Small Boat BMS/LFP with backup

[rgleason]

@Goboatingnow wrote:

"to be ISO compliant the 3 way switch should be removed ( or no common connection option allowed )"

Question: A "common connection" is the same as "no BOTH" is it not? Also "common connection" is not to be confused with the "common post" in the 3 way switch.

I had written that the 3-way switch was "locked" or "lockable". My intentions were not fully described. I am sorry about that.

So rather than asking for clarification about the existence of the "BOTH" position , you declared this configuration "not compatible" and "not a good template".

This template is still "Under Development", and a "DRAFT" (I should have so noted that and will do so from now.)
I appreciate the points and discussion.

I am going to try to digest all the posts above.

@Goboatingnow wrote

Secondly alternator protection is unnecessary if proper field coil disconnect is present and in An Li system it should always be. Alternator protection can protect the alternator but can fry attached 12v electronics

Yes, BMS advanced or early field coil disconnect is best in my opinion, you have no argument from me about that, however the less expensive DalyBMS does not provide this function. The more expensive REC BMS and TAO BMS do operate this way, however they require additional switching relays or switching FETS and thus more additional costs. I have shown this scheme in other diagrams above (for example) that is why I had suggested looking at the development in this thread.

I agree that the DC Panel will not be protected as well by the APD even with the Orion DC-DC Charger, and I would like to protect my navigation equipment. Therefore I propose putting a second APD on the DC Panel.

@Goboatingnow wrote:

I see no point to the panel associated smart switch. The bms is doing this function anyway. A low voltage Li situation should never realistically occur , if it does regularly the system is sized incorrectly.

A better situation is to break the dc panel into high and low priority loads and to only disconnect low priority loads first but it’s an unnecessary complication

Using the DalyBMS, wouldn't I need the Smart Battery Protect (even if it's BMS switching function was not used) , to prevent the DC Panel from setting off the BMS Low Voltage FET? If the system doesn't really need it and simply using the DalyBMS Low voltage disconnect is OK, that would be a further simplification and cost savings. What would I need simply an advance alarm when the DC Panel has run down the LFP too far?

The DC Panel and loads are quite simple on this boat. The major "low priority" load is Refrigeration, which is easily removed by simply switching off that breaker. I don't believe there is a need to separate and complicate the DC Panel into two panels. When a LFP or BMS issues occur, you can be sure that I will instinctively and immediately reduce loads, if I am present, and I usually leave my DC Panel "off" or at least all the switches "off" when leaving the boat, with the bilge pumps on the "always on" bus. I still have not completely resolved all these details.

I could simply connect the DC Panel directly to the SLA Battery and get a larger Orion DC-DC Charger to accommodate the DC Panel loads. The SLA would protect the DC Panel from any alternator disconnect spikes, however there would be more heat and inefficiency from the increased use of the DC-DC Charger. Then when the BMS / LFP has an issue or failure, essential navigation equipment would remain powered. For some reason, I am reluctant to do this...and prefer using the LFP and then manually switching the DC Panel and Alternator to the SLA.

@Goboatingnow wrote:

So advance disconnect alarms are a must

I would like to meet the ABYC requirements and have a "certifiable" (whatever that is) and professional installation. Given the US Marine industry's fragmentation, this relatively new LFP option and lack of real regulation, I don't believe there would be total agreement on "certifiable". I do not intend to do this until I am sure about what I am doing and why. I am taking the time to get this configured appropriately, then I have to figure out how to fit it all into a very small boat, without messing the boat up.

Doesn't the requirement to have an advanced disconnect alarm, eliminate the use of the DalyBMS?
Furthermore if there is an advanced disconnect alarm, it can probably be used to switch off the Alternator field current.
Perhaps it is necessary that I use a more expensive BMS? Or perhaps use a Victron Battery Monitor to provide the alarm?

I still have to read and digest all the other good comments and information from Wholybee and others. This is a complex subject and I appreciate all of your experience and guidance.

 

[rgleason]

Cell failure in the LA can lead to very high currents flowing into the battery from the Li with catastrophic results , especially on high capacity high current Li systems. Fault modes in either battery are excerbated when interconnected. Hence the spec bodies don’t allow them. I would not accept your comment re experience as we have not enough installation history to draw definitive conclusions.

Since the system can easily be developed to isolate the starter it’s not an issue anyway and I think medium term the starter will go Li also anyways as Li makes a better starter battery ( smaller lighter )

There is some disagreement about the dangers of connecting different chemistries, and there have been bad failures and reports of some success. ABYC says not to cross connect chemistries, so I am going to try to adhere to that requirement.

• I believe that a lockable 3-way Switch with "NO BOTH" fills that requirement, provided the switch is operated with the engine OFF. However does this meet ABYC requirements? In the event of problems, which always occur at night, in bad weather, on a lee shore, I would not like to have to undo fuses and then use a battery cable to switch over to SLA Battery. Isn't there a better way to do this?
• The last point about using an Li battery for the starter/backup battery makes a good deal of sense. Would it make the system simpler and less expensive? I think it might.

FURTHER DISCUSSION:

• I think we need some further discussion about Victron equipment (chargers: MPPT, Orion DC-DC) and Victron monitoring (Victron SmartShunt) and controls, since there is some additional cost and more wiring for a CerboGX and GXtouch50.
• Also Goboatingnow has raised the issue of having to protect the Victron chargers from spikes, when does this stop? Put an APD on each one?

We also seem to have a fundamental disagreement about the use of switching FETS and high quality switching relays amongst the experts here.

---

Wholybee's approach and LFP system described in another thread's post.

---

Incidentally, for those starting out, I have some links in first post to various resources which should help. This one shows the various typologies being discussed Dual Bus, Lead-Lithium Hybrid (not a favorite for ABYC and Goboatingnow) and Split Charging.

One link I don't have above which shows a system with three programmable Perfect Switch PowerGate "smart" FETS which are preconfigured for LFP batteries and a particular system for a boat call Mai Tai. I will try to find that link. That resulted in a very simple pre-configured installation that was not adjustable, but just worked. https://perfectswitch.com/ and the MaiTai application is here https://perfectswitch.com/applications/marine-lifepo4/ However the PowerGates are very expensive and I do not understand the programming, however now I notice in the description that the Load master and slave momentarily connect the SLA and LFP batteries on the Charge Bus to protect the Alternator.

Tao BMS has a diagram showing a way to provide backup using the SLA See the Lithium Charged by Alternator configuration and find the dashed red lines. This requires about 4 relays and will be very expensive.

 

[svsagres]

So advance disconnect alarms are a must

This is the one bit of the standard that I feel could use a little more nuance. I agree that advance disconnect alarms would be critical if the battery is being used to operate safety critical systems, namely the engine and a few other things.

In my case, my engine is powered by its own dedicated Lead Acid starter battery. If my BMS drops the load for any reason, it will not put me in a bad safety situation. Plus, I can easily override the failed BMS and force things back on again if need be.

 

[rgleason]

@svagres Thanks for advising on this. What is involved to override the failed BMS and force things back on? What would you do to start using the FLA starter battery and direct alternator charging and DC loads to that battery for an emergency solution? I know you have described this to some extent elsewhere.

 

[svsagres]

@svagres Thanks for advising on this. What is involved to override the failed BMS and force things back on? What would you do to start using the FLA starter battery and direct alternator charging and DC loads to that battery for an emergency solution? I know you have described this to some extent elsewhere.

Like I explained previously, if I lose my lfp bank, but still need those devices powered, I’ll just force the 7713 and BatteryProtect back on. In the case of the 7713, all that’s involved is pressing down the knob and manually closing the contactor. For the BatteryProtect, I’ll need to jumper the control wire, but not hard to do.

My plan is to continue to use the LFP in a safe/limp mode until I can resolve things. The pack itself is well balanced, thanks to the REC, and my loads, other than theinverter/charger are comparatively small. I have the Wakespeed and MPPT both set to target 13.2v which is around 50% charge. They will go higher than that only when directed to do so by the BMS. If I’mjust daysailing, I’ll probably not even bother with that, as my ipad and handheld vhf both have their own batteries, so I can live with those for nav and coms as I head home.

Running the whole thing without the BMS for a couple if days needed to get back to my marina is not a big deal. I’ll just need to keep a slightly closer eye on it.

This is what I like about the contactor based BMSs. If they fail, they’re easier to bypass if needed.

 

[Goboatingnow]

1. Regarding alternator protectors
APDs do not limit spikes to 12v , they are typically 28 or 32v TVS diodes hence the spike can reach that point and sometimes a little higher. Your alternator is designed to survive this.

Your other electronics that remain connected to the alternator will be exposed to upwards of 32v. Many such devices are not rated to survive that. ( there is a relevant automotive standard that mandates in car electronics survive 48v but it’s rare that marine stuff follows this )

Hence abrupt alternator disconnect needs to happen right at the alternator to protect down stream electronics. This can mean two disconnect devices

2. Smart disconnect. If u intend to use this as a disconnect device it must be next to the battery. No other devices except the bms must remain connected and provision must be contained within the bms to not discharge the li completely either. Hence as it is the smart switch is not compliant with ABYC or ISO.

3, ISO is very clear and ABYC is reasonably clear that chemistries must NOT be connected together. So a selector switch with no common is acceptable.

Personally I would ensure the Li disconnect is immediately at the positive terminal with nothing but the BMS in circuit on disconnect.

I would ensure the alternator is controlled via field coil control and remove APDs as they are imperfect solutions.

I would not use a bms that could not generate advance disconnect alarms.

I would agree that a mechanism should exist to bring the dc bus back on online either via the starter battery or via a bms shutdown override. As mentioned a relay interrupt facilitates that easily.

By the way I’m not against mosfet switches but a very robust mosfet switch that can survive and switch under arduous fault conditions is a major design challenge and is expensive.

 

[wholybee]

FURTHER DISCUSSION:

• I think we need some further discussion about Victron equipment (chargers: MPPT, Orion DC-DC) and Victron monitoring (Victron SmartShunt) and controls, since there is some additional cost and more wiring for a CerboGX and GXtouch50.
• Also Goboatingnow has raised the issue of having to protect the Victron chargers from spikes, when does this stop? Put an APD on each one?

We also seem to have a fundamental disagreement about the use of switching FETS and high quality switching relays amongst the experts here.

First, I want to express respect for @Goboatingnow and his opinion. I am not an EE, though I work closely with them, and have held an engineering position for many years, but never achieved the formal education. Undoubtedly, he has a better understanding of the details than I do.

Your further discussion questions really hit on why I hold my opinions. Making the system more robust, is a never ending spiral. We don't have BMS's on Lead Acid batteries, but they would benefit. AFAIK, ABYC/ISO doesn't require low voltage alarms on Lead Acid batteries. Or hydrogen gas detectors. The list could be endless. LiFePO4 is a different chemistry with different requirements. But LFP, like Lead Acid, like Propane systems, like Fuel systems, and so many other systems can never be 100% safe. So we draw a line somewhere. ABYC and ISO drew that line to even allow NMC. If you allow NMC, how can it be justified to require an alarm for a pending disconnect or disallow paralleling different chemistries as an emergency backup? (I digress here, that isn't the topic at hand.)

I believe the requirement is to have a system that is at least and preferably more safe and more reliable than the Lead Acid system it replaces. Forbidding a system that is safer-possibly much safer- because of a minor addition that would make it even just a tiny bit safer is plain dumb. Key in that is put a lot of time and effort into ensuring that a disconnect NEVER happens. The system should be built such that the cells never go out of balance. Mostly, that means (imho) don't buy second life or grey market cells from Alixxx, ebay, or Amazon. Discussions have been had here on that (and outright fights on Facebook), as cheap 280Ah cells are very popular. But I think its much better (safer) to spend extra money on better cells than extra money on a better BMS. (I assume no one works with an unlimited budget and compromises will be made somewhere)

And the correct LFP charging regime needs to be followed. An advanced BMS like the REC will do this, but so will Victron (and others) directly without BMS control. I will repeat here because it seems to have been missed. All of my Victron gear communicates (via bluetooth, no wires) and uses the current and voltage readings from the BMV-712. Charging stops by either tail current or voltage at the battery. Coordinated so it stops charging on every device at the same time. SOC by coulomb counting is available, low SOC alarms and low voltage alarms. Charging this way is no slower, doesn't shorten cell life, or reduce capacity compared to the REC BMS charging. It's essentially the same, save for the convenience of the REC being a single device to setup or change if you want to change settings for storage.

Lastly, a safe installation will have a well laid out and wired installation. Wires secured with no chance of chafe or inadvertent shorts, properly installed quality wire terminations, etc.

It's impossible to create a zero change of BMS disconnect. That is why the BMS is there. That is why we have an alternator protection device. But, it can be brought to very near zero change. And so, if it isn't going to happen, the arguments about FETs vs Relays, or what happens to your electronics when more than 32V is put on them, etc. have much less meaning.

Let's look for a second at a BMS disconnect where the alternator creates a large surge voltage. That surge is going to be relative to the current the alternator was generating at the time of disconnect. If the alternator is producing only a small current, the surge (if any) will be small. So damage is going to occur when the alternator is outputting a large current. When is this the case? When the battery isn't fully charged. So we are talking about not just a BMS disconnect, but a disconnect when the battery is not fully charged and charging at a high rate. Certainly, that can (and does) happen, but reasonable system design should be able to eliminate that. It won't be a pack HV disconnect because the battery isn't fully charged yet. It could be a single cell out of balance and over voltage, but that goes back to my starting with high quality cells, and ensuring they stay in balance. It could be an overcurrent if you have a high output alternator and didn't size it appropriate to the BMS, or if you didn't account for the Alternator plus 50A of solar at the same time.

Focus on what can cause a BMS disconnect, and eliminate those events from happening.

Anecdotally, I have helped several friends repair destroyed LFP installs that had been professionally installed. One was a 123/SmartBMS which controlled a pair of relays, one for charging and one for load. The boat had been unattended on the hard for 4 months. The BMS had burned, several transistors failed and caught fire. The load relay was stuck on, the charge relay stuck off, and his cells were severely bloated and at 0V. I wasn't able to definitively determine the exact cause, but his solar charger and Balmar regulator were both still on the factory defaults for AGM batteries. Solar had been active for those 4 months. A week of that there was snow, and solar wouldn't have been active at that time. Maybe a REC BMS that controlled charging would have saved him. Relays instead of FETs certainly did not as both of them failed. And I have a very low opinion of the 123/SmartBMS after my experience with it. Highly likely that correctly programming his charging devices would have prevented it.(or preferably if leaving the boat, charge to 50% and remove the battery entirely)

Regarding the alarm before disconnect. I actually think it is a good idea, but I don't think it should be required, and it isn't in my budget to buy a REC BMS just for that feature. I am just beginning a project using an Arduino that would connect to the UART port on a FET BMS and provide alarm and relay outputs to make them ABYC compliant.

And, as I expect my LFP battery to last a VERY long time, I will probably upgrade to a REC BMS at some point as well. But it wasn't in my budget at the time I converted. I foresee forgoing a separate staring battery completely and having a single highly reliable LFP battery that powers everything. I know that is controversial because everyone thinks a backup source is needed, but I think that comes from history of a relatively unreliable battery system. I think the future is a single battery.

I like the idea of remote control and monitoring from the CerboGX. I'm looking at them closely, but out of my budget for now. I see it as a luxury item/toy/convivence rather than something that fulfills any requirement of LFP.

 

[rgleason]

@wholybee You have good points and experiential perspective which is valuable. Can you advise what these mean?

• NMC.
• It's impossible to create a zero change of BMS disconnect.

Also, since both you and Goboatingnow have expressed the need for the batteries to be of high quality with matched cells "don't buy second life or grey market cells from Alixxx, ebay, or Amazon" . ... Yes, I was considering EVO 280ah cells from a Chinese supplier that has distribution from Texas! LFP batteries are not sold in local stores, so what would you suggest I do to be assured that the batteries are good quality with all good cells?

I certainly can relate to your cautious approach with regard to value, and simplicity, as this system will be expensive in any case. Thank you for giving us some ideas for a staged approach.

LATER
mittiempo wrote:

Myself as well. 8 Calb CA180 cells from LythBattery. All over 200 AH.

I have had 3 orders from LythBattery so far and zero issues with all. Great to work with, better than most companies I have bought from in North America.
They are a very large distributor in the same area as Calb, Luoyang City a manufacturer as well as distributor. Based on the service I received they are first class whether you buy a few cells or a few thousand.

Top lithium iron phosphate battery supplier in China - LYTH

LYTH is top supplier & manufacturer of LiFePO4 battery cells in China, Highest standards of safety, performance, and durability for RV, marine, UPS, golf cart and solar energy storage.Best LiFePO4 battery source.

www.lythbattery.com

If you contact LythBattery the sales manager is Rita Sun.

 

[rgleason]

1. Regarding alternator protectors
APDs do not limit spikes to 12v , they are typically 28 or 32v TVS diodes hence the spike can reach that point and sometimes a little higher. Your alternator is designed to survive this.

Your other electronics that remain connected to the alternator will be exposed to upwards of 32v. Many such devices are not rated to survive that. ( there is a relevant automotive standard that mandates in car electronics survive 48v but it’s rare that marine stuff follows this )

Hence abrupt alternator disconnect needs to happen right at the alternator to protect down stream electronics. This can mean two disconnect devices

2. Smart disconnect. If u intend to use this as a disconnect device it must be next to the battery. No other devices except the bms must remain connected and provision must be contained within the bms to not discharge the li completely either. Hence as it is the smart switch is not compliant with ABYC or ISO.

3, ISO is very clear and ABYC is reasonably clear that chemistries must NOT be connected together. So a selector switch with no common is acceptable.

Personally I would ensure the Li disconnect is immediately at the positive terminal with nothing but the BMS in circuit on disconnect.

I would ensure the alternator is controlled via field coil control and remove APDs as they are imperfect solutions.

I would not use a bms that could not generate advance disconnect alarms.

I would agree that a mechanism should exist to bring the dc bus back on online either via the starter battery or via a bms shutdown override. As mentioned a relay interrupt facilitates that easily.

By the way I’m not against mosfet switches but a very robust mosfet switch that can survive and switch under arduous fault conditions is a major design challenge and is expensive.

@Goboatingnow
Thank you for your points above. Is "Smart disconnect" the Smart Battery Protect 65?

The only discharge device is the DC Panel, so I don't understand why the "Smart disconnect" must be next to the battery and "No other devices except the bms must remain connected and provision must be contained within the bms to not discharge the li completely either". Perhaps I have something confused here? Can you explain? Maybe this is about the Alternator Disconnect?

I think you would like to see a typical dual bus configuration with relays or hefty fet switches.

It looks to me like the REC BMS with Wakespeed WS500 and Victron "Smart" equipment and monitoring is probably the best way to meet these goals. (Thank you svsagres for publishing your system).

 

[wholybee]

@wholybee You have good points and experiential perspective which is valuable. Can you advise what these mean?

• NMC.
• It's impossible to create a zero change of BMS disconnect.

Also, since both you and Goboatingnow have expressed the need for the batteries to be of high quality with matched cells "don't buy second life or grey market cells from Alixxx, ebay, or Amazon" . ... Yes, I was considering EVO 280ah cells from a Chinese supplier that has distribution from Texas! LFP batteries are not sold in local stores, so what would you suggest I do to be assured that the batteries are good quality with all good cells?

I certainly can relate to your cautious approach with regard to value, and simplicity, as this system will be expensive in any case. Thank you for giving us some ideas for a staged approach.

NMC are another Lithium chemistry other than Lithium Iron Phosphate. NMC are the ones that are highly flammable/dangerous.
Typo. "It's impossible to create a zero chance of BMS disconnect." I am acknowledging that even though the BMS should never disconnect, and we can do things to make it very improbable that it will happen, it still might.

 

[wholybee]

Also, since both you and Goboatingnow have expressed the need for the batteries to be of high quality with matched cells "don't buy second life or grey market cells from Alixxx, ebay, or Amazon" . ... Yes, I was considering EVO 280ah cells from a Chinese supplier that has distribution from Texas! LFP batteries are not sold in local stores, so what would you suggest I do to be assured that the batteries are good quality with all good cells?

At some point you have to make a judgement call. Talk to the distributor about what you get. What test documents do they come with? How is it packed? I bought CALB cells from a distributor in California local to me at quite a premium over the EVE/CATL etc. cells. They came packed in a crate (not a box) silk screened from CALB, with all the hazardous shipping labels and documents. Delivered to me via a semi-truck freight, not Fed-ex etc. Opened the crate, and in it were my 12 cells, with factory test sheets showing the mfgr date of each cell, and the results of various tests. My crate originally had 16 cells, but because I only ordered 12 cells, 4 were removed from it. Otherwise, it was a shipping crate from the CALB factory.

Buyers of the 280Ah cells are happy when they test at 280Ah. Often they accept them if they are close but not quite. Every one of my cells tested at 20% *over* rated capacity. They still do, 2 years later. I'm confident they are much better cells than the 280Ah and similar. I would expect the same from Sinopoly or Winston, bought from a US supplier. Mine are also the older plastic case cells. I've heard these are discontinued, if so I probably got some of the last ones made.

There was a youtube video on here a while back from a US distributor of EVE cells. He was showing the correct documents grade A cells ship from the factory with, and that MUST be included with EVE cells if they really are grade A. It consisted of a spreadsheet listing the S/N of each cell and the test data. The S/N of each cell being a QR code that is etched (not a sticker) onto the cells. There might have been specific markings on the shipping box as well, I don't recall. But it would be good to look for it if you are considering those cells.

All this stuff can and is forged by suppliers on Alixxx though, so that's part of the judgement call. If you have a US contact and a phone number and not just email or buying though a website, then you are much more likely to get what you are being told you are getting.

 

[Goboatingnow]

@Goboatingnow
Thank you for your points above. Is "Smart disconnect" the Smart Battery Protect 65?

The only discharge device is the DC Panel, so I don't understand why the "Smart disconnect" must be next to the battery and "No other devices except the bms must remain connected and provision must be contained within the bms to not discharge the li completely either". Perhaps I have something confused here? Can you explain? Maybe this is about the Alternator Disconnect?

I think you would like to see a typical dual bus configuration with relays or hefty fet switches.

It looks to me like the REC BMS with Wakespeed WS500 and Victron "Smart" equipment and monitoring is probably the best way to meet these goals. (Thank you svsagres for publishing your system).

Iso and ABYC standards require that a safety LI disconnect disconnects all devices from the battery bus including charge sources , nothing should remain powered by the bus expect the BMS. provision should be then made to reconnect the charge sources but only if the Li has not discharged completely.

Yes as far as battery isolators go my preference is a big high interrupt capability relay right next to the batttey positive , preferably arranged in fail safe mode.

 

[Goboatingnow]

can I just summarise my perspective

if you are designing for a boat , then ABYC and or ISO standards for Litiums apply , these are the baseline specs , ie , There's no point arguing!
in summary these are

1. no electrical paralleling of differing chemistries
2. a BMS ( or more correctly a cell level monitor ) to monitor LVC HVC, Over temp and over current IS required
3. A battery disconnect driven by the BMS is required and must completely isolate the Li battery if faults outlined in (2) occur
4. Advance alarms of such safety disconnect must be provided
5. standards on battery compartments etc are setout but easily achieved

If a BMS cannot so this (2,3,4 ) , then I would not touch it

this is your baseline , avoiding ABYC or ISO will must likely make your system uninsurable, and any professional design will almost certainly follow it
NOTE: cell balancing is NOT a safety issue, neither specification requires it , once you have cell level monitoring then out of balance cells are merely a SOC capacity loss issue, as I have opined elsewhere there is little or no need for incircuit cell balancing in fractional C system typical on a boat . make you cell balancing an offline annual event and you will be fine.

Li safety disconnect systems are "last ditch ", they are an indication that something has gone seriously wrong ( which is why ABYC and ISO requires alarms)

so the next steps are preferential but personal

My view is , based on 30 years in this and other sectors

1. Battery BMS driven safety disconnects should be high interrupt capacity Relay/contactor based, immediately adjacent to the positive terminal , wired in a "fail safe" config , ie will drop out if the BMS fails . A manual re-connect system should exist to restore either dc loads or re-connect charge sources . the triggering on such disconnects is a MAJOR system fault and should be treated so . Standard MOSFET disconnects have failure modes that render them less then suitable , unless they are very specialised ( 400v 400A) types , and these are very expensive and still require a power supply/ control signal to activate disconnect

2. Any other disconnects , ie DC panel , storage disconnect ( ie when you leave the boat ) should be in addition to the Li safety disconnect

3. The system should be designed so that all charge sources are compatible with and directly connected to the Li bank

4. By preference all charge sources should be controlled by the BMS , but I except that today that's complex to arrange

5. Immediately before ( based on max voltage setpoints ) the BMS triggers the safety disconnects it should trigger all charge sources to stop including the alternator

5. In support of (3) above the alternator should have a regulator compatible with Li charge profiles and support temperature protection of the alternator and provide charge shutdown via field coil current control

6. Alternator protection devices are designed to protect the alternator , on activation, pulses upto 32V can remain present and hence all electronics attached to the alternator during this phase need to be capable of surviving pulse to 32V , or only buy equipment qualified for car installation !!!. people think APD devices are there to protect 12v electronics, they are NOT . APD devices are not necessary in a proper system designed to the principles above ( note treat the commercial APD specs with a grain of salt as they do not provide the necessary response graphs to evaluate them ) ( my own APD design has such documentation , but its not a commercial offering)

7. Fusing and or other protection devices should be installed as per specification and the nature of the system being installed , no wire should be capable of being subjected to fault currents its not designed to carry

8. The system should be optionally fitted with selective disconnects as prevent LVC cutout and preferably a DC load disconnect device programmed above the LVC point should be fitted . The BMS LVC should never normally get triggered

In my view anything that deviates from the above is a compromise , I understand why people propose such compromises , usually to save money or to avoid dealing with systems that are awkward to modify ( the alternator is a classic) but I an of the view its a compromise

the issue of so called grade A or B cells has no impact on safety in my opinion once the chemistry is LFP AND the principles followed above are implemented , its entire subjective as to what value you ascribe to branded or unbranded LFP cells , once you have proper cell level monitoring, the system is as safe as practical anyway

to finish up , yes all this is way more then we typically do for 12v lead acid , partially because LA is there a long time and primarily because the fault modes of LA are not as troublesome to deal with as LFP ( LFP and LTO technology are the ONLY current Li technologies Id allow on a boat in big capacity configurations). Li is not as safe as Lead Acid precisely for the reasons we install it , ie low current demand over the capacity of the battery as well as sensitivity to very low SOC and overcharging

yes safety is a compromise , but pragmatic but comprehensive design can make the system as reasonably as safe as possible AND standards compliant

My own Lithium install is partially developed sitting on my lab bench , my next step is a alternator regulator design , my Li system embodies the principles I have outlined here ( it goes well beyond them actually )

 

[Goboatingnow]

to illustrate the issue with APDs

heres a typically TVS diode for such clamping requirements



note the very high power handling , but for short times, now to build a APD device that meets ISO 16750-2 load dumping pulse energies , actually requires several of these and in fact is a difficult design challenge especially if you want 12V electronics to survive ( heres an application note for a 24V system https://www.maximintegrated.com/en/design/technical-documents/app-notes/7/7084.html)

as you can see , even though selecting a 18V device ( most designers would go higher to avoid inadvertent activation , can result in a 37V spike in certain cases !!!. I know my MPPT controller fell over when subjected to that !!

 

[rgleason]

@Goboatingnow wrote

Iso and ABYC standards require that a safety LI disconnect disconnects all devices from the battery bus including charge sources , nothing should remain powered by the bus expect [except] the BMS. provision should be then made to reconnect the charge sources but only if the Li has not discharged completely.

Is "expect" intended to be "except"?
Is svsagres system compliant with this requirement?

• This seems like an additional requirement that has not been discussed, where does it come from?
• Also, why is it necessary?
• Just to understand better, All devices (charging and DC Panel loads) must be disconnected? The LFP battery must be totally isolated (except the BMS)? This would leave my nav equipment in the dark suddenly, even though the battery may be fine for discharge and powering Nav and essential loads. I am really confused about this.
• "LFP safety disconnect systems are "last ditch " Perhaps this is the distinction. There are two levels of disconnection. Control of charge devices like the alternator, solar, 120v charger, and control of discharge DC panel which all have to happen in advance of the "LFP safety disconnection". Goboatingnow wants all of this to be controlled by the BMS which is a pretty tall order.
• The disconnection spikes occur when the alternator is running, so integration of the Wakespeed with the BMS such that the Alternator field is disconnected well before the LFP safety disconnect occurs, would solve that problem.
• If small solar panels and Victron MPPT that are using a smart BMV or Smart shunt to provide current information are also part of the system, but not controlled by the BMS, is that a problem?
• Does svsagres system REC BMS control all charge and discharge devices?

provision should be then made to reconnect the charge sources but only if the Li has not discharged completely

Wouldn't you want to recharge the LFP if it was discharged? Why this requirement?

Svagres provided a small circuit diagram for wiring the REC Precharge that is probably not needed for this system which will not have a Multiplus with big capacitors.

It's become clear now that there are really two choices (some would say just one in the final analysis):
Fully compliant and More Expensive: REC BMS ($500) with 7713 ML-RBS relay ($225)and SmartBattery Protect ($60) total $850
Not compliant and Less Expensive: 200a 4S 12v DalyBMS common port with FETs total $150
Both systems: Wakespeed WS500 Regulator and shunt, and various Victron "Smart" bluetooth devices, various levels of monitoring and control.

I did see a DalyBMS with a Canbus connection and I wonder if those are able to send an advance disconnect signal, and if anyone has had success using this connection. Also the DalyBMS is more compact and self-contained than the fully compliant system, a very real consideration for my small boat with very little room for this equipment.

The cost difference is about $700 to comply.

Note: I have been intending on charging 200ah - 300ah LFP batteries at about .5C (actually less than 100a) occasionally when needed, using the Alternator. Is this considered partial charging rate?

Also for a DalyBMS 4S 12V 200A it appears that the max charge rate is 100A and it appears to me that the wires are too small.

Despite the need to get good LFP, I like the fact that REC BMS has active 2A cell balencing.

 

[Goboatingnow]

@Goboatingnow wrote

Is "expect" intended to be "except"?

yes sorry, typo

Is svsagres system compliant with this requirement?

No as far as I see there are devices that remain connected to the Li bank after disconnect by the BMS , this is not ISO or ABYC compliant

• This seems like an additional requirement that has not been discussed, where does it come from?
• Also, why is it necessary?
• Just to understand better, All devices (charging and DC Panel loads) must be disconnected? The LFP battery must be totally isolated (except the BMS)? This would leave my nav equipment in the dark suddenly, even though the battery may be fine for discharge and powering Nav and essential loads. I am really confused about this.

Yes after disconnection all devices except the BMS must be disconnected from the battery ( unless specified by the battery manufacturer )

ISO
"6.7 Lithium-ion battery connections — No electrical connections should be made directly to a lithium ion battery that would bypass the BMS or the protection relays, unless specified by the battery system manufacturer’s instructions."

ABYC is similar to 6.7 ISO ( the ISO spec was largely derived from the ABYC one and is part of the ABYC ISO convergence programme )

An alarm must be generated in advance and communicated , ie

ISO

"4.8 If a shut-down condition is approaching, a BMS or system should notify the operator with a visual and/or audible alarm, clearly perceptible from the main helm position, prior to disconnecting the battery from the DC system.
4.9 In normal operation, different battery chemistries should not be connected in parallel or in series. Combining/automated charging relays should not be used between systems using different chemistries.
4.10 The main battery switch or other means of manual disconnection should be used as an isolation switch, but not as a primary protection device in the output from a lithium-ion battery bank.
4.11 The system should communicate on the delivered systems communication network with all installed charging components as an integrated system. NOTE Small self-contained systems might not have internal communication capabilities
4.12 Multiple contactors are permitted (HVC, LVC, plus main), each providing specific protection from high voltage, low voltage and load isolation. A single main contactor is permitted, if the control system provides for protection from all conditions.

Note the intent of 4.11 , ie BMS controls everything , with a small system " getout" ( ie largely self contained which most boat Li system are not)
ISO does recognise that a "dead ship " is an issue and allows manual reconnection systems

"6.10 Lithium-ion battery disconnect — The operator should be able to safely operate a lithium-ion battery system disconnect switch or switches when the batteries are in an overheating situation. The lithium-ion battery system disconnect switch should be readily accessible without reaching over the lithium-ion battery. Consider locating the lithium-ion battery disconnect outside dedicated battery enclosure. Multiple disconnect switches are recommended to prevent a dead ship in the event of a LVC/HVC occurrence."

• "LFP safety disconnect systems are "last ditch " Perhaps this is the distinction. There are two levels of disconnection. Control of charge devices like the alternator, solar, 120v charger, and control of discharge DC panel which all have to happen in advance of the "LFP safety disconnection". Goboatingnow wants all of this to be controlled by the BMS which is a pretty tall order.
• Does REC BMS control all charge and discharge devices?

ISO ans ABYC essentially view Li disconnect as a safety system , its not regarded as a system that should operate under normal conditions of use ( hence the requirement for alarms prior to disconnection )

The svsagres system has a Vicron GX system in the mix, this is " technically " able to control all compatible charges sources and be fed info from the BMS, but I accept that today the ability to centrally control All charge sources is difficult
having said that the wakespeed alt has remote shutdown as do Victron MPPt controllers and some of its battery chargers , Hence its entirely practical , within limits , to have central BMS control of charging start/stop points

I would suggest that ANY decent Li install , has provision for selective shutdown on non priority DC loads , this could be as simple as a manual procedure ( triggered by a BMS setpoint notification ) or by using a number of electronic "battery saver" style switches . The goal is a LVC event is not a normal event

Wouldn't you want to recharge the LFP if it was discharged? Why this requirement?

Yes but if Li gets severely discharged you cannot simply apply full power recharging , you have to go through a "pre-qualifying " charge cycle at vert low values until the battery gets to about 15% SOC, hence most BMSs regard 20% as "dead" to prevent the need to execute a pre-qualifying charge, as most marine lithium chargers do not have pre-qualifying stages

But the risk exists that the Li battery is completely discharged , Hence my view that if a LVC disconnect is triggered , any charge source reconnection MUST be manual not automatic ( this is also true in low temperature situations (=)

Svagres provided a small circuit diagram for wiring the REC Precharge that is probably not needed for this system which will not have a Multiplus with big capacitors.

I cant tell enough from that but its not electrically correct for a start as far as I can see

It's become clear now that there are really two choices (some would say just one in the final analysis):
Fully compliant and More Expensive: REC BMS ($500) with 7713 ML-RBS relay ($225)and SmartBattery Protect ($60) total $850
Not compliant and Less Expensive: 200a 4S 12v DalyBMS common port with FETs total $150
Both systems: Wakespeed WS500 Regulator and shunt, and various Victron "Smart" bluetooth devices, various levels of monitoring and control.

Note that MOSFET total battery disconnect is fine from the standards point of view, but personally , unless its very specialised high end MOSFET switching, I would prefer relay/contactor solutions, there as industrial solutions that are cheaper then the 7713

I did see a DalyBMS with a Canbus connection and I wonder if those are able to send an advance disconnect signal, and if anyone has had success using this connection. Also the DalyBMS is more compact and self-contained than the fully compliant system, a very real consideration for my small boat with very little room for this equipment.

The cost difference is about $700 to comply.

Note: I have been intending on charging 200ah - 300ah LFP batteries at about .5C (actually less than 100a) occasionally when needed, using the Alternator. Is this considered partial charging rate?

In my view cell level monitoring for LVC HVC, over temp and over current with total high side Li disconnect ( either MOSFET or Relay ) , and advance alarms , is sufficient to be compliant . I certainly would not consider any BMS that did not meet these fundamental basics. I certainly would think twice before installing a non ABYC or ISO compatible system in a boat as it could render issues with insurers ( who have stated , Pantenious in my case , the system must be compliant and professional )

 

[rgleason]

@Goboatingnow Very helpful. Thank you. There is a lot to get right!

 

[Goboatingnow]

@Goboatingnow Very helpful. Thank you. There is a lot to get right!

yes we have several issues

* a lack of installed base expertise and hence good custom & practice
* emerging standards
* hodge podge of commercial equipment, a lot of which is not designed to co-operately work together
* users trying to leverage existing equipment , sometimes less then optimally
* conflicting advice

 

[svsagres]

yes sorry, typo

No as far as I see there are devices that remain connected to the Li bank after disconnect by the BMS , this is not ISO or ABYC compliant

At least in my system, the only devices that are connected after disconnect are a couple of bilge pumps. So technically not compliant, but I’m fine making an exception for the bilge pumps as they’re critical safety systems. Otherwise, everything is controlled either by the 7713 or the BatteryProtect, which in turn is controlled by the BMS. There is the precharge circuit as well, but that‘s mostly because the REC makes that external.

The svsagres system has a Vicron GX system in the mix, this is " technically " able to control all compatible charges sources and be fed info from the BMS, but I accept that today the ability to centrally control All charge sources is difficult
having said that the wakespeed alt has remote shutdown as do Victron MPPt controllers and some of its battery chargers , Hence its entirely practical , within limits , to have central BMS control of charging start/stop points

In the case of the Wakespeed, it’s controlled directly via CAN BUS from the BMS. I actually have it directly wired to the BMS to reduce points of failure in the system. If it loses the messages from the BMS, it will drop down into a limp mode, and set its voltage set point to 13.2v. Otherwise, it operates under command of the BMS. The WS itself doesn’t necessarily have a remote shut down, depending on how you have the feature-in wire configured. I have it to limit my alternator to around 40% so that I don’t put too much load on my itty bitty engine while underway, but I can flip the switch and run it flat out if I want to charge quickly at anchor.

The cerbo, in turn, also connects to the CAN BUS, and then also controls the MPPT and Multiplus.

 

[Goboatingnow]

At least in my system, the only devices that are connected after disconnect are a couple of bilge pumps. So technically not compliant, but I’m fine making an exception for the bilge pumps as they’re critical safety systems. Otherwise, everything is controlled either by the 7713 or the BatteryProtect, which in turn is controlled by the BMS. There is the precharge circuit as well, but that‘s mostly because the REC makes that external.

I like The REC units, I reread the manual and dont see where it supports pre-qualifying charge . The touble being that no marine chargers I know of are so equipped and there s no way to trigger a prequalfying charge cycle ( typically cell at 2.14V and then charge is <1/10-1/20 C)

personally I see no advantage in leaving the bilge pumps connected around the BMS , The battery is close to discharged when the BMS triggers anyway , then the bilge pump's will flatten and ruin your Li bank , a better way would be a dedicated LA battery for the pumps, fed by a DC DC from the Li !!!

ISO aludes this
4.5 Consideration should be given to providing power for critical systems (e.g., engine starting, navigation lights, etc.) if a BMS shuts down the battery.

Certainly if you leave this setup as is, do not activate recharging automatically after LVC disconnect

In the case of the Wakespeed, it’s controlled directly via CAN BUS from the BMS. I actually have it directly wired to the BMS to reduce points of failure in the system. If it loses the messages from the BMS, it will drop down into a limp mode, and set its voltage set point to 13.2v. Otherwise, it operates under command of the BMS. The WS itself doesn’t necessarily have a remote shut down

well if does if you configure one of its digital inputs to do so ( Im very familiar with the wakespeed as its the VSR open source project turned commercial and I have built the VSR units )