LI battery tuning

sodamo
Forum Emperor
Posts: 1140
Joined: Mon Apr 03, 2006 1:55 am
My RE system: Totally off grid - Hamakua side, Big Island, Hi

36 Trina 280 & 16 Phono 250 in 48 volt array (fixed) 14080w
Radians GG8048/GS4048 w GSLCs
4 FM 80 charge controllers
Hub 10.3
FNDC w/3 shunts
Mate 3s
OpticsRE - MMKL - Ninole Hi
12 SimpliPhi 3.8
14Kw Kohler 14RESA Propane

Charging Only
2FX3048T
Hub
FNDC. - shared shunt feeds from Radian FNDC
Mate 3s
10Kw MEP 803a Military gen

Honda EU7000i

Davis VantagePro2 Wx Station
On line at: https://www.wunderground.com/dashboard/pws/KHININOL2
Location: Ninole, Hi

Re: LI battery tuning

Post by sodamo »

Will be interesting to set the settings you decide on Raysun.

So with this Warranty language, seems lessor voltage might be in order.

Hopeto post some interesting observations from my data dump later today.
David
Please visit http://vacation.ninolehawaii.com
raysun
Forum Emperor
Posts: 4581
Joined: Tue Jul 26, 2016 5:57 am
My RE system: Flexpower Two: (2) FXR3048A, (2) FM80, MATE3s, FlexNetDC
SimpliPhi 48-3.8 (6 @ 48v)
Outback IBR3 battery enclosure
REC Alpha 440W panels - 2 arrays: each of 4 strings of 2 in series
Honda EU7000is gas fuel generator

Re: LI battery tuning

Post by raysun »

I think I'm going with the 54.4 @ 2H
JRHill
Forum Emperor
Posts: 1519
Joined: Thu Feb 17, 2011 11:12 am
My RE system: VFXR3648, FM80, 3k panels, Mate3s, FNDC, 4ea 3.8 PHIs. EU7000is AGS.
Location: PNW
Contact:

Re: LI battery tuning

Post by JRHill »

I'd like repost this from another thread:

viewtopic.php?p=78210#p78210

In particular:

https://www.powerstream.com/lithium-pho ... oltage.htm

For those that might have missed it.

Thanks Raysun.
sodamo
Forum Emperor
Posts: 1140
Joined: Mon Apr 03, 2006 1:55 am
My RE system: Totally off grid - Hamakua side, Big Island, Hi

36 Trina 280 & 16 Phono 250 in 48 volt array (fixed) 14080w
Radians GG8048/GS4048 w GSLCs
4 FM 80 charge controllers
Hub 10.3
FNDC w/3 shunts
Mate 3s
OpticsRE - MMKL - Ninole Hi
12 SimpliPhi 3.8
14Kw Kohler 14RESA Propane

Charging Only
2FX3048T
Hub
FNDC. - shared shunt feeds from Radian FNDC
Mate 3s
10Kw MEP 803a Military gen

Honda EU7000i

Davis VantagePro2 Wx Station
On line at: https://www.wunderground.com/dashboard/pws/KHININOL2
Location: Ninole, Hi

Re: LI battery tuning

Post by sodamo »

Continuing my data dive

Observations:
The OpticsRE periods are average values.
While my Max voltage is 54.6, it is seems to be a soft hit like Raysun suggests. 112 occurrences in 15 min report vs 2 in hour report, then it falls. Didn’t find an obvious way to adjust for load (From Battery).
Average (mean) voltage over whole year 53.211 - nice I think
Both Median (equal number data points above and below) and Mode (most frequently occurring value) are 53.1 and account for 55.8% of hourly, 62% of 15 minute data being above.
Of course not a significant amount of low voltage by choice, generator corrected 😁

Threw in look at Jim’s Warranty point
So, not sure what else I can get out of this.
Being mindful of Jim’s 54.4 Warranty point and trying to absorb Raysun’s observations I would like to adjust settings accordingly. Hitting Parameters a concern.
Any other observations or suggestions what to explore welcome.
Attachments
F7CFEE2C-8586-4C0B-9D4D-77BFCFEA1D98.jpeg
David
Please visit http://vacation.ninolehawaii.com
raysun
Forum Emperor
Posts: 4581
Joined: Tue Jul 26, 2016 5:57 am
My RE system: Flexpower Two: (2) FXR3048A, (2) FM80, MATE3s, FlexNetDC
SimpliPhi 48-3.8 (6 @ 48v)
Outback IBR3 battery enclosure
REC Alpha 440W panels - 2 arrays: each of 4 strings of 2 in series
Honda EU7000is gas fuel generator

Re: LI battery tuning

Post by raysun »

By my reconing, the battery has spent most all its life in the mid to upper part of the discharge curve.

That 3.3VPC range is like driving across the Great Plains - you go a long way without knowing exactly where you're at. Could be Nebraska, could be Iowa, could be Minnesota.

Im continuing to think we need better V/SoC tables. Until we have them, and even when we do, I sill find "coulomb counting" easier to consume.
sodamo
Forum Emperor
Posts: 1140
Joined: Mon Apr 03, 2006 1:55 am
My RE system: Totally off grid - Hamakua side, Big Island, Hi

36 Trina 280 & 16 Phono 250 in 48 volt array (fixed) 14080w
Radians GG8048/GS4048 w GSLCs
4 FM 80 charge controllers
Hub 10.3
FNDC w/3 shunts
Mate 3s
OpticsRE - MMKL - Ninole Hi
12 SimpliPhi 3.8
14Kw Kohler 14RESA Propane

Charging Only
2FX3048T
Hub
FNDC. - shared shunt feeds from Radian FNDC
Mate 3s
10Kw MEP 803a Military gen

Honda EU7000i

Davis VantagePro2 Wx Station
On line at: https://www.wunderground.com/dashboard/pws/KHININOL2
Location: Ninole, Hi

Re: LI battery tuning

Post by sodamo »

Raysun, I, for one need schooling on coulomb counting 101 as I think I under what are trying to find out, just not a clue as to how or where, methodology, etc. So if you don’t mind and assuming I’m the only one so lacking I can PM you my email if you choose not to post.
David
Please visit http://vacation.ninolehawaii.com
provo
Forum Czar
Posts: 720
Joined: Thu Jul 27, 2006 4:34 pm
My RE system: Sixteen Evergreen EC-120
(4 strings, total 1920W)
Eight Rolls S-550 (2 strings, total ~800Ah @ 24V)
One FM80
One VFXR3524A
Hub 10
Mate3s
FNDC and Trimetric
Honda EU3000is generator
Location: Sierra foothills

Re: LI battery tuning

Post by provo »

sodamo wrote: Sat Dec 05, 2020 2:37 pm
So if you don’t mind and assuming I’m the only one so lacking I can PM you my email if you choose not to post.
You're not the only one so lacking :grin: ..... I would love to see a post from raysun on coulomb counting, if you're willing, raysun!
JRHill
Forum Emperor
Posts: 1519
Joined: Thu Feb 17, 2011 11:12 am
My RE system: VFXR3648, FM80, 3k panels, Mate3s, FNDC, 4ea 3.8 PHIs. EU7000is AGS.
Location: PNW
Contact:

Re: LI battery tuning

Post by JRHill »

sodamo wrote: Sat Dec 05, 2020 2:37 pm Raysun, I, for one need schooling on coulomb counting 101 as I think I under what are trying to find out, just not a clue as to how or where, methodology, etc. So if you don’t mind and assuming I’m the only one so lacking I can PM you my email if you choose not to post.
Yup, I agree. If it was 1 watt in and one watt out it would be perfection. With the PHIs, 1 watt in and .98 out in would be perfection. But we don't live in that perfect world. I'm still trying to figure out efficiency of the new batteries. They are much better than the worn out Trojans (and me thinks much better than when they were new).

But efficiency is hard to determine. Just ask the dear wife.
raysun
Forum Emperor
Posts: 4581
Joined: Tue Jul 26, 2016 5:57 am
My RE system: Flexpower Two: (2) FXR3048A, (2) FM80, MATE3s, FlexNetDC
SimpliPhi 48-3.8 (6 @ 48v)
Outback IBR3 battery enclosure
REC Alpha 440W panels - 2 arrays: each of 4 strings of 2 in series
Honda EU7000is gas fuel generator

Re: LI battery tuning

Post by raysun »

20201205_124634.jpg
Absorb completed in 1:28 today at 54.4V.

That works.
JRHill
Forum Emperor
Posts: 1519
Joined: Thu Feb 17, 2011 11:12 am
My RE system: VFXR3648, FM80, 3k panels, Mate3s, FNDC, 4ea 3.8 PHIs. EU7000is AGS.
Location: PNW
Contact:

Re: LI battery tuning

Post by JRHill »

I'm on a battery rhythm not sunlight right now. So what the heck, I know a generator run already carries me for a day+. So I set the absorb back to 54.0 and I'll just let the system it do its thing with AGS at 51.2. It might be less than once per day so less cycle counts below 54.4. But if I get a partial charge with diffused/cloud edge solar is that another cycle count even though it didn't get to cut off voltage? Roll the dice....

I don't want to be anal but this is what happens when I try to out guess fuzzy warranty "recommendations."
raysun
Forum Emperor
Posts: 4581
Joined: Tue Jul 26, 2016 5:57 am
My RE system: Flexpower Two: (2) FXR3048A, (2) FM80, MATE3s, FlexNetDC
SimpliPhi 48-3.8 (6 @ 48v)
Outback IBR3 battery enclosure
REC Alpha 440W panels - 2 arrays: each of 4 strings of 2 in series
Honda EU7000is gas fuel generator

Re: LI battery tuning

Post by raysun »

JRHill wrote: Sat Dec 05, 2020 7:06 pm I'm on a battery rhythm not sunlight right now. So what the heck, I know a generator run already carries me for a day+. So I set the absorb back to 54.0 and I'll just let the system it do its thing with AGS at 51.2. It might be less than once per day so less cycle counts below 54.4. But if I get a partial charge with diffused/cloud edge solar is that another cycle count even though it didn't get to cut off voltage? Roll the dice....

I don't want to be anal but this is what happens when I try to out guess fuzzy warranty "recommendations."
I'd think a cycle is when the battery is discharged to some degree, and then fully charged. The PSoC days, frankly, put less wear and tear on a lithium battery.
JRHill
Forum Emperor
Posts: 1519
Joined: Thu Feb 17, 2011 11:12 am
My RE system: VFXR3648, FM80, 3k panels, Mate3s, FNDC, 4ea 3.8 PHIs. EU7000is AGS.
Location: PNW
Contact:

Re: LI battery tuning

Post by JRHill »

Well, I'm done experimenting for the time being. Its time to get used to things as set and gather some data and averages. I still have some curiosities about the FNDC relay behavior if the low voltage threshold is reached. It shouldn't occur unless I have a failed AGS start or run. A programmable audible alarm would be cool. The email from OpticsRE doesn't work with our weak cell service. I'm sure there is an easier solution that I'll come across.
sodamo
Forum Emperor
Posts: 1140
Joined: Mon Apr 03, 2006 1:55 am
My RE system: Totally off grid - Hamakua side, Big Island, Hi

36 Trina 280 & 16 Phono 250 in 48 volt array (fixed) 14080w
Radians GG8048/GS4048 w GSLCs
4 FM 80 charge controllers
Hub 10.3
FNDC w/3 shunts
Mate 3s
OpticsRE - MMKL - Ninole Hi
12 SimpliPhi 3.8
14Kw Kohler 14RESA Propane

Charging Only
2FX3048T
Hub
FNDC. - shared shunt feeds from Radian FNDC
Mate 3s
10Kw MEP 803a Military gen

Honda EU7000i

Davis VantagePro2 Wx Station
On line at: https://www.wunderground.com/dashboard/pws/KHININOL2
Location: Ninole, Hi

Re: LI battery tuning

Post by sodamo »

Jim
What about a simple alarm triggered by either Aux or FNDC? I’m thinking maybe use Aux to trigger just ahead of your AGS setting.
David
Please visit http://vacation.ninolehawaii.com
JRHill
Forum Emperor
Posts: 1519
Joined: Thu Feb 17, 2011 11:12 am
My RE system: VFXR3648, FM80, 3k panels, Mate3s, FNDC, 4ea 3.8 PHIs. EU7000is AGS.
Location: PNW
Contact:

Re: LI battery tuning

Post by JRHill »

sodamo wrote: Sun Dec 06, 2020 12:46 pm Jim
What about a simple alarm triggered by either Aux or FNDC? I’m thinking maybe use Aux to trigger just ahead of your AGS setting.
Hmmm. I have a really annoying AC buzzer mounted to the main floor in the crawl space. Its been years but when it triggered* it went through the whole house. The FNDC is tripping the VFX cut out for the PHIs but you just answered the question I had posed. Why didn't I think of that? Don't answer, please.

OK, its easy. I have the relay, the AC wart and the buzzer. (update, not so easy)

*BTW, early on when I went from Abs to Float I used this buzzer to let me go out and use power freely. I used it for a while and it worked great. But the buzzer was profoundly annoying. Like watching the news or what I didn't get done that day. So I cut the wires and taped them off.

Nope, this is more complicated than expected as the AGS is triggered by the inverter and the inverter shuts down from the FNDC aux trigger from the low/high battery voltage and the FM is set to remote to control the AGS at the inverter on/off. I think I have used up my programmable inputs.

A few more auxiliary triggers and a few more shunts and I think I could like really, really, like get things in order. Seriously, when the OB folks did all this stuff, integrating the various components, etc, etc, they did a marvelous work.
sodamo
Forum Emperor
Posts: 1140
Joined: Mon Apr 03, 2006 1:55 am
My RE system: Totally off grid - Hamakua side, Big Island, Hi

36 Trina 280 & 16 Phono 250 in 48 volt array (fixed) 14080w
Radians GG8048/GS4048 w GSLCs
4 FM 80 charge controllers
Hub 10.3
FNDC w/3 shunts
Mate 3s
OpticsRE - MMKL - Ninole Hi
12 SimpliPhi 3.8
14Kw Kohler 14RESA Propane

Charging Only
2FX3048T
Hub
FNDC. - shared shunt feeds from Radian FNDC
Mate 3s
10Kw MEP 803a Military gen

Honda EU7000i

Davis VantagePro2 Wx Station
On line at: https://www.wunderground.com/dashboard/pws/KHININOL2
Location: Ninole, Hi

Re: LI battery tuning

Post by sodamo »

You have AGS triggered by both the VFX inverter and FM?
David
Please visit http://vacation.ninolehawaii.com
raysun
Forum Emperor
Posts: 4581
Joined: Tue Jul 26, 2016 5:57 am
My RE system: Flexpower Two: (2) FXR3048A, (2) FM80, MATE3s, FlexNetDC
SimpliPhi 48-3.8 (6 @ 48v)
Outback IBR3 battery enclosure
REC Alpha 440W panels - 2 arrays: each of 4 strings of 2 in series
Honda EU7000is gas fuel generator

Re: LI battery tuning

Post by raysun »

20201206_121126.jpg
Here's a scenario I absolutely dreaded would occur with the "long Absorb" charging profile, and it took only three days to realize it.

The image is showing 0 PV harvest because the FM80 charge controllers are Silent as they transition from Absorb to Float - in this case 8 minutes into the Absorb phase.

I'm using the FNDC for charge termination, as typical for advanced battery charging using the Outback infrastructure, and specified in the SimpliPhi integration guide for Outback.

I'm using 54.4V Absorb Voltage and 2H Absorb Time.

On the FNDC, Charge Termination is enabled and the Battery Charged parameters are:
Charged Voltage: 54V
Charged Time: 5Min
Charged Return Amps: 9A

The Charge Termination parameters are typically met at the end of the Absorb phase, at about 1:30 elapsed time.

In this case, they were met during a different set of circumstances. Just at the point the charge cycle entered Absorb, clouds started covering the sun. Charging current dropped, but due to light inverter loading, the battery voltage stayed above 54V. Charge current dropped below 9A for 5 minutes, and voila! Charge Complete! at 81.5% SoC. The FNDC resets the SoC to 100% when it thinks charging has completed. It does not appear to do any sort of 'sanity check', it simply notes the Charge Complete flag and resets SoC. In this case, SoC is now instantly off by 18.5%. No bueno.

Fortunately, the Victron BMV700 is unaffected by the FNDC shenanigans, so I have a read on SoC. I unplugged the FNDC, and restarted the charge cycle.

Glad I have the BMV700 in the circuit. Reminds me of the old joke: He bought a Toyota so he'd have something to drive while his Ferrari is in the shop.

Looking for that FNDC-Li now more than ever. I have another feature to add.

*EDIT* Part is this issue was the FNDC voltmeter reading a couple of tenths of a volt high on average. After this event, I calibrated it to be in agreement with the charge controller voltage readings. This should give a more faithful evaluation of Charged Voltage. It doesn't, however, change the fact lithium batteries have different voltage behaviors than lead acid, and the charge termination function is really "threading the needle", and a very fine gauge needle at that, when evaluating Charged Parameters Met.
Last edited by raysun on Mon Dec 07, 2020 4:29 am, edited 1 time in total.
JRHill
Forum Emperor
Posts: 1519
Joined: Thu Feb 17, 2011 11:12 am
My RE system: VFXR3648, FM80, 3k panels, Mate3s, FNDC, 4ea 3.8 PHIs. EU7000is AGS.
Location: PNW
Contact:

Re: LI battery tuning

Post by JRHill »

sodamo wrote: Sun Dec 06, 2020 2:40 pm You have AGS triggered by both the VFX inverter and FM?
Cant find my notes but I know the FM had to be set to remote. The actual AGS trigger is the VFX Aux.

Update: I kept thinking about this last night and was determined to map it out. At one time I had used the FM80 to toggle the AGS. That is why its AUX was still in remote but it is not being used. During that time using the FM80's Aux I had a few weird AGS events and several folks urged me to go with the VFX Aux which I did.

But last night, with ironic timing, the generator had begun warming up from an AGS signal from the inverter and the AC throughout the house flickered once. That momentary interruption ended the generator warmup/run. Hmmm. That's not good because it creates an error in the Mate3 which has to be cleared. I cleared the error and fired it manually as it was getting late. I did check the battery softkey and the lowest voltage (51.3) was no where near the FNDC's highest setpoint for low voltage relay action of 48vdc. But I had the delay set at zero so I moved it to 1 minute. Anyway, I now have on my "To Watch List" for any further instances of this. It could be a good case for going back to the FM80's Aux for AGS control.

The above to me this still falls within the LI Tuning subject as getting the system tuned to work with the PHIs is quite interesting.
JRHill
Forum Emperor
Posts: 1519
Joined: Thu Feb 17, 2011 11:12 am
My RE system: VFXR3648, FM80, 3k panels, Mate3s, FNDC, 4ea 3.8 PHIs. EU7000is AGS.
Location: PNW
Contact:

Re: LI battery tuning

Post by JRHill »

Yup, it is the FNDC: "Batt Low Warning FNDC-2 Active Acked 13:36 12/07/20" Corrected from above message the low set point is 49.5 on the FNDC. It's the inverter low set point at 48. My mistake.

The AGS is acting up. The AGS works great in manual and even does the full absorb. But it is not reliable in Auto.
sodamo
Forum Emperor
Posts: 1140
Joined: Mon Apr 03, 2006 1:55 am
My RE system: Totally off grid - Hamakua side, Big Island, Hi

36 Trina 280 & 16 Phono 250 in 48 volt array (fixed) 14080w
Radians GG8048/GS4048 w GSLCs
4 FM 80 charge controllers
Hub 10.3
FNDC w/3 shunts
Mate 3s
OpticsRE - MMKL - Ninole Hi
12 SimpliPhi 3.8
14Kw Kohler 14RESA Propane

Charging Only
2FX3048T
Hub
FNDC. - shared shunt feeds from Radian FNDC
Mate 3s
10Kw MEP 803a Military gen

Honda EU7000i

Davis VantagePro2 Wx Station
On line at: https://www.wunderground.com/dashboard/pws/KHININOL2
Location: Ninole, Hi

Re: LI battery tuning

Post by sodamo »

Well, Jim
The good news is you have something interesting to keep your mind active. Me, I’m my own worst enemy. Wife often tells my I make work for myself. 😁

So sounds like the FM aux may be available for your alarm.
David
Please visit http://vacation.ninolehawaii.com
raysun
Forum Emperor
Posts: 4581
Joined: Tue Jul 26, 2016 5:57 am
My RE system: Flexpower Two: (2) FXR3048A, (2) FM80, MATE3s, FlexNetDC
SimpliPhi 48-3.8 (6 @ 48v)
Outback IBR3 battery enclosure
REC Alpha 440W panels - 2 arrays: each of 4 strings of 2 in series
Honda EU7000is gas fuel generator

Re: LI battery tuning

Post by raysun »

sodamo wrote: Mon Dec 07, 2020 3:30 pm Well, Jim
The good news is you have something interesting to keep your mind active. Me, I’m my own worst enemy. Wife often tells my I make work for myself. 😁

So sounds like the FM aux may be available for your alarm.
My wife often tells me I fake making work for myself. Sheesh! I get no respect.
gtarolli
Forum Guru
Posts: 419
Joined: Thu Dec 21, 2017 10:39 am
My RE system: Outback: off-grid
- 1 GS8048A, Mate3, FlexNetDC, Hub10.3
- 2 FLEXmax 80
- 1 FLEXmax 60
- 30 (10x3) 300w panels (9000w total)
- 6 SimpliPhi 3.4 kWh LI batteries (400ah , 20kWh)
- 1 Honda EM5000 generator + one EU3000
Location: Wainiha, Kauai, HI

Re: LI battery tuning

Post by gtarolli »

Regarding the issues below, I disable Auto Charge Termination Control on my Mate3 so that 100% SOC doesn't stop the FM chargers. I have two issues with FNDC and Charge Parameters Met. One is what Raysun experienced, where the system doesn't account for what the PV could produce if it was sunny. So the return amps could be low just due to low sun and not SOC. Also you have to have charged voltage very close to Absorb if you use low settings. At 56v absorb you will not get within .4v until SOC is very high, but at 54.4v Absorb you can get there at under 80% SOC with some sun. This can cause premature charge params met. And yes, you have to make sure all your voltages are calibrated, or at least relative to each other at a minimum.

On the other side, from my observations the return amps monitor doesn't seem to take into your loads, that is how much is being put into the batteries because the inverter is drawing from them. So if you have a constant 25 amp load, and return amps is < 25 amps, you will never meet the condition. I push my charged voltage very close to Absorb and keep return amps high. Sometimes I hit the params met condition early, e.g. mid 90's in SOC, sometimes I hit it a few times a day. I disable charge termination control, so if I hit it early, it doesn't matter. You can watch the actual net PV production and it will decrease as the batteries approach 100% SOC. I do like to hit 100% SOC so that SOC doesn't drift over a few days.
raysun wrote: Sun Dec 06, 2020 2:48 pm 20201206_121126.jpg

Here's a scenario I absolutely dreaded would occur with the "long Absorb" charging profile, and it took only three days to realize it.

The image is showing 0 PV harvest because the FM80 charge controllers are Silent as they transition from Absorb to Float - in this case 8 minutes into the Absorb phase.

I'm using the FNDC for charge termination, as typical for advanced battery charging using the Outback infrastructure, and specified in the SimpliPhi integration guide for Outback.

I'm using 54.4V Absorb Voltage and 2H Absorb Time.

On the FNDC, Charge Termination is enabled and the Battery Charged parameters are:
Charged Voltage: 54V
Charged Time: 5Min
Charged Return Amps: 9A

The Charge Termination parameters are typically met at the end of the Absorb phase, at about 1:30 elapsed time.

In this case, they were met during a different set of circumstances. Just at the point the charge cycle entered Absorb, clouds started covering the sun. Charging current dropped, but due to light inverter loading, the battery voltage stayed above 54V. Charge current dropped below 9A for 5 minutes, and voila! Charge Complete! at 81.5% SoC. The FNDC resets the SoC to 100% when it thinks charging has completed. It does not appear to do any sort of 'sanity check', it simply notes the Charge Complete flag and resets SoC. In this case, SoC is now instantly off by 18.5%. No bueno.

Fortunately, the Victron BMV700 is unaffected by the FNDC shenanigans, so I have a read on SoC. I unplugged the FNDC, and restarted the charge cycle.

Glad I have the BMV700 in the circuit. Reminds me of the old joke: He bought a Toyota so he'd have something to drive while his Ferrari is in the shop.

Looking for that FNDC-Li now more than ever. I have another feature to add.

*EDIT* Part is this issue was the FNDC voltmeter reading a couple of tenths of a volt high on average. After this event, I calibrated it to be in agreement with the charge controller voltage readings. This should give a more faithful evaluation of Charged Voltage. It doesn't, however, change the fact lithium batteries have different voltage behaviors than lead acid, and the charge termination function is really "threading the needle", and a very fine gauge needle at that, when evaluating Charged Parameters Met.
raysun
Forum Emperor
Posts: 4581
Joined: Tue Jul 26, 2016 5:57 am
My RE system: Flexpower Two: (2) FXR3048A, (2) FM80, MATE3s, FlexNetDC
SimpliPhi 48-3.8 (6 @ 48v)
Outback IBR3 battery enclosure
REC Alpha 440W panels - 2 arrays: each of 4 strings of 2 in series
Honda EU7000is gas fuel generator

Re: LI battery tuning

Post by raysun »

gtarolli wrote: Thu Dec 10, 2020 9:38 am Regarding the issues below, I disable Auto Charge Termination Control on my Mate3 so that 100% SOC doesn't stop the FM chargers. I have two issues with FNDC and Charge Parameters Met. One is what Raysun experienced, where the system doesn't account for what the PV could produce if it was sunny. So the return amps could be low just due to low sun and not SOC. Also you have to have charged voltage very close to Absorb if you use low settings. At 56v absorb you will not get within .4v until SOC is very high, but at 54.4v Absorb you can get there at under 80% SOC with some sun. This can cause premature charge params met. And yes, you have to make sure all your voltages are calibrated, or at least relative to each other at a minimum.

On the other side, from my observations the return amps monitor doesn't seem to take into your loads, that is how much is being put into the batteries because the inverter is drawing from them. So if you have a constant 25 amp load, and return amps is < 25 amps, you will never meet the condition. I push my charged voltage very close to Absorb and keep return amps high. Sometimes I hit the params met condition early, e.g. mid 90's in SOC, sometimes I hit it a few times a day. I disable charge termination control, so if I hit it early, it doesn't matter. You can watch the actual net PV production and it will decrease as the batteries approach 100% SOC. I do like to hit 100% SOC so that SOC doesn't drift over a few days.
raysun wrote: Sun Dec 06, 2020 2:48 pm 20201206_121126.jpg

Here's a scenario I absolutely dreaded would occur with the "long Absorb" charging profile, and it took only three days to realize it.

The image is showing 0 PV harvest because the FM80 charge controllers are Silent as they transition from Absorb to Float - in this case 8 minutes into the Absorb phase.

I'm using the FNDC for charge termination, as typical for advanced battery charging using the Outback infrastructure, and specified in the SimpliPhi integration guide for Outback.

I'm using 54.4V Absorb Voltage and 2H Absorb Time.

On the FNDC, Charge Termination is enabled and the Battery Charged parameters are:
Charged Voltage: 54V
Charged Time: 5Min
Charged Return Amps: 9A

The Charge Termination parameters are typically met at the end of the Absorb phase, at about 1:30 elapsed time.

In this case, they were met during a different set of circumstances. Just at the point the charge cycle entered Absorb, clouds started covering the sun. Charging current dropped, but due to light inverter loading, the battery voltage stayed above 54V. Charge current dropped below 9A for 5 minutes, and voila! Charge Complete! at 81.5% SoC. The FNDC resets the SoC to 100% when it thinks charging has completed. It does not appear to do any sort of 'sanity check', it simply notes the Charge Complete flag and resets SoC. In this case, SoC is now instantly off by 18.5%. No bueno.

Fortunately, the Victron BMV700 is unaffected by the FNDC shenanigans, so I have a read on SoC. I unplugged the FNDC, and restarted the charge cycle.

Glad I have the BMV700 in the circuit. Reminds me of the old joke: He bought a Toyota so he'd have something to drive while his Ferrari is in the shop.

Looking for that FNDC-Li now more than ever. I have another feature to add.

*EDIT* Part is this issue was the FNDC voltmeter reading a couple of tenths of a volt high on average. After this event, I calibrated it to be in agreement with the charge controller voltage readings. This should give a more faithful evaluation of Charged Voltage. It doesn't, however, change the fact lithium batteries have different voltage behaviors than lead acid, and the charge termination function is really "threading the needle", and a very fine gauge needle at that, when evaluating Charged Parameters Met.
On the other side of the coin, I rely on Charge Termination Control 100%. I had one "unique" event, and it really turned out to be operator error. Rather, calibration error. Now that the new FNDC is "dialed in" the scenario i described has a very narrow window in which to occur. An unlikely event.

On the other hand, Charge Termination for the LiFePO4 battery is borderline mandatory to save wear and tear on the anode. My charge often (like nearly always) reaches 100% before Charged Parameters Met, and the chargers just keep on charging. About 3AH later, charge is terminated, and the top green LED eventually starts blinking.
gtarolli
Forum Guru
Posts: 419
Joined: Thu Dec 21, 2017 10:39 am
My RE system: Outback: off-grid
- 1 GS8048A, Mate3, FlexNetDC, Hub10.3
- 2 FLEXmax 80
- 1 FLEXmax 60
- 30 (10x3) 300w panels (9000w total)
- 6 SimpliPhi 3.4 kWh LI batteries (400ah , 20kWh)
- 1 Honda EM5000 generator + one EU3000
Location: Wainiha, Kauai, HI

Re: LI battery tuning

Post by gtarolli »

raysun wrote: Thu Dec 10, 2020 9:45 am On the other hand, Charge Termination for the LiFePO4 battery is borderline mandatory to save wear and tear on the anode. My charge often (like nearly always) reaches 100% before Charged Parameters Met, and the chargers just keep on charging. About 3AH later, charge is terminated, and the top green LED eventually starts blinking.
I guess I don't understand - how do you know you are at 100% before charge parameters are met? Does your PV production drop way down, or are you relying on the SOC reading? I have my efficiency set to 94%, maybe yours is too high? I set mine so that during the Absorb phase I see the Solar minus Load approach zero as SOC approaches 100%. And as the Net Battery Ah gets close to 0. If efficiency is set too high you will reach 100% SOC (FNDC reading) before you are really there - I trust reaching the asymptote in net battery charging as the true measure. At the start of Absorb I can see 5kw go into the batteries, at the end it drops way down to under 1kw telling me the batteries are full. The other experiment is to restart a bulk charge after Absorb ends, and you will see the voltage go very quickly back to Absorb if your batteries are full.

But maybe I am wearing out the anode? From last year I seem to have lost about 15% in battery storage (I lost none the first 2 years). I'm not sure if that is normal, or due to changing my Absorb from a fast and furious 56v for .1 hour, to a slow and steady 55v for an hour. I am going back to 56v for a week to see if it matters.
raysun
Forum Emperor
Posts: 4581
Joined: Tue Jul 26, 2016 5:57 am
My RE system: Flexpower Two: (2) FXR3048A, (2) FM80, MATE3s, FlexNetDC
SimpliPhi 48-3.8 (6 @ 48v)
Outback IBR3 battery enclosure
REC Alpha 440W panels - 2 arrays: each of 4 strings of 2 in series
Honda EU7000is gas fuel generator

Re: LI battery tuning

Post by raysun »

gtarolli wrote: Thu Dec 10, 2020 10:05 am
raysun wrote: Thu Dec 10, 2020 9:45 am On the other hand, Charge Termination for the LiFePO4 battery is borderline mandatory to save wear and tear on the anode. My charge often (like nearly always) reaches 100% before Charged Parameters Met, and the chargers just keep on charging. About 3AH later, charge is terminated, and the top green LED eventually starts blinking.
I guess I don't understand - how do you know you are at 100% before charge parameters are met? Does your PV production drop way down, or are you relying on the SOC reading? I have my efficiency set to 94%, maybe yours is too high? I set mine so that during the Absorb phase I see the Solar minus Load approach zero as SOC approaches 100%. And as the Net Battery Ah gets close to 0. If efficiency is set too high you will reach 100% SOC (FNDC reading) before you are really there - I trust reaching the asymptote in net battery charging as the true measure. At the start of Absorb I can see 5kw go into the batteries, at the end it drops way down to under 1kw telling me the batteries are full. The other experiment is to restart a bulk charge after Absorb ends, and you will see the voltage go very quickly back to Absorb if your batteries are full.

But maybe I am wearing out the anode? From last year I seem to have lost about 15% in battery storage (I lost none the first 2 years). I'm not sure if that is normal, or due to changing my Absorb from a fast and furious 56v for .1 hour, to a slow and steady 55v for an hour. I am going back to 56v for a week to see if it matters.
Sorry, I wasn't clear.

When, near the end of the charge cycle, FNDC reports 0AH Corrected Net Battery, and resets to 100% SoC. The battery/charger sense additional charging needed and continue doing so. When the Charged Parameters Met, SoC is again set to 100%, if Charge Termination Control is enabled, the Mate terminates charging by all OB chargers.
gtarolli
Forum Guru
Posts: 419
Joined: Thu Dec 21, 2017 10:39 am
My RE system: Outback: off-grid
- 1 GS8048A, Mate3, FlexNetDC, Hub10.3
- 2 FLEXmax 80
- 1 FLEXmax 60
- 30 (10x3) 300w panels (9000w total)
- 6 SimpliPhi 3.4 kWh LI batteries (400ah , 20kWh)
- 1 Honda EM5000 generator + one EU3000
Location: Wainiha, Kauai, HI

Re: LI battery tuning

Post by gtarolli »

What do you have the charge factor (efficiency) set to? That could cause a premature 100% SOC reading. I agree with everything you say, I just am wondering what is wrong with allowing the Absorb settings to "terminate" charging and switch to Float mode. And if you've tuned the charge factor so that SOC and charging (PV-loads) asymptotically level off at the same time.

Sometimes my charge parameters get met around 95% SOC, SOC jumps to 100%, and I just let charging continue , another few Ah usually ends the Absorb cycle for me and goes to Float mode (which is what charge termination would do right?) . Since charge params were met early (at 95%) I think its safe.
Post Reply