Designing yet another BMS (yaBMS)

[s/v Jedi]

Originally Posted by tanglewood

Right. It’s about contactor close, not open. If you read the data sheets from contactor manufacturers like GigaVac and TE, all describe the need to do pre-charge.

Can you explain the issue again for me? You are talking about the contactor of the battery on the charge bus? While the inverter (charger?) is off?

[nebster]

Big inverters create huge inrush on their DC inputs if they are being powered up from a cold start. To avoid welding an expensive relay (on any bus), it can be necessary to precharge the input caps with a high impedance path first.

Then it’s safe to close the big relay.

[CatNewBee]

I know, but here's the catch.

You have indeed a high current inrush especially on first installation, when the unit comes from the factory and the caps are completely discharged. There are procedures described in the manual how to connect such a beast properly. You wire it up, you wire up the fuses, and when everything is done you engage the circuit by the MANUAL power switch. Combi units can precharge themselves when first connected to shore power btw, before connecting them to the battery by turning on the battery switch AFTER turning on the mains fuse, inverters of course not. Any way, the relays of the BMS are already closed and operational at this time, so there is no spark on the relay contacts. The capacitors are charged almost instantly. On subsequent disconnects and connects they stay charged for a long time, so the inrush is much smaller.

The BMS OVP / LVP relays are not meant to switch all the time, they are fuses that engage if something goes wrong with a rogue charge source or with a deep discharge due to forgotten loads. The contacts remain like new on the first install.

The BMS (at least mine) communicates with the inverter via the 2 wire BMS macro - other use CAN bus for this, and controls the inverter charger by communication, not by cutting the power supply to the box, so the capacitors stay charged on switching on and off, and the unit is off long before the OVP or LVP relays engage.

So yes, I am aware of the issue, but after some thought, chose to take care of it myself in those rare occurances when the battery is depleated and needs a jump start, instead of installing a precharge unit, that adds up to stand by current draw of the whole system and may add up to this situation in the first place.

[nebster]

Originally Posted by CatNewBee

There are procedures described in the manual how to connect such a beast properly.

I wish my manual had covered this. Instead, I learned the hard way, by melting stuff upstream.

the circuit by the MANUAL power switch. Combi units can precharge themselves when first connected to shore power btw, before connecting them to the battery by turning on the battery switch after turning on the mains fuse, inverters of course not.

Yes, with shore power available that is true for many units, but not all.

On subsequent disconnects and connects they stay charged for a long time, so the inrush is much smaller.

This I have not found to be true with my capacitors. They are drained down to empty within 10 minutes, possibly faster.

The BMS (at least mine) communicates with the inverter via the 2 wire BMS macro - other use CAN bus for this, and controls the inverter charger by communication, not by cutting the power supply to the box, so the capacitors stay charged on switching on and off, and the unit is off long before the OVP or LVP relays engage.

I think everyone wants to exert digital control over their inverter-chargers. But the issue is when the BMS has to force the relay open, then there is a problem. Not when regular, day-to-day charge disable instructions are being sent.

So yes, I am aware of the issue, but after some thought, chose to take care of it myself in those rare occurances when the battery is depleated and needs a jump start, instead of installing a precharge unit, that adds up to stand by current draw of the whole system and may add up to this situation in the first place.

The issue could occur on a depletion or an overcharge scenario.

Personally, I have the same opinion as you: a HV or LV disconnect scenario is one where I do not want an automated restart. I want to stop and see what happened before anything is allowed to touch the battery again. And manually precharging is not hard to do in that (extremely rare) case, since I will already be "hands-on" with the battery anyway.

I don't think a proper automated precharge design would have standby power demands. A minimal implementation would just be a simple timed dry pair that engages ahead of the close-relay dry pair. (The user needs to calibrate the precharge with the right resistor anyway, so I can't imagine it would be something that could be entirely built in to the BMS.)

This feature doesn't make the top of my list for many of the same reasons you discuss. But it would still be nice to have in the "ultimate" BMS. And, I don't think it's that complicated.

[CatNewBee]

Look ar REC d.o.o. (REC ABMS) website, in the accessories shop, they sell precharge units for the Kilovac solenoids.

Maybe you can get some clue how they have done it...

http://www.rec-bms.com/datasheet/Use...echargeNew.pdf

Its a timer, relay driver and relay and some 33 ohm / 7W resistors to limit the current, protection diodes...

[CatNewBee]

Also think a little out of the box.

A 7W resistor 33Ohm leads to a charge current of 12V/33 = 0.36A and burns 12V * 0.36A = 4.36W.

When you switch on the precharge unit, you power up not only the capacitors of your inverter, but the whole house installation, like fridges, freezer, bilge pump whatever, this devices are beefy and simply will suck up all this energy coming in from your precharge unit, not much left for the capacitors, so pretty useless in this setup. You need to power on everything in sequence, first the house, than the capacitors of the inverter by the precharge unit, than the inverter. Pretty complicated, you need a separate solenoid for the inverter to be efficient.

[s/v Jedi]

I think that pre-charge of ELCO’s in inverters is not the task of a battery. It should be part of the inverter. This may be another project... a solenoid on inverters that pre charges their ELCO’s before switching onto the bus. I think that may be what REC is selling? It doesn’t even need a uC. I used to build these for big audio amplifiers where their transformers surged the h@ck out of the mains AC to build up their magnetic field

I do not expect to ever see the battery solenoid disconnect from the bus bars, just like I never saw the main fuse pop. I can’t see the contacts wear out, not even when there is no pre-charge of capacitors.

I’m starting to like the idea of a solenoid on the inverter battery terminals with pre-charge feature. If an old fashioned RC network for timing is considered obsolete, I’d rather use a Tiny85 than a 555 timer for it

[Lake-Effect]

Interesting discussion re inverter precharge.

I'm wondering why more manufacturers of larger inverters haven't incorporated some solution such as precharging and "soft start" into their own control circuitry. Complete with the appropriate contactor inside.

(I'm not that familiar with big inverters, so maybe some have already thought of this?)

[s/v Jedi]

Originally Posted by Lake-Effect

Interesting discussion re inverter precharge.

I'm wondering why more manufacturers of larger inverters haven't incorporated some solution such as precharging and "soft start" into their own control circuitry. Complete with the appropriate contactor inside.

(I'm not that familiar with big inverters, so maybe some have already thought of this?)

Yes, they pre charge using the AC input when it’s an invert/charger unit. Or so I understand. With my Victron units I always have a spark when first installing them and I believe I actually read the manual once so believe I did everything as they instructed.

I don’t remember ever dealing with this after initial installation so I don’t think it’s an issue. But then again I never had LiFePO4 batteries disconnecting from the busbars... I’m still brainstorming about the whole thing with the simple old ways vs separate charge and discharge busses, surges, transient spikes etc.

[CatNewBee]

If I would design a new BMS I would do it modular, as all other manufacturers do, separate the BMS logic from the power switching interfaces, so the system can scale.

[s/v Jedi]

Originally Posted by CatNewBee

If I would design a new BMS I would do it modular, as all other manufacturers do, separate the BMS logic from the power switching interfaces, so the system can scale.

I will. But my first tests with the BlueSea Systems solenoid show that 7A is needed just to control those beasts. I’m thinking to use MOSFET’s for that but then there’s the wish for different solenoids that need non-pulsed control signals. A small relay would be better for that. Control of inverters, chargers, alternator regulators, solar chargers etc. all need relay contacts as well. Before we know it, it’ll look like a ‘60’s telephone exchange with relays everywhere

[tanglewood]

Most has been covered, but to reiterate, the issue is on initial closure of the Main BMS contactor, not when it opens. A depowered inverter will present a very large capacitive load on contact closure, and massive initial surge current. I had a 500A gigavac contactor weld on me with no DC load and the inverter turned off. And the evil thing is that you don’t know it’s welded until you try to open it, and it doesn’t. For me, that was when the defective inverter was inflames.

The gigavac datasheet describes it well. For a home brew system you can manage it manually, but for any production system you can’t rely on operational procedures

It’s all rather counter intuitive since our contractors cycle infrequently, if ever, and are rated for many thousands of cycles. But read the fine print about. Capacitive loads.

Jedi mentioned a similar issue with initial charge of large transformers. A friend had exactly this issue with a 12 kw isolation transformer that was tripping small (6 to 10A) shore breakers on initial closure. Same issue, 180 deg out of phase.

[rvlandlubber]

Originally Posted by s/v Jedi

But my first tests with the BlueSea Systems solenoid show that 7A is needed just to control those beasts.

The open and close controls for the 7700 rbs need less than 100ma. The 7A switching current is supplied via the red wire and is not switched.

The auto release version (7713) needs around 30mA on the control wire to switch. The 7A coil current is via the A or B terminal.

Bluesea dont really spell this out very clearly imo.

[nebster]

Originally Posted by s/v Jedi

Yes, they pre charge using the AC input when it’s an invert/charger unit. Or so I understand. With my Victron units I always have a spark when first installing them and I believe I actually read the manual once so believe I did everything as they instructed.

Their manuals are pretty sucky.

Also, smaller Victrons have smaller buffer caps. There will be a size that's small enough that the inrush doesn't cause damage as quickly, or at all. Bigger inverter, bigger caps.

The ones in my 15kVA unit are enormous and will make a very exciting "spark" that looks more like a plasma cutter. :O

[omsray]

It seems to me that the Precharge circuit belongs in the power feed to the Inverter along with its own contactor to turn power on to the Inverter, not in the BMS cutoff Contactors. Then when power is switched back on by the BMS Contactor, the Inverter Precharge circuit then gets power and goes through its sequence and safely powers up the Inverter.


OMSRAY