diy solar

diy solar

4000 watt Inverter on 12 V?

Devin82m

New Member
Joined
Dec 15, 2020
Messages
94
Hey everyone, I know it's preferred to have anything over 2000 watts on a 24 v or 48 v system. I plan on building a 200 Ah battery pack using Fortune cells and was going to use a 4000 Watt inverter with it. Is there anything I need to look out for, like do I need large gauge wires, a special fuse size, more heat generation? I want to be able to power one of our three full sized refrigerators, router, modem, and a couple laptops and tablets if the power goes out in our home. I don't have a specific time requirement other than maybe a day at the most, but most like under 6 hours.

I was going to go with a Giandel 400 watt inverter, but I thought I better check here first. Will's wiring offerings put 4000 watts in the 24 v category, so do I need to buy 24 V cables for my 12 V system? I know people who run high wattage inverters on 12 V just fine, but those same people tend to rig everything too, so I want to be realistic, but don't have to be perfect or absolute ideal best practice. What is safe and reliable within reason?
 
4000 ac watts / .85 conversion factor / 12 volts low cutoff = 392 dc amps
392 dc amps / .8 fuse headroom = 435 fuse amps
Your bms should also be rated for 435 amps
4/0 awg cable with insulation rated to 105C is good for 440 amps.
 
Last edited:
Ouch, OK so the OverkillSolar BMS I was going to buy won't handle that much current, it's only rated 120 amps.

So if I understand this right, even for a 200 watt system I would need a BMS that can handle 245 amps?
 
Assuming you mean 2000 watt inverter...
2000 ac watts / .85 conversion factor / 12 volts low cutoff = 196 dc amps
196 dc amps / .8 fuse headroom = 245 fuse amps
Your bms should also be rated for 245 amps
 
OK, so I see Will has the Daly 12 V BMS that can handle 250 Amps, so that could work for at least 2000 Watts right?

To power a refrigerator with a potential to ramp up to 3000 watts when starting (I was told by someone on this forum) I would need a 48 V system, to power a 4k watt inverter? Otherwise I would need to do the parallel battery pack setup with 4 batteries/BMSs to handle 4k Watts and I wouldn't be increasing my AH doing that either?
 
You could put 2x overkill bms in parralel.

smoothJoey, so do those numbers still apply if I made it a 24 volt system, or are the number different? I mean the .85 and .8 bits. Same for 48 V?

I ask because I see people on here using a 24 V battery with a 100 Amp OverkillSolar BMS and a 3000 watt inveter and they seem to be doing fine. I used the numbers above but replaced 12 with 24 and I came out to 245 Amps that those people should be pulling through a 100 Amp BMS. Am I missing numbers or some concept?
 
smoothJoey, so do those numbers still apply if I made it a 24 volt system, or are the number different? I mean the .85 and .8 bits. Same for 48 V?
The .8 de-rate applies across across system voltages.
The .85 conversion factor is a good efficiency for the low end high frequency inverters.
They get a little more efficient at 24 and 48 volts but not significantly.
Higher quality high frequency inverters will be ~90% efficient.
High quality low frequency inverters will be ~90 efficient as well but we do the math differently to account for their substantial surge capability.
I ask because I see people on here using a 24 V battery with a 100 Amp OverkillSolar BMS and a 3000 watt inveter and they seem to be doing fine.
I used the numbers above but replaced 12 with 24 and I came out to 245 Amps that those people should be pulling through a 100 Amp BMS. Am I missing numbers or some concept?
3000 ac watts / .85 / 12 = 294 dc amps
294 dc amps / .8 = 367 bms amps
3000 ac watts / .85 / 24 = 147 dc amps
147 dc amps / .8 = 183 bms amps

Others may do their math differently.
If you think that driving a bms to 147% of its rated capacity is a good idea then go for it.
As I've said before I de-rate quality parts by .8 and "ebay specials" by .66 to .5.
There are very few components that I would run at 100% of spec.
What evidence do you have that these folks are "doing fine"?

The thing about fet based bmss is transistors are susceptible to heat stress.
The other thing about fet based bmss is the transistors have a habit of failing closed.
That means you won't notice a problem until it fails to protect your battery.
 
The .8 de-rate applies across across system voltages.
The .85 conversion factor is a good efficiency for the low end high frequency inverters.
They get a little more efficient at 24 and 48 volts but not significantly.
Higher quality high frequency inverters will be ~90% efficient.
High quality low frequency inverters will be ~90 efficient as well but we do the math differently to account for their substantial surge capability.

3000 ac watts / .85 / 12 = 294 dc amps
294 dc amps / .8 = 367 bms amps
3000 ac watts / .85 / 24 = 147 dc amps
147 dc amps / .8 = 183 bms amps

Others may do their math differently.
If you think that driving a bms to 147% of its rated capacity is a good idea then go for it.
As I've said before I de-rate quality parts by .8 and "ebay specials" by .66 to .5.
There are very few components that I would run at 100% of spec.
What evidence do you have that these folks are "doing fine"?

The thing about fet based bmss is transistors are susceptible to heat stress.
The other thing about fet based bmss is the transistors have a habit of failing closed.
That means you won't notice a problem until it fails to protect your battery.

OK, that's good info and makes sense. So by quality you mean something like Victron? Are there other "quality" brands? I believe in the case of the Overkill Solar BMS that would be 127% of it's rated capacity since the BMS is 120 Amp. I'm not sure what other BMS options are out there that are as good of quality as Overkill Solar that handle more amperage. I'm basing this on Will's suggestions only.

As for the case of a refrigerator, the surge seems to be the problem. Would the surge draw enough current to go over the max amperage of the BMS? Otherwise I read most refrigerators run at 650 watts on average after the milliseconds long surge of the compressor motor starting.

Are there non-transistor inverters? Like my linear power supply for my ham radio gear?

As for people claiming they are fine, I see one guy ran a 1500 watt space heat for 2 hours with no issue on 3000 watts 24 V. Another guy boiled a bunch of water with a hot plate and power other things at the same time on 3000 watts on 24V, I would assume those draws are large.

I just got my Fortune cells in 30 minutes ago. Those are my first prism sized cells.
 
My 20+ years old 25cu ft, 120VAC 6.5A as printed on the tag, draws about 26 ~ 27A when compressor starts up, then it will drop down to about 180 ~ 200W, when defrostor comes on it uses about 600W; I have to use 12V 3000W HF inverter (Reliable Power brand) to run the fridge. It will probably starts fine with LF 2000W inverter which will be the replacement on my next Inverter purchase.
 
OK, that's good info and makes sense. So by quality you mean something like Victron?
Yes
Are there other "quality" brands?
My fave is Samlex Evo

I believe in the case of the Overkill Solar BMS that would be 127% of it's rated capacity since the BMS is 120 Amp.
100 / 120 * 147 = 122.5%

As for the case of a refrigerator, the surge seems to be the problem. Would the surge draw enough current to go over the max amperage of the BMS? Otherwise I read most refrigerators run at 650 watts on average after the milliseconds long surge of the compressor motor starting.
Get an ac/dc clamp meter with peak capgture and a line splitter.
Knowledge is power.
Are there non-transistor inverters? Like my linear power supply for my ham radio gear?
Do you mean non-transistor BMSs?
If yes, yes but they are expensive.
As for people claiming they are fine, I see one guy ran a 1500 watt space heat for 2 hours with no issue on 3000 watts 24 V.
That is 50% load for 2 hours.
I am not impressed.
I've run a "1500" watt space heater on my 1500 watt high frequency inverter for 6+ hours.
BTW "1500" watt inverters actually draw 1440 watts.
15 amps * 120VAC * .8 de-rate = 1440 watts. ;)

I just got my Fortune cells in 30 minutes ago. Those are my first prism sized cells.
Congrats, those are reputed to be some tough cells.
 
How much power do you really need and for how long?

If you really did need 4,000 watts, that 200 amp hour 12 volt battery would not even run it for 30 minutes. But if you have a device like a fridge or A/C with a compressor, and it needs a surge to start, that 4,000 watt inverter may be needed to do the job, even if your constant draw is only 300 watts. I have a cheapo 700 watt 12 volt modified sine inverter, and it will run my fridge, but barely. A 1,000 watt would probably do the job just fine. And a good sine wave unit would be better for the motor, it will make a lot less heat and be more efficient.

Before you spend a lot of money on a good inverter, do a little math on the loads you want to run. Maybe invest in a Kill-O-Watt meter and see what they really draw. Anything with a motor, expect it to need about 5 times the power to get started. Low frequency inverters can usually handle double surge power (200%) for a few seconds. High frequency inverters are more like 150% surge, but only for a fraction of a second. My 700 watt inverter is just able to start my 230 watt fridge. Resistive loads like a hot plate and space heater are easy for an inverter, there is not surge. Large motors, especially starting a pump into pressure is where you have problems. Some older microwave ovens also had some pretty harsh start up load, but newer ones seem a lot better.

If you truly need 1,000 watts, and some of that is motors, I can see needing the 4,000 watt inverter to be reliable at getting everything started. A BMS is just like anything else. They have continuous ratings, and surge ratings. Not all of the specs will be published. My 200 amp JK BMS (also sold as Heltec) is rated to 350 amp surge current. In the setup app, I can set how much current it will take before it turns off on long term, and how many seconds I will let it pull more current before shut off. For 12 volt systems, the math is pretty easy. Figure 10 watts for every amp. Gives a decent derate for the efficiency and losses. 100 amps for 1,000 watts. 200 amp hours in a 12 volt system should run 1,000 watts for about 2 hours to complete discharge. A fridge should cycle, and only pull full power about 1/3 of the time, less if you don't open the door much, more if it is a pig like my old one here. Using the Kill-O-Watt, it can give you the real consumption over a period of time. I ran mine for 4 days to get a decent average, and realized how bad it is. But a new $2,000 fridge still won't pay off for 5 years.

Your constant run current should be under 50% of the rated current on a BMS if you want it to last. And your startup surge current should never exceed 200% of the constant rating under any conditions. Many do not list a surge rating. And those that do are rarely 200%. My JK Being rated to 350 amps on a 200 amp unit is 175% surge capacity. Most electronic components take time and heat up and fail. So they will usually take a short overload without permanent damage. It takes a second or so for the device to heat up. Even the fastest fuses have a time before they blow. The weak link in a BMS is the mosfets in the protection switch circuit. I just picked a random power FET from infineon and opened the data sheet. The IRFR120Z is rated for 8.7 amps continuous, but 35 amps pulsed. So it can take about 4 times the current for a short period of time. This is pretty typical of quality mosfets. Even cheap ones will do double. So if they actually use a bank of mosfets rated for 100 amps constant, it should survive a start surge of 200 amps. If you do go with the 4,000 watt inverter, you should have at least 200 amps of BMS to be on the safe side as the inverter could easily try to pull 400 amps when loaded up. 2,000 watts will pull 200 amps. My personal rule is that your constant current load should not exceed 100 amps. Yes, you can go higher, but the size of the wire, the fuses, and connectors, and any losses, just become a lot worse as the current climbs past 100 amps. At 1,000 watts a 12 volt cable will have 4 times the loss of the same cable running the same 1,000 watts at 24 volts. At 12 volts, not only do you have 100 amps of current, but a 1 volt drop is 1/12th of the voltage. At 24 volts, the current falls to 50 amps, so the same cable would only drop 0.5 volts, or just 1/48th of the system voltage. 100 amps at 1 volt is losing 100 watts of power. Losing 0.5 volts at 50 amps is just 25 watts of lost power.

Hope this helps
 
Ouch, OK so the OverkillSolar BMS I was going to buy won't handle that much current, it's only rated 120 amps.

So if I understand this right, even for a 200 watt system I would need a BMS that can handle 245 amps?
No. You need 400 to 500 amps to properly drive a 4000 watt inverter on 12 volts. 4/0 wire is marginal at best. Probably looking at 2x 4/0 or 350MCM. Far better to go 48 volts... or at least 24.
 
Why is everybody talking about running all of this massive current through the BMS?
Why not connect the inverter directly to the cells with a relay that the BMS can trigger? I thought that was the standard way of hooking up really big loads?
 
OK, you guys are helping me understand this better. Thanks a lot
How much power do you really need and for how long?

If you really did need 4,000 watts, that 200 amp hour 12 volt battery would not even run it for 30 minutes. But if you have a device like a fridge or A/C with a compressor, and it needs a surge to start, that 4,000 watt inverter may be needed to do the job, even if your constant draw is only 300 watts. I have a cheapo 700 watt 12 volt modified sine inverter, and it will run my fridge, but barely. A 1,000 watt would probably do the job just fine. And a good sine wave unit would be better for the motor, it will make a lot less heat and be more efficient.

Before you spend a lot of money on a good inverter, do a little math on the loads you want to run. Maybe invest in a Kill-O-Watt meter and see what they really draw. Anything with a motor, expect it to need about 5 times the power to get started. Low frequency inverters can usually handle double surge power (200%) for a few seconds. High frequency inverters are more like 150% surge, but only for a fraction of a second. My 700 watt inverter is just able to start my 230 watt fridge. Resistive loads like a hot plate and space heater are easy for an inverter, there is not surge. Large motors, especially starting a pump into pressure is where you have problems. Some older microwave ovens also had some pretty harsh start up load, but newer ones seem a lot better.

If you truly need 1,000 watts, and some of that is motors, I can see needing the 4,000 watt inverter to be reliable at getting everything started. A BMS is just like anything else. They have continuous ratings, and surge ratings. Not all of the specs will be published. My 200 amp JK BMS (also sold as Heltec) is rated to 350 amp surge current. In the setup app, I can set how much current it will take before it turns off on long term, and how many seconds I will let it pull more current before shut off. For 12 volt systems, the math is pretty easy. Figure 10 watts for every amp. Gives a decent derate for the efficiency and losses. 100 amps for 1,000 watts. 200 amp hours in a 12 volt system should run 1,000 watts for about 2 hours to complete discharge. A fridge should cycle, and only pull full power about 1/3 of the time, less if you don't open the door much, more if it is a pig like my old one here. Using the Kill-O-Watt, it can give you the real consumption over a period of time. I ran mine for 4 days to get a decent average, and realized how bad it is. But a new $2,000 fridge still won't pay off for 5 years.

Your constant run current should be under 50% of the rated current on a BMS if you want it to last. And your startup surge current should never exceed 200% of the constant rating under any conditions. Many do not list a surge rating. And those that do are rarely 200%. My JK Being rated to 350 amps on a 200 amp unit is 175% surge capacity. Most electronic components take time and heat up and fail. So they will usually take a short overload without permanent damage. It takes a second or so for the device to heat up. Even the fastest fuses have a time before they blow. The weak link in a BMS is the mosfets in the protection switch circuit. I just picked a random power FET from infineon and opened the data sheet. The IRFR120Z is rated for 8.7 amps continuous, but 35 amps pulsed. So it can take about 4 times the current for a short period of time. This is pretty typical of quality mosfets. Even cheap ones will do double. So if they actually use a bank of mosfets rated for 100 amps constant, it should survive a start surge of 200 amps. If you do go with the 4,000 watt inverter, you should have at least 200 amps of BMS to be on the safe side as the inverter could easily try to pull 400 amps when loaded up. 2,000 watts will pull 200 amps. My personal rule is that your constant current load should not exceed 100 amps. Yes, you can go higher, but the size of the wire, the fuses, and connectors, and any losses, just become a lot worse as the current climbs past 100 amps. At 1,000 watts a 12 volt cable will have 4 times the loss of the same cable running the same 1,000 watts at 24 volts. At 12 volts, not only do you have 100 amps of current, but a 1 volt drop is 1/12th of the voltage. At 24 volts, the current falls to 50 amps, so the same cable would only drop 0.5 volts, or just 1/48th of the system voltage. 100 amps at 1 volt is losing 100 watts of power. Losing 0.5 volts at 50 amps is just 25 watts of lost power.

Hope this helps

OK, you guys are helping me understand this better as we go, I really appreciate you all being patient with me. Sometimes I reask questions or reword them to make sure I understand it from all angles.

I went ahead and bought a Kill-A-Volt and will start collecting data.

I will explain a couple things for you coming late to the game.

1. I want to be able to power a normal sized LG refrigerator (it's about 4 years old), I would love to power our other refrigerator and our full upright freezer too, but know I don't have the money to do that.

2. I thinking 2 hours or so would be good enough. Power outages around here last 1 to 4 hours in the last 3 years we have lived here. We have a 7000 Watt generator I've never used, still new, but I want to convert it to propane before running it the first time (avoid gumming and fuel storage issues)

3. The only other stuff I want to power along with one refrigerator is a router, modem, possible a 46" LCD TV we bought new a few months ago, and 2 laptops. We can live without the TV and one laptop, we just have an ideal load.

4. I wanted to go with 12 V because it simplified things (in my mind), but it sounds like that won't work. I have given into the idea of 24 V, but 48 V is a little scary because of the voltage. Also the higher the voltage the less I get with my Fortune cells, I will need to buy a lot to make a 48 V system and that will defeat the purpose of a hand truck emergency setup from a size perspective.

5. I want the system to be safe, I'm not talking hold my hand safe, but reasonably reliable, easy to work on, and not a electrocution, fire hazard, or damage my appliances/electronics.

6. I don't have to go with a particular brand of anything, any I mentioned above are just ones Will mentioned, he is all I know as an authority and I know people in the forums disagree with him sometimes, so lay it on me. I'm very open to suggestions.
 
Why is everybody talking about running all of this massive current through the BMS?
Why not connect the inverter directly to the cells with a relay that the BMS can trigger? I thought that was the standard way of hooking up really big loads?
Not sure, I usually do advise people to do that.
I went ahead and bought a Kill-A-Volt and will start collecting data.
I'm pretty sure a kill-a-watt meter won't show you the locked rotor amps of the refrigerator which appears to be the determining factor for the inverter sizing.
 
4. I have given into the idea of 24 V, but 48 V is a little scary because of the voltage. Also the higher the voltage the less I get with my Fortune cells,.....
You do not lose any actual energy or capacity. You draw half the amps and twice the voltage (same wattage) for the same amount of time.
 
What is going to charge the batteries? Is this solar or just grid charge to have standby for an outage?

There are pros and cons to running the current through the BMS. In my case with the JK active balance "smart" BMS it is nice to be able to monitor the battery state, see the full charge and discharge current, have an accurate count of Amp Hours remaining, etc. When you bypass the BMS with the contact for the main load, you won't have that monitoring, but it will still protect the battery from discharging too low.

That's true, the Kill-O-Watt does not see the startup surge. A 2 year old LG Fridge should easily fit in the 5 x run current range. Many newer fridges have inverter compressors which are far less of a start surge. But 4 years old is a question.

A 12 volt system with four 200 AH cells will only provide 1,000 watts for 2 hours. That might be enough, but it will be close with the fridge, TV, PC's, etc. Going to 8 cells gives you double the power. 2,000 watts for 2 hours, or the same 1,000 watts for 4 hours. . You can run them in 8S series for 24 volt, or 4S2P and stay 12 volt, but with the higher current and thicker wire etc. If you have devices that will run directly off 12 volts, then you do have some reason for staying at 12. If it is all going to be running off of the inverter, then there is no big reason to stay at 12 and just go to 24.
 
If you have 8 100Ah Fortune then definitely make a 24V 8S pack.

12V 2P4S or 24V 8S are the same 2400 watt hours.
Hey everyone, I know it's preferred to have anything over 2000 watts on a 24 v or 48 v system. I plan on building a 200 Ah battery pack using Fortune cells and was going to use a 4000 Watt inverter with it. Is there anything I need to look out for, like do I need large gauge wires, a special fuse size, more heat generation? I want to be able to power one of our three full sized refrigerators, router, modem, and a couple laptops and tablets if the power goes out in our home. I don't have a specific time requirement other than maybe a day at the most, but most like under 6 hours.

I was going to go with a Giandel 400 watt inverter, but I thought I better check here first. Will's wiring offerings put 4000 watts in the 24 v category, so do I need to buy 24 V cables for my 12 V system? I know people who run high wattage inverters on 12 V just fine, but those same people tend to rig everything too, so I want to be realistic, but don't have to be perfect or absolute ideal best practice. What is safe and reliable within reason?
 
Back
Top