How does the battery controller figure out when to stop bulk and proceed with float?

[diymarek]

I am running a tiny setup with 6x 12V 200Ah gel batteries (huawei) - 3 parallel banks of double batteries. I am just starting (second month) and trying to figure out how to configure my solar controller to properly power these batteries and avoid sub-50% discharge. And I fail miserably day after day...

Running chinese inverter 3500W branded as Raggie (this one). Using collective Internet knowledge I managed to customize pipsolar setup to retrieve all tech data live, bypassing the "wifi dongle" they provided (based on desmonitor.com solution).
I have a 0.5V drop between the battery terminal and the voltage measured at the inverter, so I assumed I have to take this into consideration when setting parameters to my inverter. The bulk charge as per battery spec is set to 29.2V (I configured it to 29.7), float 27.4 (27.9). From what I noticed, the device never enters the float mode (it did only once - below is the result). It reaches 29V which corresponds with the peak sun time between 11 to 4PM, the amps going down, load goes down and at the end of the day when I sum up Ah delivered, it often reaches 550-580Ah, which makes no sense to me at all (it would suggest a totally drained battery to begin with and I simply don't have enough solar capacity to charge it, but the math does not add up).

And then one day - out of nowhere - I see the charging during the day never reaches 29V, and is stuck flat at 27.4V (float mode indicated as ON by the data reader) and battery is being barely charged, solar panel output is within 60-70% of the max output during that specific time, there is no heavy load. Pure float...

I am doing something wrong here. Can you suggest what I should do next to standardize my daily charge cycles?



^ this is the example of the battery voltage and charging current reading on the day when float engaged (left, charging current below 70Amps peak, significantly lower than during bulk) vs the day when the battery was charged using maxed current all day long, reaching over 550Ah by the time, never existing bulk phase (75-90Amps, depending on the load of other devices in the system).

Living in Dubai - reaching almost ideal sun conditions except during the days when the air is dense with sand, but this was not one of these days...

Thoughts?

 

[yodamota]

it looks like to me that the batteries on aug 28th did get discharged very far. So the charge controller had less to charge the next day. It looks like on the 29th the bateries got down below 24v, causing the charge on aug 30 to charge up to absorption charge.

DO you have an absorption end amp setting?

How much solar do you have? how panel panels? what is the wattage of the panels?

 

[diymarek]

thanks for your comment. I do not have the absorption end amp setting configurable in my controller, I can play around with voltages only and define float / bulk charge voltage. Last night the voltage did not drop below 24V, and today's charging was also very similar - just peaking at 29.9, not switching into float mode despite the charge current dropping below
The installed setup is 4400W nominal, peaking consistently at 3200W.

Just when typing this message I realized that today my inverter actually got into float mode by the end of the day (4:18PM - yellow mark on top of the chart) and continued it until the sunset when discharge started, which means that this device actually had some long (1 month?) learning curve and is finally beginning to work like it should.

 

[yodamota]

Well do you have a setting for how long the batteries should be in absorption? it looks like your batteries absorped for like 5 hours on aug 29th.

My rough math tells me
0.38 x 600ah / 90 charging current = 2.53 hours for your absorption cycle.

 

[diymarek]

Well do you have a setting for how long the batteries should be in absorption?

This device does not have such setting - but from what I see your math rougly corresponds with what's happening - the amount of amps flowing into the battery drops significantly after 2-3hours of max charge - despite the capacity on the solar side - the load is reduced. My only worry was, that it never went back into bulk mode (it kept on charging with 29.2V). But it appears that since last couple of days it is now working (almost) like it should. Thank you very much for your comments. I'll keep observing how it behaves...

 

[yodamota]

Your welcome.

I like the graphs you are showing. Ill admit our system usually is in float by 12:30pm And We usually have 3-5 hours of float. So I do wonder based on the graph why it takes so long to get to absorption? 4400w is plenty to charge those batteries up to float.

How are your panels oriented? How are they set up?(series? parallel?)

 

[diymarek]

but that's exactly what I am observing here... it appears that the "float" should kick in around noon (this is where the charge counter achieves 350-400Ah), and I usually don't have enough load to maintain the beautiful panel output curve because of that - I use additional loads (AC in common room) to maintain it, but otherwise the system is bit overscaled. The only problem is that this controller does not drop the battery voltage nor reports back the FLOAT mode.

That's my power output curve. I can see from the amp output that mid-day is when the bulk charge occurs. Then demand drops once batteries are full (despite peak power I see the amp dropping, daily counter already passing 400Ah), so I switch one of ACs from grid to solar, in order to maintain the load, but as you see it's still not enough to meet the panels output capacity:



Panels are aligned like this. Currently almost flat, at less than 10 degrees (summer in the Middle East), but with possibility to angle down as much as needed in the winter (around 30 degrees, or to 90 for servicing). Azimuth-aligned east-west, because I'm OCD...



And with perfect sunny conditions, no clouds and no rain - I get 360 days of sunshine here. Perfect conditions to study temperature dependencies on these panels (summer is brutal, but winter is great to achieve the STC or even NOCT conditions)

 

[yodamota]

so the more I look at the specs on your solar inverter the more I am confused. Its a 24v inverter and a 24v charge controller. its capable of 80 amps at 24volts. 24v x 80amp = 1920 watts. Where are you getting this 3000 watts of solar charging power from? Is that not the 230v ac load?

3000watts / 24v = 125 amps which the solar inverter is not capable of based on the specs

Unless that website is really mismanaged. I am seeing that you linked to a 3500w solar inverter. and it shows the specs for a 5000w 48v solar inverter!!

 

[yodamota]

so I found a more accurate specification list i hope here https://www.raggiepower.com/cpzx/info_27_itemid_576.html

Its says max charging current of 100amps.

so max the charge controller can output is 100 amps based on those specs. 24vx100a = 2400watts.


I still think that graph you showed with the PV charging power is maybe the 230VAC output of the inverter.

 

[diymarek]

Its says max charging current of 100amps.

yes, it's configured to 100 amps, and it's set to 100 at the moment. But max I've ever seen from it was around 95 (it's the self-reported value, as I do not have the measurement shunt). The device does not have any advanced settings - you can choose the max charging current, bulk voltage and the float voltage, but nothing that would allow it to calculate the battery size. That's why I'm guessing it's learning these values with x number of cycles and only after that it will be effective in switching from bulk to float. I'll report back in the afternoon to see if anything changes today...

I still think that graph you showed with the PV charging power is maybe the 230VAC output of the inverter.

this is the total power output from the panel array (again - self reported)
I do measure output load on several occasions independently (self reported, external meter, individually at devices' level) and it aligns pretty accurately with what the device measures.

 

[Substrate]

Question - if those are true gel's, than no more than 2.35v / cell. (14.1v for a nominal 12v gel) In that hot clime, I'd drop it to 2.33v / cell. Takes longer to charge though.

Be careful here - some manufacturers label agm's as "gel", and visca-versca. See your battery spec sheet for what the CV limit is. (AGM will be 14.4 to 14.7 - temp compensated). So-called hybrids - anyone's guess between engineering and marketing...

 

[diymarek]

Question - if those are true gel's, than no more than 2.35v / cell. (14.1v for a nominal 12v gel) In that hot clime, I'd drop it to 2.33v / cell. Takes longer to charge though.

Be careful here - some manufacturers label agm's as "gel", and visca-versca. See your battery spec sheet for what the CV limit is. (AGM will be 14.4 to 14.7 - temp compensated). So-called hybrids - anyone's guess between engineering and marketing...

it is gel, not AGM. The battery label says "deep cycle gel", but besides that I can't find out any spec sheets, so I picked the values based on what I found about similar gel setup from Huawei... I adjusted the charging voltage to 28.2V (float 27.2) on the device and will keep observing.
Thanks for the input!