fhorst1
commented
2 years ago Good morning Jochen from Thailand!
Quite a long story to read.
To summarize, you found that the charge settings of the Victron isn't matching your LiFePO4 cells and balance.
And you are correct. Not negatieve intended, you need to research more.
Balancing should start at 3.5 or earlier. 3.4, 3.45v. 3.65v is the maximal adviced voltage. At this point it's 100% charged. That doesn't mean that as soon as it reaches 3.65v it's fully charged. As during charge, voltage is higher than actual voltage. (Depending on the amperage used) LiFePO4 have amazing wide margins. It can be charged up to 4.2v without damaging, and discharge 2.0v without damaging.
Many are aware of the advised 2.5 as normal low setting and 2.0 as "absolute" low point.
But not aware of the advised 3.65 and the "absolute" top of 4.2.
Going beyond 3.65 during a charge cycle won't damage your cells, unless you push it too far.
Probably many will disagree, those need to research more also :-) Fear based knowledge isn't knowledge. Facts are. A little journey on Google will educate.
For LiFePO4, longest lifecycle is obtained by charging up to 90-95% and discharge down to 10%.
Say use 80% capacity.
Yes, it would be nice if the charger (usually solar) would slow down near the top.
Realistically? They usually don't. LiFePO4 isn't lead acid, and the reduction in absorbing the energy is much steeper curve then lead acid.
LiFePO4 get stuck at 3.35 - 3.45v, depending on the amperage and brand of lifepo4. It seems to hang there during charge "for ever", up to the point the cell has 90% (95%?) absorbtion. Then the last part goes fast. After the "knee", it can be minutes till the 3.65 us reached and passed.
As most chargers don't listen to individual cell voltage, it can't reduce charge capacity based on this.
With a result that the BMS needs to step in, and stop the charge, burn of the top, so the rest can catch up.
A well balanced set hardly ever has this problem. All the cells reach the knee roughly at the same time, and the overall battery voltage is at a that high level that the charger does reduce its amperage.
Even a Daly BMS with just 0.08A balance (burn off) capacity is able to keep up and balance the cells.
If the DIYBMS isn't powerful enough with +/- 1A capacity... You can increase with adding a heatsink. Even add 2 or more cell modules at the "acting up" cells. (Program to discharge/ burn off at 3 55v)
Well... In that scenario you do have balance problem with your cells that DIYBMS won't fix.
Did you top or bottom balance the pack before building it? As the pack / battery should need only like 0.08A (or less) to keep the pack balanced....
Needing over 1A burn off balance shows something else is wrong.
While I'm at it :-) Temperature? To keep it simple about the same as Humans can, without thermal protection 0 to 45-50c (You won't be happy at 0c without thick clothes, nor at 45-50, but won't die. Most happy at 10-35c. Same same for LiFePO4)
Basic knowledge about LiFePO4.
Nothing to do with any brand of charger or versions of BMS.
Not even the Victron or DIYBMS. I have my voltage set (S16) at 54.75 top and 2.7v low. With a balanced pack, no problems at all. I don't need to reach 58.4v, the last 3.5-4v will provide just a few (2-3) % additional charge, and pushing it to it's limits is a dangerous game. Lead acid is happy at 100% charge. LiFePO4 not so. Let is hover around 90-95% and it lasts a lot longer. Lead acid will get damaged below 50% SOC. LiFePO4 doesn't, again.. limits are a dangerous game... Keep it around 10% SOC as max low.
Why? If one or more cells act out, it will go beyond 3.65 and under 2.5, even beyond 4.2v and under 2.0 the damaging voltage.
(4.2 is during charge, not constant 4.2v. after charge stops, the cell will set itself to or under 3.65 in a few hours time, same for the 2.0, it might even dip under, but after the discharge stops, the cell will settle up to around 2.5v)
Balancing at 3.65 is way to late. It should start after the knee. Depending on your battery, this will be on, under or over 3.5v
I hope this helps, as the problem you describe isn't Victron, DIYBMS related but to the build of the battery.
It can be a less optimal connection of a busbar between the cells. Clean, free of oxidation and enough clamping force (4-5 Nm) (assuming you have the standard +100Ah rectangular LiFePO4 cells with or without studs) This happens a lot. Especially if you have one or more cells "acting strange"... The contact isn't optimal. Clean, remove oxidation (sandpaper, sponge) remove contamination (alcohol/acetone) use oxiguard (or similar) and work fast. Aluminium oxidises in 30 seconds to a level that it can provide probelms. 5 minutes after cleaning it will provide problems.
Especially with not laser welded studs in the terminals, you are in for a Wilde ride :-) It's a thin line between clamping down enough and stripping out the 6mm thread in the aluminium terminal. When you use loctite with primer (aluminium needs primer) you safe yourself a lot of problems. (And wait 24 hours before clamping, let the loctite harden)
It can be that the cells didn't bottom or top balance before installing. (Charge or discharge each cell to top / bottom before installation, place them in parallel for a few days to set even the last bit.
At top balance, you can keep a charger at 3.65v for days (in parallel setup) after each individual cell had reached 3.65v You will find that without tickling it to the top after several days the charge isn't equal.
Top/Bottom balance is something that takes several days, maybe even weeks to do it right.
280Ah cell, 1% imbalance would need to burn off 2.8A for an hour.
With 800ma (DIYBMS) that is 3.5 hours...(210 minutes) Where the protection sets in after minutes... (As charge continued, it will reach 3.65 fast)
Just 1% will take several weeks for DIYBMS to burn off, even at start on 3.5v for balancing.
All those weeks the BMS will stop regularly the charge.
Real life... Charge on solar. Discharge at night. Fully charged after how many hours? Probably 3 or 4 hours on a sunny day time left for balancing? After this the electricity consumption is higher then the solar panels will provide. And here battery is starting to be discharged.
Not all days are sunny.
We talk months before the DIYBMS has balanced the set / battery.
Till that time... Stop-start for many many days :-)
That is with 1% imbalance at 280Ah.
I have 4 sets of 260Ah. One Eve, 3 BYD. With active balancer BMS (jikong) because I like the way it works, and 4 X DIY BMS who acts as additional safety and information provider.
I have 80 cell modules, a few spare. 64 in use.
After playing with LiFePO4 for several years, living off grid, being fully depending on it, and have a pile of China (and western) crap of defective or less functional BMS, I know how important BMS is, and how easy they don't work when needed.
I have over 30 dead cells to prove this (bye bye $$$) where chargers doesn't stop at their max 58v, or BMS don't stop at 3.65v or combination.
Victron probably would not have charged beyond 58v. EASUN ... They do.. (don't buy)
I have some experience :-)
On Mon, Mar 28, 2022, 22:50 HerrFrodo1 @.***> wrote:
Hello diyBMSv4 fans and those who want to become one :-)
First of all you need to know my setup: DIYBMS ESP32 release 26.02.2022, 16x Module V4.4 (9,6K) - Roundtrip: 880ms 2x 1m twisted cables to modules on cells. 8-10cm cables module to module. DC/DC power supply 48/5V Meanwell, UTP cable Victron CANBUS to Victron Cerbo GX DC/DC power supply 48/24V, Contactor LEV200A5NAF(broken) Victron Cerbo GX (master unit between Victron Quattro II, diyBMS, smart shunt 500A and more) 70mm2 Batterie cables from Batteries to Quattro II. Smart shunt in minus and 300A fuse and contactor in +. The Cerbo, contactor coil, diyBMS gets their battery power after the shunt and before the contactor. Each one has a circuit breaker for DIN rail. The Victron Quattro II has to be setup by an MK3 USB adapter for general settings. After that you can view and make some settings with the Cerbo GX on a webserver - I use LAN for it. The diyBMS communicates with the Quattro II and controls the charging power. Some diyBMS values are displayed in the remote console and in Victron's VRM portal - I only have a smart shunt from Victron (at the moment)
The Victron Quattro II is identical to the Victron Multiplus II (best-selling) - the Quattro II only has one more AC input (e.g. for a generator) The ESP32 transmits the CVL, CCL and DCL via the CANBUS over the Cerbo GX to the Quattro II. The Quattro charge with these values. I can confirm that. But the Quattro charge according to its own charging curve. As soon as a battery cell enters the bypass threshold area, the diyBMS transmits the balancing data (CVL, CCL, DCL). So far so good. The only problem is that the Quattro then does not reduce the charging power quickly enough. There is no middle area - something like a few mV before... the point of balancing is the last one. At the moment it's not bad because I balance to 3.52V. I had tried it with 3.65V and then it is dangerous for the cells - especially since the balancing current is not sufficient. I connected an active balancer, so it's much better. The BMS always tries to balance exactly on the bypass threshold value, so it must be the last step. If it balance earlier, then all cells go into full balancing mode and heat up - the Quattro then doesn't know the right energy to bring each cell to 3.65V. In addition, the BMS tries to reach the lower value and a lot of energy is lost. That's why I've now set the bypass threshold to 3540mV and the maximum CVL to 56.4. This value is better for the cells and brings more charge-discharge cycles. Not much should be lost in terms of capacity. The Quattro now charges up to 56.4V with full power (what the cells are ready to absorb) and as soon as a cell reaches the bypass threshold of 3540, the Quattro switches to balancing mode via CAN from the diyBMS. Now the performance is reduced and if cells are now well above 3540, that's not a problem because the cells are still far away from 3650. Then the end-of-charge voltage of 56.4V is reached and the Quattro regulates down to 0.1A or even 0A. Now the BMS balances much more relaxed and the resistors don't get hot. I flashed the ESP32 and the V4.4 modules to the latest firmware. In addition, the modules to 9.6K. Roundtrip 880ms, before 3300 - that's great :-) There doesn't seem to be any glitches on 16S. The bus cable from the controller to the first and last module is about 1m long and twisted. the cables between the modules are about 9-10cm long. All in all, the communication from the BMS to the Cerbo/Quattro is ok if, like me, you first balance at 3540mV and set the maximum charging voltage to 56.4V. If you want to get to 3.65V per cell, then we still have to work on how to do it better. In order to be able to balance more passively, you have to change the whole board or plug/connect an additional balancer board to a DIYBMS module. That is certainly a lot of effort and you burn more energy unnecessarily. Either you develop a separate active balancing board or you take one from China. I think if Mr. Pittaway should add another line to the Victron settings and call it "approach voltage". Absolute end-of-charge voltage is 58.4V and 3.65V per cell (LiFePo4) - you have to balance on that - but you have to balance from below 3.65V, if the cells won´t get over 3.65V. So first set a total voltage up to which the Quattro will charge full power and then one from which it will reduce. If the Quattro has then reduced the power and the cells are balanced to eg 3550mV (+/- a few mV), then continue charging with low energy (0.2-0.5A - depending on the capacity) and the cells up to high 3.65V/cell charge further. LowTopUp :-) Is that possible?
I enabled DVCC and took a screenshot of my settings. If you activate LiFePo4 in the Victron settings, a fixed charging curve is automatically set. Maybe I can change it - I have to study it. In addition, I entered the maximum charging voltage of 56.4V in the DVCC. I had experimented with the charging current, but it didn't work properly. The data from the DIYBMS seems to have priority. Nevertheless, he uses his charging curve - I have to do more research on that.
Whats your opinion?
Greetings from bavaria, Jochen
— Reply to this email directly, view it on GitHub https://github.com/stuartpittaway/diyBMSv4/issues/113, or unsubscribe https://github.com/notifications/unsubscribe-auth/ADVGBDCUAJLF34GFSPP5HOLVCHINTANCNFSM5R3S466A . You are receiving this because you are subscribed to this thread.Message ID: @.***>
HerrFrodo1
Hello diyBMSv4 fans and those who want to become one :-)
First of all you need to know my setup: DIYBMS ESP32 release 26.02.2022, 16x Module V4.4 (9,6K) - Roundtrip: 880ms 2x 1m twisted cables to modules on cells. 8-10cm cables module to module. DC/DC power supply 48/5V Meanwell, UTP cable Victron CANBUS to Victron Cerbo GX DC/DC power supply 48/24V, Contactor LEV200A5NAF(broken) Victron Cerbo GX (master unit between Victron Quattro II, diyBMS, smart shunt 500A and more) 70mm2 Batterie cables from Batteries to Quattro II. Smart shunt in minus and 300A fuse and contactor in +. The Cerbo, contactor coil, diyBMS gets their battery power after the shunt and before the contactor. Each one has a circuit breaker for DIN rail. The Victron Quattro II has to be setup by an MK3 USB adapter for general settings. After that you can view and make some settings with the Cerbo GX on a webserver - I use LAN for it. The diyBMS communicates with the Quattro II and controls the charging power. Some diyBMS values are displayed in the remote console and in Victron's VRM portal - I only have a smart shunt from Victron (at the moment)
The Victron Quattro II is identical to the Victron Multiplus II (best-selling) - the Quattro II only has one more AC input (e.g. for a generator) The ESP32 transmits the CVL, CCL and DCL via the CANBUS over the Cerbo GX to the Quattro II. The Quattro charge with these values. I can confirm that. But the Quattro charge according to its own charging curve. As soon as a battery cell enters the bypass threshold area, the diyBMS transmits the balancing data (CVL, CCL, DCL). So far so good. The only problem is that the Quattro then does not reduce the charging power quickly enough. There is no middle area - something like a few mV before... the point of balancing is the last one. At the moment it's not bad because I balance to 3.52V. I had tried it with 3.65V and then it is dangerous for the cells - especially since the balancing current is not sufficient. I connected an active balancer, so it's much better. The BMS always tries to balance exactly on the bypass threshold value, so it must be the last step. If it balance earlier, then all cells go into full balancing mode and heat up - the Quattro then doesn't know the right energy to bring each cell to 3.65V. In addition, the BMS tries to reach the lower value and a lot of energy is lost. That's why I've now set the bypass threshold to 3540mV and the maximum CVL to 56.4. This value is better for the cells and brings more charge-discharge cycles. Not much should be lost in terms of capacity. The Quattro now charges up to 56.4V with full power (what the cells are ready to absorb) and as soon as a cell reaches the bypass threshold of 3540, the Quattro switches to balancing mode via CAN from the diyBMS. Now the performance is reduced and if cells are now well above 3540, that's not a problem because the cells are still far away from 3650. Then the end-of-charge voltage of 56.4V is reached and the Quattro regulates down to 0.1A or even 0A. Now the BMS balances much more relaxed and the resistors don't get hot. I flashed the ESP32 and the V4.4 modules to the latest firmware. In addition, the modules to 9.6K. Roundtrip 880ms, before 3300 - that's great :-) There doesn't seem to be any glitches on 16S. The bus cable from the controller to the first and last module is about 1m long and twisted. the cables between the modules are about 9-10cm long. All in all, the communication from the BMS to the Cerbo/Quattro is ok if, like me, you first balance at 3540mV and set the maximum charging voltage to 56.4V. If you want to get to 3.65V per cell, then we still have to work on how to do it better. In order to be able to balance more passively, you have to change the whole board or plug/connect an additional balancer board to a DIYBMS module. That is certainly a lot of effort and you burn more energy unnecessarily. Either you develop a separate active balancing board or you take one from China. I think if Mr. Pittaway should add another line to the Victron settings and call it "approach voltage". Absolute end-of-charge voltage is 58.4V and 3.65V per cell (LiFePo4) - you have to balance on that - but you have to balance from below 3.65V, if the cells won´t get over 3.65V. So first set a total voltage up to which the Quattro will charge full power and then one from which it will reduce. If the Quattro has then reduced the power and the cells are balanced to eg 3550mV (+/- a few mV), then continue charging with low energy (0.2-0.5A - depending on the capacity) and the cells up to high 3.65V/cell charge further. LowTopUp :-) Is that possible?
I enabled DVCC and took a screenshot of my settings. If you activate LiFePo4 in the Victron settings, a fixed charging curve is automatically set. Maybe I can change it - I have to study it. In addition, I entered the maximum charging voltage of 56.4V in the DVCC. I had experimented with the charging current, but it didn't work properly. The data from the DIYBMS seems to have priority. Nevertheless, he uses his charging curve - I have to do more research on that.
Whats your opinion?
Greetings from bavaria, Jochen