Announcement

Collapse
No announcement yet.

So, lets spec out and set up a system of LiFePo4 batteries and inverter

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #46
    Originally posted by inMichigan View Post
    My charging strategy was attempting to always remain low enough to avoid the bleeder function kicking in. However, had I ever crept over that limit (or if the limit is not a on/off function), that is certainly a good way to explain my spread of voltages. Now that I've removed the boards, time will tell.

    Can anyone comment if the bleeding function is a discrete switch or analogy by-pass filter behavior for Mini-BMS boards?

    inMichigan
    It could be the boards, but some voltages are up and some are down, that I don't understand. I am not sure that I ever saw Micro Cycles addressed ( where you cycle rapidly between charge and discharge or doing both at the same time at top of your range ) in LiFepo4 cells. Your charge and sell voltages are close together. You mentioned how small your bank is, I know that FLA users had problems with ripple currents with a small bank, that doesn't seem like your issue.

    The guy at Clean Power Auto is named Dimitri, he could tell you anything you want to know about the boards, LFP is his wheel house.

    I'd keep a eye on them and see if it corrects it.

    Comment


    • #47
      Originally posted by inMichigan View Post
      My charging strategy was attempting to always remain low enough to avoid the bleeder function kicking in. However, had I ever crept over that limit (or if the limit is not a on/off function), that is certainly a good way to explain my spread of voltages. Now that I've removed the boards, time will tell.

      Can anyone comment if the bleeding function is a discrete switch or analogy by-pass filter behavior for Mini-BMS boards?

      inMichigan
      Not certain which Vampire Board you have, but most are Passive Dumb Boards. They are either engaged or disengaged but never off. No two boards draw thee exact amount of power, and they are the cause of imbalance. They are parasites that sap the energy out of your battery at different rates.

      Unless you spend the big money on Centralized BMS like Orion, you over charge every battery except one. They are dumb stupid boards. In order to really work, they must communicate with your charger. All of them are limited to how much current they can bypass, and none of them can Balance and out of Balance pack in any reasonable time.

      They all have a fixed Bypass Current. Most are very small and can only bypass 1 amp or less. There are two very expensive ones I know that actually work good. They can bypass up to 3 amps and communicate with your charger to control it. Very few if any use them. Mostly commercial EV's. I know of no DIY that can aford or use them.

      Anyway you most likely have a Dumb Vampire Board. They trun on at 3.5 to 3.6 volts and can only bypass 1 amp or less. Typically is .15 to .5 amps. They turn off once the voltage is bled down to a set point lower than they turn on. Thus why they are called Vampires and Parasite. So here we go charging say at 10 amps on a 100 AH battery. Eventually the first board turns on and bypasses say .5 amps around the cell, but still allows 9.5 amps to flow in the cell over charging it cause it to heat up and loose capacity. Then the next cell reaches the set point and that board turns on shunting .5 amps around the cell leaving 9.5 amps flowing through th ecell over charging it. This continues until you get to the last few cells, and finally the charger starts to taper current off toward 0 amps. Finally th elast vampire board turns on and the last cell is now fully charged up and current finally stops. All but one or two cells is severely over charged. The Vampire boards remain on until the voltage is bled off to safe levels and finally turns of. Turns off is not the right word. It just removes the Bypass circuit. They never turn off. Only time they turn off is if you disconnect them or they fail. Most when fail fail shorted out and quckly discharge the battery to destruction. If you are lucky they will burn themselve up amnd go open circuit before destroying the battery.

      I don't know about you, but I see two major flaw here. Do you know what it is? The big obvious one is the boards do not control the charger. Only a very select few expensive Centralized BMS have that capability, and those are made for a EV only and customized AC LFP charger made to work with it. All of those chargers use CAN BUSS protocol developed for the EV industry.When that first cell reaches 100% or whatever level you choose must communicate with the charger to cut charge current back to whatever value the Bypass circuit can bypass. So if you have a typical .15 to .5 amps, charge current shall be cut back to that value so you do not overcharge any cell.

      That should make you ask and see the next flaw. On 10/24 you were severely out of balance. Some 50% capacity. You were just a few Minutes from a melt down on th elow cells when they had ZERO capacity and most still had over 25%. Those cells with capacity left will destroy the adjacent cells with no capacity left in them. If your LVD did not operate in time, you would have destroyed cells with no warning unless you were monitoring it yourself. But that is not my point. Let's say from high to low you had 10% spread or 10 AH difference on a 100 AH pack. How many days would it take a Solar system to Balance your pack at .15 to .5 amps. Simple you would never get there if you regulated current. It could work quickly if the Vampiire Boards could bypass say 5 amps, but none you can buy can do that. Only Commercial EV's have that capability.

      I also bet you were were completely unaware of that 5 minutes ago. Am i right? Don't feel bad, very few people know that. Only experienced professionals know that. It is a Secret manufactures do not want you to know. Slick Willy has no clue.
      MSEE, PE

      Comment


      • #48
        Testing Balancing boards

        Originally posted by inMichigan View Post
        Removed the boards this past weekend... If I run for months and do stay in balance, I would then conclude with my own data that vampire boards must not be installed on banks using a bottom balance strategy.

        In hindsight, once I decided to not proceed with a contactor circuit setup for a low voltage disconnect system, I should have taken them off. I also regret not serializing each board with a number to match up against my batteries (tracked by serial number)....one of those boards is really not like the others.

        inMichigan
        I wonder if it is worth testing the balancing boards to see if you can get to the cause of the problem. I would think connecting them each to a bench power supply and applying different voltages to each individual board, say 3.0 volts and 3.4 volts and measuring the current consumption of the board, then keep increasing the voltage until the board starts balancing and note this voltage.

        Simon
        Off-Grid LFP(LiFePO4) system since April 2013

        Comment


        • #49
          Originally posted by Willy T View Post
          It could be the boards, but some voltages are up and some are down, that I don't understand. I am not sure that I ever saw Micro Cycles addressed ( where you cycle rapidly between charge and discharge or doing both at the same time at top of your range ) in LiFepo4 cells. Your charge and sell voltages are close together. You mentioned how small your bank is, I know that FLA users had problems with ripple currents with a small bank, that doesn't seem like your issue.
          Now that the Mini-BMS's are removed, if the spreading voltage goes away, that's the source. If it reappear, ripples currents would be my next suspect. I don't have an oscilloscope. On the +/- bus of the batteries, if I switch my Fluke DVM into AC mode, I recall I only saw 0.2 volts AC riding on top of the 57 VDC. I should know by Christmas...

          The +/- deviation in mV is relative to the average. It could be the deviations are all negative or all positive, it's just that the amount of drift is not constant, so relative to the average, it appears like a + & -.

          I could lower my sell voltage, but my charge voltage is already close enough to the top, especially with the unknown drift lurking. Here is the normal ebb/flow over 3 days with this delta between charge and sell:

          I realize the charge ripple is a 120 HZ critter, not related to the above graph spanning 3 days.

          If I lowered the sell, I'd just see this 'rescale'. What do you think the benefit would be?


          I'll mention the disadvantage of going too low, because this happened to me once. I was not selling to the grid because net metering was not switched on (so my meter was charging me to export or use power), so the sun usually topped off the batteries. Tech support changed my sell voltage to something like 52V... a few days later, when I re-engaged 'sell', the batteries sitting at 58V tried to dump all that power (58V - 52V) via the two 8kw inverters. The 200A fuse on the battery cables popped. This was in early July.

          inMichigan
          Attached Files

          Comment


          • #50
            Originally posted by karrak View Post
            I wonder if it is worth testing the balancing boards to see if you can get to the cause of the problem. I would think connecting them each to a bench power supply and applying different voltages to each individual board, say 3.0 volts and 3.4 volts and measuring the current consumption of the board, then keep increasing the voltage until the board starts balancing and note this voltage.

            Simon
            I thought about that as well. I'm not sure if I can measure power consumption at that level. Let's assume the bleeding used enough current to shift the voltage of a single 100AH cell shift by 0.2 volts over 3 months. Trying to get a sense of the scale, let's say 80AH was stored between 3.5 and 3.1 volts, that 80AH / 0.4 volts, so that 0.2 shift was worth 40 AH.
            40 AH = 48*24 Amin = 48*24*60 Asec = 70,000 Asec
            3 months is 89 days, or 8M seconds

            70,000 Asec / 8M sec = 0.010 amps....

            assuming that's a good assumption, 10 mA... hmm, that doesn't seem too bad. And in hindsight, those pretty blinking red LEDs.... how many mA's used by those? (how much variation from light to light)

            Thoughts on my rule of thumb sense of scale?

            To be honest, at this point, I don't plan to put them back onto the batteries. I tasted the BMS coolaid, but only partially swallowed... but I feel a science project when winter weather drives me inside.

            inMichigan

            Comment


            • #51
              Originally posted by Sunking View Post
              Not certain which Vampire Board you have, but most are Passive Dumb Boards. They are either engaged or disengaged but never off. No two boards draw thee exact amount of power, and they are the cause of imbalance. They are parasites that sap the energy out of your battery at different rates.
              I 'had' these: http://www.electriccarpartscompany.c...-BMS_c_55.html
              They're now in the drawer of parts.

              Originally posted by Sunking View Post
              Unless you spend the big money on Centralized BMS like Orion, you over charge every battery except one. They are dumb stupid boards. In order to really work, they must communicate with your charger. All of them are limited to how much current they can bypass, and none of them can Balance and out of Balance pack in any reasonable time.
              Maybe EV chargers even have this as a possibility, but so far, I haven't noticed a Home RE inverter with communication to a BMS. I never use the charger function of my inverter, as it makes no sense to use grid power to charge my batteries except for testing purposes. My charge come from the PV FlexMAX 80 chargers, and certainly have no advanced communication. They essentially use bank voltage for everthing.

              Originally posted by Sunking View Post
              Anyway you most likely have a Dumb Vampire Board. They trun on at 3.5 to 3.6 volts and can only bypass 1 amp or less. Typically is .15 to .5 amps. They turn off once the voltage is bled down to a set point lower than they turn on. Thus why they are called Vampires and Parasite. So here we go charging say at 10 amps on a 100 AH battery. Eventually the first board turns on and bypasses say .5 amps around the cell, but still allows 9.5 amps to flow in the cell over charging it cause it to heat up and loose capacity. Then the next cell reaches the set point and that board turns on shunting .5 amps around the cell leaving 9.5 amps flowing through th ecell over charging it. This continues until you get to the last few cells, and finally the charger starts to taper current off toward 0 amps. Finally th elast vampire board turns on and the last cell is now fully charged up and current finally stops. All but one or two cells is severely over charged. The Vampire boards remain on until the voltage is bled off to safe levels and finally turns of. Turns off is not the right word. It just removes the Bypass circuit. They never turn off. Only time they turn off is if you disconnect them or they fail. Most when fail fail shorted out and quckly discharge the battery to destruction. If you are lucky they will burn themselve up amnd go open circuit before destroying the battery.

              I don't know about you, but I see two major flaw here. Do you know what it is? The big obvious one is the boards do not control the charger. Only a very select few expensive Centralized BMS have that capability, and those are made for a EV only and customized AC LFP charger made to work with it. All of those chargers use CAN BUSS protocol developed for the EV industry.When that first cell reaches 100% or whatever level you choose must communicate with the charger to cut charge current back to whatever value the Bypass circuit can bypass. So if you have a typical .15 to .5 amps, charge current shall be cut back to that value so you do not overcharge any cell.

              That should make you ask and see the next flaw. On 10/24 you were severely out of balance. Some 50% capacity. You were just a few Minutes from a melt down on th elow cells when they had ZERO capacity and most still had over 25%. Those cells with capacity left will destroy the adjacent cells with no capacity left in them. If your LVD did not operate in time, you would have destroyed cells with no warning unless you were monitoring it yourself. But that is not my point. Let's say from high to low you had 10% spread or 10 AH difference on a 100 AH pack. How many days would it take a Solar system to Balance your pack at .15 to .5 amps. Simple you would never get there if you regulated current. It could work quickly if the Vampiire Boards could bypass say 5 amps, but none you can buy can do that. Only Commercial EV's have that capability.

              I also bet you were were completely unaware of that 5 minutes ago. Am i right? Don't feel bad, very few people know that. Only experienced professionals know that. It is a Secret manufactures do not want you to know. Slick Willy has no clue.
              Not only was I out of balance on the 24th, even back on Oct 2, I was heading into the weeds. In reality, the problem began even earlier. I specifically was avoiding the > 3.4 v/cell to avoid the issues you mentioned in detail. I have no measurements of bank voltage that make me think the cells were ever exposed to more than 3.4 v/cell. So, I should have avoided that problem, however, the data shows that something happened. I think that 'one' vampire board was defective or mis-calibrated. It was bleeding like you described. If you throw that '1' cell out, recompute the average, and look at the spread of the remaining cells, now the variation is more like ~<25 mV. This is why I really wish I had
              labeled them as I took them off....

              Because the solar charge controllers are 'dumb', another potential source of damage is that I don't have the ability to limit the charge 'into' the batteries. I have a coulumb counting shunt that is monitored by a FlexNET DC (FN-DC) but it's not really designed for lithium and doesn't have the control functions needed. Let's say it's a beautiful sunny day, the net output of the panels could be 240 ADC at 58 VDC. Normally, the 16kw of AC inverter can slurp up that power, and the batteries sit at a constant voltage of 58V. But if the grid goes down at night, the system drops into battery mode and the voltage drops, in the morning the sun comes out, and the batteries are charged at a reasonable rate as the sun rises slowly out of the morning murk until they reach 58V; all is well. But, if the grid goes down at night, we drain the batteries, and it's heavily raining in the morning until around noon, and all of the sudden the sky clears; wack, my batteries are going to be exposed to 240 A of charging.... 2.4C. This is my other risk of an undersized bank, but few catch on that.

              I'll say the one thing I did learn in the last few days is that even though I knew to avoid the shunting effect of the BMS boards, it didn't sink in (and I've been lurking on this board since 2012) that the presence of the board creates enough of a parasite loss that might explain what I've measured. I interpreted the parasite and vampire (mentioned many times in many posts) as one and the same, as the shunting current. Really, they refer to two DIFFERENT losses. Assuming my 'divergence' clears up and the measurements of the Mini-BMS's find one totally out to lunch board, I'd like to think these postings with data & excellent discussion of the last few days will give confidence to other users to do you you've strongly advocated since early this Spring, let your batteries be naked.

              inMichigan
              Last edited by inMichigan; 10-28-2015, 08:30 PM. Reason: spelling

              Comment


              • #52
                There is no end to ridiculous explanations about balance boards on here. All you have to do is read the instructions included with them. They are not designed to balance a pack per se. They are certainly not designed to be used by the bottom balancers to trim their top end. They are designed to remove a few .0X's of a volt from a cell in a balanced pack of cells at the end of a cycle to keep them in balance. If you think they will fix a cell balance that is 3.55v and trim it till a 3.3v cell catches up, it's probably not going to happen.

                Since I am not going to 3.55v mine have not been tested yet. People that use them every cycle say it only takes a few minute's or so if used each cycle to do the whole pack. I was kind of trying to figure how long I could go without needing to balance. Once I see that more than .02 difference I will.

                Comment


                • #53
                  Originally posted by Willy T View Post
                  If you think they will fix a cell balance that is 3.4v and trim it till a 3.1v cell catches up, it's probably not going to happen.
                  Personally, I did not expect any balancing action at 3.4 V/cell. That's one of the reasons I close a lower voltage. All I am trying to clarify for lurking readers is that during the 2 months I pondered if I should finish the connections and wire up a low voltage disconnect circuit, my cells became very unbalanced in a way I had not anticipated. We then began speculating (without data presented) on ways to investigate further which I'm interested in try and reporting the results.

                  In my world, good data is supreme. Previously, I only measured resting cell voltage when the bank was isolated from the system. To do this, I have to shut down all sensitive electronics in the house, throw the master DPDT, shutdown all the equipment.... the family does not appreciate the disruption. I am going to try something that I don't have confidence, maybe you have data to indicate if there's a chance it will work, or useless. About 4 hours after the sun drops into the murky sky, the bank voltage is very stable until the following morning when the sun comes up. All the Outback equipment is fed from the battery. That is, without the battery being present, even the grid AC pass-thru in the inverters won't work. During the night, in grid-tie mode, all the house AC is from the grid passed thru the two inverters. So, the batteries are not resting with zero load attached, however, the load is very small (~100 watts) compared to the normal systems loads. Usually, from 9pm to 8am, I might see the bank drop by 0.1 V on average. Since, I'm interested in the +/- relative to the average at the time of measurement, how do you think this semi-resting state will compare to better conditions? That is if the average cell voltage is 3.40, will that cell that was higher by 5 mV be the same cell that is also higher by 5 mV if I disconnect the bank and let it sit overnight? We measured cell voltages last night around 9pm. I'll do this a few times in the coming weeks. Thoughts?

                  inMichigan

                  Comment


                  • #54
                    Yeah, I to have Outback equipment, Inverter, Mate, hub, FNdc, + some of my own stuff and it's at least a parasitic drain. As you said it's at least .08 higher at rest after 6-12 hours. I really didn't realize it was that high until recently and haven't thought about how it effects all the set perimeters. Since I don't go close to the knees I had decided to ignore it. I can see where you are collecting data it's a factor to consider. I'd foot note the data for sure. Your sheets look like mine, all my own terms and notes.

                    You mentioned board drain, here is what I found for my boards.

                    Less than 4 mA operating current to prevent bleeding cells over long time period when not in
                    daily use. When cell reaches LVC, its cell board goes to sleep mode and quiescent current drops
                    to 0.1 mA, preventing further bleeding of the cell.
                    That 3.4v in my previous post was a bad example, because the boards do not kick in until 3.55v, I fixed it.

                    Comment


                    • #55
                      Originally posted by inMichigan View Post
                      I thought about that as well. I'm not sure if I can measure power consumption at that level. Let's assume the bleeding used enough current to shift the voltage of a single 100AH cell shift by 0.2 volts over 3 months. Trying to get a sense of the scale, let's say 80AH was stored between 3.5 and 3.1 volts, that 80AH / 0.4 volts, so that 0.2 shift was worth 40 AH.
                      40 AH = 48*24 Amin = 48*24*60 Asec = 70,000 Asec
                      3 months is 89 days, or 8M seconds

                      70,000 Asec / 8M sec = 0.010 amps....

                      assuming that's a good assumption, 10 mA... hmm, that doesn't seem too bad. And in hindsight, those pretty blinking red LEDs.... how many mA's used by those? (how much variation from light to light)

                      Thoughts on my rule of thumb sense of scale?
                      Has your Fluke got a mA and an Amp range on it? If so you can use that. I would use the mA range to measure the currant at 3.0 volts and 3.4 volts and the amp range to see where the balancing switches on, would also be good to see what voltage it switches off at. PM me if you would like more details.

                      A 0.2 volt rest voltage difference between 3.35 volts and 3.15 volts equates to around 80% SOC, whereas 0.2 volts between 3.35 and 3.55 and 3.15 and 2.95 volts equates to only around 10%.

                      I am very interested as I am sure you are to find out what caused this imbalance. There maybe a fault in one of your LFP cells or it just might be the battery settling down rather like a Lead Acid battery takes time forming the plates.

                      Simon
                      Off-Grid LFP(LiFePO4) system since April 2013

                      Comment


                      • #56
                        Variability of relative voltages for an LFP battery

                        Originally posted by inMichigan View Post
                        In my world, good data is supreme. Previously, I only measured resting cell voltage when the bank was isolated from the system. To do this, I have to shut down all sensitive electronics in the house, throw the master DPDT, shutdown all the equipment.... the family does not appreciate the disruption. I am going to try something that I don't have confidence, maybe you have data to indicate if there's a chance it will work, or useless. About 4 hours after the sun drops into the murky sky, the bank voltage is very stable until the following morning when the sun comes up. All the Outback equipment is fed from the battery. That is, without the battery being present, even the grid AC pass-thru in the inverters won't work. During the night, in grid-tie mode, all the house AC is from the grid passed thru the two inverters. So, the batteries are not resting with zero load attached, however, the load is very small (~100 watts) compared to the normal systems loads. Usually, from 9pm to 8am, I might see the bank drop by 0.1 V on average. Since, I'm interested in the +/- relative to the average at the time of measurement, how do you think this semi-resting state will compare to better conditions? That is if the average cell voltage is 3.40, will that cell that was higher by 5 mV be the same cell that is also higher by 5 mV if I disconnect the bank and let it sit overnight? We measured cell voltages last night around 9pm. I'll do this a few times in the coming weeks. Thoughts?

                        inMichigan
                        I have found that the individual cells voltages move relative to each other all the time, depending on load, charging, SOC of the battery and maybe even phase of the moon, i don't know. I don't think it is worth trying to keep track of how the cell voltages relate to each other. You can usually pick trends, but there are so many variables with this. I think what is important is to make sure that the cells don't diverge too far from each other, i would start getting alarmed if the divergence was more than 0.1 volts regardless of when it was measured. The other thing is to make sure that none of the cells go outside the range you have picked to get the lifespan you want. In my case for normal operation this is 3.5 to 3.0 volts with an occasional deviation to say around 3.6 volts or 2.8 volts.

                        If you want to measure SOC I don't think you can go past an Ah meter. Due to the very high coulomb (current) efficiency of the LFP batteries you can get SOC readings to within a few % with relative ease. This is not the case trying to deduce SOC from voltage, which I would think is even more inaccurate than trying to do it with Lead Acid batteries.

                        Simon
                        Off-Grid LFP(LiFePO4) system since April 2013

                        Comment


                        • #57
                          Originally posted by inMichigan View Post
                          I 'had' these: http://www.electriccarpartscompany.c...-BMS_c_55.html
                          They're now in the drawer of parts.
                          I and many others know those POS Vampire boards very well and are the source of your problems. Thousands of them in Landfills including the 16 I had at one time. I found a sucker to buy them. Was it you?

                          Originally posted by inMichigan View Post
                          Maybe EV chargers even have this as a possibility, but so far, I haven't noticed a Home RE inverter with communication to a BMS. I never use the charger function of my inverter, as it makes no sense to use grid power to charge my batteries except for testing purposes. My charge come from the PV FlexMAX 80 chargers, and certainly have no advanced communication. They essentially use bank voltage for everthing.
                          Of course EV chargers and most chargers for Lithium have that capability. Solar has nothing for Lithium, thus why BB is the best option for a solar users on a budget. There are no charge controllers made for Lithium or any Balance Boards for Solar use. Just repackaged EV parts. The mini-BMS you have has a NC comm link for the Head-End Board. All it can do is it has a HVS that uses a NC relay for you to buy a large solenoid to operate and to open or close the AC or DC power to the charger. It is what a Farmer would do to Jury Rig and make it work, but poorly. A HVS is completely worthless. It is the equivalent of a doctor where you have a heart attack, go to the hospital, doctor tells you that you will die shortly because you smoked for 40 years. Too damn late to quit smoking, have another one before you die. Another example a plane without a fuel guage. You take off, get to cruising altitude, the fuel idiot light comes on, then your engine stops and you die. .

                          In the specs for your Vampire Boards lies the answer to your problems. You, Willy, and Karrak, cannot spot it, but it sticks out like a sore thumb to anyone who knows a little about electrical fundamentals and Lithium batteries. It explains in great detail what your problem is. Have you found it yet? I will withhold the answer and give you time to find it. I guarantee you Willy and Karrak have no clue what it is and cannot find it.

                          Don't sweat it out because you got rid of the Vampire Boards and all your problems will go away now that you have them removed. But I must warn you if you are going to use BB, do it correctly, and so far you have not. So if you are going to BB, remove your batteries, wire them in parallel, and discharge them to 2.5 volts and let them sit 24 hours. You want your final resting voltage to be no lower than 2.4, and no higher than 2.6 volts. Reconnect them in series and charge them with constant current by setting your
                          charger for 57.6 volts. When the first cell reaches 3.6 volts, terminate the charge. Note which cell it is. Now lower the voltage of the charger to around 54.4 volts, and cycle the batteries normally. What you are doing is finding the right Voltage Set Point where the weakest cell (the one that got to 3.6 volts first and you noted) is roughly 3.4 volts rested slightly after being charged up. All others will be slightly lower voltage. In the end your voltage set point will be around 54 volts. Set your Inverter Low Voltage Disconnect to 48 volts. When done you will be able to sleep at night.
                          MSEE, PE

                          Comment


                          • #58
                            Originally posted by Sunking View Post
                            I and many others know those POS Vampire boards very well and are the source of your problems. Thousands of them in Landfills including the 16 I had at one time. I found a sucker to buy them. Was it you?
                            When I was researching CALB LFP batteries, I watched a Youtube video where a guy had a test setup in his house. He was BB some prismatic LFPs, I think around 200Ah. He was bemoaning having ruined a few cells - I think a swollen one was visible, and blamed balancing boards. In one view, a rats nest of loose wires and BMS boards could be seen on the floor of a closet. He was getting ready to pitch them into the trash. YMMV - just saying what I saw.

                            Comment


                            • #59
                              Originally posted by Willy T View Post
                              Yeah, I to have Outback equipment, Inverter, Mate, hub, FNdc, + some of my own stuff and it's at least a parasitic drain. As you said it's at least .08 higher at rest after 6-12 hours. I really didn't realize it was that high until recently and haven't thought about how it effects all the set perimeters. Since I don't go close to the knees I had decided to ignore it. I can see where you are collecting data it's a factor to consider. I'd foot note the data for sure. Your sheets look like mine, all my own terms and notes.

                              You mentioned board drain, here is what I found for my boards.
                              Less than 4 mA operating current to prevent bleeding cells over long time period when not in
                              daily use. When cell reaches LVC, its cell board goes to sleep mode and quiescent current drops
                              to 0.1 mA, preventing further bleeding of the cell.
                              Thanks, What you quoted is a great bit of info.... essentially on the order of my back of the napkin estimate. And, since I've never had a low voltage period in 89 days, if mine work the same way, mine have 'never gone to sleep'.... therefore, it's quite possible my 100's of mV of deviation are just differences in the idle draw of the BMS board.

                              This last set of voltage measurements were made after the bank was take apart, bolts/straps/BMS put into a bin, batteries put on the ground, and allowed to rest.

                              And, with that jogging of my memory, something I forgot to mention.... Back in May or June, I had everything setup on plywood as a giant bread-board for testing... Anyway, as I was taking it apart, there was one BMS that was not 'blinking'... I set it aside, ordering a replacement and spare. Meanwhile, decided not to proceed with building a low voltage disconnect system. GUESS which battery did NOT have a board on it!

                              Since I needed to see what the bank looked like before diassembly.....I happened to have a high resolution photo...

                              #13... my problem cell....




                              Recalculating the difference from the average (without including #13 which now can be discounted), I get:


                              This spread is just the difference in power consumption of each of the individual BMS boards... mixed in with the normal difference in cell to cell.

                              A range of 70 mV or so, let's blame 20 mV on actual cell performances issues, divided by the 216 consumed by a board (which I know now having found what wasn't missing in #13), about 25% range in power draw by the little boards, seems feasible....

                              Thanks to all the simulating posts that helped me realize the high voltage cell was all my fault, and really quite obvious, even to a 3rd grader.

                              inMichigan

                              Comment


                              • #60
                                Originally posted by Living Large View Post
                                When I was researching CALB LFP batteries, I watched a Youtube video where a guy had a test setup in his house. He was BB some prismatic LFPs, I think around 200Ah. He was bemoaning having ruined a few cells - I think a swollen one was visible, and blamed balancing boards. In one view, a rats nest of loose wires and BMS boards could be seen on the floor of a closet. He was getting ready to pitch them into the trash. YMMV - just saying what I saw.
                                I bought new... I now have 17 -1 + 2 = 18 for sale as well...

                                Well, I have the 17th cell for a reason. If one cell had to die (and I don't think it has), I've learned a valuable lesson and have left enough data in this thread for others to draw there own conclusions.

                                As I mentioned a few posts ago, the voltage measured under a very light load at 9 pm on the 28th.:


                                #13 is now the 5th from the top (sorted the list since originally posted). I would assume (anybody have data) that if I had damaged #13, it would not be charging/discharging in lock-step with the other 16 cells. It's only been a few cycles.... Thoughts?

                                inMichigan

                                Comment

                                Working...
                                X