Will this work for pre-charging inverter capacitors?

[MisterSandals]

So if you have 24V and a switch good for, say, 16Amps you simply use Ohm's Law.

I'm confused, if Krby has a 24v 3000w inverter, shouldn't the switch be rated for 125+ amps?
I think Krby later said he was planning on 30A.

What am i missing?

 

[gnubie]

The switch that carries the bulk current to the inverter would. A separate precharge switch would only have to deal with the current limit the resistor imposes.

 

[Keith C]

As gnubie sez. This thing would be operated only during the time BEFORE the inverter is even turned on. If you turned ON the inverter first it wouldn't even start-up because of the limit imposed by the resistor.

So,
1) Check inverter is OFF.
2) Check main disconnect is OFF.
3) Press the button to feed the inverter inputs for 5 seconds.
4) Turn the main battery disconnect ON.
5) Turn ON the inverter.

 

[MisterSandals]

Okay, that makes sense.
Would using a 4 position 200 amp battery switch (Off / 1 / 1+2 /2 ), with a 20 Ohm 50 watt resistor on position 1, require the diligent startup sequence as stated by Keith C?

Assume in Off position to start for this scenario.

1) Check inverter is OFF. (necessary? what happens if on?)
2) Turn to position 1, pause for a few seconds.
3) Turn to position 2. (turn past either Off or 1+2)
4) Turn on inverter.

 

[SCClockDr]

Okay, that makes sense.
Would using a 4 position 200 amp battery switch (Off / 1 / 1+2 /2 ), with a 20 Ohm 50 watt resistor on position 1, require the diligent startup sequence as stated by Keith C?

Assume in Off position to start for this scenario.

1) Check inverter is OFF. (necessary? what happens if on?)
2) Turn to position 1, pause for a few seconds.
3) Turn to position 2. (turn past either Off or 1+2)
4) Turn on inverter.

In theory yes BUT with the momentary push button one can not forget to move on to the next position as the return spring enforces the disconnect.
Much heat will be generated should the switch be left in Pos. 1 with potential for greater danger.

 

[Keith C]

SCClock is right, though I'm betting the inverter would go into conniption fits about a dead battery as that's what it would look like. What would bother me most though is that this is a "boundary condition" you'd be putting a trying-to-operate inverter into. Those realms are often poorly dealt with during design and the result of messing around the boundaries is a risk of damaging the inverter.

Hence the make sure the inverter is OFF.

 

[krby]

Hey everyone, others have covered it in various posts, but just to clarify where I ended up for version 1 and how it fits into my overall setup:

• This is for a mobile backup electric generator all contained in a box or two. So inverter would be on, but no permanent load connected. AC out on inverter is connected to a couple NEMA 5-15 outlets.
• The inverter/charger is a Victron Multiplus 24/3000. I have the power save/load sense feature configured on, so with the physical switch "on" it draws hardly anything then up to like quick burst of 1A (I measured it a couple of weeks ago, forget the actual value, but I do remember it was nothing to worry about thru the resistor) every one or two seconds.
• Red 4pos disconnect dial/switch from blue sea. Pos 1 in the lower current pre-charge circuit, pos 2 is main the 2/0AWG connection to the inverter

Using it will look like (this is pretty much what MisterSandals wrote)

1. Red disconnect switch is on OFF. Inverter is on, nothing connected to inverter output.
2. Wheel backup box where I need it
3. Turn switch to 1, count to five, turn switch thru 1+2 to pos 2.
4. Plug in load.
5. Smile.

I don't have this built yet, still waiting on the battery. But this is the plan. In a v2 of this, I'll probably use the programmable relay on a timer or if that doesn't work out for some reason, the momentary switch.

 

[backyardshane]

That will certainly do the pre-charge. Use a meter to watch the voltage rise the first time to gauge how long the button should be pressed. If the inverter is off, most of them are soft off these days, the charge on the capacitors won't be depleted significantly over a few seconds between you releasing the button and moving the switch to on.

You could also use a 3 position switch - off, 1st on position is precharge, 2nd on position is full current available to inverter.
[/QUOT

Thanks all for the replies, I ended up with a 20ohm 50W resistor for just a few dollars. Once everything shows up, I'll measure the inductance and post back here so others can find it.

that is the same resistor i ordered yesterday. hope the turn out to work.

 

[Keith C]

Sounds pretty clean and it should work great krby.

We await a couple of pictures sometime.

 

[FilterGuy]

Interesting discussion.

In his videos Will generally 'expects the spark' and just hooks up the inverter. That seems to work but, in at least one video, it ended up tripping the over-current of his BMS.... so he used a resistor first.

That begs a question: If your BMS doesn't trip, is there any problem with the surge? It certainly isn't good for a switch but how bad is it? (The answer to that is dependent on the switch)

Hey Will: A video on pre-charging might be fun.... Or maybe make this one of the things you discuss in your upcoming BMS reviews?

 

[krby]

He has one

Hey Will: A video on pre-charging might be fun.... Or maybe make this one of the things you discuss in your upcoming BMS reviews?

He has one already:

 

[FilterGuy]

He has one already:

Thanks!!! I had not seen that one. Will is amazingly prolific!

 

[Fred]

Can anyone show how they were able to fully automate the precharge process so that it will happen when the BMS sends a signal to an external contactor to reconnect LOAD to BATTERY BANK?

Sounds like the solutions discussed so far have been to basically use a parallel circuit with a resistor for a short time then switch to the main circuit.

Instead of doing that, is there a way to add a component in series with the main positive lead that will attenuate the in rush current? What if you place a large capacitor between the inverter + and the relay... one capable of taking the inrush current? (This would essentially be wired such that the capacitor + was connected to the inverter + and the capacitor - to the inverter -.

Theoretically this large capacitor would pre-charge and while doing so would attenuate the spike going into the inverter. (I have no idea if this could work)

 

[DThames]

Can anyone show how they were able to fully automate the precharge process so that it will happen when the BMS sends a signal to an external contactor to reconnect LOAD to BATTERY BANK?

Sounds like the solutions discussed so far have been to basically use a parallel circuit with a resistor for a short time then switch to the main circuit.

Instead of doing that, is there a way to add a component in series with the main positive lead that will attenuate the in rush current? What if you place a large capacitor (or inductor) between the inverter + and the relay... one capable of taking the inrush current? Theoretically it would pre-charge and while doing so would attenuate the spike going out to the inverter. (I have no idea if this could work)

A capacitor will work in a series circuit only when the signal is AC. Otherwise the capacitor will charge up and the current will stop flowing. An inductor will cause a delay in the current, but not really reduce it. However, the built in resistance of an inductor will cause a bit of an curve in the amp increase and you will be stuck with that resistance on series. You want to make that less, then it will have less affect on the current rise rate.

I have not given this enough though to say it is a good idea but leaving the inverter caps energized by a trickle path would seem to make sense. It will cost a tiny power but it would keep it hot. This is assuming my inverter was switched off at the inverter's on off switch. Then engage the contactor and turn on the inverter. No better than any of the other solutions, just less switching. I am still building my standby system and I have not settled on what I want to do. It will be a manual switch over for now.

 

[krby]

If you want simple, I think you either have to have a parallel "trickle" circuit as @DThames says (either on all the time or on a switch) I've heard that some products use an NTC Thermistor
From Wikipedia:

With NTC thermistors, resistance decreases as temperature rises. An NTC is commonly used as a temperature sensor, or in series with a circuit as an inrush current limiter.

I don't think anything else will work in series.

If the parallel going to be automatic, you need some "signal" to control when to switch on the main low-resistance circuit. Off the top of my head, those could be:

• Time: a relay that kicks over in a set period of time, someone posted on one on this thread.
• Current: when the pre-charge path current drops to near zero, switch on the main circuit.
• Voltage: When the voltage across the pre-charge path drops to near zero, switch on the main circuit

Anything automatic is going to be have to use one of these signals, I think.

 

[Supervstech]

Um... I should post what I did.

I wired in a lamp with a 60watt bulb in it. Tied it to a 2 bank battery switch, set it to 1 and the batteries to 2. I turn to 1, watch the bulb glow, then when it dims, I switch to 2, and fire up the inverter.

 

[DThames]

If you want simple, I think you either have to have a parallel "trickle" circuit as @DThames says (either on all the time or on a switch) I've heard that some products use an NTC Thermistor
From Wikipedia:


I don't think anything else will work in series.

If the parallel going to be automatic, you need some "signal" to control when to switch on the main low-resistance circuit. Off the top of my head, those could be:

• Time: a relay that kicks over in a set period of time, someone posted on one on this thread.
• Current: when the pre-charge path current drops to near zero, switch on the main circuit.
• Voltage: When the voltage across the pre-charge path drops to near zero, switch on the main circuit

Anything automatic is going to be have to use one of these signals, I think.

Yes, something along those lines. If you start your cap charge cycle, you can then look at the voltage on the caps to trigger the coil of the power connector. Odds are you would want a buffer circuit so that it was a clean Off then On rather than the contactor coil actually riding the cap charge ramp voltage. I have some little cheap pilot relay boards that will take high or low going signal to trigger the relay. You could bias that signal with a voltage divider off of the cap voltage if the cap voltage was too high too soon for the relay trigger signal. Add to the front end of that your charge timer and that might work. Of course a simple time delay would work as well but just me being me, I would rather trigger off of the voltage rise as that is what you are waiting for. That way if you didn't charge, the sequence would halt in a fault condition.

 

[FilterGuy]

If you want simple, I think you either have to have a parallel "trickle" circuit as @DThames says (either on all the time or on a switch) I've heard that some products use an NTC Thermistor
From Wikipedia:


I don't think anything else will work in series.

If the parallel going to be automatic, you need some "signal" to control when to switch on the main low-resistance circuit. Off the top of my head, those could be:

• Time: a relay that kicks over in a set period of time, someone posted on one on this thread.
• Current: when the pre-charge path current drops to near zero, switch on the main circuit.
• Voltage: When the voltage across the pre-charge path drops to near zero, switch on the main circuit

Anything automatic is going to be have to use one of these signals, I think.

With the pre-charge path, the voltage at the inverter input will start low and ramp up, consequently you could change the 3rd bullet to:

• Voltage: When the voltage across the pre-charge path drops to near zero, or the voltage at the inverter rises to near the battery voltage, switch on the main circuit.

As an example, if you were on a 12 volt system and switched on at 10 volt, you would still have a small surge but it would not be large enough to cause a problem.

 

[Fred]

A capacitor will work in a series circuit only when the signal is AC. Otherwise the capacitor will charge up and the current will stop flowing. An inductor will cause a delay in the current, but not really reduce it. However, the built in resistance of an inductor will cause a bit of an curve in the amp increase and you will be stuck with that resistance on series. You want to make that less, then it will have less affect on the current rise rate.

I have not given this enough though to say it is a good idea but leaving the inverter caps energized by a trickle path would seem to make sense. It will cost a tiny power but it would keep it hot. This is assuming my inverter was switched off at the inverter's on off switch. Then engage the contactor and turn on the inverter. No better than any of the other solutions, just less switching. I am still building my standby system and I have not settled on what I want to do. It will be a manual switch over for now.

Series wasn't the best description so I revised my post. The idea was to have a large capacitor with the + lead on the +inverter lug and the - lead on the - inverter lug. When the contactor closes this would effectively connect the capacitor + to the battery + and the inrush current would charge this capacitor. The leads would have to be able to take the current and the cap would have to be properly sized.

Capacitors are used to attenuate (smooth out spikes and dips) voltage in signals, power supply voltage, etc. so this would seem to be an appropriate application.

How much energy does a large inverter take before the caps are charged?
Current through a 25 ohm resistor at 50 volts would be 2 amps. If 10 seconds is sufficient then the amount of coulombs would be C= 2 amps x 10 seconds = 20 coulombs.
The capacitor size then would : F=C/V = .33/50 = .4 farads

That is a pretty big capacitor. 10 seconds at 2 amps was a guess though. Does anyone have a better measurement of the amount of power it takes to precharge a large inverter?

 

[FilterGuy]

Series wasn't the best description so I revised my post. The idea was to have a large capacitor with the + lead on the +inverter lug and the - lead on the - inverter lug. When the contactor closes this would effectively connect the capacitor + to the battery + and the inrush current would charge this capacitor. The leads would have to be able to take the current and the cap would have to be properly sized.

If I understand the proposal correctly, this would make the problem worse. The problem is caused by the large capacitors across the + and - inside the inverter. However, if you had a large capacitor that was already charged and you switched it in at turn on, it could provide the inrush current in order to protect the BMS. However, you would need to think about how this new capacitor gets charged.... if you aren't careful, you would just move the problem.