LiFePO4 Batteries: Discussion Thread for Those Using Them as House Banks

[goboatingnow]

I'll not continue this debate as you simply don't understand. Nor are you applying ohms law properly, the purpose of my original resistor is to provide the alternator with a small load then disconnection occurs.

A transzorb can be used to clamp the voltage spike to Less then 20 V , and anyway most IC based regulators already handle load dumping spikes.
Since there is no other electronics on the alternator charge lead , there nothing else other then the alternator to protect

I'm not disagreeing that disconnecting the field is the best way. It's that it requires modifications to the alternator

On a boat, with a separate alternator charge path , there is no other electronics.

The load dump on disconnecting a battery while retaining some load ( the power resistor ) results in a peak spike around 120 V for upto 400ms. This is easily clamped to 20v by a TSV or trsnszor. And these devices are readily available.

Since there is nothing else on the charge bus there nothing to protect

Protecting from regulator failure is irrelevant , disconnecting the battery and it's loads has already achieved that , what I was saying related to minimising load dump spikes on disconnects.

But away you go , absolutely mod the alternator. Having actually done one or two, I can tell you some are not at all easily done.

Designing a circuit to disconnect an alternator output can be done and is not difficult or expensive.

Ultimately a HVC must kill the charge circuit one way or the other

Adding a circuit to protect against regulation failure is a separate thing entirely


Dave


Sent from my iPad using Tapatalk

[OceanSeaSpray]

Originally Posted by goboatingnow

I'll not continue this debate as you simply don't understand. Nor are you applying ohms law properly, the purpose of my original resistor is to provide the alternator with a small load then disconnection occurs.
...
Since there is nothing else on the charge bus there nothing to protect

Protecting from regulator failure is irrelevant , disconnecting the battery and it's loads has already achieved that , what I was saying related to minimising load dump spikes on disconnects.
...
Dave

My friend, not only this is an argument with a variable geometry as you keep adapting what you are saying, but what you are saying is inept.

Your "small resistor" is not in the load path, referring to the documentation you posted yourself... and if it was it would get blown away.
"There is nothing to protect", well what about the alternator rectifiers???

It is more than Ohm's law you might want to study I am afraid, because it is not even the point here. Unbelievable. If you only tried to implement your "solutions" first, the amount of noise would be greatly reduced. Muffled by the smoke too.
Enough time wasted.

[goboatingnow]

I was originally referring to the load resistor solution

Furthermore I have done a lot of work in the past in power electronics and motor drives and I built two alternator based chargers.

The thing is to separate the " practice " or sometimes " myths " from the theory and its practical application.

I also work with Li virtually every day on various forms ranging from small to the smaller large form factors.

I have pointed out it is possible to disconnect alternators and build a survivable System , equally I have pointed out the ridiculousness of , for example switching AC on battery chargers .

The fact remains HVC events on a Li setup on a fractional C environment are very rare events. Proper cell monitoring will in fact detect imbalances and should warn you well in advance of such events.

Hence HVC disconnects should be regarded as " last gasp" scenarios where protecting the alternator is of somewhat secondary importance.

In fact given that unattended alternator charging is not recommended, an alarm alerting the user , that precedes the actual HVC disconnect would allow the engine to be shut down before the emergency HVC disconnect occurs.

This would remove any requirement to modify the alternator

I am developing on a the side a BMS for Li cells , when I'm finished I will publish the circuit as a OSH project. You can evaluate it then

In the meantime why not get an alternator a power resistor ( 100w is about $9 ) and a few suitable TVS devices. Get out your scope and actually see what happens with load dumping and its suppression. I have

( pS firstly a 14A 100A alternator is irrelevant, no one is shorting its output , secondary the load resistor is around 0,5 -1ohm , it's merely a load , dissipation of around 14-30 W , and in my experience all modern regulators will handle the un suppressed spike and will certainly handle the suppressed one , but at no point is there " kilowatts " of anything )

Dave

[poiu]

Originally Posted by gael

After 4 years using my Li banks, one concern I have is about the alternator.
As the Li batteries can accept an high C charge during almost all the charging time , the alternator become pretty hot and my concern is its lifetime.
It is a 75Amp 24V Leece neuville "heavy duty", feeding 2 *130Ahbatteries and I derated the alternator to 60Amp max. Added a blower to pick up "cold " air from the bildge, but still pretty hot.
Any suggestion how to size an alternator to keep a raisonable charging time and a good lifetime. I do understand that all alternators are NOT designed to run at 80% of the nominal capacity.... but ...
cheers
gael

I use the Mastervolt charge regulator and they recommend an alternator is sized to a maximum of 30-50% of the battery capacity to prevent overloading, so you are outside their guideline limits. I realise that recommendation will assume lead acid though and you have limited its output.

There is no current control limit on external regulators like the Mastervolt, which I have (or Balmar I believe). It is possible to work around the problem. I have found that with normal low engine revs the output of the alternator is about 60% of max anyway, so mostly output is fine. Also what I did was dial down the voltage output such that at about 80% battery charge the alternator output was at the low end of the knee voltage. The result is it outputs typically something like 65 to 30% of its current capacity at normal i.e. slow engine revs for a state of charge of nearly flat to nearly full. I do have a mental note if the battery is flat not to use high engine revs for long without checking the alternator It is only with very high engine revs that I get full alternator output. At full engine revs it guzzles fuel in a biblical way so is very rarely done.

An automatic current regulating control would be a good idea so I could forget the issue and dial it up to make a little more use of the alternator. As the genset is the main charge device and I see the alternator as a back up, things work OK like this.

[Maine Sail]

Originally Posted by poiu

I use the Mastervolt charge regulator and they recommend an alternator is sized to a maximum of 30-50% of the battery capacity to prevent overloading, so you are outside their guideline limits. I realise that recommendation will assume lead acid though and you have limited its output.

There is no current control limit on external regulators like the Mastervolt, which I have (or Balmar I believe). It is possible to work around the problem. I have found that with normal low engine revs the output of the alternator is about 60% of max anyway, so mostly output is fine. Also what I did was dial down the voltage output such that at about 80% battery charge the alternator output was at the low end of the knee voltage. The result is it outputs typically something like 65 to 30% of its current capacity at normal i.e. slow engine revs for a state of charge of nearly flat to nearly full. I do have a mental note if the battery is flat not to use high engine revs for long without checking the alternator It is only with very high engine revs that I get full alternator output. At full engine revs it guzzles fuel in a biblical way so is very rarely done.

An automatic current regulating control would be a good idea so I could forget the issue and dial it up to make a little more use of the alternator. As the genset is the main charge device and I see the alternator as a back up, things work OK like this.

Balmar regulator do have current limiting and newer models within the last two years have adaptive alt temp sensing, which finds the sweet spot, not a 100% or 50% cut of alt output. Current limiting in a Balmar is called "Belt Manager" it used to be called "Amp Manager"..... This is a very, very effective tool for LFP or large AGM installs. I use alt temp sensing as the back up to a well set up current limited alt...

The Mastervolt alts are made by Leece-Neville, last time I checked..

You may want to check your alt temp at low revs. Many high output alts can develop a tremendous amount of their current at low revs but they then lack the fan speed for effective cooling. While they are at less output due to RPM they are also lacking fan speed to cool as ideally as one might expect. Some of these alts are rated to 15,000+ ALT RPM we are generally running them at much lower speeds thus less effective fan cooling. tight engine spaces also contribute. Moving exhaust fan ducts to the top of the engine space helps too.

With many high output alts low revs actually wind up being where the alt gets hottest, due to slow fan speed, especially with LFP or large AGM banks...

When I set up alts I artificially load the bank so the alt is in bulk. Beats draining the customers bank.. I then hot run it with the vessel under loaded cruise RPM for about 40 minutes to get the engine room up to temp. I keep the engine bay closed and use a remote temp sensor and monitor alt temp. I then return to the dock and run at fast idle and monitor alt temp. More often than not the alt temp creeps up at low revs while loaded to bulk but this depends on the particular alts low RPM performance. Some are not as good as others in this regard and won't get hotter at low revs because their low RPM performance is poorer... I then adjust the regulator current limit accordingly to maintain temps below 220F....

A Sabre I was working on last week had the Balmar AT165 developing 118-122A at dead idle when hot, fast idle bumped it to 135A. These alts cool well so it only wound up at Belt Manager level 1... Some alts wind up at Belt Manager level 4-5 depending upon engine spaces & pulley ratios etc..

My own boat runs a custom built Mark Grasser DC Solutions 160A alt set to run at 115-120A. At 115-120A in my engine bay it stays below 220F and can run in bulk from 80% DOD to 100% full without bumping into temp limiting. I may eventually throw a 200A alt in, but it is low on the priority list....

Best bet I've found is to artificially load the alt, inverter and small space heater works great for most alts, and then run the boat hard and monitor alt case temp...

Try keeping the alt below 220F and it will live a happy life... When AGM's first hit the market I saw many melted down factory alts. Still see them from time to time. They were getting "cook eggs" hot because the internal reg just stayed in bulk/CC for far too long with no way to limit it. The alts were also very often undersized which also kept them in bulk/cc for far longer than their indented duty could handle...

[poiu]

Originally Posted by Maine Sail

Balmar regulator do have current limiting and newer models within the last two years have adaptive alt temp sensing, which finds the sweet spot, not a 100% or 50% cut of alt output. Current limiting in a Balmar is called "Belt Manager" it used to be called "Amp Manager"..... This is a very, very effective tool for LFP or large AGM installs. I use alt temp sensing as the back up to a well set up current limited alt...

The Mastervolt alts are made by Leece-Neville, last time I checked..

You may want to check your alt temp at low revs. Many high output alts can develop a tremendous amount of their current at low revs but they then lack the fan speed for effective cooling. While they are at less output due to RPM they are also lacking fan speed to cool as ideally as one might expect. Some of these alts are rated to 15,000+ ALT RPM we are generally running them at much lower speeds thus less effective fan cooling. tight engine spaces also contribute. Moving exhaust fan ducts to the top of the engine space helps too.

I checked again. I looked at one of the less advanced versions before and as you say the belt manager on featured on some models does indeed look like a good feature.

There is another feature on some of the Balmar regulators too I noticed that relates to this issue - temperature sensing of the alternator, where the field voltage is reduced if the alternator exceeds a user set temperature threshold. A useful protection.

Thanks for the tip on slow running. I will check it. It is a Leece-Neville 175A. 2000-8000rpm, but geared down to give 150A at about 4000rpm on the alt at max engine cruise speed and 2300rpm on the alt at slow engine cruise speed.

[goboatingnow]

BMS Discussion

As Im working ( on the side) on a simple BMS ( monitoring not management) for Li Cells, I wanted to put my thoughts down on paper , given there is always a debate about HVC and LVC and BMS etc


1.A BMS is a battery protection system, not in any way a charge control system

2. All monitoring should be a cell level, battery level monitoring is effectively useless as you are not really measuring what is going on, This include HVC, and LVC events which should be monitoring for at cell level. ( This precludes battery voltage sensing relays etc)

3. There are 4 voltage thresholds of interest , in deceending voltage, namely ( a) HVC (b) Upper charge limit, ( c) Lower charge limit and ( d) LVC

4. A BMS will user alarm on (b) and (c ) and alarm and perform either charge or load bus disconnects at ( a) and (d) on HVC and LVC respectively

The battery remains connected to the load in HVC ( this is a good thing) , and isolated from load at LVC, ( see below)

4. HVC and LVC require manual resets ( debate for LVC)

5. HVC is an anomalous event, mainly caused by cell imbalance, rather then charging failure , LVC could be common event.

6. any BMS should perform cell monitoring continuously and alert the user is a cell imbalance is detected at any part of the cycle.

7. HVC is a battery protection process, not a charge protection process. Charge control/disconnect is a separate matter

8. In normal operation, the user upper and lower charge voltages are within th control of the charge regulators not the BMS. i.e. the BMS is not there to compensate or correct for incorrect Li charge regulation settings

9. The BMS itself must survive any battery transients that occur.

PS , in the case of where for example a second typical LA battery is in the system, like typically the engine start , then HVC disconnects can simply remove the Li bank from all charge sources by simple disconnect, other wise precautions have to be taken, depending on charge source.

Anyone like to add or debates features ?

The hardware is currently ATMEL AVR based, with 4 12bit ADCs and outputs for contractors, indicator lights and an alarm buzzer.

The current software implements

Upper charge voltage reached, 4 beeps and an indicator light
Lower charge voltage reached , slow intermittent beep and an indicator light
HVC, continuous alarm and an indicator light - charge bus contactor triggered
LVC, fast intermittent beep and an indicator light - load bus contactor triggered
There is a push button to silence the alarms and for manual reset of HVC and LVC
An output can be dedicated to alarms for cell imbalance, during normal operational voltage limits ( light , etc)

There is ( well there will be ) a failsafe mode, where if any cell falls below the LVC, the BMS opens both contactors , and goes into a sleep mode. In essence pre-qualification charging maybe needed on the affected cell(s) .

I might stick the whole project up on indigogo later in the year?

Comments, features, missing things?

dave

[Roy M]

My new Lehr 15 hp propane outboard came with a LiFePO4 starter battery. I would like to install an additional small battery to run some low draw electronics on the dinghy. Can I simply parallel the additional bank in with my system? Or would a combiner to the second LiFEPO4 battery be a smarter way to achieve this?

[poiu]

Originally Posted by goboatingnow

BMS Discussion

As Im working ( on the side) on a simple BMS ( monitoring not management) for Li Cells, I wanted to put my thoughts down on paper , given there is always a debate about HVC and LVC and BMS etc

snip

I might stick the whole project up on indigogo later in the year?

Comments, features, missing things?

dave

Shooting from the hip:

Why not make a combined charge control/management system? Automation would be useful and I can't see it would reduce safety/protection. This could be an option if the buzzer was through a relay so the user could choose to use it to activate his charger.

Needs to be very low power so can be left on for extended trips away. The Sterling units for example have an almost zero consumption in sleep.

logging and network transmission of data would be nice. A panel display too (the Junsi is a pain to read).

[mrm]

Originally Posted by goboatingnow

BMS Discussion

As Im working ( on the side) on a simple BMS ( monitoring not management) for Li Cells, I wanted to put my thoughts down on paper , given there is always a debate about HVC and LVC and BMS etc
[...]
4. HVC and LVC require manual resets ( debate for LVC)

[...]

Anyone like to add or debates features ?

The hardware is currently ATMEL AVR based, with 4 12bit ADCs and outputs for contractors, indicator lights and an alarm buzzer.

[...]

Comments, features, missing things?

dave

Dave, good idea.

Re 4, LVC reset, I think it would be good if the user had a choice (a jumper or a setup option) to either make LVC reset manual or make it automatic with a fixed or adjustable hysteresis.

Independent small current relay contact pairs (or maybe just OC drivers) for all four events (LVC, LVW, HVW, HVC) would increase flexibility for users/installers.

Atmel AVR is a good choice and very popular, and with house bank capacities I don't think it is power hungry, but if power draw will become a consideration, there is a MSP430 line of microcontrollers from Texas Instruments, which are really low power. However, programming them is a pain when compared with Atmel AVR (IMHO).

Your design objectives look OK to me. How many cells in series maximum?

[durundal]

Originally Posted by goboatingnow

Anyone like to add or debates features ?
...
An output can be dedicated to alarms for cell imbalance, during normal operational voltage limits ( light , etc)

There is ( well there will be ) a failsafe mode, where if any cell falls below the LVC, the BMS opens both contactors , and goes into a sleep mode. In essence pre-qualification charging maybe needed on the affected cell(s) .

I might stick the whole project up on indigogo later in the year?

Comments, features, missing things?

dave

Sounds like you're making something I would like to buy, here's my laundry list of features:

-Encapsulated electronics (potted or at least conformal coated)
-Mounting features that don't require diving boarding it directly off the cell lugs (a flat back and holes for screws/bolts would be my preference)
-Ability to use outputs to drive latching relays (or other non-continuously energized triggers) like this during HVC/LVC: ProLatch-R - Zero Volt drop Latching Relay, 160 amp marine grade relay
-Wire interfaces large enough to use crimped connections (18ga or bigger)
-if keeping costs down removed displays, pulse the lights to indicate which cell caused the problem for a HVC/LVC
-Dedicated power LED that will always be lit if power is at the terminals
-Run on 12V input

[FlyingCloud1937]

Bill,

A simple thing to think about is taking a page for the early adopters that went off grid with big wind machines.

At the first HVC alarm mode, switch in a 12 volt HW immersion heater into the domestic HW tank. Now we are not wasting any power. If this isn't enough to bring the pack voltage down, then the 2nd HVC event can trigger the charge disconnect. The immersion heater will adsorb the current/voltage spike allowing time to cut out the alternator, and or other charging sources.

Lloyd

Originally Posted by goboatingnow

BMS Discussion

As Im working ( on the side) on a simple BMS ( monitoring not management) for Li Cells, I wanted to put my thoughts down on paper , given there is always a debate about HVC and LVC and BMS etc


1.A BMS is a battery protection system, not in any way a charge control system

2. All monitoring should be a cell level, battery level monitoring is effectively useless as you are not really measuring what is going on, This include HVC, and LVC events which should be monitoring for at cell level. ( This precludes battery voltage sensing relays etc)

3. There are 4 voltage thresholds of interest , in deceending voltage, namely ( a) HVC (b) Upper charge limit, ( c) Lower charge limit and ( d) LVC

4. A BMS will user alarm on (b) and (c ) and alarm and perform either charge or load bus disconnects at ( a) and (d) on HVC and LVC respectively

The battery remains connected to the load in HVC ( this is a good thing) , and isolated from load at LVC, ( see below)

4. HVC and LVC require manual resets ( debate for LVC)

5. HVC is an anomalous event, mainly caused by cell imbalance, rather then charging failure , LVC could be common event.

6. any BMS should perform cell monitoring continuously and alert the user is a cell imbalance is detected at any part of the cycle.

7. HVC is a battery protection process, not a charge protection process. Charge control/disconnect is a separate matter

8. In normal operation, the user upper and lower charge voltages are within th control of the charge regulators not the BMS. i.e. the BMS is not there to compensate or correct for incorrect Li charge regulation settings

9. The BMS itself must survive any battery transients that occur.

PS , in the case of where for example a second typical LA battery is in the system, like typically the engine start , then HVC disconnects can simply remove the Li bank from all charge sources by simple disconnect, other wise precautions have to be taken, depending on charge source.

Anyone like to add or debates features ?

The hardware is currently ATMEL AVR based, with 4 12bit ADCs and outputs for contractors, indicator lights and an alarm buzzer.

The current software implements

Upper charge voltage reached, 4 beeps and an indicator light
Lower charge voltage reached , slow intermittent beep and an indicator light
HVC, continuous alarm and an indicator light - charge bus contactor triggered
LVC, fast intermittent beep and an indicator light - load bus contactor triggered
There is a push button to silence the alarms and for manual reset of HVC and LVC
An output can be dedicated to alarms for cell imbalance, during normal operational voltage limits ( light , etc)

There is ( well there will be ) a failsafe mode, where if any cell falls below the LVC, the BMS opens both contactors , and goes into a sleep mode. In essence pre-qualification charging maybe needed on the affected cell(s) .

I might stick the whole project up on indigogo later in the year?

Comments, features, missing things?

dave

[goboatingnow]

Originally Posted by poiu

Shooting from the hip:

Why not make a combined charge control/management system? Automation would be useful and I can't see it would reduce safety/protection. This could be an option if the buzzer was through a relay so the user could choose to use it to activate his charger.

Needs to be very low power so can be left on for extended trips away. The Sterling units for example have an almost zero consumption in sleep.

logging and network transmission of data would be nice. A panel display too (the Junsi is a pain to read).

Thanks , interesting stuff, I might do a serial output of data , that would allow others to add display and or user interface stuff

Dave


Sent from my iPad using Tapatalk

[Maine Sail]

Originally Posted by goboatingnow

Thanks , interesting stuff, I might do a serial output of data , that would allow others to add display and or user interface stuff

Dave


Sent from my iPad using Tapatalk

Dave,

I have a few items:

*HTC - High Temperature Cut Individual cell temp monitors for each cell.

*LTCC - Low temp CHARGING cut. Loads still allowed but no charging. Same monitors but no charging allowed below 0C/32F..

*I would also prefer a device that can drive latching relays such as the Blue Sea 7712 etc..

*Can incorporate charge HVC / Load LVC cuts at a warning level and emergency level.

*Can have LVC and HVC levels adjustable as well as hysteresis on warning level

*user configurable warning and emergency level outputs selectable normally closed or normally open (gives more options when choosing relays & contactors)

*Any balancing must be done manually and preferably "attended" just like equalization charging. A manual override of warning level HVC would be required to drive the bank to shunting levels.

I know I have more and will think of them the minute I am away from the computer.....

[goboatingnow]

Originally Posted by durundal

Sounds like you're making something I would like to buy, here's my laundry list of features:

-Encapsulated electronics (potted or at least conformal coated)
-Mounting features that don't require diving boarding it directly off the cell lugs (a flat back and holes for screws/bolts would be my preference)
-Ability to use outputs to drive latching relays (or other non-continuously energized triggers) like this during HVC/LVC: ProLatch-R - Zero Volt drop Latching Relay, 160 amp marine grade relay
-Wire interfaces large enough to use crimped connections (18ga or bigger)
-if keeping costs down removed displays, pulse the lights to indicate which cell caused the problem for a HVC/LVC
-Dedicated power LED that will always be lit if power is at the terminals
-Run on 12V input

Conformal coating maybe, interesting the justification for conformal coating is not really justified on basically all digital circuitry. Good gold plated boards are often better

Mounting. Personally I prefer running a fused wire to each cell and a single BMS board that can be mounted outside the battery box.

The alternative idea is a two board solution, with the analog sensing on a small board at the battery box, with a serial link to a board mounted behind the panel, hence easier to wire in indications and button to a panel

Latching relays good idea , kgtechnologies make a good line , sterlings is very dear , there are several manufacturers that provide low power operation by using magnets.

The board will have on board small signal relays to all driving a wide range of contactors/relays including a relay that operates before the contactor relays , this can used for make before break , or things like field disconnect of a alternator

Wire interface , absolutely , no Mickey Mouse wires.

Currently not proposing a display , but the system could easily drive a cheap 2x16 LCD

The board would provide outputs for the various lights ( HVC, upper limit, lower limit, LVC , imbalance ) two buttons ( alarm silence , reset contactors )

I will probably put the LEDs and buttons on the board directly and publish a panel cad to allow people to make the panels ( front panel express etc )

Dave


Sent from my iPad using Tapatalk