Epever 6415AN doesn't respect my settings

more normal function today. The system was below the undervoltage but the warning didn't show this am. Spent the day in boost but only made it to 75%SOC. Not surprising since I upped the constant load yesterday to more than my panels can make. I plan to install 4-445w panels tomorrow morning so thought i'd take the rare opportunity to overload my current production, bring the battery down some, then use the new panels to resolve.

jnrhome said:

more normal function today. The system was below the undervoltage but the warning didn't show this am. Spent the day in boost but only made it to 75%SOC. Not surprising since I upped the constant load yesterday to more than my panels can make. I plan to install 4-445w panels tomorrow morning so thought i'd take the rare opportunity to overload my current production, bring the battery down some, then use the new panels to resolve.

Click to expand...

Did you get another controller? I had mine go to under voltage warning a couple days ago. I had my relay disconnected from the inverter so it went very low before the inverter turned itself off. The transfer switch move the load to building power at 24v but there is still a draw from the inverter at idle...and the controllers and the RPi. My software detects when the system switches to building power and turns off the inverter so it doesn't continue to draw down the battery. I've since fixed that. There isn't a lot of charge between 24 and 22 volts in LiFePO4 batteries.

no. Still deciding if this is more a matter of my dumb settings than a controller problem. Being in the Virgin Islands, receiving a replacement could take weeks. got down to 49.02v this am. Controller did report undervoltage this time. I'm thinking it might take this controller a day to propagate any config changes. While annoying, I can live with that if it's consistent. I agree with you about lithium in that way over and way under voltage is much easier than lead because of the near vertical ends of the curve. I'm narrowing my range on top and bottom to allow for more error. I never needed a low voltage cutoff with lead because it was a steady drop. These lithiums run away fast over 3.5v and fall off a cliff below 3v.
Added 15' of rails and 70' of 1' conduit to accommodate my 4 gauge upgrade this am, but so hot I had to stop. Will add 4 445w canadian solar mono panels and wire in tomorrow. Will also keep a small load since there's no chance I will catch up today.

So it would appear that configuration changes need an overnight to take hold. I've made incremental changes in my settings day after day. They don't fire on the day I make them, but seem to kick in the following day. Odd, but workable when you know. Spent the weekend upgrading my solar panels from 1,080w to 2,860w. Not a moment too soon as my fancy new lithium bank was low when i started. After being disconnected for a whole day, they made it down to 49.02v. Finally patched in the bigger array on larger gauge wire early this morning. Before, my peak was around 1,150w/hr. Now, I've been up to 2,500w on a fairly overcast day. I'll be adding to my load now and see just how much I can use and replenish.

Sounds like you are getting there. I think you are right. I did a test today that gives some insight into how these things think. Last night I changed the parameters on my controllers to make 27v the target. I adjusted everything else accordingly. This morning both arrays proceeded to shoot for the target and by 8am they had reached 27v. They held that for about a half hour then backed off slightly. The amps dropped a little and the voltage dropped to 26.9v. By about 9:50am they shut off..zero amps produced!

The batteries proceeded to drain down to 26.2v which is where I had the boost reconnect voltage set. After that the first array started to produce weakly and the second remained off for another hour before it started to produce. After lunch I went down there and changed the boost reconnect from 26.2v to 26.5v. Both arrays came back stronger. The voltage went up to 26.5v. It reversed the draining of the battery but each was only producing ~250w. By 3:30pm the batteries were teetering on the zero point.

I decided to go down there. I turned off then on controller #2. When it came back, it did so strongly again with about 500w. The day is pretty hazy but it is varying between 400w and 500w and the voltage is now up to 26.8v. Controller #1 is back to idle. I didn't reboot it. I wanted to see what it would do. It's producing 8w with 1200w of panels.

So I'm still thinking about what this tells me. I think these things only execute one charge cycle a day regardless of the conditions. I don't like this behavior of it turning off with a load on it. I'm going to leave things as-is until tomorrow. Could the settings change help prevent what I saw today? Don't know. I'm unsure of the importance of the boost reconnect setting. It hit it this afternoon but stayed throttled. Doesn't make sense to allow the batteries to drain while there is sun to compensate for the loss.

Behavior today was different for an unknown reason. Started in boost until around 10am, then switched to float, at the voltage i set two days ago. It floated for an hour or two, then went back to boost for the rest of the day. I made another set of configuration changes today because the stronger panel array has a larger voltage gap between what the charge controller says and what the balancer says. Raised the max, boost and float voltage and boost duration to compensate for the wider disparity. I think i should give it two full days of cycle before the latest changes in configuration should be honored and I can finally understand how this damn thing functions.

Where is your boost duration? If it only held your 27v for thirty minutes before floating, you might up the boost duration and see if that works better

Boost is set to 120 minutes. My float is much lower 26.4v until I adjusted things upwards at noon (26.6v now). Perhaps if I raised it? Not sure. According to the Tracer rules list it can be as high as the boost voltage.

no respect for current configuration today. Going back, the apparent configuration being used is from march 5...
hmm...


I found both of these videos enlightening regarding boost, float settings
I've adopted the guidance in my settings and am waiting to see if they perform well for my set up. Now that i have the larger array, i should be able to see if they are actually causing the performance i'm looking for.

How are you making the changes? I assume you are using the software, right? You hit update, it says save successful, right? Have you tried to reload them to see if they took? Have you looked on the MT50 to see if they are reflected there? I'm also concerned that I'm doing it right as well as I've had my own weirdness.

hehe...
Changes made through the software. Hit update, got changes successful message. I know that's true because i occasionally have set things wrong and got a update failed and why message. I have re-read them after i got a update successful message and the new numbers are reflected accurately. I even confirmed them directly on the controller. Since i use the USB/software connection, the MT50 is a hassle to connect so i've mostly abandoned that. I did make a couple adjustments yesterday and will not change anything now until I see them implemented. Even if it takes two or three days, or two or three charge cycles to adopt, I can live with that once I know for sure the implementation schedule.
Crazy that this has to be decoded...

another day...
Hit float by 10a and still using the 5 day old configuration. By 1p, i couldn't stand it any longer. All this great sun being throttled by float...
Shut down the panels, then shut down the battery bank. Waited a moment, then fired it all back up. Now... not sure if a new boost cycle or new settings have taken hold but i've been above the old boost voltage since the reboot and am still in boost after the 120 boost duration. Battery is being measured well above the old setting, but not yet at the new setting, so can't yet tell if its a new boost cycle or if it's honoring the new settings. Another day should tell the tale. I now suspect that you have to shut down panels and battery to load updated configuration settings.

I'm learning something about these controllers. The settings are not well suited to systems with a continuous load on them. They work fine if you use your batteries at night and recharge them during the day.

Here's how I'm seeing it now, let me know what you think. First the boost only lasts so long. Mine hit full about 9:30am similar to yours. It will then keep it up for the boost time which I've left at 120 minutes. After that it goes into freefall until it hits the float voltage. As you say, it is painful watching all that good sunshine go to waste! Mine then stays that way until sunset. One controller might back off to near zero if the other is handling the back and forth of supporting the load and maintaining the float.

If you would happen to have enough load come on that the voltage would drop to the boost reconnect, then that will kick in. Since my load is constant, not very likely unless it gets cloudy in which case it won't do me any good anyway.

So for this scenario which I think you have as well, float is very very important. You need to set float high enough that you maintain enough battery charge to support your loads when there isn't any sun (night, cloudy day, whatever).

I'm trying to cycle my batteries about 80% for longevity so I'm not going all the way to the top. I'm monitoring the battery with a shunt so I can see what is going on. The controller must just kiss my boost voltage before it stops. I never get a record AT boost voltage but usually 0.1v less. So you have to assume the boost will end before it hits your target and adjust slightly upwards. I did mine about 0.2 volts. My boost voltage is 27.2v. I have the float set to 26.7v and boost reconnect at 26.4v.

As soon as the solar energy starts to weaken in the afternoon and can't keep up with the load, the voltage drops quickly from 26.7v to 26.5v. Within an hour I'm down to 26.3v although it does jump to 26.4v when the plant lights kick off at 7pm. I have a fig and a hibiscus I keep inside for the winter. I think the float voltage is very important to maintain the batteries while there is load on the system. I may raise mine more. I would check what you have set. I wouldn't go above what people recommend as a maximum float but I wouldn't be afraid to go that high since you do have a load on your system too.

I completely agree. The vast majority of configuration advice is centered around intermittent and surge load, With constant load, wider boost, float, reconnect settings are bad. Also, the bigger the battery, the longer it needs to sit at voltage (CV), for that voltage to soak in (CC).

I would suggest using much narrower settings than what is normally recommended for exactly the reasons you outline. Instead of boost being your target voltage, what if you set boost slightly above target, float to the target and set boost recon barely below that? I see the same concern with float tapering in the afternoon and boost recon not bringing them up to where they should be in time. I too wish to run my battery below max(~54v), and well above min (>51v). I would like the battery voltage to be as close to 54v at sundown as possible, which dictates that they are approx 55v when afternoon charge begins tapering.

I've begun thinking of boost voltage as max voltage, float voltage as target voltage and boost recon as below target voltage. The closer these three settings are, the less sun you waste and the closer to target voltage you end up at when charge stops. This also helps in cloudy points while in float. If the clouds take more than a short time, boost kicks back in.

I may get the settings honored today, after power cycling yesterday. If that happens, I now have the (convoluted) method for implementing configuration changes. IF that's true, I can finally begin to tune all this in. If my theory is correct, I'll end up raising all three voltages until I float at charge target, and end up at overnight target. With more panels and bigger battery, the charge voltage to battery voltage delta is much more than my old system, so I'm having to increase voltages to compensate. Starting today at 75% SOC so should top off around noon and finally get a solid reading based on current config. If that happens, I can begin tuning for optimal.

Hope all this is illuminating for you as well. I think we may be on the right path to getting both our systems optimized.

mid-morning update:
Hit 54.8v (boost voltage)around 10:45a. If old settings are applied, it will be 60 minutes to float. If current settings are applied, 120 minutes to float. At heavy charge, there is a 1.3v difference between controller voltage and balancer voltage so that has to be taken into consideration when determining proper set voltages in the controller. Outback has a calibrate feature that you can use to accommodate this offset. Too bad EpEver doesn't. But, it's a simple adjustment once you know.

past 60 minutes and still in boost...

Sounds like we are on the same page. I like your idea for setting boost and float. After a few sunny days, today is cloudy with rain forecast this afternoon. The transfer switch moved to building power at about 8:45am after the plant lights kicked in. The battery dropped to 24.4v. I know the transfer switch reads a lower voltage than measured. I will have to adjust it. I have it set for 24.2v. The batteries likely will not recover today.

Calculating my average daily usage I'm at close to 80% of the battery capacity. I wanted to have a limited start with LiFePO4s until I was comfortable with them. After tax season I will likely buy another 8S to add in parallel. I'm going to give my current settings a few days to see what happens with the cloudy weather then adjust.

120 boost duration has come and gone but still in boost. Sun was blasting until about 12:15 and is now partly cloudy. Still mostly above boost voltage tho. I'll be bumping up on the charge limit voltage soon so will see if that stops it. If so, i feel safe with the top end of charging and can relax. If not...WTF again...

Your situation allows detailing the low end discharge stuff. Both need to be attended to. Funny that we're in opposite ends of the charge range, but only a thousand miles apart and really close to the same longitude. Entirely unrelated weather patterns tho

180 minutes... still no float
broke 10KW for the first time tho

Mine switched back to the inverter. I checked at lunch the transfer switch is set to 27 and 24 respectively. I'll have to adjust for it not reading the voltage right. I disabled my relay to control the transfer switch when I started to have issues. I'll have to go back. I don't have the problem like I did with lead acid batteries where the charge voltage is higher than the full voltage, but I still want to fill the batteries as much as possible during the day. We had a brief lightening to moderate overcast and the panels started to produce (about 14a each). It's getting darker again so the switch back was counter productive. I still need to work on bulletproofing my software to properly deal with "charge events".

You hit 10K, cool! Best day I had was about 9K mid-summer. I might do better this year once I get my third array up. I'm gearing everything for winter sun. I even remounted one array to 60 degrees. The other is at about 50. The third one I will build a similar frame at 60 degrees. The 60-60-60- triangle is also very stable for dealing with wind. No concrete for me at the moment. I'm a tinkerer. I have array #2 pointing more west but its voltage is lower than array #1, so they cut out about the same time even though the sun is behind array #1 in the late afternoon. I may reverse the two in terms of facing.

I'm looking into pulling data directly from the charge controllers into my Raspberry Pi via an RS485 interface. As mentioned before the charge controllers have a very optimistic view of what they see in terms of voltage and current. I'd like to compare real time what my other sensors see and what the charge controllers think they see. Even the battery voltage isn't the same. It's the never ending project.

damn... over charge limit and still banging in juice at boost. I've never had the charge limit above 56 and yet, there it is, still with no float.
Up to 12.5KW for the day so far. My panels are at 10 degrees facing west. They should be at 17 degrees facing south but our sun comes so directly overhead that even facing west at 10 degrees, i occasionally exceed the rated max output of the array.

F@#$%ing charge controller... right past charge limit and pounding in juice at boost. Now i have two cells overcharged and have shut down the solar and disconnected the battery until the balancer fixes it. Only takes a moment or two for the balancer to knock down the high ones, but damn..

When lithium goes sideways, it happens fast. At least I was watching it this time and could intervene. Got the balancer running and have reconnected everything. Re-saved settings and dropped boost duration to 10 minutes. I'm going to try to use float as CC instead of boost duration for a little more flexibility.

Having to leave the balancer run, and alternate panels on and off. Charge limit voltage is ignored. I've even changed it slightly and updated, to no avail.
I have traversed the value multiple times with no response from the controller. Every time the sun comes back out (cloudier now) the amperage shoots up and i have to run for the disconnect. grr...

whew... what an afternoon. I must have cycled the panels 20 times to keep things from running away. I did keep it in check and managed to partially top balance along the way. Not sure why, but the system finally dropped into float and is still in float now. Perhaps it will actually work properly tomorrow.

That is strange. Mine seems to overreach as soon as it there is a hint at hitting the target voltage. How out of whack do your cells get? I have a little cell monitor that I connect to check but don't leave it attached. Mine get about 30mV out at most and come down to 8 to 10mV when charged. Are you using an active balancer? The nice thing about those is they can move charge around faster than the discharge passive type. I have the wires on mine for one but haven't had to use it. I did try it a bit when I was initially testing and balancing my cells and it is pretty neat.

I'll be very interested in performance tomorrow. Maybe it's ok now? I sure hope so. The controller stayed in float until the panels cut out, never returning to boost like usual. I top balanced to 2mv delta prior to putting in service. Was a bit OCD about it as many sources online stress near perfect top balance as a substitute for a BMS. I don't trust things to just "work" (clearly, HA) so I have a 2A active balancer connected; drilled and soldered the leads directly to the busbars. I have left it turned off and am using it as a cell gauge to monitor cell voltages only. When at the bottom end of the charge, the pack went out of balance by about 50mv, but I let that go, noted the lowest <5mv delta voltage, then called that "empty", 48.2v. Throughout the charge cycles since, the delta varies between 0-6mv, but couldn't correlate any specific voltage to a higher delta, so have assumed its normal variability. My cell wire resistance is between .106 and .109 ohms so a small variability in delta is to be expected. It actually ran much closer to dead even then I expected so have been quite pleased, with no need for activating the balancer. In fact, just prior to exceeding 3.5 on the highest cell, I had a delta of 8mv which was the highest i had seen since the initial balance (above 48.2v). Once the cells got above 3.5v, the delta started growing, getting increasingly worse until the first cell went over, then it started going way out of balance (>200mv delta). Shutting down the panels and using the active balancer brought the high cells down very quickly, just minutes, with the delta decreasing more slowly along the way (as expected). by 3.5v, the delta was around 50mv, so I started up the panels again, let them run until the highest cell was ~3.6v with the balancer running, then shut them back down and let the balancer continue. Below 3.5 and fire them back up... lather, rinse, repeat about 20 times. Charge stopped about an hour and a half ago and I am at 53.20v with a 2mv delta on the balancer. Controller says 53.11v. I may see a better top end performance tomorrow, or that might have been a waste of effort. Guess we'll see. I did have the best solar day ever with 13.19KW even with all the on and off from 3-6pm so that's something anyway. Pic of my battery attached. Old sealed lead acid bank below.

right off the bat, as soon as the panels began producing, the controller started in float. I shut down the panels and battery, then brought it back up. Now it's in boost and below all trigger voltages. I dropped the boost duration to 10 so the timer won't affect stage transition. Changes in charge stage should be affected only by voltage, I hope...

another white knuckle day so far...
Fought with it all morning as the settings i wanted were simply ignored. No amount of cycling, array or battery or both affected it at all. I finally gave up and reset the configuration to libattery and lifepo4 and hung on for dear life as the voltage shot up even higher. When 5 cells were between 3.69 and 3.71, the wattage started throttling back and dropped into float. Now its seems to be honoring the DEFAULT lithium settings, though they are so high that they overcharge almost half the cells before settling back. At least it's reacting to the config values, though the voltages are far higher than I want. Balancer is just running now, but float voltage is manageable and the cells are within normal voltage. I guess I'm going to have to accept that EpEver knows best and tough titties if I want to be more conservative. I'm debating letting it run in the default mode for a few days or attempting single voltage adjustments.