Hi all,
I'm new to Solar as well as the Midnite Solar Classic 150.
I've had this system running for a few weeks now, very happy except I haven't got the charging working as well I think it should. The robustness of this device is superb! I just cant figure out how to make it work.
I'll state my situation, then provide a lot of system information further below as well as some attachments.
Here's what I THINK, maybe, that I know...
1) Classic 150 evolved as a Lead Acid charger which is now also commonly used for LiFePo4 applications.
2) The settings for controlling LiFePo4 charging basically include:
>> Absorb Voltage (55.4V @ 3 minute max). This effectively triggers transition from Bulk MPPT to Float.
>> Float Voltage (53.4V)
>> Rebulk Voltage (51V) triggers return to Bulk MPPT if battery voltage falls too low.
Here's the concern I'm having...
Under varying sunlight conditions, I've been unable to consistently charge.
>> In strong sunlight, Bulk MPPT really goes to town, up to 80 amps, but creates a higher voltage in doing so. This voltage reaches Absorb Volts and ends Bulk MPPT prematurely, then, once in Float, the battery may only be at 75% SoC. Float (delivering about 1 to 3 amps) is way too slow at getting the job completed.
>> In cloudy weather, Bulk MPPT produces a smaller charging current (and resulting voltage) allowing Bulk to go longer that it should. Once it finally switches to Float, it takes an hour or two to bleed current to get reduce 'over' charge.
In other words, the process appears to be only related to voltage and does not pay attention to charging current.
I'm sure I am misunderstanding some of this and hope someone can help me to understand.
Also, If there exists a detailed description of the strategy that the Classic uses, I haven't found it yet.
Thanks, Tim, Jackson, TN USA
BTW, my charge profile graph is generated from data collected by an Arduino datalogger which saves A, V, and SoC once per minute. This chart shows what happened today, cloudy / rainy, premature end of Bulk at about 76% SoC.
My background:
Retired, EET degree from 70s, proficient with electronics and power systems.
System Components:
Panels: 14 ea Trina 415W Mono Panels, 7 strings of 2
Charge Cont: Midnite Solar Classic 150
Batteries: 4 ea KiloVault LiFePo4 2.4kWh 200Ah 12V HLX+
Inverter: Magna-Sine MS4448PAE
NOTE: Detailed schematic and equipment list is attached
System Goals:
Off Grid to reduce Grid load while providing backup.
Currently supplying to Water Heating, Refrigerator, Freezer, and Water Heater
No connection to grid in any way
Charge daily to about 90-95 % SoC
Classic 150 Charge Controller Setup
VOLTS
Equalize Voltage 56.4 Used only for Lead Acid batteries
Absorb Voltage 55.4
Float Voltage 53.4
CHG TIME
Absorb Time 3 minutes
Equalize Time 0 minutes Used only for Lead Acid batteries
T-COMP
Enable / Disable Disabled Used only for Lead Acid batteries
EQ
Start / Stop Stopped Used only for Lead Acid batteries
ADVANCED
Ending Amps 2.0 Amps Used only for Lead Acid batteries
ReBulk Volts <51.0 V
Skip Days 0 Days Used only for Lead Acid batteries
LIMITS
Output Amps 88 A
Input Amps 99 A
Skip Used only for Lead Acid batteries
T-Comp (min) 52.8V Used only for Lead Acid batteries
T-Comp (max) 58.4A Used only for Lead Acid batteries
Hi Batt Temp 170 deg C Used only for Lead Acid batteries
NOTE: Setup Sheet pdf is attached
Hi Tim, welcome to the Forum,
You mentioned,
"Here's the concern I'm having...
Under varying sunlight conditions, I've been unable to consistently charge.
>> In strong sunlight, Bulk MPPT really goes to town, up to 80 amps, but creates a higher voltage in doing so. This voltage reaches Absorb Volts and ends Bulk MPPT prematurely, then, once in Float, the battery may only be at 75% SoC. Float (delivering about 1 to 3 amps) is way too slow at getting the job completed.
>> In cloudy weather, Bulk MPPT produces a smaller charging current (and resulting voltage) allowing Bulk to go longer that it should. Once it finally switches to Float, it takes an hour or two to bleed current to get reduce 'over' charge".
When MPPT CCs are in Absorb, and Float (and for LA batts, EQ), it is the battery Charge Acceptance that determines the battery charge current.
If you feel that Bulk ends too early, you can increase the Absorb voltage (if this is OK with your batts), and the same, for the current in Float.
Forgot to look, are you using the WbJr battery current monitoring accessory? This allows monitoring the current going into the batts.
I know, essentially nothing of Lithium batteries.
Many others here know much more than I. BTW, you might want to keep the EQ voltage set to the same value, as Abosrb, and some folks with LiPO batts, DO seem to use EA, on occasion ...
EDIT, to add: Here is a Link to a reseller for the WhizBang Jr monitoring device, (which you may have already):
https://www.solar-electric.com/midnite-solar-whiz-bang-jr-current-sense-module.html
Later, Vic
Hi Vic,
Thanks for your quick response!
I've worked with adjusting both Absorb volts and Float volts, trying to get the Bulk MPPT to coast into float at the right point. Unfortunately, the optimal adjustments for today, based on sunlight supply, user load, and starting SoC, will not be right if any of the conditions change.
I'm hoping that there's a lever I haven't discovered yet to make the system a little more adaptive to conditions.
Regarding battery current, I'm using an eddy current sensor which measures battery current, however, I only use it for Arduino datalogging and SoC calculation. No connection to MNS CC.
You have to be careful when a retired geek starts playing with Arduinos, no limit to what can be done with these things ;)
I wonder if having the WB Jr installed changes the way the CC works?
Thanks Again! Tim
Quote from: SunTim on September 11, 2022, 04:20:47 PM
Hi Vic,
Thanks for your quick response!
I've worked with adjusting both Absorb volts and Float volts, trying to get the Bulk MPPT to coast into floatat the right point. Unfortunately, the optimal adjustments for today, based on sunlight supply, user load, and starting SoC, will not be right if any of the conditions change ...
... I wonder if having the WB Jr installed changes the way the CC works?
Thanks Again! Tim
The WbJr will need to mount onto a 50 mV, 500 A Shunt, that is in the negative battery lead. This shunt probably exists in the Magnum DC panel.
The Classic CC can be told to use the Wb battery current to end Absorb (it is an OR function twix the Absorb timer, and the Wb reaching a setting [EA] that you choose), SO, YES, the Wb charge current measurement can change the end of Absorb, based on time, OR the EA setting.
Wb Manual"
https://www.midnitesolar.com/pdfs/10-264-1_REV_F.pdf
Later, Vic
Hey Vic,
Thanks for the eye opener!
I was aware of the WB Jr, but had no idea it was a needed add-on for reliable LiFePo4 charging. I already had plans for battery current and SoC tracking so I did not consider the WB Jr.
I bought all of the major equipment from altE last year and was provided a 'whitepaper' describing how the Classic would work with the KiloVault LiFePo4 battery set. I may be missing something but I never understood that this additional device would be needed for charging.
whitepaper attached
I'll need to do some studying to figure out what to do next. Ideally, I'd like to use my controller to tell Classic when to go to float, have no idea whether or not that's feasible.
Thanks Vic, you've helped me a lot already!
Tim,
Did you tweak the voltage your batteries are at with the voltage the Classic is seeing ?
You want to do this for the high charge point ( Absorb Voltage ) .
Do your Kilovault batteries have the built in bluetooth and do you use their app to monitor the voltage and also the SOC ( if it shows that ) ?
You have a couple of things to consider.
One is the batteries built in BMS . It may have been programmed to cut off charging at certain voltages. I have heard that some lithium companies keep their top voltage lower so they don't get into balance issues. Not sure if Kilovault is one of them . If the cells have not been balanced at the top , then there could be some that are higher than others.
My friends have Kilovault batteries and the app for them does not show each individual cell voltage so it is tough to know what is going on exactly . Each of your batteries could also be a different voltage which of course will all add up to the total when you have them in series.
You can raise the absorb time if you want them to charge longer , or raise the absorb voltage too, though it may not do much if the bms is cutting off input at above a certain voltage.
Assuming you are going by the Kilovault settings you may just have to recalibrate your SOC to say it is 100% when the batteries reach the 55.4v you have in there.
You probably want to raise the Float voltage . The idea is that you want the controller to supply power to the loads as soon as possible after Absorb , but lithium battery manufacturers don't want you to keep charging in float. If you observe what is going on with your system , you can find that point and it is probably a higher voltage than 53.4v . I have lithiums and have mine set to 54.6v . It takes awhile for lithium voltage to drop so between the time you are waiting from Absorb to Float you may see a resting period ( maybe reason for resting on the Classic will show 38) . In that time you using some of your stored power from the lithiums.
I agree with Vic that you need a shunt and Whizbang on your system to know much more of what is going on. That is the only way to know how much current is going into loads vs how much is going into batteries for charging. The Classic SOC calculations are very accurate .
Larry
Quote from: SunTim on September 11, 2022, 05:07:40 PM
Hey Vic,
Thanks for the eye opener!
I was aware of the WB Jr, but had no idea it was a needed add-on for reliable LiFePo4 charging. I already had plans for battery current and SoC tracking so I did not consider the WB Jr.
I bought all of the major equipment from altE last year and was provided a 'whitepaper' describing how the Classic would work with the KiloVault LiFePo4 battery set. I may be missing something but I never understood that this additional device would be needed for charging.
whitepaper attached
I'll need to do some studying to figure out what to do next. Ideally, I'd like to use my controller to tell Classic when to go to float, have no idea whether or not that's feasible.
Thanks Vic, you've helped me a lot already!
According to the white paper you posted - you don't have your absorb voltage high enough.
Your white paper says 56.4v and you say you have your absorb set for 55.4v
That would be one reason your batteries are maybe not filling up all the way.
Larry
Hi Tim, Larry,
I was hoping that Larry would reply, as he is much farther along in the LiPO, than I.
As I read it, sometimes ReBulk (which can start another Bulk charge, based on battery voltage) has been used by some LiPO users to help get the charge results they are looking for.
Also, there is the Open Source Topic area on this Fourm that you might peruse. Arduinos seem to have been kind of supplanted by the Raspberry Pis, it seems to me, here is a start, if you have not yet seen this:
http://midniteftp.com/forum/index.php?board=33.0
Later, have Fun discovering all of the great functions that the Classics, has as built-ins. The Wb is a very inexpensive addition, especially considering the additional info, and control of the end of Absorb.
Later, back to work, here, Vic
Quote from: Vic on September 11, 2022, 05:24:32 PM
Hi Tim, Larry,
I was hoping that Larry would reply, as he is much farther along in the LiPO, than I.
As I read it, sometimes ReBulk (which can start another Bulk charge, based on battery voltage) has been used by some LiPO users to help get the charge results they are looking for.
Also, there is the Open Source Topic area on this Fourm that you might peruse. Arduinos seem to have been kind of supplanted by the Raspberry Pis, it seems to me, here is a start, if you have not yet seen this:
http://midniteftp.com/forum/index.php?board=33.0
Later, have Fun discovering all of the great functions that the Classics, has as built-ins. The Wb is a very inexpensive addition, especially considering the additional info, and control of the end of Absorb.
Later, back to work, here, Vic
Vic - actually most of these batteries are LiFePO4 lithium iron phosphate
I think the designator LiPO ( lithium polymer I think ) usually refers to a different lithium chemistry ( and those are the ones more prone to catch on fire whereas LiFe are safer) .
Larry
Hi Larry,
Thanks for the correction.
Was trying to use more abbreviations, and forgot the correct one for LiFePO4.
This Laptop keyboard is flaking out, and the more that is typed, the more re-typing that is needed, as more, and more keys do not work reliably, makes touch-typing difficult ...
Thanks, Glad that you chimed into this conversation.
73 Take care, Vic
Hi Larry and Vic,
Tons of info thanks!!
**Larry Said...
Did you tweak the voltage your batteries are at with the voltage the Classic is seeing ?
You want to do this for the high charge point ( Absorb Voltage ) .
>> I have not done this yet. I have a 10V precision reference IC on order, should be here next week. Right now I see deviation between my DVM (12 year old Omega), The CC Voltage reading, and the Magnum Battery Voltage display. Typical deviations .1 to .2 volts. Once I get the reference I'll try to get everything trued up.
**Larry Said...
Do your KiloVault batteries have the built in Bluetooth and do you use their app to monitor the voltage and also the SOC ( if it shows that ) ?
>> Yes, they have a Bluetooth app which reports: SoC, terminal voltage, as well as the individual cell voltages. I used to check the individual cells but they were always within 10mV. Probably need to circle back on that one.
**Larry Said...
You have a couple of things to consider.
One is the batteries built in BMS . It may have been programmed to cut off charging at certain voltages. I have heard that some lithium companies keep their top voltage lower so they don't get into balance issues. Not sure if KiloVault is one of them . If the cells have not been balanced at the top , then there could be some that are higher than others.
>> I'll start keeping a closer eye on individual cells to verify there's not a problem there.
Also, I don't believe I've ever caught the KiloVaults interrupting charge current. I imagine this would be hard to miss and the CC would freak out.
**Larry Said...
You can raise the absorb time if you want them to charge longer , or raise the absorb voltage too, though it may not do much if the bms is cutting off input at above a certain voltage. Assuming you are going by the KiloVault settings you may just have to recalibrate your SOC to say it is 100% when the batteries reach the 55.4v you have in there.
>> I have tweaked the Absorb and Float Voltages in a .3 V range to try to get more consistent charging. As of now, I've put them back at recommended settings.
**Larry Said...
You probably want to raise the Float voltage . The idea is that you want the controller to supply power to the loads as soon as possible after Absorb , but lithium battery manufacturers don't want you to keep charging in float. If you observe what is going on with your system , you can find that point and it is probably a higher voltage than 53.4v . I have lithiums and have mine set to 54.6v . It takes awhile for lithium voltage to drop so between the time you are waiting from Absorb to Float you may see a resting period ( maybe reason for resting on the Classic will show 38) . In that time you using some of your stored power from the lithiums.
>> Yes the float voltage has been interesting. altE Tech Support told me to set float at 53.4 rather than the 55.20 shown on the white paper. I believe he was concerned about not overcharging/damaging the KiloVaults. I have played around with float voltages, looks like 53.4-53.6 floats with no net amps through battery. More work needed here, though. My biggest challenge is that the settings which seem to work change day to day based on conditions.
**Larry Said...
I agree with Vic that you need a shunt and Whizbang on your system to know much more of what is going on. That is the only way to know how much current is going into loads vs how much is going into batteries for charging. The Classic SOC calculations are very accurate .
>> I'll look at the WB Jr but I have been quite happy with my homebrew SoC meter. The current measurement is surprisingly accurate and consistent. OTOH, If I have to add WB Jr to get proper charge management then I'll bite the bullet.
Also, I wonder if the Aux2 input which forces float would be of any use. Does it just switch the mode to float (good?) or does it hold it in float (bad?). I could easily trigger it with my device, but I don't want to jump the gun.
**Larry Said...
According to the white paper you posted - you don't have your absorb voltage high enough.
Your white paper says 56.4v and you say you have your absorb set for 55.4v That would be one reason your batteries are maybe not filling up all the way.
>> I had lowered it a couple days ago after the charge cycle overshot and blew past the float value and had to discharge for a while. I have been a little gun-shy on doing much with the Absorb settings. I've kept the time at the minumum as I was cautioned by altE tech support. He considered this essential to avoid damage to the batteries.
So I have not used absorb for any period above 3 minutes. I'm still trying to understand how this thing is supposed to work.
Let me try to state it in my words and please correct me as needed...
1) Sunrise, Bulk MPPT begins at a couple amps and builds to 70-80 amps if there is enough sun.
2) When voltage reaches Absorb voltage level, mode changes to Absorb but only for 3 minutes (I wonder if this short time has any appreciable effect on charge?)
3) When Absorb 'times out' we switch over to Float. It seems like the system does not often coast into float, its generelly too empty or too full at that point.
This seems like a very rough process which will only succeed when I'm lucky, hence constant tweaking Absorb & Float voltages. I'm chasing my tail.
**Vic Said...
Also, there is the Open Source Topic area on this Fourm that you might peruse. Arduinos seem to have been kind of supplanted by the Raspberry Pis, it seems to me, here is a start, if you have not yet seen this:
http://midniteftp.com/forum/index.php?board=33.0
>> Vic, I'll definitely dig into this, thanks! Yup, the Raspberry is a more powerful device, but so far Arduinos have always given me the performance I needed.
Tim
Hi Tim,
Regarding Aux 2, the WbJr needs to have Aux 2 comitted for its exclusive use, as Aux 2, can be used as an Analog input, whereas, Aux 1, is not able to do Analog.
FWIW, Vic
Quote from: SunTim on September 11, 2022, 07:29:32 PM
Hi Larry and Vic,
Tons of info thanks!!
**Larry Said...
>> I had lowered it a couple days ago after the charge cycle overshot and blew past the float value and had to discharge for a while. I have been a little gun-shy on doing much with the Absorb settings. I've kept the time at the minumum as I was cautioned by altE tech support. He considered this essential to avoid damage to the batteries.
So I have not used absorb for any period above 3 minutes. I'm still trying to understand how this thing is supposed to work.
Let me try to state it in my words and please correct me as needed...
1) Sunrise, Bulk MPPT begins at a couple amps and builds to 70-80 amps if there is enough sun.
2) When voltage reaches Absorb voltage level, mode changes to Absorb but only for 3 minutes (I wonder if this short time has any appreciable effect on charge?)
3) When Absorb 'times out' we switch over to Float. It seems like the system does not often coast into float, its generelly too empty or too full at that point.
This seems like a very rough process which will only succeed when I'm lucky, hence constant tweaking Absorb & Float voltages. I'm chasing my tail.
Tim
Tim, ]
Not sure I follow you.
Lithium stay at a pretty constant voltage while they are charging. It is only when they get in the 90% full range that the voltage on them starts to increase towards the Absorb setting.
Depending on bms the cells may also start balancing sometime in this range . I can see it in the software I have but doubt there is any balance indication in software you have.
So once the voltage gets to the 56.4 they call for , every cell will ( or should ) be 100% full.
So charging is done at that point ( unless a very low amount of current was being put in and balancing happening - the bms only has limited balancing current its resistors for each cell can handle.
On a lead acid battery the transition to float works right away . But on a lithium the top voltage may stay awhile ( depending on loads ) and during this time the Classic will show resting and won't go to float until the lithium battery voltage drops down to or below the Float voltage setting.
So if you watch your meter and don't see any current going into the battery - you want to put the float voltage up higher than lower so that PV will supply the loads as soon as possible instead of batteries stored power.
The bms should protect your battery from any over voltages , over currents, under voltages.
This guy has a lot of videos with lots of info on all his testing using all kinds of bms .
There was one video he made where he explained what he learned watching the cells charge but I can't find that one now
Anyway take a look - he is always tinkering
https://www.youtube.com/c/OffGridGarageAustralia/videos
Larry
Thanks for the response Larry.
I noticed that the charge profile I attached in first post seems to be chopped off, making it impossible see the voltage values.
I'm now attaching a pdf which will hopefully do better.
I will study your comments and the videos to see if I cant get a better handle on what's happening.
Thanks again for your help. Have a nice evening.
Tim
I went out and measured some voltages...
To be continued...
Thanks, Tim
Quote from: SunTim on September 11, 2022, 10:14:03 PM
Thanks for the response Larry.
I noticed that the charge profile I attached in first post seems to be chopped off, making it impossible see the voltage values.
I'm now attaching a pdf which will hopefully do better.
I will study your comments and the videos to see if I cant get a better handle on what's happening.
Thanks again for your help. Have a nice evening.
Tim
How are you calculating SOC ?
I don't understand - according to your graph you only got a bit over 55 v . If that is what you have the absorb set for that is too low.
MyMidnite is capable of graphing all the data also - though without a Whizbang it won't show the system amps ( what is actually going into the batteries) .
But it will show the classic charging state.
With the MyMidnite graph you can choose which data you want to show on the graph , and then drag your mouse across it to show the data at any particular time.
Maybe you can do that with yours.
But I don't understand how the voltage on your batteries would jump up so high if their SOC is so low.
I will have to look at your schematic again.
Larry
Where on your wiring is the current sensor located ?
With a whizbang - the shunt would be located on the negative wire between the battery and all the other negatives.
There would only be one connection from battery to one side of the shunt. Every other negative in the system would be on the other side of the shunt .
That way it can tell exactly what is going into and out of the battery .
It counts the SOC down as the power is drawn out , and counts it back up as the power goes in. None of the power going to a load directly from the charge controller will be going through the shunt.
Does your bms data show the amount of current charging or discharging in each battery ?
Larry
good luck figuring out the best charging practices for lifepo4 batteries.
The forum topic linked below seems to have been 2 years of numerous people debating how to interpret everything with voltage and current
https://diysolarforum.com/threads/lifepo4-voltage-chart.3156/
The Australian garage guy video I watched he said he couldn't tell by voltage and had to watch the current to truly tell for each cell.
I will try to find that video.
Seems to be too much info and a lot of depends on bms or amount of charge or discharge going on or if cells are at a rest for awhile.
Larry
Hi Tim.
So, as far as the Classic not outputting any current to the battery for a while after Absorb, that is OK since the battery voltage is above the Float set point.
There is no place for the power to go at that time. Not until the battery voltage drops below the float voltage. If your system had a larger load on it, then that voltage should drop faster. No need for that though.
If, in Bulk, you think there is not enough power going out, please take note on the PV input voltage at that time and also the Voc of the array. The open circuit voltage. With no partial shading on the PV array, that MPPT voltage (Vmp) will be somewhere around 80% of Voc.
For Absorb and Float, it could be almost anywhere above the Vmp all the way up to the Voc.
If you have the Absorb time set to 3 minutes and it does NOT do any Absorb, try either changing the ending amps to zero (disabled) or raising the 3 minute time to, say, 10 minutes and see if it Absorbs. Last I checked, Absorb should work fine at 3 minutes.
Lettuce know what you find :)
boB
Quote**Larry Said...
According to the white paper you posted - you don't have your absorb voltage high enough.
Your white paper says 56.4v and you say you have your absorb set for 55.4v That would be one reason your batteries are maybe not filling up all the way.
>> I had lowered it a couple days ago after the charge cycle overshot and blew past the float value and had to discharge for a while. I have been a little gun-shy on doing much with the Absorb settings. I've kept the time at the minumum as I was cautioned by altE tech support. He considered this essential to avoid damage to the batteries.
So I have not used absorb for any period above 3 minutes. I'm still trying to understand how this thing is supposed to work.
Let me try to state it in my words and please correct me as needed...
1) Sunrise, Bulk MPPT begins at a couple amps and builds to 70-80 amps if there is enough sun.
2) When voltage reaches Absorb voltage level, mode changes to Absorb but only for 3 minutes (I wonder if this short time has any appreciable effect on charge?)
3) When Absorb 'times out' we switch over to Float. It seems like the system does not often coast into float, its generelly too empty or too full at that point.
This seems like a very rough process which will only succeed when I'm lucky, hence constant tweaking Absorb & Float voltages. I'm chasing my tail.
I really think the main reason why you are at 75% is that you are under charging your batteries (due to the 55.4V absorp setting). You mentioned that the voltage shoot pass the float voltage, which is 2V less than what you set the absorb voltage at. Did it shoot pass the absorp voltage of 55.4 that you have set.
In my view it is either you trust the midnite CC to curtail charging at or around 55.4V or you dont. Even if you dont trust the CC the BMS in the battery should open the circuit and automatically stop receiving charge from the CC. Either way my advice would be to set the Absorp voltage as the white paper suggest, 56.4V, and trust that the CC and the BMS will protect the battery from over charge.
When you set the voltage so low in Lithium it has a benefit of extending the life cycle of the batteries tho so the system will still work. It is whether or not you are storing enough capacity for loads you wish to maintain. you can stay at 55.4V/75%, but may have to increase capacity by installing more batteries if your loads are not being maintained.
I must also add that you have a lot of charging and a relatively small bank. As a comparison I currently have 5.8kW of panels charging a 25.5kwh Lithium bank. I have it programme to only charge up to 85-90% of capacity to extend the life of the bank. I had to offset the reduction of capacity, as a result of under charging, by increasing the battery bank size because I have prioritized cycle life.
Hope this helps.
Damani
PS
I should also add that the kilovaults seems to be on the very low side of trypical LifePO recommended absorp voltages. If you research and read enough on this topic you will discover that most absorp voltage for this chemistry sits around 3.60-3.65V, so even going with the lowest at 16cells that would be 57.6V vs 56.4V for the kilovaults. this suggest that they have been very conservative in their design and probably did this so that they could reach the desired 7yr warranty without any issues(sorta like what the auto manufacturers are doing with their warranties on electric vehicle battery banks, where they warranty for 10yr and dont let the bank be charged at more than 80% hence ensuring that they meet the desired 10yr lifecycle). The result is that decommissioned e-vehicle batteries still have alot of life left in them even after service life, as evidenced by the very robust, popular and in high demand 2nd life EV lithium batteries that have flooded the market in the past 5 to 7yrs.
Thanks for all the help so far.
I found two videos at Off-Grid Garage which were very helpful...
LiFePO4 charging and discharging curve explained. How far to go?
https://www.youtube.com/watch?v=uNFIkNtsd1c
Takeaways:
1) Absorption Setpoints: 3.45 >> 55.2
2) Float Range: 3.35-3.375 >> 53.6-54.0
Finding 20% and 80% SOC with LiFePO4. Why charging current and charging speed matters!
https://www.youtube.com/watch?v=qZg4Jxftw-c
Takeaways:
1) The achieved SoC at the reaching of Absorption varies greatly based on how much charge current is involved.
2) This problem can be reduced by using the optimal absorb voltage (somewhere around 3.45V)
3) Since CC does not actually regulate the charge current during the Bulk MPPT phase, this variation is not going away. It needs to be compensated by adjusting the Absorb time
So, here's where I am...
1) I ordered the WB Jr last night. I now understand that you can't know when to end absorption cycle unless you monitor Absorb Current.
2) Until I get this installed I'll return to Midnite Solar / KiloVault recommendations while holding the absorb time to the 3 minute minimum. While this doesn't give good charge consistency, it is conservative and battery friendly.
Thanks, Tim
** Larry Said...
How are you calculating SOC ?
>> I measure battery voltage and current using a non contact 200A Hall Effect Current Sensor (LEM HASS 200-S)
The sensor is located in the jumper between two of the batteries so it is insured to capture true battery current flow.
SoC is derived by measuring and totalizing Ah in and out.
I routinely check my SoC against that reported by the KiloVault app, they track pretty well.
** Larry Said...
I don't understand - according to your graph you only got a bit over 55 v . If that is what you have the absorb set for that is too low.
MyMidnite is capable of graphing all the data also - though without a Whizbang it won't show the system amps ( what is actually going into the batteries) .
>> Whizbang on order now, I considered triggering float using my current measurement and the AUX2 input but didn't want to deviate from standard CC operating modes.
** Larry Said...
Does your bms data show the amount of current charging or discharging in each battery ?
>> The KiloVault app indicates battery current in real time, that data seems to be in agreement with what I measure in the battery circuit.
** boB Said...
So, as far as the Classic not outputting any current to the battery for a while after Absorb, that is OK since the battery voltage is above the Float set point.
There is no place for the power to go at that time. Not until the battery voltage drops below the float voltage. If your system had a larger load on it, then that voltage should drop faster. No need for that though.
>> Hey Bob, thanks for joining in. This makes sense, but if the voltage is higher than it ought to be then I already did something wrong.
** boB Said...
If you have the Absorb time set to 3 minutes and it does NOT do any Absorb, try either changing the ending amps to zero (disabled) or raising the 3 minute time to, say, 10 minutes and see if it Absorbs. Last I checked, Absorb should work fine at 3 minutes.
>> That's interesting, implies that the battery should already be nearly fully charged before it enters Absorb. So I've got a dumb question...
Why Absorb at all, why not just jump to float as soon as the correct Absorb voltage is sensed?
** Damani Said...
I really think the main reason why you are at 75% is that you are under charging your batteries (due to the 55.4V absorp setting). You mentioned that the voltage shoot pass the float voltage, which is 2V less than what you set the absorb voltage at. Did it shoot pass the absorp voltage of 55.4 that you have set.
In my view it is either you trust the midnite CC to curtail charging at or around 55.4V or you dont. Even if you dont trust the CC the BMS in the battery should open the circuit and automatically stop receiving charge from the CC. Either way my advice would be to set the Absorp voltage as the white paper suggest, 56.4V, and trust that the CC and the BMS will protect the battery from over charge.
When you set the voltage so low in Lithium it has a benefit of extending the life cycle of the batteries tho so the system will still work. It is whether or not you are storing enough capacity for loads you wish to maintain. you can stay at 55.4V/75%, but may have to increase capacity by installing more batteries if your loads are not being maintained.
I must also add that you have a lot of charging and a relatively small bank. As a comparison I currently have 5.8kW of panels charging a 25.5kwh Lithium bank. I have it programme to only charge up to 85-90% of capacity to extend the life of the bank. I had to offset the reduction of capacity, as a result of under charging, by increasing the battery bank size because I have prioritized cycle life.
>> Hi Dasani, thanks for responding. I think your right that I had set the Absorb voltage too low. I was trying to avoid what I thought was overshooting of voltage prior to going into float. I'm learning.
Up til now my goal was to try and get the batteries above about 90% SoC by end of each day. My rationale is that if I fail to keep the batteries near full then I could be caught short if / when I need them for backup during outage.
When you refer to "programme to only charge up to 85-90% of capacity to extend the life of the bank" can you tell me how you accomplish that? That's what I'd like to be doing, I just haven't figured out how to get there.
** Damani Said...
I should also add that the kilovaults seems to be on the very low side of trypical LifePO recommended absorp voltages. If you research and read enough on this topic you will discover that most absorp voltage for this chemistry sits around 3.60-3.65V, so even going with the lowest at 16cells that would be 57.6V vs 56.4V for the kilovaults. this suggest that they have been very conservative in their design and probably did this so that they could reach the desired 7yr warranty without any issues(sorta like what the auto manufacturers are doing with their warranties on electric vehicle battery banks, where they warranty for 10yr and dont let the bank be charged at more than 80% hence ensuring that they meet the desired 10yr lifecycle). The result is that decommissioned e-vehicle batteries still have alot of life left in them even after service life, as evidenced by the very robust, popular and in high demand 2nd life EV lithium batteries that have flooded the market in the past 5 to 7yrs.
>> I completely agree that KiloVault would like me to baby the battery, that makes good sense. OTOH, I am being careful not to do anything reckless that could affect battery life.
Thanks for all the help!!
Tim
Here's a little additional info...
Updated Schematic including Battery Monitor
A few photos
Have a great day! Tim
Tim,
QuoteWhen you refer to "programme to only charge up to 85-90% of capacity to extend the life of the bank" can you tell me how you accomplish that? That's what I'd like to be doing, I just haven't figured out how to get there.
I am using 2nd life lithium ion (nmc) and using 4v per cell (translate to 32V for a 24V or 8cell system) and programmed the midnite CC accordingly (i.e. absorp set to 32V for 10mins and float set at 31.6V, so that the PV and CC are carrying the loads of the house up until resting). So the end effect is that the system voltage at the end of the charging day is as close to 32V as possible and discharges down to about 29V which is approximately opperating between 85/80% down to about 35% ideally to at least get a doubling of life cycle.
Review the below link at battery university. it is such a good resource to learn about all types of batteries. It is a highly recommended read.
https://batteryuniversity.com/article/bu-808-how-to-prolong-lithium-based-batteries
Damani
Tim,
It is possible to program the Classic to limit just the charging current to the battery while leaving it also capable of providing full current to any loads.
That probably wouldn't benefit you though.
Larry
Thanks Damani,
It appears that there is a pretty severe life-cycle penalty for hot-rodding these batteries, I had read that but hadn't looked at test data.
A question... The article refers to Li-Ion batteries rather than LiFePo4, I wonder how similar they are from the life cycle point of view?
Hey Larry,
To accomplish the limiting of battery current, does the WB Jr needs to be in place?
I had seen the limit parameters for input and output amps, but it didn't seem very useful if it doesn't differentiate between inverter and battery juice. It would certainly help protect the CC or PVs, but not much else.
I first balked at having to add the WB Jr, but I'm warming up to the idea. ;D
Thanks for the help, the knowledge found on this forum is a great asset!
Tim
Hi Tim, I am not Larry,
But, because the MidNite Classic CCs are the by far most versatile on this planet, there is a function that will Limit battery charge current, and still supply loads (up to the total available available from the input to the Classic. Think that Larry alluded to this.
Had a bit of trouble finding the White Paper on it, but, this is, at least an early version of the paper, attached.
The WbJr is the mechanism that allows the Classic to work this magic, let's see if this works.
Later, Vic
Thank You Vic,
Its now in my file folder waiting for the new toy to arrive.
Seems like the WB Jr will allow a lot of additional capability.
I had noticed, in my equipment search last year, that there were very few CC that incorporated a Battery Monitor. Perhaps that's due to the way that the LiFePo4 technology arrived and turned the tables.
I think I remember seeing a Chinese label, maybe HiWatt? that incorporates CC, Inverter, and Battery Monitor in one unit. However, I was scared of committing to equipment where user support might be a problem.
Thanks! Tim
The current limiting info is also in the Knowledge Base along with a lot of other info you might be interested in
http://www.midniteftp.com/support/kb/faq.php?cid=1
Larry
If it were me I'd raise the charging setpoints. Your absorb setpoint is netting you 3.462 Vpc.(volts per cell). I tried using conservative setpoints early in my LiFePo4 battery life. I found without getting a full charge I was ending up with a pretty low Soc. in the morning. Raised my setpoint closer to 3.65 Vpc = 58.4 and corrected that problem.
** Larry said...
The current limiting info is also in the Knowledge Base along with a lot of other info you might be interested in
http://www.midniteftp.com/support/kb/faq.php?cid=1
>> Thanks Larry, I hadn't visited the KB yet. Looks like lots of useful info. I can imagine that the current limit should make the charge process much more repeatable.
A question, I have not been able to figure out which Firmware rev I am running. Since my Classic is less than a year old, Id assume that my firmware is correct for the WB Jr, but I'd like to verify.
** littleharbor2 said...
If it were me I'd raise the charging setpoints. Your absorb setpoint is netting you 3.462 Vpc.(volts per cell). I tried using conservative setpoints early in my LiFePo4 battery life. I found without getting a full charge I was ending up with a pretty low Soc. in the morning. Raised my setpoint closer to 3.65 Vpc = 58.4 and corrected that problem.
>> Hi littleharbor2, thanks for the info, I'll be doing several trials once I get the WB Jr current controller installed. I know that for right now, as I run at 56.4 V Absorb, I'll see charging current hitting 80 amps, about .4C. I'd be a little nervous about going beyond that.
Thanks All!
Hi Tim,
Regarding finding the FW Revs that the Classic is using;
From the MNGP display, from the Main Status screen, press the small round, grey button on the left about six times, and you should find the Cl FW, MNGP, and, IIRC, the Net FW versions.
From the Local Status APP, that runs on Win machines, it is probably also available, under the Info selection ... do not have a Classic available here, am in town.
Larry, thanks for the info on the bat charge current limiting info being on the Knowledge Base, was using the Forum Search (from the top, Index page, and the KB info, was generally not there, probably because the subject name was not my search term ... thanks.
FWIW, Vic
Quote from: SunTim on September 13, 2022, 01:41:36 PM
** Larry said...
The current limiting info is also in the Knowledge Base along with a lot of other info you might be interested in
http://www.midniteftp.com/support/kb/faq.php?cid=1
>> Thanks Larry, I hadn't visited the KB yet. Looks like lots of useful info. I can imagine that the current limit should make the charge process much more repeatable.
A question, I have not been able to figure out which Firmware rev I am running. Since my Classic is less than a year old, Id assume that my firmware is correct for the WB Jr, but I'd like to verify.
** littleharbor2 said...
If it were me I'd raise the charging setpoints. Your absorb setpoint is netting you 3.462 Vpc.(volts per cell). I tried using conservative setpoints early in my LiFePo4 battery life. I found without getting a full charge I was ending up with a pretty low Soc. in the morning. Raised my setpoint closer to 3.65 Vpc = 58.4 and corrected that problem.
>> Hi littleharbor2, thanks for the info, I'll be doing several trials once I get the WB Jr current controller installed. I know that for right now, as I run at 56.4 V Absorb, I'll see charging current hitting 80 amps, about .4C. I'd be a little nervous about going beyond that.
Thanks All!
You can see the firmware version number on the MNGP - one of the screens shows it
Are you using Local Status app ? It is good tool to backup and change your settings . Anyway it says version number on a screen on there too.
Your 56.4 should be okay .
I am glad you asked about the current limiting of the Classic because I had never used it before - but I did today and it is pretty much what I need.
I have some cells out of balance but when they are charging full current from PV voltage raises too high too fast and balancing won't work right.
So today when the batteries were showing close to 100% full I experimented by turning down current. It let the voltage rise much slower and I could watch when balancing started and nothing overshot too much.
My batteries are a bit different than yours - I have three 48v and they communicate with each other to keep the voltages of each battery close . The BMS will turn on or off current as needed to do this. So everything works much better with lower current - I started with 15 amps for all three but then lowered it to around 3 amps. I think there is a minimum that it won't go below even though the setting goes lower - maybe it is 5 amps ? Anyway I will need to experiment to find out how this works out.
By the way - when you get into the current limit - if you hold the top left button you can scroll through numbers faster - not sure if that is in the instructions or not. I just figured it out.
Larry
Thanks for the Firmware locations, I'll check it out.
BTW, is there a way to update firmware without putting Classic on line? Can I just download it to me PC then connect to Classic later and install?
I don't have wireless in the barn, I suppose I could wire up a router out there but I'd like to avoid that.
I'm attaching chart pdf from today's charge cycle...
>> Absorb 56.4V Float 53.6V
>> No clouds today, charge current was steady at about 40A
>> Charge went from 72% to 100% (56 Ah) took about 1.5 hours, so the numbers work out pretty well.
>> Voltage not exactly what I expect, I'll feel better when I get all the voltages calibrated. USPS has my Standard Reference IC about 60 miles from me, so that means I might have it another week ::)
Downside is that the charge got jerked right to 100% SoC, I want to control that better.
I'd like to see it coast into about 90% SoC without the day to day variation.
I'm starting to understand that just tweaking voltages won't get me there.
This is getting fun! Tim
Yes you can download firmware update and just do it from that download.
That is how I do it.
Just make sure you do both the classic update and the mngp update since they work together.
I have been thinking of getting a nice power supply that is programmable for constant voltage, constant current and will switch over automatically.
That would make getting the cells in my battery balanced much easier. Yeah you would think they might come balanced but that isn't the way things work when you don't spend the $$ .
Larry
You will have to figure out how to get the individual cell data graphed - that would tell a lot.
Larry
It is also important to understand when a controller transitions from Absorb to Float it can not pull the battery voltage down so it has to back of or even go to resting until the voltage sags down to the float voltage. This is not an over shoot it is just how batteries work:
So if absorb is 56.4 and Float is 55 when the classic times out and goes to float it will make 0 amps until the battery drops to 55, This is not considered absorb still in the battery world as we are now discharging not charging.
Can you elaborate on "Blew past Float voltage" as the classic is very fast at regulation so this seems odd?
Ryan
Good morning Ryan, Thanks for chiming in.
Glad you mentioned that concern as I was getting ready to discuss a related observation.
IR Drop is a consideration which has been associated with DC Motors forever, it is a voltage drop that exist 'inside' the motor. The IR Drop is considered to be in series with another voltage which is called Counter EMF. The CEMF is a function of motor speed and field strength. So in a fixed field situation, the CEMF will be exactly proportional to speed. So, when you want to control the speed of a motor, controlling the armature voltage works poorly because the varying IR drop will cause the CEMF will not be constant and the motor speed will also vary. OTOH, a "CEMF Regulator" calculates IR drop and makes appropriate adjustments to the motor voltage, resulting in fairly constant speed over a range of loads.
OK, sorry for that side road, and apologies if I'm covering material that everyone already knows.
When I search for info about LiFePo4 IR drop concerns I find very little. Battery Internal Resistance is covered pretty frequently, which appears to be about 0.1 ohms. So, in my mind, battery IR drop must be considered for LiFePo4 batteries just as with DC motors.
Here's the reasoning...
Example:
56.4V Absorb for a battery with 0.1 Ohm Internal Resistance.
50 Amp charge current just before Absorb Voltage is reached.
At that time we are reading 56.4V
However, the actual voltage is 56.4 - IR Drop, or 56.4 - (50A x 0.1 ohm), or 56.4V - 5V, or 51.4 V.
So once the charge amps are stopped at end of Absorb, in this case, the voltage would immediately drop to 51.4 V.
If Float is set to say 53.6, then the battery voltage is shy and will be 'slowly' brought up by the float voltage.
I believe that a 'Correct" Absorb will have the battery voltage very close to the desired float voltage as soon as its done.
Practical Implications...
If you have a particular Absorb voltage set:
On a sunny (High charge amp) day, high current IR drop will cause the final voltage to be lower than expected.
On a cloudy (Low charge amp) day, low IR drop will cause the final voltage to be higher than expected.
Based on the above, there is no solid relationship between absorb voltage and SoC. For this reason, I believe folks opt for higher Absorb Voltage to be sure it can cover the range of IR Drop.
So I conclude that charging LiFePo4 batteries with no control of the charging current is a significant compromise.
When I get the WB Jr installed, I'll probably try to do a daily charge rate of 15 or 20 amps.
Interested in others thoughts on this. I'm a newby on these batteries so may be misunderstanding things. Also, I know nothing about Lead Acid batteries, so I have no comments on them.
Thanks, Tim
Tim
The Classic was designed primarily for lead acid batteries.
When charging lead acid there is no lag between absorb and float. The absorb time for lead acid is hours , and once done the voltage on the battery will drop fast enough to go right to float.
So you kind of have to look at the Classic as being repurposed to use for Lithium.
Once you get to know your battery and what you want to do , you can certainly tweak all the settings to make it work pretty decently.
Of course there would be other features that would be nice to have for lithium and even if possible , those would have to be designed and tested with firmware revisions.
Where lead acid are very rugged and can take some abuse, lithium can not be abused . So they have a bms but every variation of bms act a bit differently .
Anyway the Classic will do the job once tweaked for specifics.
Larry
Hi Larry,
I hope my post didn't seem like a slam of the Classic 150, that was not my intention at all. I like the Classic, robust design and a lot of flexibility.
As you stated, one has to learn how to use that flexibility. ;) I think, once the WB Jr is in place, the Classic should do a fine job.
With all that said...
My reseller who helped me select the system did not have much to say about what types of challenges the charging would present.
I think it would be great if MN let their resellers know that the WB Jr is a very valuable and "recommended" add on for LiFePo4 clients.
Have a great day! Tim
No I didn't think you were slamming the Classic.
They work pretty good and I haven't ever had a failure with mine in 10 years.
Larry
SunTim said:
Downside is that the charge got jerked right to 100% SoC, I want to control that better.
I'd like to see it coast into about 90% SoC without the day to day variation.
The Classic with WB Jr. monitoring the actual battery current and SOC% calculation has a sanity check, "reset" to 100% SOC when it goes from Absorb to Float. This is because the SOC% can be calculated without doing Absorb cycles, BUT it is assumed that when an Absorb cycle finishes, that the battery has been as fully charged as it can ever be.
That is why it goes to 100% at that time. IF the Classic always seems to be going to 90% (ish) SOC% and then goes to 100% when going from Absorb to Float, then it is possible that the amp-hour efficiency in the Classic's WB-Jr. settings are set too low. This is normally especially possible with lithium batteries.
boB
I got the monitoring of one of my batteries working - thanks to Graham who has done all the ClassicDIY monitoring stuff.
I have esp32 plugged into battery which gets the individual cell data. I graphed that on Grafana and also added some data from the Classic so we know what is happening with that.
I had full current going into battery but then it clouded up - and I also went out and set current limiting on the Classic ( to charging ) to 30 which is 3 amps. I am not sure the Classic can reliably limit current this low but it is okay what I have observed . It keeps the current low enough so that the the battery bms can do its job.
I had figured out that the battery bms was set to balance at 3.5v and I think 30mv differential between cells. I talked to Jakiper and they remote connected to my computer and then changed the balance threshold to 3.3v and 20 mv differential. The bms will only start to balance cells when both those conditions are met. And the bms will only balance as voltage is rising - soon as it see zero current or a negative current the bms stops balancing. Anyway the imbalance I was seeing happened as the voltage neared 55.2v when battery was about 100% SOC.
The new monitoring lets me watch this now historically instead of having to sit in front of the computer watching it real time .
Here is the latest graph where you can see the cell voltages start to diverge as voltage comes up toward absorb setpoint.
Larry
Here is the transition this morning to absorb voltage where you can see the cell voltages diverge. There was a 30 minute absorb time but the cells do not balance the whole time since the bms cuts off the current and balancing stops. Towards the end I raised the absorb voltage to 56v and you can see that rise on the graph. At the end it times out absorb and goes to float voltage of 54 but charge controller goes to resting( provides no current) until battery voltage drops to 54v . That is why it is showing negative current from battery to supply loads.
Larry, so you have gotten an ESP32 to work ?
I have some of these but I hear from everybody that they do NOT work ?
I always wondered how they could sell so many of these if they did not work ?
boB
Hi Larry,
Great work on getting the BMS data graphed. Looks very nice, we can all learn from the work that you are doing.
73, and, have more more battery fun! Vic
Quote from: boB on September 23, 2022, 06:40:28 PM
Larry, so you have gotten an ESP32 to work ?
I have some of these but I hear from everybody that they do NOT work ?
I always wondered how they could sell so many of these if they did not work ?
boB
Yes Bob , it is using an Esp32 to get that battery data - four wires connect to esp32 pins , 3.3v , ground, tx, rx . Those go to RS485 to ttl converter board which then connect A and B to RS485 jack on the battery. The wifi on the esp32 connects to my router .
I have esp8266 and esp32's running stuff around here like my energy ac and dc energy monitors , and I have another one with some one wire temperature sensors on them. They work great . If you use the Tasmota project , you just flash their firmware and they are all ready to support lots of different kind of sensors and devices . The webpage showing data is built in , so you just connect to them and see data. But they also let you easily configure to send the data out as mqtt to a broker. So that is what I do - send it to my raspberry pi running mosquito broker , and then that gets the data to the graphana software also running on the pi. It has all been running very stable for a long time.
So to answer you question they do work , but I can see why someone said they didn't work. When I first started messing around with them it was more a pain , but with the Tasmota flashed in them it makes everything very easy to do.
The one that I am using for the battery is more of a custom firmware code that Graham put together. He said he just adapted the esp project that he had for the Classic from modbus tcp to rtu . Though I think he has done a bit more tinkering with it . But worked for me first try.
I am trying to learn how to use VS Code with PlatformIO extension in it because that seems to be how everyone with github projects does this stuff these days. You can get the projects right from github and use platformio to write and debug the code and then send the revisions back. I only scratched the surface on how to make all this work, and ran into a few issues. But at least I can see the code and try to understand how it all works.
Oh and by the way - the Tasmota firmware is so easy to flash because you can use Chrome and go to Tasmota webpage and it will flash it on the esp32 right from there - amazingly easy. But if you want you can also flash the firmware with ArduinoIDE or the PlatformIO but there is a little bit more messing around to do.
Here is a screen shot of a Tasmota manger program running on my Pi - but not all those esp devices are active right now.
Larry
That s awesome, Larry !
Glad to hear that all these ESP32's I bought (cheap) may have a future after all !
Update on addition of the WB Jr...
Installed a few days ago, everything worked exactly as described and as covered in the instructions. The design, using a single wire interface is a clever and effective method. The software is well integrated including using the yellow LED to indicate current limit and an intermittent pop up of SoC value on the main status display. Kudos on a well implemented after market addon.
I've set the current limit at 30A and am seeing a consistent charge cycle each day (93-95% SoC). If left to its own, Bulk current will vary between 20 and 60 amps based on sun. As mentioned earlier, this variation directly affects the battery terminal voltage at end of Absorb cycle. My goal is to have it complete at about 90% daily.
It'll be interesting to see how the WB Jr likes this routine of daily charging to something less than 100%. I believe it will self calibrate when a 100% SoC condition is detected, not sure whether it will interpret my transition to float as a 100% charge even though it is something less than that.
Having fun, Tim
PS Larry, thanks for sharing the serial communication developments, I'll definitely be following your lead, somewhere down the road.
Beautiful mounting job, SunTim ! THAT's the way it's done !
Yes, I really liked the one-wire power and bi-directional method.
Well, it's really TWO wires including the negative cable but don't tell anybody !
boB
Hi All,
Back again with a few more questions about the Classic 150 CC.
I have been trying to get consistent daily charging performance at a predictable charge around 90%
My current settings are...
Absorb 55.8V at 3 minutes
Float at 53.6
Battery current limit at 30A via WBjr
Classic 150 voltage calibration set at +0.3V (based on battery measurements with DVM)
My results...
In general, I get a daily charge in the 92-97 % range.
Some days, however, things go to float too early resulting in charge anywhere from 75 - 90%.
Today I was able to watch this early absorb-float situation.
CC displayed voltage was fluttering in the 55.1 to 55.4 range. (Absorb set at 55.8 V)
Mode display started jumping between Bulk MPPT and Absorb for about 3 or 4 minutes.
At only one point during the mode toggling, I saw voltage quickly jump to 55.8, at the same time the internal fan shut off.
Then CC went to float.
The charge was at about 90% (as reported by Kilovault battery app)
So, I forced back to Bulk mode by raising ReBulk setpoint to 54V. This worked fine, but within about 5 minutes the above events just happened again.
Tried a second ReBulk, same result again.
So, it looks like we are determined to get to absorb even though the displayed voltage is not there yet.
Questions...
1) Can battery voltage jitter enough to be triggering Absorb even though the displayed voltage is about 0.5V shy?
2) Does the Classic have any filtering settings to slow the reaction to transient voltages.
3) Is it normal for mode to toggle back and forth between Absorb and Bulk MPPT?
4) Is there any quick way to return from Float to Bulk without having to change the ReBulk setpoint.
5) Has anyone tried putting a scope on the battery voltage to see how it behaves? I will probably give this a try, but I'll need round up an isolation transformer for the scope first. I wonder if battery BMS operation can cause voltage to fluctuate quickly.
Thanks for the support!
Tim
Quote from: SunTim on October 14, 2022, 01:42:49 PM
Hi All,
Back again with a few more questions about the Classic 150 CC.
I have been trying to get consistent daily charging performance at a predictable charge around 90%
My current settings are...
Absorb 55.8V at 3 minutes
Float at 53.6
Battery current limit at 30A via WBjr
Classic 150 voltage calibration set at +0.3V (based on battery measurements with DVM)
My results...
In general, I get a daily charge in the 92-97 % range.
Some days, however, things go to float too early resulting in charge anywhere from 75 - 90%.
Today I was able to watch this early absorb-float situation.
CC displayed voltage was fluttering in the 55.1 to 55.4 range. (Absorb set at 55.8 V)
Mode display started jumping between Bulk MPPT and Absorb for about 3 or 4 minutes.
At only one point during the mode toggling, I saw voltage quickly jump to 55.8, at the same time the internal fan shut off.
Then CC went to float.
The charge was at about 90% (as reported by Kilovault battery app)
So, I forced back to Bulk mode by raising ReBulk setpoint to 54V. This worked fine, but within about 5 minutes the above events just happened again.
Tried a second ReBulk, same result again.
So, it looks like we are determined to get to absorb even though the displayed voltage is not there yet.
Questions...
1) Can battery voltage jitter enough to be triggering Absorb even though the displayed voltage is about 0.5V shy?
2) Does the Classic have any filtering settings to slow the reaction to transient voltages.
3) Is it normal for mode to toggle back and forth between Absorb and Bulk MPPT?
4) Is there any quick way to return from Float to Bulk without having to change the ReBulk setpoint.
5) Has anyone tried putting a scope on the battery voltage to see how it behaves? I will probably give this a try, but I'll need round up an isolation transformer for the scope first. I wonder if battery BMS operation can cause voltage to fluctuate quickly.
Thanks for the support!
Tim
Hi Tim...
It is somewhat normal to see going from Absorb to Buk and back and forth but not always.
Do you have Ending Amps set to greater than zero amps ? (0A = Disabled)
The voltage below Absorb that the Classic can go down to and still count down the Absorb time is about -0.3V so -0.5 0r -0.4V should not be allowed.
We are working on changing this to be less value for lithium batteries though.
I would think that if there was a BMS issue where it makes the Classic go UP in voltage, it would go higher than what you were seeing.
I don't suppose you have any battery temperature compensation enabled in the Classic since you are using lithium ? Is there a temp sensor plugged in ?
What is your battery charge efficiency set to ? Should be near 100% or a few % below that.
If you are able to get a video of the MNGP screen, that might help see what is going on.
boB
Hi Tim,
There is the ability to force the Classic, into doing a new Bulk, or force it into Float:
Tweaks > More, and Right Arrow over about three selections, and up arrow ... I forget, the exact position of these two items.
Toward the back of the Classic Manual, there is a very complete Menu Map that will show these Deluxe Menu options.
To me, the Classic has, by far the most comprehensive number of actions that the Classic will perform, often available from the MNGP, or from the MN Locat app. Forcing a new Bulk, or Float is available in the local app.
Vic
Did you tweak the Classic voltage for the high part of the charge ?
Larry
Hi all, thanks for the responses, you guys are great!
from boB's post...
Do you have Ending Amps set to greater than zero amps ? (0A = Disabled)
>> Hi boB, thanks! Ending amps currently set to 2.0 Amps. It seems that current stays close to around the 30 A current limit value, so I suspect the End Amps should not be coming into the picture.
I don't suppose you have any battery temperature compensation enabled in the Classic since you are using lithium ? Is there a temp sensor plugged in ?
>> No sensor connected
What is your battery charge efficiency set to ? Should be near 100% or a few % below that.
>> Set at 94%. Can you tell me what effect this could have.
If you are able to get a video of the MNGP screen, that might help see what is going on.
>>I'll take a look at doing that. I'll need to research how I'd upload a video.
from Vic's post...
There is the ability to force the Classic, into doing a new Bulk, or force it into Float:
>> Thanks Vic, I had not discovered that menu feature, I'll try it next time.
from Larry's post...
Did you tweak the Classic voltage for the high part of the charge ?
>> Hi Larry, I calibrated battery voltage offset during a low amps 'Float' period.
Interesting point, though, as I think I have observed an increasing offset error as the battery reaches full charge. Makes me wonder if any correction or compensation is built into the displayed voltage.
I wonder if the moonscope display would have any benefit here. What data and scaling does it display. It is probably not fast enough though, I know how quickly those features can eat up all your resources!
Thanks again for the help. I'm going to see about putting my scope across the battery to see just what's going on with voltage. If I'm successful I'll post an update.
Tim
If you are using lithium, the charging efficiency is closer to 100%. Like, 98% I hear a lot.
If your charge is finishing at around 90%, I was just thinking that raising the efficiency a bit would better reflect the state of charge of your batteries.
Not sure but maybe ? That is, if you were expecting them to be higher SOC ?
boB
I just had the same thing happen with my batteries and Classic.
The Classic went to Float before it reached it's absorb setpoint.
I need to double check some things like the tweaks but I don't have temp compensation enabled or ending amps set.
I had Absorb setpoint set for 55.2v .
But I have the pv input tweaked up quite a bit at 4v . I wonder if that has something to do with it.
I am going to reconfigure the pv again to be 3 series instead of 2 series .
You can see in the graph below where I forced ti back to absorb and after a bit it dropped back to float.
My batteries were only at around 85%
Larry
update - reconfigured pv so input voltage is higher- started forced absorb cycle again and now it has been staying on bulk - will see how that goes when it gets closer to absorb setpoint.
Quote from: ClassicCrazy on October 15, 2022, 02:30:46 PM
I just had the same thing happen with my batteries and Classic.
The Classic went to Float before it reached it's absorb setpoint.
I need to double check some things like the tweaks but I don't have temp compensation enabled or ending amps set.
I had Absorb setpoint set for 55.2v .
But I have the pv input tweaked up quite a bit at 4v . I wonder if that has something to do with it.
I am going to reconfigure the pv again to be 3 series instead of 2 series .
You can see in the graph below where I forced ti back to absorb and after a bit it dropped back to float.
My batteries were only at around 85%
Larry
update - reconfigured pv so input voltage is higher- started forced absorb cycle again and now it has been staying on bulk - will see how that goes when it gets closer to absorb setpoint.
I'm sure that it did reach the Absorb voltage but that data log did not catch it since it can't log all the time... It must have ?
Could part of it be the battery temperature compensation reducing the voltage ?
How long is the Absorb time set for ? Have you checked the target voltage in the CHARGE menu ?
Bob
after I reconfigured my PV , the only thing I changed was to take the Tweak voltage adjust for PV back to zero.
I forced it into Bulk using Local Status app and it started charging and kept charging .
I had 15 minutes on the absorb initially so should have been enough time to spot the absorb voltage I would think.
It never did get to the absorb setting after I changed the PV to 3s but got the batteries 99% to 100% full. I had some loads on it too and am loosing light on the pv now .
So tomorrow if the sun is out again I will be around and can watch it better.
The temperature compensated target voltage is the same as the setpoint when I just checked.
It was cold in the room so if anything I would have thought the temp compensation would have raised the target voltage instead of lowering it.
I will look again tomorrow and see if target voltage changes.
Larry
I actually had the Classic Target voltage reporting to my graph and was able to go back and look at it. It never changed it's target voltage until it went to float and then used the float setpoint as target voltage.
WELL. I guess we will just have to see about that !
Will check it out here, too in the next day or two.
Turns out that I am actually working on some Classic software things.
boB
I'd suspect that battery voltage carries some noise, spikes, etc, that result from battery BMS actions. I'm rigging up an analog isolated amp which should allow me to get a closer look. I'm a little paranoid about putting scope on a floating circuit.
I agree, the classic is likely doing its job and acting on voltages that don't necessarily show up in digital sampling for displays or data logging.
In any case, I should have some data in the next few days.
Thanks, Tim
One thing I have noticed with Classics is they can be in a state 0.4 or even 0.5 volts lower than target. If memory serves the classic MUST hit target on its instantaneous battery voltage NOT averaged voltage then it is allowed to fall 0.4 volts and stay in that state.
Now back int he day with only lead acid that worked as the absorb timer played out the current dropped and the voltage rose to target allowing a nice full charge. But with Lithium its such a small voltage window this can be temperamental. It seems to be the harder it is pushing to get to target the more it drops below target.
I wonder if Larry can log instantaneous voltage instead of averaged?
I know this is one of the things boB is currently working on, he is diving deep into the classic now to make a few code updates, Mostly to better support lithium. For example we will be able to ignore temperature compensation now but still use the BTS to do cold stop charge on Lithiums
Interesting stuff.
I put a scope on battery voltage during Bulk MPPT phase. Voltage does in fact jump around in the neighborhood of 1 V P-P. The dither of noise tends to look pretty random with a rough frequency 2 - 4 cycles per second. Its really hard to characterize the noise since its too fast for a meter to report accurately and too random for the scope to trigger on. The one thing that's pretty visible is the magnitude of about 1 volt. Also noticed that a few times a minute the voltage will drop about .5 to .75 volt then return very quickly. Maybe this is an MPPT artifact?
During Float, battery voltage is pretty quiet, as expected.
One thing I want to study a little closer is whether using WBjr current limit feature contributes to noise. I would suspect that a current regulator function such as this would definitely add a little cyclical noise.
All this seems to correspond to Absorb starting about .5 V early.
I hope this info is helpful, You guys do a commendable job of 'continuous improvement' on your product.
Related question...
I had planned to use an Analog Devices isolated (1500 volt) dc amp from my stash to scope the voltage from the battery. I tested the amp in the house, it worked fine. It never worked again once I connected to battery. This is the third time I have destroyed some sort of isolation device when I connect it to the battery. Seems to be the initial hookup that is the killer, static charge?
I have a Midnite MNSPD-115 across the PV at the combiner and a MNSPD-300AC across the inverter AC output. I have nothing across the battery buss. At first glance that would not seem necessary, but if the CC and the inverter are both transformer isolated devices, I wonder if my battery circuit is still susceptible to static?
BTW, I lugged the old line voltage isolating transformer to the barn to achieve an isolated scope measurement.
Thanks, Tim
Quote from: FNG on October 18, 2022, 08:04:01 AM
One thing I have noticed with Classics is they can be in a state 0.4 or even 0.5 volts lower than target. If memory serves the classic MUST hit target on its instantaneous battery voltage NOT averaged voltage then it is allowed to fall 0.4 volts and stay in that state.
Now back int he day with only lead acid that worked as the absorb timer played out the current dropped and the voltage rose to target allowing a nice full charge. But with Lithium its such a small voltage window this can be temperamental. It seems to be the harder it is pushing to get to target the more it drops below target.
I wonder if Larry can log instantaneous voltage instead of averaged?
I know this is one of the things boB is currently working on, he is diving deep into the classic now to make a few code updates, Mostly to better support lithium. For example we will be able to ignore temperature compensation now but still use the BTS to do cold stop charge on Lithiums
Ryan - is there a way to log that on the MNGP ?
If it is a modbus number that isn't in the program code I use , I would have to redo some of the magic monitoring things to make it work and not sure i can do that.
But yes today I again noticed that the Absorb was lower than the setpoint.
I increased the setpoint from 55.2 to 56 when it was already in absorb and you can see the resulting voltage increase in the attached screenshot.
I am not sure how much more absorb time was left when I changed it .
Larry
Follow up...
I disabled the WBjr current limit by raising the value over 100A.
The voltage dither reduced significantly.
Put CL back to 30A and the wiggle returned.
The current controller probably need to be a little more sluggish to avoid hunting.
Hope this helps.
Thanks, Tim
Quote from: SunTim on October 19, 2022, 10:49:27 AM
Follow up...
I disabled the WBjr current limit by raising the value over 100A.
The voltage dither reduced significantly.
Put CL back to 30A and the wiggle returned.
The current controller probably need to be a little more sluggish to avoid hunting.
Hope this helps.
Thanks, Tim
Yes !
I believe that the default WB Jr current limit is set to some insanely high limit to do just what you are saying.
boB