I am using a Ji Kong aka Heltec BMS with the 2 amp active balancing and it has been working great. Only 4 months in, so I can't so how long it will last, but so far no issues at all.
The way the Heltec balancer works is that it pulls current from the highest voltage cell. Only one cell at a time. But if there are a few that are high, it will switch between them and pull from a few before it switches to push mode where it pushes the current into the lowest voltage cell. And if there are 2 or 3 that are low but close to each other, it will switch between those, but for the most part, I see it pushing into 1 cell on each cycle. It is very effective and accurate, but it is not super fast. Even at 2 amps, it is only doing one cell at a time. If you have 5 cells that are all high, it is only pulling 2 amps from each high cell for only 1/10 of the time. That makes the net balance current just 200 milliamps out of the 5 cells. That is close to the newer high power Daly passive balance units. But the flip side is, this is able to take that power and push it to low cells. The passive balancers just turn it into heat. In theory, the active balancing is more efficient, but not much faster. If you only have 1 or 2 high and low cells, then it is pretty quick, but I don't think speed is really that important as long as it keeps the cells safe.
I totally agree that a high power balancer is no replacement for good cells that are well matched. For most solar storage systems, a good top balance is going to help a lot. When we have sun to charge, we want to get all of the cells full at the same time. Even if the cells do not match, a good top balance will go a long way to making the system behave.
Lets say you have an 8S LFP pack. One of the cells is only 80 AH while the rest are a real 100 AH. If you top balance, all of the cells are at the top knee at full capacity. As the inverter pulls them down, the one low capacity cell will drop faster and reach a lower state of charge. But as long as the weak cell is still in a safe range, this is not too big of a problem. You basically have to treat the entire pack as just 80 AH because of the one weak cell. This is true no matter how good your balancer might be. A 5 amp active balancer might gain you a tiny bit, but don't bet on it. When the sun comes out the next day and starts charging, the weak cell started lower, but since it is a lower capacity, charging at the same current, it will come up faster as well. And in the end, all of the cells will get up to the full charge knee at about the same time. And efficiency difference is likely to be pretty small, and that is what the balancer should deal with. If a cell comes up too fast, the balancer can bleed off some charge current to let the others catch up. As long as the absorption charge voltage is a little below the protection voltage setting, the charge current will fall off and let the balancer do it's job and re top balance the cells each time they come up to full charge. An active balancer will use the bled off energy to also help pull up the lagging cells, but we are talking a small difference.
The better the cells match, the less work a balancer will need to do. The BMS's more important job is monitoring the cells and making sure they are all staying in the safe range. For 6 years, it may not need to do any balancing, but it can if needed, but when one cell starts to fail, the BMS will see it and shut things down before it can cause a big problem. A cell that becomes weak will climb in voltage much faster than the other cells. It will become over charged and could fail in a bad way. A proper BMS will shut it off and give you a chance to fix the problem before a cell bursts. Even though LFP may not catch fire, they can still bulge and burst and spew liquid and smoke. And if you use Li NMC cells like I am, yes, they could turn into a torch aimed at the next cell over. With a good BMS and a cabinet protecting the cells from shorts and damage, I am not concerned about them catching fire, but without a BMS, they can easily become a time bomb. With the BMS, people are parking cars full of them in their garage. My pack is about 30% of what is in a Chevy Bolt. The car only uses a passive balancing BMS.
Since I am only running my cells from 50% to 90% state of charge, I never actually did a full top balance. I had planned on it, but when I did the first charge up to 90%, the active balancing pulled all of the cells to within 3 millivolts. All of the cells track even through the "S" curve that happens at about 3.76 volts per cell. So they balanced up great and the active balancing basically has been sitting idle for 2 months. I never expected these surplus cells to be this well matched, but I am glad they are.
When I build another pack (I do plan on adding more capacity), I will just let it charge up at a reduced current with the active balancing turned on and let it get the cells all charged up and balanced at the top. I will then manually check them at the cell terminals and ensure they did balance at the top. If any cells are still out, I will set the current even lower to give the balancer more time to pull them in. I would only resort to separate cell or parallel manual top balancing if the BMS has to shut it down for going too far out of balance with a cell hitting high voltage cut off. If I was starting with more "questionable" cells I might be a bit more cautious, but when they arrive all between 40% and 60% charge, and they track together as they charge, I am not worried about them. My initial charge was at just 10 amps with a 2 amp balancer, so it could handle a 10% error in capacity and make up for it. And if it was worse than that, the BMS could still shut it down. The best part about the Heltec (JK) balancer/BMS is that it can keep pulling the high cells down and pushing the low cells up even when it turns off charging for one cell getting too high. With no charge current going in, it very quickly starts bringing the cell voltages together.
Yes, LFP cells have a VERY flat voltage curve and measuring the voltage is not a good indicator of their state of charge, but when all of the cells are the same type and at the same temperature, having them at the same terminal voltage will still put them very close to the same state of charge. The best state of charge balance will still occur at the top knee, but having the active balancer pull the voltages together even in the middle of the flat discharge curve is not hurting the pack at all. And if the curve is flat enough, you can easily set the balance threshold so that it won't balance as long as they stay close. I have mine set to 0.006 volts, but you could open it up to 0.02 and it probably would not do anything until you charge up near full and a few cells start hitting the knee before the others. It will pull current to slow those high cells and help pull up the ones that are lagging. This is never a bad thing.