LVD/BMS redundancy and apparent contradictions

[sprucegum]

I'm a bit confused why there's a universal recommendation for an LVD in a LFP bank with a BMS. The only argument I hear repeated over and over is in the form of a truism: "your BMS *should* be the last resort for protection."

But, what exactly is the technical reason for this? As far as I can see, there's little difference between the often suggested Victron Battery Protect (microcontroller with FET-based switches) and a good quality BMS (also a microcontroller with FET-based switches) in terms of the nature of how or why they might fail.

On these forums and elsewhere, I often hear it argued not to trust FETs on a BMS -- with recommendations to oversize "cheap" BMSs (with no specification whether this includes the popular "high quality" BMSs the community recommends like JBD/Overkill and JK). At some point, after oversizing the BMS by 125-200%, setting your voltage cutoffs in a super-conservative range and every other conservative practice you can think up, it seems it should be okay to then RELY on using the equipment to do what it was designed to do. No one is saying to oversize the Victron BP by 125-200%, so I'm confused about the patterns of trust and distrust. If these BMSs are really awfully unreliable, why use them or recommend them? Or, if there's higher quality BMS options, why not put your money there instead of into redundancy?

Then there's the weird loophole created for certain other solutions: I have heard it argued that BMS's that simply signal the inverter, LVD or SCC to turn off in over/under voltage situations are completely acceptable (or even preferred in terms of safety). Yet, in this instance there is only one point of failure, not two, and it's still the same technology: an Electrodacus SBMS0 simply off-loads it's critical LVD responsibility onto a Victron BatteryProtect on the argument that putting FETs in the BMS is not reliable. Does anyone else find this to be a huge contradiction?

It seems to me that safety devices can be a bit of a slippery slope. Unless each one is different in *nature*, I personally think that including multiple points of failure adds to system complexity. All else equal, complexity is often not a good thing for safety. Seems to me you should first design the system so it's very unlikely to need a safety, then rely on one well-made safety device and minimize complexity. Particularly when the only danger in this case is loss of equipment, not a fire or other human-safety risk.

So what am I missing? If the BMS should be the "last resort", why not have a third or fourth protection device?

 

[BentleyJ]

In Basic terms a BMS is only a safety switch. Just like an oil level sensor in a generator or a pressure sensor your HVAC condenser is NEVER used as a means of control. Rather they disable the system to prevent damage when the normal operation of the equipment is compromised for whatever reason. The values in the Inverter & charge controller(s) user set up menus are intended to be the means of controlling the system. More recently, some BMS may communicate SoC and other information to the Inverter but its still the Inverter firmware and set points that control everything not the BMS. Finally, a BMS may have active cell balancing and-or current limiting features. Smart BMS? I don't like the use of that term, electronics are not smart.

Its a well known fact that most low cost, budget electronics from China tend to have overstated performance data, especially current capability. They are in common usage for same reason we don't all drive a Maybach. I believe, others may disagree, FET's and SSR's are fine for lower power systems, 100-150A and should be derated compared to the specification sheet. For large, stacked inverter systems that require high amperages a BMS that operates an external contactor are much better in my opinion. Although if you have 6 x 100Ah rack batteries a, 450A draw is still only 75A from each one.

In a certain sense you are not wrong about too many saftey devices though. I've changed out far more automotive coolant sensors, oil pressure senders and fuses than ever had a failure of the system itself. On the other hand, I've also replaced the coolant level sensor in my Mercedes several times but was only needed once, so far. That one time saved thousands of dollars in repairs and down time. Likewise, batteries are the most expensive part of the system and they are permanently damaged by high & low voltage, high & low temp. as well as high amperage. So yes BMS are necessary even though a BMS can itself be a point of failure.

 

[TacomaJoe]

In my use case (a trailer), I want my inverter to cut out itself and the BMS to keep the low current (lights etc.) running.

 

[sprucegum]

I get if there's a criticism of FETs in some specific high amp bank configurations. What I don't get is if people recommend a FET in an emergency LVD device (Victron BP) to avoid a FET in another LVD device (BMS)... or worse, the Electrodacus eschewing a reliance on FETs and then relying FULLY on the Victron BP with a DC load.

I'm curious if part of the safety philosophies I see developed here come hand in hand with very large banks. From an insurance perspective, the greater the investment in cells, the more irrelevant the cost of the other devices. Also, the more dangerous the object is to begin with, the more neurotic the safety plan can become (48v vs 12v, 100Ah vs 1000Ah).

When you have a huge bank, an expensive class T fuse or an LVD seems trivial in cost.

In my case, I've only got a 100Ah 8s bank. The cost of the cells was $400. My array was about $600 (1200W). Both are about 2x larger than needed but sized to be future proof for 25 years. Also, I would know at least a day ahead of time if I was getting anywhere close to LVD value, and my shunt would give off an annoying alarm at 5%. Plus there are not any loads I run when not at home. It's quite hard to argue myself into paying $50 for a Victron SBP simply because "I'm not supposed to" use the $100 JK BMS I bought with the same parts serving the same purpose as the SBP. Plus, the BMS manufacturers have no issue with me using it as intended and don't tell me to put an LVD in front of their LVD. It's just these forums that recommend that (I'm guessing even Victron doesn't recommend an LVD in front of their BMS).

As for inverters, that's a different department, because I believe they aren't just behaving as a switch but have more complex circuitry and also may suffer from sudden power loss, as well as providing a capacitor surge on re-connection to the BMS. Still, if an LVD event is supposed to never happen in your well-scaled system, it seems reasonable to have the BMS turn off the Inverter too (since it may only happen once or twice in a decade).

I can see if your system is specifically designed to have LVD events on a weekly basis, then perhaps the risk changes slightly. But I think the majority of off-grid solar setups are designed specifically around avoiding EVER having an LVD, which means a total shut down of the entire house.

 

[sprucegum]

@TacomaJoe Since those dc lighting loads are going to be low current, it seems to me that especially in that application one wouldn't be stressing the BMS very much by setting the BMS LVD in-line with your voltage safety margin you want for the other LVD. If the lights are, say, 20A (that's a lot of LEDs), and the BMS is, say, 100A, it's a very low stress to the components.

 

[TacomaJoe]

@TacomaJoe Since those dc lighting loads are going to be low current, it seems to me that especially in that application one wouldn't be stressing the BMS very much by setting the BMS LVD in-line with your voltage safety margin you want for the other LVD. If the lights are, say, 20A (that's a lot of LEDs), and the BMS is, say, 100A, it's a very low stress to the components.

It is only if my inverter is sucking 150A out of my bank and my SOC is low, I don't want my BMS to put me in the dark. I want the inverter to do its LVD above the point of the BMS so that I can have some lights while I figure out what's up.