Inverter Capacitor Pre-charging

[mrfusi0n]

I've watched Will Prowse and other's on Youtube pre-charging the capacitors on their inverters before connecting them to the battery. Generally, they use a high power resistor to ease the current in without a big spark. That's a great idea but it got me thinking... the right kind of incandescent light bulb could do the same thing but also give a visual indicator of what's happening!

I tried it out with my 2x Growatts (SPF 3000TL LVM-ES) and a 48v lifepo4 battery and it works great! The one I picked up was a 48v, 25w bulb intended for truck/machinery https://www.amazon.com/gp/product/B0046IQMNG. Once I was ready to connect my battery to the Growatt, I first touched the wires to the light bulb; one to the side of the bulb and one to the "button" on the bottom of the bulb (ideally, I would have had a matching socket for the bulb with wires but I don't have that yet). The light lit up brightly for about one second and then faded out over another 2 seconds as the voltages equalized!

I really loved being able to visually see the power moving. Good confirmation that the capacitors were charging as well as showing when they were done (or at least close).

PS. I've also used this bulb when I need to restart my Daly BMS. Normally you temporarily short P- and B- but I like to put the bulb between to prevent the spark.

Hope this helps someone

 

[Diysolar123]

using filament bulbs is a common current limiting technique used for ages in many application; for example, its very common in overcurrent RF input limiters for SDR's in ham radio...
It is reallllll important that it be a filament bulb, NOT an led bulb hehe

The reason you use a resistor is that unlike a filament with a short lifecycle, a power resistor will probably last your lifetime.

 

[GXMnow]

I have used light bulbs many times for things like this. It works great. But have to agree, they can burn out. If it does not light, try another bulb.

 

[mrfusi0n]

using filament bulbs is a common current limiting technique used for ages in many application; for example, its very common in overcurrent RF input limiters for SDR's in ham radio...
It is reallllll important that it be a filament bulb, NOT an led bulb hehe

The reason you use a resistor is that unlike a filament with a short lifecycle, a power resistor will probably last your lifetime.

I would think a filament bulb would last plenty long enough with such minimal usage - years, I would imagine. But, like GXMnow mentions, if it burns out, it will be obvious since it won't light when you try to use it. Just have a backup bulb or power resistor handy, I guess.

The "problem" with power resistors IMHO, is that it isn't very obvious to newbies what value resistor to buy and how long to apply it. Imagine being new to this sort of thing and connecting up the resistor. It doesn't look like anything happened, so, did it work??? If use use a bulb and it doesn't light up, or it stays on, or it fades very slowly, or flashes quickly one time (or multiple? ?), or whatever it does, it just gives you so much more insight into what just happened. To me, that information is invaluable.

And just to be clear, I have NO problem at all with using power resistors! They work great and do last forever. Using a bulb is simply my preference and I really did enjoy seeing it light up and fade out ?

 

[Diysolar123]

The "problem" with power resistors IMHO, is that it isn't very obvious to newbies what value resistor to buy and how long to apply it.

valid point so the easy answer is...everybody use a 400W 2ohm resistor and count to 60 hehe

naaaa....lets try some math to see what we get!
This should not be this hard hehe

we should have a simpler way to help answer this recurring question.
You are trying to limit the inrush current without damaging the limiter, or the circuit associated with the capacitors.

The truth is that is does not need to be a perfect charge, you are just looking at reducing the capacitor charge current rush.

I just hooked my lcr meter to my all-in-one and it read an input capacitance of 20farads, lets call that "the standard capacitive load".
checking with my ESR(equivelent series resistance) meter I get 0.36ohms (lets call that the "standard short circuit resistance").
That ESR means if I were to "short" a 48V lithium battery to that capacitor it would suddenly take 48/.36 = 134amps!! (hence the spark, even at 24v that is still 67Amps)
so we want to slowww down that current to something a little nicer...maybe just 20amps; but we also want a simple rule???

The simple equation to fully charge a capacitor is 5*r*c. (it takes about 4-5 time constants to fully charge, a time constant is just R*C).
So a simple precharge resistor of 2ohm and 20farad cap bank would mean a full charge time of
4 * (0.36+2) * 20 = 188seconds!! so much for counting to 60...but wait...

here is the catch... we really do not need to fully charge the cap!! (even the bulb goes out before the cap is fully charged)...

basically the capacitor will charge to 63%V after just 1 "time cycle"...which is R*C, in this case 40seconds
the current inrush will be cut to 37% of max in just 1 time cycle!
SO now our "max" current at the start of our PRE-charge cycle is:
24V / (2+0.36) = 10amps!!
48V/(2+0.36) = 20amps!!

well look at that, we could just count to 10 and all is well as we are already in our safety zone, during PRECHARGE...so lets continue.
We could countdown to the full first cycle (the 40seconds) which would mean the current is just 37% of the starting current

lets double check the resistor power disipation...
I*I*r == 20*20*2 = so at the 48V battery that poor resistor needs to survive an 800W burst for about 20seconds, ouch.
for the 24V battery it is jsut 10*10*2 or an easy 200W...easy...

consider this as a candidate for the "easy" rule:
your precharge resistor ohms should be your battery pack voltage / 10
Your resistor wattage should be your battery voltage * 10
and you count to your battery voltage (elasped seconds) for your precharge duration!

lets test the assumptions for a 48V system which would mean:
precharge resistor is 48/10=4.8(lets go with 5ohms)
wattage is 48*10 or at least 480watts
RC constant would be 5*20 = 100seconds
48v/5 = 9.6A inrush during precharge
power disipation is 9.6^2*5 = 460watts
so far so good...

after 48seconds we stop precharge and direct connect, which means back to 0.36ohms...but our cap is now half charged!
we will have an inrush of less than 50% of our peak current so thats a (48V/2) = 24V/0.36=66A pop...hmm high but the current(I) will be dropping off like:


Its not ideal, but it does work out reasonbly well and is "easy"....

thats kind of simple, right, the precharge 10 10 Vbat rule???

 

[Horsefly]

I just hooked my lcr meter to my all-in-one and it read an input capacitance of 20farads, lets call that "the standard capacitive load".

Good analysis, but... 20 farads?! Really? I have a hard time believing it is that big! That's getting up into the super capacitor range!

Now you got me wanting to open up my Schneider CSW4024 and see....

 

[ddanley]

I've watched Will Prowse and other's on Youtube pre-charging the capacitors on their inverters before connecting them to the battery. Generally, they use a high power resistor to ease the current in without a big spark. That's a great idea but it got me thinking... the right kind of incandescent light bulb could do the same thing but also give a visual indicator of what's happening!

I tried it out with my 2x Growatts (SPF 3000TL LVM-ES) and a 48v lifepo4 battery and it works great! The one I picked up was a 48v, 25w bulb intended for truck/machinery https://www.amazon.com/gp/product/B0046IQMNG. Once I was ready to connect my battery to the Growatt, I first touched the wires to the light bulb; one to the side of the bulb and one to the "button" on the bottom of the bulb (ideally, I would have had a matching socket for the bulb with wires but I don't have that yet). The light lit up brightly for about one second and then faded out over another 2 seconds as the voltages equalized!

I really loved being able to visually see the power moving. Good confirmation that the capacitors were charging as well as showing when they were done (or at least close).

PS. I've also used this bulb when I need to restart my Daly BMS. Normally you temporarily short P- and B- but I like to put the bulb between to prevent the spark.

Hope this helps someone

I love this idea. I am getting the 3000TL 24V system. Would a 24V 25Watt bulb work for this?

Thanks!

 

[Pyrofx]

Hedges did so investigating on our SMA stuff


did a couple tests with SI 5048.
Breaker off, battery terminals show 3uF
Breaker on, <overload> uF
Inside, capacitors are 18x 2700uF = 0.050 F (or amps-seconds/volt)
That capacitor could carry 2300A for 1 millisecond, 23,000A for 0.1 millisecond charging to 48V.
Energy 1/2 C V^2 = 56 joules (at 48V); this is what's available to damage a contact.



I attempted to measure inrush with my 100A current transformer.
It captured a 382A peak, with 0.4ms rise time.
The waveform doesn't begin to fill a rectangle 600A high x 4 ms wide (2400 amp-milliseconds, my calculation), maybe 1/4 to 1/2 of that area.

I figured frequency is higher than 60 Hz, might not saturate even well above 100A.
But, no idea the frequency bandwidth of the core.
60 Hz is 16 ms period, 4 ms rise time, so what I captured is only 10x higher.

What I was able to measure doesn't even reach the steady-state current rating of your relay.
At that sort of current and 48V data sheet shows you should get 10,000 cycles.
My measurement may be way off. (or not; I haven't looked up frequency response of the caps.)

 

[740GLE]

Are you guys making these connections through your shunts? My small 1500w PSW hardly makes any spark with my 300a cheap Amazon shunt monitor. Maybe that’s a sign my cheap shut is more a heater than a shunt. Hahah

Either was I do like the ideal of a bulb, a regular 60w 120v should be plenty robust.
If the bulb burns out, no light and the inverter won’t turn on.

 

[Pyrofx]

1500 watts doesn't have much in it . The sma units have something like 16 giant caps. The 6500 watts can surge to 11k with no issues