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

[T1 Terry]

Originally Posted by mrm

Ah, ok, so the logger alarm port is an open collector type drive.

Understood, this means some logic will need to go to the 'HVC interlock' block.

And I see a potential problem here.

Scenario:

0. Lead acid voltage detector (the triangular item on the left, next to LA battery) has a hysteresis built in. This is required to prevent contactor being repeatedly bounced ON-OFF when charging current hits LiFePO and causes a voltage drop due to higher load on charging source(s). Let's assume Von = 13.2V and Voff = 12.9V;

1. charging starts, Lead acid goes > 13.2V ==> relay goes ON;

2. charging continues, voltages rise, charging stops before HVC levels, so no HVC condition;

3. since there was no HVC, LA and LiFePO remain connected, through a bidirectional high current capable path;

4. since there is hysteresis (as described in 0.), relay will stay ON till voltage in BOTH batteries drops below 12.9V (or some other hysteresis set-point).

Now, it does not look too good to me, but maybe it is OK?

I added more descriptions to the diagram: http://infinity.eti.pg.gda.pl/tmp/LA-LiFePOv2.png

Thanks for the more detailed drawing, all makes sense now. You raise a very good point, one possible solution would be a solenoid driven by the engine being switched on, this could even be a leg to the voltage comparator so it would effectively see zero volts and open the circuit. this would need an over ride if mains/generator charging was to be used in place of alternator charging, a manual switch would probably be the simplest method with a warning light to indicate the system was energised, probably handy for either charging system so the operator knew when the anchor which battery reached fully charged, a quick reference to the cell logger would give a further clarification as to why the charging system had shut down, an alarm like a smoke alarm could be also connected across the cell logger alarm terminals as a back up warning if the system failed to disconnect.
I think a suitable relay complete with voltage comparator is currently available from Redarc in Adelaide Aust, they call it a VCR (voltage sensing relay) they are commonly used in motorhome electrics. That just leaves the HVC interlock block to figure out.

T1 Terry

[deckofficer]

Originally Posted by T1 Terry

This battery is roughly 3 yrs old, it showed no signs of being past it'd use by date till this test.

T1 Terry

Drives home the point that LiFePO4 is the way to go for house banks. Most people would never run A/C off the inverter with LA, but sized right it would be child's play for LiFePO4.

[Maine Sail]

Originally Posted by T1 Terry

Is this alternator dedicated to house battery charging? Do you plan to use a Junsi logger to drive a charge disconnect relay if a cell goes over 3.6v?

T1 Terry

I have a BMS with HVC relay for the regulator. I also have a Junsi ordered but more for visual and tracking.. Alternator is charging direct to the house bank but the HVC relay can cut it. I don't want to get to the HVE point though and would prefer to keep the voltages out of HVC cut range.

[Lagoon4us]

Originally Posted by deckofficer

Welcome to the club. What is the 5 volt charger for, did I miss something?

Woops that should read 3.65Volt apologies.

[deckofficer]

Originally Posted by Lagoon4us

Woops that should read 3.65Volt apologies.

Is this a special charger for balancing LiFePO4 cells? If so, what kind of amps can it push? Maybe a link to it? Is it this one at 20 amps? http://www.ev-power.eu/Chargers-Sing...ell.html?cur=1

[Lagoon4us]

EV-Power | Charger 3.6V/6A for LiFePO4 cells (1 cell)

Pavel at ev-power recommended the 6 amp unit, and only top balance as these batteries are brand new.

Cheers

[Lagoon4us]

ev-power offer the most basic of BMS units that give the ability to shut the cells down should H or LV occur, for someone like myself (probably typical yachtie) moving ahead with installation it becomes simpler to wire these proprietry units in rather than construct a board from the beginning.

Unless consensus is reached here on design and someone here makes up control boards then i think it best for the likes of me to opt for a proven and available system.

[mrm]

Originally Posted by T1 Terry

Thanks for the more detailed drawing, all makes sense now. You raise a very good point, one possible solution would be a solenoid driven by the engine being switched on, this could even be a leg to the voltage comparator so it would effectively see zero volts and open the circuit. this would need an over ride if mains/generator charging was to be used in place of alternator charging, a manual switch would probably be the simplest method with a warning light to indicate the system was energised, probably handy for either charging system so the operator knew when the anchor which battery reached fully charged, a quick reference to the cell logger would give a further clarification as to why the charging system had shut down, an alarm like a smoke alarm could be also connected across the cell logger alarm terminals as a back up warning if the system failed to disconnect.
I think a suitable relay complete with voltage comparator is currently available from Redarc in Adelaide Aust, they call it a VCR (voltage sensing relay) they are commonly used in motorhome electrics. That just leaves the HVC interlock block to figure out.

T1 Terry

I modified my drawing to allow for any number of 'request relay ON' inputs. Active high.

If we drop the request for HVC latching, then HVC interlock circuit is pretty simple. Again, any number of 'force relay OFF' is possible. Active low.

If you think that HVC latching is a must, then a place to latch the signal is marked on the diagram. I can draw a very simple solution for that block if you like.

The more I think about it, the less I like the voltage sensing method. Too many ifs. Dedicated 'charging present - OK to switch relay ON' signals seem much more robust.

However, there is one more scenario I am not comfortable with:

1. you are forced to pull anchor;

2. engine does not work (for whatever reason) so you will sail away;

3. pulling anchor draws windlass LiFePO bank to near empty (maybe several attempts to free anchor were needed);

4. you sail away, repair your engine, fire it. Soon 'Relay ON' signal is active, relay goes ON;

5. LA bank is full, LiFePO bank is empty. Both are now connected with a high current capable link....

Something may start to smoke...

Marius

[ebaugh]

Originally Posted by Maine Sail

I have a BMS with HVC relay for the regulator. I also have a Junsi ordered but more for visual and tracking.. Alternator is charging direct to the house bank but the HVC relay can cut it. I don't want to get to the HVE point though and would prefer to keep the voltages out of HVC cut range.

I think I remember you are using one of the Clean Power Auto BMS systems. If you don't clip the balance resistors, some value below where they are activated. If you leave them, they may be useful for an occasional observed balance if ever needed.

I'd set bulk to 3.45V or 13.8, but that has to be measured at the battery at max output after the alternator temperature regulation has kicked in. It will likely be a higher value setting. Set bulk time to .8xC/charge rate. Set Absorption and Float to 13.4, also measured at the battery, but this one will measure about the same at the battery as the alternator. Max time of absorption to zero if it takes that. Remove battery temperature sensor, it's not needed. No equalization allowed. I did these from memory, Internet too slow to download the manual for you alternator controller today.

The idea is max rate, then quit. It will take the full current, until charged. Make sure the alternator controller switches to Float based on voltage, you should never hit the time setting. You could hold charge at the 13.8 for awhile using the absorption settings, but this value is hard to set since the voltage will rise as the current falls. might be worth experimenting a bit. But I have not done that. At 100A, the voltage at the battery is .4 V less than at the charge source. The charger is 10 feet away from the battery, via 4/0 cable, 1 battery switch, 1 fuse, 1 EV200 Solenoid and 1 Shunt. So there is some resistance in all those connections causing the voltage drop.

[Maine Sail]

Originally Posted by ebaugh

I think I remember you are using one of the Clean Power Auto BMS systems. If you don't clip the balance resistors, some value below where they are activated. If you leave them, they may be useful for an occasional observed balance if ever needed.

I'd set bulk to 3.45V or 13.8, but that has to be measured at the battery at max output after the alternator temperature regulation has kicked in. It will likely be a higher value setting. Set bulk time to .8xC/charge rate. Set Absorption and Float to 13.4, also measured at the battery, but this one will measure about the same at the battery as the alternator. Max time of absorption to zero if it takes that. Remove battery temperature sensor, it's not needed. No equalization allowed. I did these from memory, Internet too slow to download the manual for you alternator controller today.

The idea is max rate, then quit. It will take the full current, until charged. Make sure the alternator controller switches to Float based on voltage, you should never hit the time setting. You could hold charge at the 13.8 for awhile using the absorption settings, but this value is hard to set since the voltage will rise as the current falls. might be worth experimenting a bit. But I have not done that. At 100A, the voltage at the battery is .4 V less than at the charge source. The charger is 10 feet away from the battery, via 4/0 cable, 1 battery switch, 1 fuse, 1 EV200 Solenoid and 1 Shunt. So there is some resistance in all those connections causing the voltage drop.

I am not using the battery temp sensor just the alt temp sensor.

Are you suggesting that I sense the voltage at the regulator vs. the battery? I was going to sense the charge voltage at the battery so as not to over shoot my absorption settings...

Max rate then quite is what I am after and was thinking of a lower "float" than 3.35V per cell. That way the reg would shut down until the voltage dropped down into the 13.0V Bank / 3.25 VPC range..?

I have been running my "average" loads on the bench and the bank holds voltage above 3.30V per cell under a 10A load which would keep my alt running continuously.

[ebaugh]

Originally Posted by Maine Sail

I am not using the battery temp sensor just the alt temp sensor.

Are you suggesting that I sense the voltage at the regulator vs. the battery? I was going to sense the charge voltage at the battery so as not to over shoot my absorption settings...

Max rate then quite is what I am after and was thinking of a lower "float" than 3.35V per cell. That way the reg would shut down until the voltage dropped down into the 13.0V Bank / 3.25 VPC range..?

I have been running my "average" loads on the bench and the bank holds voltage above 3.30V per cell under a 10A load which would keep my alt running continuously.

If you can sense voltage right at the final positive terminal at the battery that's the best place to do it, My inverter chargers won't do that, hence the earlier discussion. You may be able to avoid this entirely with your alternator controller.

In reference to float, you have the test setup to actually get us some SOC vs Charge to Voltage values here. I figure my 3.35V per cell float at zero charge current represents about 90% capacity. But it's just a reasonable guess. By 3.25V per cell you will be below 50%, I think? I have single stage taper alternators and can only set a max float voltage and I use 3.35V, but I'm also a powerboat in the Caribbean occasionally doing 2-3 days at sea with both alternators running. At marinas I do reduce that to more like 3.33 for long floats. At anchor, I almost never get to 3.45, unless my wife has a lot of baking to do on the generator (all electric boat no solar or wind).

When you shut down the engine after a long run (if you do that), where do you want the SOC? At anchor, you just kill the engine if using it for charging when complete to your satisfaction.

[deckofficer]

Originally Posted by ebaugh

If you can sense voltage right at the final positive terminal at the battery that's the best place to do it, My inverter chargers won't do that, hence the earlier discussion. You may be able to avoid this entirely with your alternator controller.

In reference to float, you have the test setup to actually get us some SOC vs Charge to Voltage values here. I figure my 3.35V per cell float at zero charge current represents about 90% capacity. But it's just a reasonable guess. By 3.25V per cell you will be below 50%, I think? I have single stage taper alternators and can only set a max float voltage and I use 3.35V, but I'm also a powerboat in the Caribbean occasionally doing 2-3 days at sea with both alternators running. At marinas I do reduce that to more like 3.33 for long floats. At anchor, I almost never get to 3.45, unless my wife has a lot of baking to do on the generator (all electric boat no solar or wind).

When you shut down the engine after a long run (if you do that), where do you want the SOC? At anchor, you just kill the engine if using it for charging when complete to your satisfaction.

If nobody has done this, I think I will. At the 3.7 amp draw of the load I've been using allowed my 100 a-hr cells to produce 165 a-hr. I could divide discharge into a group of 10 stages, after each stage allow the cells to rest, then take a voltage reading. Since over the discharge curve our cells don't give up potential (volts) the way that LA does, it could be interesting.

So, before I do it, has anyone already done it?

[ebaugh]

Originally Posted by deckofficer

If nobody has done this, I think I will. At the 3.7 amp draw of the load I've been using allowed my 100 a-hr cells to produce 165 a-hr. I could divide discharge into a group of 10 stages, after each stage allow the cells to rest, then take a voltage reading. Since over the discharge curve our cells don't give up potential (volts) the way that LA does, it could be interesting.

So, before I do it, has anyone already done it?

I think you need to charge for each test to measure accurately. Im also a bit concerned about the 165 Ah figure out of 100 Ah cells. It should be more, but 165% is better than anything I would expect.

[deckofficer]

Originally Posted by ebaugh

I think you need to charge for each test to measure accurately. Im also a bit concerned about the 165 Ah figure out of 100 Ah cells. It should be more, but 165% is better than anything I would expect.

I agree. On the last draw that I got the 165 a-hr figure when cell voltages were 3.14, 3.02, 3.12, 3.16, so I stopped the draw down due to the 3.02 volt cell.

What do you think it should be at a 3.7 amp draw? Maybe 180 a-hr?

So it doesn't interfere with my daily routine and sleep, I'll assume 185 a-hr at 3.7 amp draw, and do it in 20% steps instead of 10% which will work out to (5) sessions of 10 hours draws. This will allow me to time the start of the test so that it will end during evening hours while I'm still awake and can charge right back up in under 2 hours. I just can't leave cells at a deep discharge for any length of time, it is against my battery religion.

[T1 Terry]

Originally Posted by mrm

I modified my drawing to allow for any number of 'request relay ON' inputs. Active high.

If we drop the request for HVC latching, then HVC interlock circuit is pretty simple. Again, any number of 'force relay OFF' is possible. Active low.

If you think that HVC latching is a must, then a place to latch the signal is marked on the diagram. I can draw a very simple solution for that block if you like.

The more I think about it, the less I like the voltage sensing method. Too many ifs. Dedicated 'charging present - OK to switch relay ON' signals seem much more robust.

However, there is one more scenario I am not comfortable with:

1. you are forced to pull anchor;

2. engine does not work (for whatever reason) so you will sail away;

3. pulling anchor draws windlass LiFePO bank to near empty (maybe several attempts to free anchor were needed);

4. you sail away, repair your engine, fire it. Soon 'Relay ON' signal is active, relay goes ON;

5. LA bank is full, LiFePO bank is empty. Both are now connected with a high current capable link....

Something may start to smoke...

Marius

The switch on point that the relay closes is 13.2v? Is that correct? If so, the exhausted lithium battery with no load will be at 12v plus, this gives a 1.2v differential, the lithium battery will increase it's terminal voltage very quickly to around 13v, there may be a high inrush current, but it wouldn't last very long, a few minutes at the very most. I've never tested what the inrush current would be, I would imagine that the house battery voltage would drop with such a sudden current draw, so the voltage differential may as low as the 12.9v relay cut off voltage and 12v at the lithium battery, that would leave only a 0.8v differential, the current would be low. I think, between the resistance within the cables etc and the voltage drop at the house battery, the current would be self regulating. I would expect the house battery and lithium battery would climb to 13.8v together so the max continuous current flow would be the max charging capability, if the cabling and components were rated to an additional say 10%, they would all be capable of handling the brief inrush current.

I'm interested in how the latching part works, I know how to do it with mechanical DPST relay, switch the negative and loop the N/O terminal to the negative driver terminal, negative to common and load to N/C, when the relay switches so the common is linked to N/O terminal, the relay is self powered and will stay in the position till either the positive is cut or the loop between the N/O terminal and negative driver terminal is opened, the relay then resets.

T1 Terry