Specific gravity low for given voltage and SOC, where is the problem?

[Mainewoods]

Yesterday evening we noticed that our bank voltage was lower then it should have been for the stated WBJR SOC (data below).  I tested the SG of every cell and it corresponded with the voltage based off of Rolls specs - but not with the WBJR SOC.  Okay, so likely the WBJR needs some fine tuning.  I thought I would monitor it and check again at WBJR 100% SOC to confirm, here is where it really gets weird.  This afternoon I checked and it was at 88% or so, a few hours later it was at 100% (it was cloudy and we were charging around 30Amps - it should have taken twice that or more to get us to 100%) - so maybe the classics hit the float amperage and went to float earlier then the WBJR calculated - not strange.  What is strange is when I tested the SG at WBJR 100% SOC I got an average of 1.200 (rolls states 100% SOC at 1.255-1.275) BUT the battery voltage (with classics in OFF mode) was 2.15VPC - which is 100% charged according to Rolls specs - even with a 20amp load on the batteries.  This is weird, I am going to talk with Rolls but want to get a few more 100% SOC worth of data before starting the conversation. 

Note: When I received the batteries, two of the batteries had older production date stamps - by 2 months, both of those batteries had lower then 100% SOC via SG.  Rolls told me that they would come up and not to worry about it - although I was/am worrying about it.

Batteries are Rolls Surrette (L16) S-550, two strings of 8 each for 16 batteries. Totaling 856Ah at 48vdc.  They are one month old.  The bank has new 4/0 cables - overkill but I had the 4/0 laying around.

When tested yesterday evening:
SG of all cells (see exceptions below) was 1.187 - which amounts to ~56% SOC according to Rolls specs.
SG of two older batteries was 1.150 and 1.175 respectively.
Voltage was 48.5VDC (no load) - which amounts to ~56% SOC according to Rolls specs.
WBJR thought the bank was at 71% SOC
Last 100%SOC according to WBJR was 27 hrs prior to testing.
Bank was at 20 deg Celsius.

When tested today:
SG of random cells was 1.200 - which amounts to ~60% SOC according to Rolls specs.
Voltage was 51.5VDC (20amp load) - which amounts to 100% SOC according to Rolls specs.
WBJR thought the bank was at 100% SOC
Bank was at 25 deg Celsius.


Charging perimeters (which were vetted by Rolls):

Bulk 60.0vdc
Absorb 60.0vdc
When end amp based 10.7Amps (classic default use).
When time based 2.75Hrs (classic backup - assuming 130amps).  This is only setup if end-amps does not work well.
When time based 6Hrs (inverter/charger - assuming 60amps). This will be infrequent.
Float 54.00vdc
Equalize 63.0vdc

WBJR is set to
80% battery efficiency.
And is temp compensated via specs from Rolls.

WBJR classic is set to limit total charging +amps to 130amps - so that the batteries don't charge too fast.

I have low voltage shutdown set at 48vdc - I don't ever want the bank to drop below 50% - yesterdays 56% was the lowest it has ever been.


Any ideas what is going on?  Maybe I have two problems - that would make this easier to understand.

Maybe my SG meter is bad? - I only have one but it has seemed correct in the past.

[TomW]

Have you calibrated The Classic's idea of what voltage is on its battery input terminals with to agree with voltage at the battery?

"SG Meter" could be anything and some are simply junk from the get go. Maybe test it against distilled water? Should be 1.00... with distilled water.

Just a couple thoughts to start.

Tom

[Mainewoods]

Hi Tom, thanks for the quick reply.

Yep, I have calibrated the bank voltage at the classics - I wish my inverters had the same feature.

You nailed it, I should have thought of testing the meter with distilled water...  I just tested it, using store bought distilled water and then reverse osmosis water just to be sure.  It was low in both tests - by at least .025 - could be more, hard to say for sure because the scale does not go below 1.00.

Any idea which meter I should buy?  I would like to have a accurate one.

Thanks,
Travis

[Vic]

Hello Maine..,

First,  the WBjr SOC numbers are like any other battery monitoring device  --  just rough estimates.   When the WB is set up well,  the SOC indication at the transition to Float will often be accurate.

Using battery voltages as an indication of SOC  can be accurate,  or quite far off the mark.   It can be useful to measure the voltage of each battery,   as an indicator of which batteries are lagging.

Battery voltage readings need to be temperature compensated,  when trying to guess approximate SOC.   Battery voltage is a very good indication of SOC,   when temperature compensated,   and the battery/ies have been Rested (NO charge/NO discharge,  at all whatsoever)  for a number of hours.

Would suggest that you buy two or three glass Hydrometers.   Ones at the Auto Parts store are often acceptable.   Compare them,   make notes in your battery Logbook on each Hydro's differences.   Mark each with some identifier,   and reserve one or two for later use.

Would INCREASE Absorb time to about six hours for now,  until you get  the SGs up.

The Surrette battery banks in use here have needed about 0.9% of actual 20-hour Capacity as a Shunt EA value,  when moderately cycled.   When more deeply-cycled on a daily basis,   the needed Shunt EA value will need to be increased.   As these batteries have aged (each bank is now in its 12th year),  the needed Absorb voltage has been increased to attain rated SGs from Absorb,   and this has necessitated an increase in the Shunt EA value to about 1.5% of C.

Surrette Flooded batteries need --5 mV/Cell/C as temperature compensation.

Check the Limits page,  in the Charge menu.   Make certain that you have set the Limits wide enough for your Flooded batteries.   The Default Limits seem quite narrow (probably for Sealed,  Lead/Calcium batts).   

Did you do the Surrette-recommended Commissioning Charge and EQ when you first placed these batteries in service?

Here is an article from Surrette on measuring SGs:
http://support.rollsbattery.com/support/solutions/articles/4347-measuring-specific-gravity

Please note,  that is very,  very important to rinse your Hydrometer at least three times,  with Distilled water,  to help maintain accuracy,  at the end of each SG measuring session.

Enough for now,   Vic

[Mainewoods]

Hi Vic, Thanks for a lot of great information - again:)

I will pickup at least another Hydrometer - and rinse them better after use, I have been rinsing mine but not "3 times or more", maybe this is contributing to my -~0.025 error.

I am currently using endamps to control absorb, would you switch it to the 6hrs you recommend or adjust endamps down to say 5 amps, instead of the current 10.7 amps that I am using until my SOC comes up?

It sounds like you determined that your Surrette batteries are operating at 90% rated capacity based on your use?  We lack a long baseline/history but so far we normally discharge about 15% overnight and return to 100% the following day - even if cloudy.  We did have a stretch of snow days - it snowed four days straight and totaled 4.5' deep, we knew it was coming and conserved power after 4 days we were down to 63%SOC.  Given our normal 15% discharge would you consider this moderately cycled?  With more history I should be able to calculate our actual capacity relatively accurately - and will then adjust accordingly, as you did.  Thank you, this is good advice.

I set the charge menu limits to rolls specs when we installed the new bank as follows:
Charge Low: 52.8
Charge Hi: 63.0
Max Temp 125C
Does this sound right?

I did the commissioning charge but I did not see a note on doing an EQ - I must have missed that.  It seems that it would be a good idea to do one soon (especially after seeing the two low batteries), maybe after I get the SOC back up and have a new hydrometer on hand.

BTW Vic, 12 years - not bad:)

Thanks for your time,
Travis

 

[CDN-VT]

Use your DVM and do a Voltage drop on all connections of the DCV system , from CC to banks & within . do this on neg backwards & also POS forwards . WRITE IT DOWN !!

I do mine under load always .
Im amazed I can just move a nut one (hex/flat/1/8 of a turn) and zip !! 
At 48Vdc (60Vdc charge) then check out the VD under charge to the cells .
Under use/ loaded  & no charge check the VD in battery supply .

I do this every watering of the FLA's .


VT

Edit add . I use a cheap hydrometer
http://www.midnitesolar.com/productPhoto.php?product_ID=505&productCatName=Battery%20Accessories&productCat_ID=39&sortOrder=1&act=pc

For quick testing ,& ease of seeing , but for REAL testing , 1970 imperial eastman  with included temp & range , all glass & was certified , when the tools were made & we trusted .
That is MY stardard against the rest.

DVM's are Fluke  or Pico scopes that are My standard to trust

VT

[ClassicCrazy]

.  This afternoon I checked and it was at 88% or so, a few hours later it was at 100% (it was cloudy and we were charging around 30Amps - it should have taken twice that or more to get us to 100%) - so maybe the classics hit the float amperage and went to float earlier then the WBJR calculated - not strange. 
Batteries are Rolls Surrette (L16) S-550, two strings of 8 each for 16 batteries. Totaling 856Ah at 48vdc.  They are one month old.  The bank has new 4/0 cables - overkill but I had the 4/0 laying around.



Bulk 60.0vdc
Absorb 60.0vdc
When end amp based 10.7Amps (classic default use).
When time based 2.75Hrs (classic backup - assuming 130amps).  This is only setup if end-amps does not work well.
When time based 6Hrs (inverter/charger - assuming 60amps). This will be infrequent.
Float 54.00vdc
Equalize 63.0vdc

WBJR is set to
80% battery efficiency.
And is temp compensated via specs from Rolls.

WBJR classic is set to limit total charging +amps to 130amps - so that the batteries don't charge too fast.

Any ideas what is going on?  Maybe I have two problems - that would make this easier to understand.

You said they were at 88% and a few hours later at 100% .
But how long had they been in Absorb when you saw them at 88% ?
If the Absorb time of 2.75 hours is reached before ending amps it would go to Float and show 100% .
So it sounds like you don't have Absorb time set long enough - Vic seems to suggest the same thing.

I find looking at the history via MyMidnite graphs is a good way to look back and get an idea of what happened over the course of the day since you can drag the cursor along and see battery voltage, SOC, Whizbang system amps, charging state , etc change as you go through the time.

Larry

[dgd]

That 80% battery efficiency setting for WBJr looks a bit miserable and since reported SOC is looking lower than it should be AND these are new batteries, I would suggest a closr to 90% setting for efficiency.
That may close the reported SOC closer to the real SOC
Dgd

[Mainewoods]

Thanks all,

I bought a new hydrometer - and have a 3rd (different model) coming as Vic recommended.
(and washed them both very well this time)

I changed endamps to 8.6amps - 1% of capacity is the minimum limit recommended by surrette/rolls -
as per the rolls tech I spoke with.

I changed absorb time to 4Hrs - although it was partly cloudy today and endamps triggered absorb long before 4hrs.
2.75Hrs was based on surrite/rolls recommendations - I will review this before leaving it set at 4Hrs.

Larry, to answer your questions - They had not been in absorb for more then 2.75Hrs when the WBJR went to 100%, I have noticed that endamps is
almost always the trigger to go to float with my system.

I watched my local app while working on my PC as the bank approached its old endamp setting, it was about an
additional 1.5hrs in absorb from its old setting to new - so roughly 15 additional amps.

I waited about an hour after reaching float and checked with the new hydrometer (verified it with distilled water first).
The average SG was 1.248 - which is very close to Rolls specs for 100% charge.

Due to network security concerns I had not setup mymidnite (I have some programming experience and I don't consider Drupal to be secure).
But, I just did it anyways, it should be helpful in T/S since the data is a bit better then the local export data.

I have WBJR set to 80% because that was what rolls/surrette recommended,  I agree that efficiency is likely much higher, 90% does seem about right - especially after watching its behavior.  I will get some more time and data on it after my new settings, then adjust accordingly.  Rolls is likely making a low efficiency statement so that their customers are not disappointed.

I will check the SG again in a few days (we have nice weather for the next week) and see how it goes, I will post an update then.

Thank you all for your thoughts and time on this - I really appreciate it.

Travis

[ClassicCrazy]

Thanks all,


Due to network security concerns I had not setup mymidnite (I have some programming experience and I don't consider Drupal to be secure).
But, I just did it anyways, it should be helpful in T/S since the data is a bit better then the local export data.


Travis

If you leave Local Status app on during the day it will log and  you can download some detailed data from it. John has some posts about an error in it but otherwise you can make graphs with Excel.
Larry

[CDN-VT]

Thanks all,

I bought a new hydrometer - and have a 3rd (different model) coming as Vic recommended.
(and washed them both very well this time)

I changed endamps to 8.6amps - 1% of capacity is the minimum limit recommended by surrette/rolls -
as per the rolls tech I spoke with.

I changed absorb time to 4Hrs - although it was partly cloudy today and endamps triggered absorb long before 4hrs.
2.75Hrs was based on surrite/rolls recommendations - I will review this before leaving it set at 4Hrs.

Larry, to answer your questions - They had not been in absorb for more then 2.75Hrs when the WBJR went to 100%, I have noticed that endamps is
almost always the trigger to go to float with my system.

I watched my local app while working on my PC as the bank approached its old endamp setting, it was about an
additional 1.5hrs in absorb from its old setting to new - so roughly 15 additional amps.

I waited about an hour after reaching float and checked with the new hydrometer (verified it with distilled water first).
The average SG was 1.248 - which is very close to Rolls specs for 100% charge.

Due to network security concerns I had not setup mymidnite (I have some programming experience and I don't consider Drupal to be secure).
But, I just did it anyways, it should be helpful in T/S since the data is a bit better then the local export data.

I have WBJR set to 80% because that was what rolls/surrette recommended,  I agree that efficiency is likely much higher, 90% does seem about right - especially after watching its behavior.  I will get some more time and data on it after my new settings, then adjust accordingly.  Rolls is likely making a low efficiency statement so that their customers are not disappointed.

I will check the SG again in a few days (we have nice weather for the next week) and see how it goes, I will post an update then.

Thank you all for your thoughts and time on this - I really appreciate it.

Travis

Travis , good weather & sun I'd lean towards boiling off some water OR make the settings at the higher ends to see if you can boil !!! Then decrease , You might need summer & winter settings as I have .

VT

[Vic]

Hi Travis,

Good that you have a new Hydro.

Your batteries are young.   It will take 50 - 75 cycles for them to settle in.   Cycling them down into the 60 - 70% SOC range,  on occasion can help them build Capacity and stabilize.    Using EA with young batteries can be a bit frustrating,   as sometimes seemingly similar cycles can yield differing results.

It is good that you are paying careful attention to your batteries.   This really the most Important thing that one can do,   for long and dependable battery life.

If the electrolyte temperature is close to 25 C  (77F),  1.248 average SG readings is not really fully-charged.  Personally would shoot for 1.265.  Temperature compensation for SG readings is not large  --  about 3 points per 10 F variation from 77 F.  Subtract from SG readings for temps above 77 F,   and add to readings for temperatures below 77 F.

Would suggest that you EQ the bank,   after a full-ish charge,  using an EQ voltage at the low range of the recommended voltage from Surrette.   This is often done after the initial charge when the batteries are brand new.

Try cycling the batteries a bit deeper on occasion,   and perhaps even use  fixed time for Absorb in lieu of EA,  for a week or two.

If you do not already have one,  it is probably a good idea to get a DC Clamp Ammeter.   This will allow you to monitor just how equal the charge and discharge current balance of one battery string,   verses the other.

It is not clear what are the SGs of the two older batteries.  If they differ much from the SGs of the rest of the bank,  would suggest that one of these older batteries in each of the two strings,   if this is not now the case.

IMO,  80 - 85 percent Ah efficiency is a good starting place for your Flooded batteries.

Mine are now set at 77%,   and something closer to 70%  is probably more accurate for these Senior Citizen batteries,  which are use 3 Skip days,  and a full charge on the fourth day.   On the full charge day,  am really trying to reach 1.265 SGs,  which requires a fairly long Absorb (and therefore a hit on efficiency) ...  many tradeoffs.

FWIW,  Keep up the good work.    Vic

[Vic]

And Travis,

You did ask several question previously,  which.   believe have not been answered.

First,  IMO,  a lightly-cycled battery would see discharges to about 90% SOC,  or a bit lower.   Lead Acid batteries should generally be cycled to 90% SOC or below on each cycle.   For Flooded batteries,  occasional cycling to about 60% SOC is a good idea,   IMO.

You had previously stated your Limits, ...
"I set the charge menu limits to rolls specs when we installed the new bank as follows:
Charge Low: 52.8
Charge Hi: 63.0
Max Temp 125C
Does this sound right?"   ...

Think that these are too narrow for Flooded batteries.   If the batteries are warm/hot,    52.8  may be to high a voltage.   Have set my  Low limit to about 49.5 V,  IIRC.   If the batteries are  cool/cold,   63 V  IS too low for an EQ.   This High Limit will also depend upon how well your inverter can handle high battery voltages,   but would suggest something around 65,  or 65.5 V  if the Magnum inverter can handle it.

Also,  would suggest that you set the EQ temperature compensation = YES.

Recommend that the BTSes be placed on the same battery (Classic and Magnum),  half way down the case,  on a battery in the center of the bank.  If your batteries are exposed to a fairly wide temperature variation,  you might even consider placing some foam insulation over the BTSes,   and tape it in place,  or place a ratchet strap around the battery to hold the foam in place.

FWIW,   IMO   Vic

[Mainewoods]

And Travis,

You did ask several question previously,  which.   believe have not been answered.

First,  IMO,  a lightly-cycled battery would see discharges to about 90% SOC,  or a bit lower.   Lead Acid batteries should generally be cycled to 90% SOC or below on each cycle.   For Flooded batteries,  occasional cycling to about 60% SOC is a good idea,   IMO.

You had previously stated your Limits, ...
"I set the charge menu limits to rolls specs when we installed the new bank as follows:
Charge Low: 52.8
Charge Hi: 63.0
Max Temp 125C
Does this sound right?"   ...

Think that these are too narrow for Flooded batteries.   If the batteries are warm/hot,    52.8  may be to high a voltage.   Have set my  Low limit to about 49.5 V,  IIRC.   If the batteries are  cool/cold,   63 V  IS too low for an EQ.   This High Limit will also depend upon how well your inverter can handle high battery voltages,   but would suggest something around 65,  or 65.5 V  if the Magnum inverter can handle it.

Also,  would suggest that you set the EQ temperature compensation = YES.

Recommend that the BTSes be placed on the same battery (Classic and Magnum),  half way down the case,  on a battery in the center of the bank.  If your batteries are exposed to a fairly wide temperature variation,  you might even consider placing some foam insulation over the BTSes,   and tape it in place,  or place a ratchet strap around the battery to hold the foam in place.

FWIW,   IMO   Vic

Thanks everyone for your help with this,

Sorry for my late reply - the weather has not been the best for battery work but I think I am good now.

Now I understand what Vic and ClassicCrazy mentioned about my narrow limits - I completely forgot to take temp compensation into account.  I adjusted it accordingly: 49.5-64.0vdc.  Unfortunately, 64vdc is the top of the range for my inverters - otherwise I would have gone to 65vdc for equalizing but it does seem 64vdc is enough - see below.  EQ temp comp is set to "yes".  My batteries are in my workshop utility room - which keeps them between 18-25C for the most part.

The BTS's are both glued (I had some thermal glue laying around) middle side of the center battery in the back string.  I have checked with my IR thermometer and the batteries are all about the same temp during heavy charging - although I have never seen them above 27C.

The two questionable batteries were both in the same string - this was because I wanted them in the front string for easy removal if I decided that they were a problem.  We had good weather about 4 days ago, I disconnected the string after reaching 100%SOC. Then I isolated and put the two questionable batteries on independent 6v chargers for 24Hrs - they both took a pretty decent charge.  When the bank came back up to 100% the following day, I shut everything down and rotated one of each of the questionable batteries into each string - so now I should be balanced.  Then following day I preformed an EQ after the bank hit 100% and until all the SG's were the same (from 10 sample cells, including all 6 cells of the questionable batteries).  The EQ took about 4Hrs, it started at 64.0vdc and ended at 63.0vdc, all of the sample cells ended at 1.275sg.

The day after we had 17" of snow (the upper array cleared as I am typing) and the batteries seemed to preform better overnight and through the snowstorm - remained at about the corresponding voltage as the WBJR SOC.

Currently at 100%SOC, 4 of the questionable cells are at 1.250, 2 are at 1.275 and two sample cells are at 1.275.  So, I don't feel like I am quite done but I am closer to where I need to be.

I managed to dig an old ampclamp that was given to me years ago out of a storage box.  I have never used it before, I don't have the manual (could not find one online) and it does not state if it is AC or DC.  In any case I checked the jumpers between each string (before and after rotating the batteries).  Both jumpers read the same and the main batt- cable reading was double the jumpers - both before and after rotating.  That said, the readings were about triple the WBJR readings, so I suspect it is a AC meter.

Does anyone have a suggestion for a affordable DC amp clamp?  I do feel I should have one but money is tight - despite what it may seem like I spent on the entire system...  I really should put together a "how I built my solar system on a budget" post or article someday.

Since I have been doing alot of work on the system I have not changed to a temporary fixed absorb time yet.  I will try that tomorrow and see how it goes.

Thanks again,
T

[Westbranch]

Sears has a reasonable one for ~ $60- $70.  Sorry no part #..  search the NAWS forum .  Moderator BB has the part listed several times..