fusing and BMS questions on a proposed build...

I'm working on basically a homebuilt UPS, this is by far the best forum I've been able to find so thanks to everyone who's already helped (and how much stuff I've read already)... I've got a line on some half-used Valence LFP batteries and they're cheap enough to justify playing around with for my build... but I'm still learning a bunch of the second-level stuff. (The house we're in is basically 100% shaded so the solar part will have to wait for another property) Some questions based off the very basic starting point below...

• fusing. I know I need a fuse at least on the main busbar to inverter cable (P0 in the figure). Do I need one anywhere ELSE? If I fuse right at the busbar on cable "P0" then the only parts of the system that are live with respect to ground are the individual positive busbar cables (and the + busbar itself)... yes?
• grounding... is what's shown here sufficient (the inverter AC input ground tied to utility ground)
• expansion: let's assume I'm going to start with two of these batteries in parallel. Later on, is there anything that prevents me from adding another battery or two (shown in the shaded/hashed section with P3/P4 N3/N4 cables...)
• wire sizing: 2AWG for the busbar<->inverter connections (1500w % 0.85 for derating % 12.8v) -> 137A...
  • do the battery<->busbar cables need to also be the same size? They shouldn't, as these will not carry more than X/N worth of current, where X is the total current across the busbar and N is the number of batteries, right? Would I need to assume that at some point things fail to where there's a single battery powering the whole thing? Or do I just find whatever current value the BMS will cut-out, and size the wires for that?
• wire length: I think that the cable lengths need to be {P1=P2=P3=P4) and (N1=N2=N3=N4). I don't think that P0=N0, right? Isn't the requirement that all of the positive runs need to be the same total resistance, and all the negative runs need to be the same length? Or did I miss something here as well?
• separate BMS: the individual batteries all have their own internal BMS, but does it make sense for me to add some kind of external one to try and keep the 12v batteries balanced? and if so, what sorts of ratings do I need for the BMS?

Thanks again for all the super useful info so far...

ljwobker said:

fusing. I know I need a fuse at least on the main busbar to inverter cable (P0 in the figure). Do I need one anywhere ELSE? If I fuse right at the busbar on cable "P0" then the only parts of the system that are live with respect to ground are the individual positive busbar cables (and the + busbar itself)... yes?

Click to expand...

Yes. Your analysis is correct. However.....

For most batteries, the remaining 'hot' parts of the circuit still have the BMS short circuit protection. However, as I recall, the Valence batteries count on a distributed BMS model and I don't know if the batteries have any short-circuit protection without the distributed BMS system.

ljwobker said:

grounding... is what's shown here sufficient (the inverter AC input ground tied to utility ground)

Click to expand...

Yes. As long as the system is tied to the grid ground it should be ok.

ljwobker said:

expansion: let's assume I'm going to start with two of these batteries in parallel. Later on, is there anything that prevents me from adding another battery or two (shown in the shaded/hashed section with P3/P4 N3/N4 cables...)

Click to expand...

No, but the parts of the system that remain hot become increasingly dangerous as more batteries are added. (See the comments above about the Valence BMS)

ljwobker said:

wire sizing: 2AWG for the busbar<->inverter connections (1500w % 0.85 for derating % 12.8v) -> 137A...

• do the battery<->busbar cables need to also be the same size? They shouldn't, as these will not carry more than X/N worth of current, where X is the total current across the busbar and N is the number of batteries, right? Would I need to assume that at some point things fail to where there's a single battery powering the whole thing? Or do I just find whatever current value the BMS will cut-out, and size the wires for that?

Click to expand...

You should size the wire and fuse to the inverter capability. 2AWG is way too small for PO and N0

ljwobker said:

wire length: I think that the cable lengths need to be {P1=P2=P3=P4) and (N1=N2=N3=N4). I don't think that P0=N0, right? Isn't the requirement that all of the positive runs need to be the same total resistance, and all the negative runs need to be the same length? Or did I miss something here as well?

Click to expand...

Correct. Some people will tell you P0 and N0 must be the same, but that is not true.

ljwobker said:

separate BMS: the individual batteries all have their own internal BMS, but does it make sense for me to add some kind of external one to try and keep the 12v batteries balanced? and if so, what sorts of ratings do I need for the BMS?

Click to expand...

For parallel batteries, there is no need for an additional BMS to keep the balanced. However, as mentioned above, the valance batteries are designed for a central BMS controller. When I looked at Valence a year or more ago, I decided to give them a hard pass due to the BMS situation. Please research the Valence well so you know what it does or doesn't do.

Super useful. A couple followups:

• let's assume the valence batteries are all treated as separate batteries - totally ignore the fact that they have some #additionalValenceMagic that we might use... I'll just ignore that for argument. If they're "just plan normal parallel 12v LFP batteries" -- would I need a BMS/balancer?
• I don't follow the math for "2AWG is way too small", help? My math was : a max load on the inverter of 1500W. It's 90% efficient. We're running at 12.8V nominal. This gets me to (1000W / 0.9 / 12.8) * 1.25x --> 115A. Isn't 2AWG sufficient for 115A?

ljwobker said:

let's assume the valence batteries are all treated as separate batteries - totally ignore the fact that they have some #additionalValenceMagic that we might use... I'll just ignore that for argument. If they're "just plan normal parallel 12v LFP batteries" -- would I need a BMS/balancer?

Click to expand...

No

ljwobker said:

I don't follow the math for "2AWG is way too small", help? My math was : a max load on the inverter of 1500W. It's 90% efficient. We're running at 12.8V nominal. This gets me to (1000W / 0.9 / 12.8) * 1.25x --> 115A. Isn't 2AWG sufficient for 115A?

Click to expand...

A few things:
1) Most of the smaller inverters operate at closer to 85% efficiency, but maybe you have a better one so we can stick with 90%

2) I use the inverter capability for determining the fuse and wire size on the input. It is far too likely someone will plug in more than expected. Plugging that into the formula we get:
3) As the battery voltage goes down it requires more current to supply the same wattage to the inverter. Therefore, the worst-case scenario is when the battery is low. Because of this, I use 12V for calculating the max current on the input. Plugging that into the formula we get.
The good news is you can probably get a 175A fuse that will match this calculation almost perfectly.

Now let's look up the wire size using the more aggressive boating standard:


According to this chart, the min wire size would be a 2AWG...... so, using this chart, you were correct.

However, if the more conservative NEC chart is used we get a much different answer:



With the NEC charts, the min size would be 2/0AWG. If the build will need city inspection, it will need to use the NEC chart.

Note: When using the marine charts, the wires are smaller and therefore are going to run warmer. This results in lower system efficiency and a lower safety factor.

Super useful. My application (at least for the forseeable future) is substantially over-engineered for a 1500W inverter: I literally can't come up with a use case that consumes more than 1000W, and even that is a stretch. But the process of learning what values matter and where is quite useful, so thanks for that! My particular install is in a temperature controlled environment where the AbsoluteWorstCase is about 35-40C.

For 12v inverters the biggest enemy is voltage drop which 12v systems does not have a lot of margin to afford too much of.

You are thinking wiring heating but with 12v systems you usually have to dominantly think voltage drop between batteries and inverter.

Inverters regulate AC output voltage. Lower DC input voltage means they have to draw more DC current for same inverter output power. Inverter's efficiency drops as DC input voltage drops and input DC current rises, more DC current means more DC cable voltage drop. It is a compounding problem.

Wires gauge needed also depends on battery cable length. Fuses are better for 12v inverters as they have less series resistance than breakers. High amperage breakers are in the 4 milliohm resistance range where fuse is 0.5 to 1 milliohm range.

Breakers have flexible braid wire to electrically attach movable contact joints and a short circuit trip that consist of a solenoid plunger release trigger. The solenoid electro-magnetic coil is a few turns of wire which is in series with breaker's pass current. All this 'stuff' adds series resistance to breaker. High current breakers can get very hot having in the range of 40-100 watts of internal heating so they should have reasonable air circulation around them for cooling. Think how much heat a 100 watt incandescent light bulb produces.

Fuses just have a narrow piece of fusible metal. Their resistance does vary based on pass current due to temp rising of the fusible metal link. Most fuses do not spec series resistance because it is variable depending on current through them. They more often spec voltage drop at some current level, or spec nothing at all with regards to current induced voltage drop across fuse.

For high current 12v systems, you have to be very careful of all cable terminal lug connections creating additional series resistance. A lot of attention to details is necessary when you get above 150 amps of current.

Most of the Chinese BMS's current rating are based on about 20 watts of heating from series MOSFET's resistance. MOSFET series resistance get greater as they get hot. Most BMS's do not have sufficient heat sinking to sustain 20 watts of heating for more than five or ten minutes before they trigger a thermal shut down. Figure a usable continuous current of half what their maximum current spec is.

ljwobker said:

My particular install is in a temperature controlled environment where the AbsoluteWorstCase is about 35-40C.

Click to expand...

The temperature ratings in the charts are the ratings of the insulation on the wire. Not the ambient temperature.
A wire with a higher temperature rating can get hotter before the insulation melts or breaks down.

This gives the designer another variable to play with. A higher temp-rated wire can carry more current and is smaller, but it will be less efficient and tends to be a bit more expensive when compared to the same gauge with a lower temp rating.

FilterGuy said:

Yes. Your analysis is correct. However.....

For most batteries, the remaining 'hot' parts of the circuit still have the BMS short circuit protection. However, as I recall, the Valence batteries count on a distributed BMS model and I don't know if the batteries have any short-circuit protection without the distributed BMS system.

Click to expand...

Hi,

I'm also new trying to follow and understand..a couple of questions:


1. So I'm clear..(which I'm not) if a battery BMS has short circuit protection, no addtional fuse is required for the battery itself? Correct? Only the one fuse protection is required for the inverter.

2. Is there a distance that is recommended for the inverter Class T fuse location (i.e is too far or close bad). Is 12 inches too far? Is right on top of the battery bad (if fuse blows can it do damage to the battery (catch it on fire)?

Thank you

Chaucer said:

1. So I'm clear..(which I'm not) if a battery BMS has short circuit protection, no addtional fuse is required for the battery itself? Correct? Only the one fuse protection is required for the inverter.

Click to expand...

That is the way I would build it. However, a small few people add fuses to each battery.

Chaucer said:

2. Is there a distance that is recommended for the inverter Class T fuse location (i.e is too far or close bad). Is 12 inches too far? Is right on top of the battery bad (if fuse blows can it do damage to the battery (catch it on fire)?

Click to expand...

IMHO. If all else is equal, the closer to the battery, the better. However, when I am advising people on this I usually use the phrase "as close to the battery as practical'. There are valid reasons to move it a bit away. As an example, if putting it closer to the battery makes it more exposed to physical impact or exposes more metal that could be shorted, it may be better not to have it so close. As with anything else in engineering..... it is a trade-off.

If properly sized, a class T fuse will blow with little or no effect on its external shell. It can be hard to tell the difference between a blown class t and a good one. Consequently, there is no real chance of the fuse damaging the battery when it blows. (There are fuses that 'explode' but these are typically used on the grid or in industrial high-power systems that are way beyond anything a DIY system would use)

As an aside: The fusing element in a class t fused is surrounded by non-conducting silica sand. As soon as the fusing element blows, the sand falls into the resulting gap and will douse any arc that has formed. This is why class T fuses have such a high interrupt capacity.

RCinFLA said:

For 12v inverters the biggest enemy is voltage drop which 12v systems does not have a lot of margin to afford too much of.

You are thinking wiring heating but with 12v systems you usually have to dominantly think voltage drop between batteries and inverter.

Inverter regulate AC output voltage. Lower DC input voltage means they have to draw more DC current for same inverter output power. Inverter's efficiency drops as DC input voltage drops and input DC current rises, more DC current means more DC cable voltage drop. It is a compounding problem.

Wires gauge needed also depends on battery cable length. Fuses are better for 12v inverters as they have less series resistance than breakers. High amperage breakers are in the 4 milliohm resistance range where fuse is 0.5 to 1 milliohm range.

Click to expand...

All great points! Particularly the point about voltage drops on 12V systems.
- The working voltage range for a 12V system is fairly narrow.
- 12V systems typically have higher current so the voltage drops tend to be higher.

On 12V systems you are much more likely to see an inverter cut out due to low-voltage disconnect from voltage drops even though the battery is not depleted.

This is a point that does not get discussed much but should.

And I say to use an additional 1.12 multiple when sizing fuse (and therefore wire ampacity.)
Reason is, about 100% of AC cycle appears (rectified) as ripple current on battery cable. Heating of wire and fuse goes as current squared, and RMS current of a rectified sine wave is 12% higher than mean current (which delivers power to load.)

If Filter Guy calculates 175A, I say 200A.


Don't think anyone has mentioned, but tap off busbars to feed inverter from opposite ends, or otherwise balance current draw.

Don't think anyone has mentioned, but tap off busbars to feed inverter from opposite ends, or otherwise balance current draw.

Sorry I'll have to fess up here, I have absolutely no idea what this sentence means...?

In the following picture, both positive and negative wires connect to "top" end of busbars.
There is voltage drop across each busbar, so top battery supplies most current, bottom battery supplies least.
Moving one of the wires to "bottom" end of its busbar should help spread current draw more evenly among batteries.

Here is a diagram of what @Hedges is describing as a better way to hook up the batteries

Oh cool - makes perfect sense now. Thanks.