the battery/junctioning. Unless you seal the metal box a "BANG" is unlikely.

Something as simple as dropping a spanner/wrench into the battery box is something you should consider. People go as far as to use carboard sheets like a surgeon using a cloth when operating to only expose 1 or 2 terminals at a time.

Lead Acids generate both hydrogen and oxygen in perfect ratio to be turned back into water and if that happens, the flame front is transonic (IIRC). Aka instant. BANG! There are many guides on their safety too

[venquessa]

The MPPT controller (the video with the screen showing the in/out/load currents).

I am designing building it, but it's really just learning. First lesson. Basic beginners buck converter on a breadboard has rubbish efficiency. Second lesson. Mosfets are tricky beasts.

The goal is to run the MCU (without the screen) from the solar panel and generate enough power so that it can connect to Wifi, send a blip of data once an hour and mostly do MPPT, or connect the panel to the battery and go back to sleep. The idea is, can I get it to self sustain and what size of panel and battery do I need. Starting with a 6V panel and 4.2V LiPo cell, I quickly discovered that is hardly enough voltage range to really flex the panel, so I went up to a 12V 3W panel. I got it to charge an 18650. However by the next morning it was down to 30% and by late afternoon I ended it. That was with the screen, but with the back light off.

The next version, requires a bit more detailed analouge electronics and that is to use an off-the-shelf buck converter PCB which has far, far higher efficiency that I can do and then take control of it by effectively putting my MCU into the feedback look of the buck converter. So I can not only vary it's output voltage based on MPPT power algorithms, but also on output current limits etc.

[georgia088]

The MPPT controller (the video with the screen showing the in/out/load currents).

I am designing building it, but it's really just learning. First lesson. Basic beginners buck converter on a breadboard has rubbish efficiency. Second lesson. Mosfets are tricky beasts.

The goal is to run the MCU (without the screen) from the solar panel and generate enough power so that it can connect to Wifi, send a blip of data once an hour and mostly do MPPT, or connect the panel to the battery and go back to sleep. The idea is, can I get it to self sustain and what size of panel and battery do I need. Starting with a 6V panel and 4.2V LiPo cell, I quickly discovered that is hardly enough voltage range to really flex the panel, so I went up to a 12V 3W panel. I got it to charge an 18650. However by the next morning it was down to 30% and by late afternoon I ended it. That was with the screen, but with the back light off.

The next version, requires a bit more detailed analouge electronics and that is to use an off-the-shelf buck converter PCB which has far, far higher efficiency that I can do and then take control of it by effectively putting my MCU into the feedback look of the buck converter. So I can not only vary it's output voltage based on MPPT power algorithms, but also on output current limits etc.

I’ve read and reread several times. Great info! I appreciate you sharing!

I’m a little confused about what you’re talking about when a cell under volts. If a cell under volts at 2.5 and there is not cell balancing with a bms, The solar is still collecting power, but there is no where to put it because the battery has “disconnected” So, the controller could be fried? What do the expensive chargers use to prevent this?

Thanks again!

[Ampster]

............... What do the expensive chargers use to prevent this?

It is pretty simple, when the sun is out, solar can charge a battery. It does not take expensive equpment to prevent that. It only takes simple logic. What users do to prevent that is to set the inverter to not discharge to the point of causing the BMS to disconnect. In my system I do not continue to discharge my battery below 50 percent becaus if we have a power outage at that point in time I want some reserve until the sun shines and I can charge my bateries again It is not rocket science. It is just simple logiic.

[venquessa]

Pretty much. The BMS disconnects are for emergency use, they shouldn't really be used for routine operation.

As an aside, was watching a video from Off Grid Garage on YT last night. He was discovering that when you turn off say "Discharge" on a BMS, say because you are only charging the battery and don't need discharge, that the charge side experiences about a 0.5V voltage drop and doesn't charge all the way up, the PSU say 57.60V the BMS says 57.11V. What gives?

"Beware the body diode."

The charge/discharge circuit is a bank of mosfets in a back to back config. If both are "ON" current flows in both directions unimpeded. However, if you turn "OFF" the discharge mosfets the current cannot flow "out" of the battery, but it can still flow in. The only path it has available to flow "in" while the "out" mosfet is OFF is via the body diode of the discharge mosfet. That diode has a voltage drop.

This is where a pretty intuitive, safe sounding, sensible thing can turn into something rather different. I think it's also an example where measuring voltages in more than one place is extremely important and almost entirely forgotten in many 'communities' in energy storage.

The issues here are not just the voltage drop. It's considering what it actually means. Your battery doesn't hit the full voltage? Nope. That 0.5V drop will create half a watt of heat for every amp you charge the battery with. Charge your big battery at 50A and one side of the mosfet bank will now be dropping 25 Watts. This is one of the reasons they have such large heatsinks.

Obviously that is non-ideal and if you are using the battery in anything like normal circumstances, both charge and discharge should be enabled.

I honestly don't know what exactly is or would be the failure mode of an MPPT when you very suddenly remove the battery from the system while it's running. My thoughts are that the battery is, electrically a massive block of capacitance and resistance. Suddenly removing that, well, it would be like removing the output capacitor on a buck converter, the power in the inductor has nowhere to go and will "free wheel" around. That should cause the PWM to go 0% duty cycle fairly instantly, but it will take a bit to disipate that current. However, additionally the MPPT is now completely at the mercy of the panels and whatever they want to do. Panels are just a bunch of diodes spread out onto a sheet of plastic. God only knows what kind of static voltages they can float their way to without any sizable "balast" to hold them at a sane reference. That's 90% conjecture though.

I have disconnected the battery from small MPPT controllers while the panels where in sun. It just went off. It still came back on when I reconnected the battery. So it's likely a YMMV type thing. One shot could kill it, or you could get away with it 100 times.

As a "layman" with pockets, you can buy into any number of off grid eco-systems and their varying components that will work together and give you "wizards" and pretty GUIs and apps to set up trigger points like that. Which say, shuts off the island inverter or connects it directly to grid when the battery SoC hits 35% or whatever you wish.

Without pockets, you need a few DIY skills, a familiarity with github and you can do that with most mid-range hardware. It's just not made easy.

Last night I got my new JKBMS feeding my "Grafana" dashboards with data. I'm really chuffed as that was end-to-end via about 3 hops. ESP32/ESPHome MCU/Code combo to do the bluetooth integration with the BMS and relay that via MQTT message bus (not my code). A "proxy" node of mine feeds off those, timestamps all messages, filters them and publishes them on the internal system. From there an influx_mqtt_bridge of mine takes the incoming data and inserts it into "influxdb" where of course grafana can make lovely pretty graphs and dashboards.

Grafana/influx is only one client of that data though. It's "public" on my LAN to anyone and anything. As a mickey mouse example of how that works, my automated hall and bathroom lights ask the solar panel if it's day time or not. 2 completely independent systems working together, because they all share a standardized/normalized message bus.

So, when I get the inverter wired you, you can absolutely bet the victron VEBus will be hacked and it will be switching to AC-In direct when the battery hits 20%.... OR any cell hits 3.0V.

Even better than that, I can if I choose, do actual proper active current tapering to help the BMS if it needs. I can change the MPPT output current limit dynamically based on, say, "cell_voltage_delta" or "max_cell_voltage" from the battery.

[georgia088]

"Beware the body diode."

The charge/discharge circuit is a bank of mosfets in a back to back config. If both are "ON" current flows in both directions unimpeded. However, if you turn "OFF" the discharge mosfets the current cannot flow "out" of the battery, but it can still flow in. The only path it has available to flow "in" while the "out" mosfet is OFF is via the body diode of the discharge mosfet. That diode has a voltage drop.

This is where a pretty intuitive, safe sounding, sensible thing can turn into something rather different. I think it's also an example where measuring voltages in more than one place is extremely important and almost entirely forgotten in many 'communities' in energy storage.

The issues here are not just the voltage drop. It's considering what it actually means. Your battery doesn't hit the full voltage? Nope. That 0.5V drop will create half a watt of heat for every amp you charge the battery with. Charge your big battery at 50A and one side of the mosfet bank will now be dropping 25 Watts. This is one of the reasons they have such large heatsinks.

Obviously that is non-ideal and if you are using the battery in anything like normal circumstances, both charge and discharge should be enabled.

I honestly don't know what exactly is or would be the failure mode of an MPPT when you very suddenly remove the battery from the system while it's running. My thoughts are that the battery is, electrically a massive block of capacitance and resistance. Suddenly removing that, well, it would be like removing the output capacitor on a buck converter, the power in the inductor has nowhere to go and will "free wheel" around. That should cause the PWM to go 0% duty cycle fairly instantly, but it will take a bit to disipate that current. However, additionally the MPPT is now completely at the mercy of the panels and whatever they want to do. Panels are just a bunch of diodes spread out onto a sheet of plastic. God only knows what kind of static voltages they can float their way to without any sizable "balast" to hold them at a sane reference. That's 90% conjecture though.

I have disconnected the battery from small MPPT controllers while the panels where in sun. It just went off. It still came back on when I reconnected the battery. So it's likely a YMMV type thing. One shot could kill it, or you could get away with it 100 times.

As a "layman" with pockets, you can buy into any number of off grid eco-systems and their varying components that will work together and give you "wizards" and pretty GUIs and apps to set up trigger points like that. Which say, shuts off the island inverter or connects it directly to grid when the battery SoC hits 35% or whatever you wish.

Without pockets, you need a few DIY skills, a familiarity with github and you can do that with most mid-range hardware. It's just not made easy.

Last night I got my new JKBMS feeding my "Grafana" dashboards with data. I'm really chuffed as that was end-to-end via about 3 hops. ESP32/ESPHome MCU/Code combo to do the bluetooth integration with the BMS and relay that via MQTT message bus (not my code). A "proxy" node of mine feeds off those, timestamps all messages, filters them and publishes them on the internal system. From there an influx_mqtt_bridge of mine takes the incoming data and inserts it into "influxdb" where of course grafana can make lovely pretty graphs and dashboards.

All this is interesting and things I have it plan to dabble in!! My last PCB design was basically an esp8826 (wemos d1 mini) with a “songle” type relay among other electronics. I use the esp8826 bc I haven’t familiarized my self with the much better slightly more expensive esp32 big brother. I’m not sure how much I’d have to learn about the differences from a coding standpoint? Any way my PCB is very similar to a Sonoff device but caters to my unique needs. These device are often flashed with software (ex. Tasmota) that utilize mqtt. I’m sure you are familiar with this, but I'm guessing your code is better or atleast more specific to your needs.

The MQTT is again, another area that I am familiar with, but too lazy to figure out. So I continue to use the Wi-Fi connection. My basic goal for my pcb was to be able to control light switches normally AND with Amazon Alexa. Using GitHub and some basic very clunky code. I’ve had them working for a couple weeks now and very pleased. As many “smart devices” as I have now, I am guessing the MQTT is probably the route I should go, but I’m not sure how it would work with ALEXA.

Sorry for the rant but you touched on some of the things that are similar to other projects I have going on. Back to this project. In what I am trying to do, The relays would be the primary “cut off” to prevent the sources from connecting. The mosfets and the body diode would be a safety. Hopefully never used.

Current will have to flow through the two “on” mosfets in normal use. However in normal use the current is flowing through the series diodes and both mosfets are “on” (current flowing unimpeded). But I think what I am understanding that you are saying is the voltage drop occurs when one of the series mosfets are off and current is flowing through it across the body diode. In my case in normal situations with the relays working as they should, if the current did back flow it should never reach the second source if the relay is “off”.

However, if this was the case the mosfets are not doing any good as a protection. So, is it worth doing?

I’m still a little confused. It still seems like the question is (in my diagram) does current flow across the body diode in the wrong direction when the mosfet is off? I have done more research and found where the circuit with two sets of series mosfets are used in many circuits. Battery chargers (maybe exactly what you are alluding to) and atmel uses them with multiple usb sources of inputs. I don’t completely understand the ins and outs of them. However, I feel like they must be something to them preventing back feeding. I’ll provide a few links I have found if similar circuits to what I’m proposing.

Thanks!

[georgia088]

I still feel like with the mosfets in a series if one mosfet is off. Current can only flow in one direction.

Here is some discussion I found interesting:

https://electronics.stackexchange.com/questions/79028/understanding-two-mosfet-with-sources-connected

Also, here is where an atmel data sheet shows the same circuit being used with multiple usb sources:


same basic diagram I am proposing is on pg 48 of data sheet from link below:

https://www.farnell.com/datasheets/2007959.pdf

[venquessa]

The ESP8266 is out of support. "Not recommended for new projects". Has been for a few years now. The ESP32, assuming you are using platforms and frameworks like "Arduino" and "PlatformIO" or "ESPHome" it's pretty much the same, but with extra bits. Like Bluetooth BLE, 2 cores, 240Mhz and 384K of RAM. The 2 cores things takes a bit of getting used to, especially outside of a "general purpose" OS where none of the sharp corners to cut yourself on have been softened. Usually the default set up is for Wifi and BLE to occupy one core and user code the other, so unless you go looking for them, you won't notice there are 2.

I still use 8266s myself. The EPEver Rs845 Wifi monitor is an ESP8266 running originally my code, but now someone elses as it did a better job

Mosfets are in used in many parallel/series configs for different types of protection. The Charge/Discharge circuit is I believe a variant of the "FET-Bridge" topology. The one you are after for stuff like reverse polarity protection are different. Great Scott has a video on it as do a few others. There are different types, so watch a few of those should help you decide.

Personally I'd place myself on the first decent from "mount stupid" in terms of Mosfet learning curve. I understand the basics, but if I conceive of a mosfet circuit and build and test it, the odds of it doing what I expected or something completely different is still around 50%!

Consider If you put the two mosfets such that their body diodes are facing each other. If both are off, no current can flow in either direction. You attach their gates on your + and - inputs in such a way that a higher voltage on - will shut off all current flow entirely, resulting in the reverse polarity being blocked by at least one mosfet and the body diode of the other.

[venquessa]

For your anti-parallel circuit I would look at SSRs/Relays. If you look at the pinout for them you will typically find they have at least 3 dry poles, sometimes 4. STDP and DTDP. Usually they refer to the contacts as "Normally closed" and "Normally open". They cannot be in both states (do check they are break-before-made though, there are specialist transfer relays which have "make-before-break" mechanics. You might find these in say a UPS.

If you put them mirror image on the batteries such that they are controlled with a single coil voltage and a single ON or OFF. Both relays toggle, one opens, one closes.

So it's up to you if you want to go with mosfets as additional protection or, you could consider more physical protections like reverse polarity diode and fuse. For over voltage you can do similar with zener diodes (let them go short with the fuse when an over-voltage occurs). If nothing critical is running off the system this is the cheapest protection. The only downside is... if/when you engage that protection it destroys itself pretty much in the process of saving everything else.

If you haven't yet looked into it. I'd start there. If you want to control the mosfets from the MCU you have to consider that the vast majority of mosfets require up to 10V/-10V on the gate and the gate capacitance can mean to get the stated ~nanosecond rise time through the resistive region, can require 3 Amp current spike. None of those things are things ESP32s or Arduinos like. The sure fire way (bad choice of words), is to use a tiny boost converter to take the 3.3V/5V up to 12V and then employ a mosfet driver IC.

[venquessa]

Consider this "Strawman". The big bad world would include your battery switch circuit.

The 18V zener. Adjust to taste. As long as it's low enough that connecting 24V to the 12V side will flow current via that diode and blow the fuse.... and high enough that your charge circuits etc never, ever come with 1V of that zener threshold.
It is extremely rudimentary, fails to cover many corner cases and is pretty much self destructive in failure. Not a circuit you can really test fully.

The other diodes give a low resistence path to any reverse polarity. Technically the zener will behave as a normal diode below it's zener threshold, but.. call it for "illistrative purposes".
rpp.png