High Specific Gravity Unexplained

Vic · June 2018

Hi Ako,

Your idea of waiting a few days with the present settings, so see how the bank responds sounds fine to me.

Will have a couple of questions about the data that you posted, but later for that. Yes, this is a long Thread, and there have been many, many suggestions of what you should try. This can be a bit overwhelming, but you have hung in there. Later, Thanks for your Spain location. Vic

Ako · July 2018

Well Ive done what Rolls suggested and not had much effect other than reduce the SG slightly and seeing the voltage back to 25.3 when entering night mode rather then the 25.6-25.8 im used to seeing since the SGs increased but i still have a long way to go to be at safe SG levels so im now wondering if my batteries have been irreparable damaged . Unfortunately i dont use enough power during the non charging times for the voltage to go low enough to disconnect which was the reason for replacing my 10 year old Tudor batteries of the same size , my voltage just before charging begins is between 24.1-24.6 , slightly lower than before this problem started , previously i was showing close to 25v , my amps coming in during the charging period has not declined , if anything its increased slightly to average closer to the upper limit of my systems historical records at this time of year .

I followed Rolls advisers instructions and have dropped my voltage from the originally advised 29.76 ( .24 less than the manual still states ) down to 29.16 , a reduction of .6v together with reducing the Absorption time from 5 hours to 2 ,( the minimum time recommended ) . Also did an Equalisation a few days ago . The only thing i have not done is add distilled water , the batteries are all bubbling away nicely daily during charge but the levels have only dropped slightly and still within the upper and lower levels shown in the Rolls manual , between 6mm and 13mm . The batteries are going into float now each day for several hours .

I have again read the manual for the TS60 and can see theres the option to change from PWM to ON/OFF switching , will this help ?

Ako · July 2018

Think might have found a possible causes so i have today changed the TS60 and this whats become obvious .

The voltage readings on the Batteries , TS60 battery input connectors show a voltage higher than the battery terminals 0.1-.02 but im not sure if that might have something to do with the terminals being inside the TS60 and partly reading the solar input also , some sort of slight drain back ?

The sense wires inside the TS60 show a voltage of 0.1-0.2v lower than the battery terminals , that would explain overcharging but unfortunately it wasn't possible to disconnect the batteries or but i wouldent have thought it would make a difference for the purpose of seeing a difference in voltages as put raising or lowering voltages through input or load would be uniform across either reading sources although i did reduce the output voltage to 60 watts . I understood a sense wire was to compensate for loss through cables between batteries and TS60 so should be higher , having said that i have only a maximum of 1 metre of cable between the Batteries Bus Bar and TS60 using 6 gauge copper multi strand cable so i wouldent have expected much of a loss anyway . I used the same multi meter for all readings .

.

jonr · July 2018

No meter is going to read "equivalent steady state voltage" with a pulsed/PWM DC supply and a non linear load like a battery. Roll's specs are steady state voltages. See here for the nonsense you get from even good meters:

https://meettechniek.info/multimeter-avo/measurement-deviation.html

IMO, Roll's advice is bad (for your situation) or misunderstood - I would just accept your high SG readings and not undercharge to reduce it. Why - because if you haven't reduced fill level or added acid, then the high SG reading can't represent the actual acid concentration (which is what really matters).

Ako · July 2018

I used a digital Multi meter Jonr so is it possible the difference between the readings on the TS60 and the Multi meter will be different but i was mainly concerned with the differences between Batteries , TS60 battery connectors inside the unit and the sense wire connectors inside the unit rather then what they actually were .

Dont think accepting the higher SGs would be good , Rolls themselves told me that they were bad for the batteries and from what i have read online everything else supports that and anyway , they were normal for at least a year and as far as i can tell nothing else has changed to explaine it other than the battery temperatures are 10 degrees C higher .

jonr · July 2018

My guess: your sense wire circuit has some resistance, capacitance and inductance. This will change the pulsed DC waveform shape and the meter reading.

Ako · July 2018

I have now bitten the bullet and removed the sense wire on the TS60 end, i cant see it was needed anyway as i didnt see a voltage drop between batteries and TS60 internal connectors , the only drop in voltage i see is between the batteries and the TS60 internal sense connector when it was conected where im sure i should be seeing an equal voltage . . As i see it the worst case scenario is without a sense wire undercharging is a possibility but in this case i think its possible the sense wire could actually have been causing an overcharge as its giving a lower voltage then Batteries .

My sense wire was the right gauge and no more than a meter long and im sure when i first connected it the readings were the same as the battery as it is now that its no longer conected at the TS60 end .After disconnecting the sense wire i left it connected to the battery and read the voltage with a multi meter and it was showing exactly the same on both so even if it was working correctly it wasn't needed .

Problem with the MSview is when yo disconnect the TS60 you dont have any readings for the previous 24 hours , its gobbledegook . I have been trusting MSview to give me accurate readings as its those im using for everything as i dont have a meter connected .

Will have a better idea of the effect is had in 24 hours .

mcgivor · July 2018

The voltage sense will compensate for voltage drop when under load, the higher the charge current the higher the volt drop, in reality as the current drops to minimal amounts, the voltage difference at battery and CC would be less, disconnected it would have a minimal effect on overall charging.

The process of elimination requires the change or removal of one component or setting if no change is observed then replace/restore before moving on to the next. Changing multiple things could result in other problems without revealing the root cause. The pragmatic approach would be to think along the lines of why the SG is higher than it should be, the only way it could rise would be through water loss or addition of acid, no amount of overcharging can increase the amount of acid in solution therfore it must be related to concentration meaning not enough water.
Lowering the charge voltage may eventually reduce SG but will be at the expense of lost capacity due to sulfation.

Just a collection of thoughts,

Attached pdf contains useful information

Ako · July 2018

Thanks Mcgivor , i read the document and found it very helpful , re read several points to make sure i understood so must have gained knowledge from it . You made a valid point about making changes progressively to establish what works and what doesn't , only thing that made a difference was lowering the Voltage 0.4 initially , it did reduce my SGs slightly but wasn't the answer so i did several things and now im in spitting distance of finding the solution ( no pun intended ) .

I removed the sense wire completely , my loads are relatively small and the most drop i see is 0.2v for a few seconds and the voltage shown on my batteries is identical to the Bus bar and TS60 after charging has ended , while charging its not the same as i think the panel input raises the voltage at the connectors inside the TS60 and gives a false elevated reading , i have noticed that on both controllers from the start , bit of an anomaly and im sure theres a simple explanation for it .

I also added 3 1/2 ltrs of distilled water when batteries were fully charged bringing the fluid level to 3mm under the high fill point , works out about 300ml per cell . The same night i reset the voltage back up to Rolls original settings which has always worked well 29.76 @25c .

The following day i waited until the Bulk charge had ended and put it into Equalisation at 31.2v for 2 1/2 hours as recommended , thought it wouldent do any harm , would stir up the mixture and level cells as i was showing a 0.3 variation between high and low .

The SGs for the 1st 2 days after an hour in Float are all where they should be with the exception of one cell , 1.255 - 1.265 but one cell is always showing .1 higher at 1.275 and with the temperature correction of 0.06 it would be over 1.28 so now i would like advice on how to correct the one cell . I am considering removing a hydrometer full of the liquid from the cell and adding the same amount of distilled water , is this s stupid idea or might it work ?.

The thing i cant get my head around is that all cells have been treated identically , all only filled with distilled water and taking approximately the same amount and the same branded water i used for 10 years on my Trojan and the last 18 months with the Rolls now , acid has never been added , i wouldent even know where to buy it although have read so many articles that say it should never be added anyway so i would never consider it . I really need to work out why so i can avoid it happening again .

mcgivor · July 2018

Progress is being made, be patient, nothing happens fast, the one cell's slightly higher SG may be partly due to slight difference in electrolyte level, should an SG adjustment be made? I think not at this point, do one step at a time see what happens, perhaps add a little more water to the high cell, it wouldn't hurt, see what the results are over time. As previously stated I fill water to the point of touching the neck when fully charged , which is pretty easy to see accurately, battery temperature in low 30°C range, not much expansion happens to 36°C to be of any concern, based on experience. Yes the article has useful information but not entirely for off grid applications, so reading between the lines a little is required. Hang in there and good luck.

Ako · July 2018

Yes Mcgivor i agree , will wait a while before making any more changes to see what pattern emerges . I added a small quantity of water last night but only a small amount so i doubt it will have a significant effect but will give it a few days .

I should have added that from reading Rolls manual and other sites advising on charging times that Rolls and many others either give Adsorption times or a formula for calculating Absorption time . The TS60 i always belived was not capable of being set to run for a specific number of hours/minutes because it used an algorithm that relied on a Duty Cycle which in turn relied on either Voltage or battery resistant . Through my constant endeavours i discovered that the TS60 can be set to run Absorption for however minutes a user requires , dont understand why Morningstar have no made users aware of the hack , obviously they have a lot of time and money invested in their method but ir an end user wishes to set their equipment in the same way other controllers can be set then i think they should be able to .

mcgivor · July 2018

Probably they want to prevent novice operators form making decisions which could cause harm, thereby allowing those smart enough, through gained knowledge, to make decisions based on experience, just an analogy. Of course I could be wrong.

Ako · July 2018

Or Maybe Morning star dont realise the full capabilities of the TS60 , i did after all discover it accidentally . .

jonr · July 2018

It's not clear to me that Roll's absorb algorithm is best or just an approximation for dumber chargers that only have fixed absorb timers. I'd verify by checking that charge current tapers off to around 2% (and charging continues after that to reach full charge).

Ako · July 2018

Yes drops to around 6 - 8 amps which is right ish according to Rolls Manual , about 2% of C20 which is 445 amp/hrs for the S605s . Float it drops to about 2 - 3 amps 0.05% . Think anything using the batteries natural resistance has to be better than a pre set time because an hour in winter is a lot different than an hour in summer here .

Vic · July 2018

This is a fairly long discussion, and forget some of the fine detail, and all of the suggestions that have been offered/discussed.

The Return or Finishing current in Absorb will depend upon the Absorb voltage. Likewise for battery current in Float and EQ.

Do wish that Ako had an MPPT CC, which was more settable, and info from it would be more easily understood. The TS-60 CC has always been a bit too mysterious to me.

I forget just how you determine battery current, is this inferred from Duty Cycle? A battery monitor, ... or ...?

If the :"Rolls absorb algorithm" that jonr referred to, was the one for calculation Absorb time, then, agree, IMO, this is for constant current chargers (IUI, or similar), and is not well suited for Solar charged batteries.

FWIW, each of the two Surrette battery banks here (4KS25s), each bank has had one battery (two cells) that have an SG of 1.280 (temp compensated), while the rest of the cells average 1.260 - 1.266-ish. IMO, this is not an impossibility. There are variations in the way each cell happens to turn out. There are differences in how the cells are Formed. Sere some of our batteries stored at the manufacturer, as dry charged, etc?

Had also wondered, if, perhaps some manufacturers ship batteries a bit short-filled, with 1.280 SG electrolyte, or similar (for those with target SGs of 1.265), hoping that the lower electrolyte level might result in less spillage during shipping (for those shipped, wet), Then the dealer/customer tops-up with Distilled Water, resulting in nominal 1.265 SG electrolyte, when at the correct fill level.

Couple of other things;
In a single string battery bank, each cell will have a voltage across its terminals that represents the SOC of the battery. Cells that are more fully charged, will have a bit higher voltage than the lagging cells. When we add water duing Absorb for example (my favorite time to do so), then the cells with the highest terminal voltage will be gassing more vigorously than the lagging cells. Gassing (as well as heating during charge) raises the apparent electrolyte level. SO, generally with healthy batteries, the cells that electrolyze the most water, and need the most make-up water tend to get the least, when they are filled to a certain level. Taller batteries probably tend to exaggerate this difference.

Just more opinions, FWIW, Vic

MrM1 · July 2018

When my charge current drops to 2% I am still not at - 100% SG, unless I use hi absorb voltages

Ako · July 2018

Vic , i dont think the Rolls formula for calculating asorb time is for constant current chargers as it recognises the input current tapers off and voltage rises and factors that into the equation , if it were for a constant current charge then the calculation would be simpler .

I have found the end current voltage to be reasonably accurate using the TS60 algorithm and terminates around the figure Rolls suggest at around 2% of C20 . After that Float kicks in and current in gradually reduces from around 8 amps to 2 amps .

Vic · July 2018

Ako said:

Vic , i dont think the Rolls formula for calculating asorb time is for constant current chargers as it recognises the input current tapers off and voltage rises and factors that into the equation , if it were for a constant current charge then the calculation would be simpler .

I have found the end current voltage to be reasonably accurate using the TS60 algorithm and terminates around the figure Rolls suggest at around 2% of C20 . After that Float kicks in and current in gradually reduces from around 8 amps to 2 amps .

Ako,

Fine on that, but the Surrette formula appears to treat the available charge (max) current as a constant -- 50% of the maximum current available, They mention,

" ... for example,

2 strings of 6 Volt 6 CS 25P models 20 hr AH rate = 820 AH x (2 strings) = 1640 AH I = 10% of 1640 AH = 164 Amps. If charger output is 120 Amps max, 120 is used T = 0.42 x 1640/164 = 4.2 hrs OR T = 0.42 x 1640/120 = 5.75 hrs ... "

It appears the "charger" is not a PV system, but some Grid-connected box. A consistent current value available. For most of our PV charged systems, the Max current varies, form day-to-day.For Solar-charged batteries, the Max available current is NOT constant (IMO), bur, rather, a variable amount, depending upon weather, other loads, etc.The experience with SUrrette banks here (as well as those from other manufacturers), the Absorb time is not a constant, but varies depending upon the Depth Of Discharge In the previous discharge. Deeper discharges usually require longer Absorb times, and so on.It is possible that that Surrette formula assume a consistent DOD, to 50% DOD, or something similar.One could say, that the recharge during Bulk returns the SOC of the battery to 80.000%, but there are variables, like the rate of discharge, rate of charge, battery age, adequacy of EQing, etc, that this formula seems to neglect.Have found NO perfect way to predict the length of Absorb, for off-grid PV-charged systems. Still, using Shunt measured EA still seems to be the closet thing to Nirvana ...Just MHO. FWIW, Vic

jonr · July 2018

> algorithm ... is not well suited for Solar charged batteries.

I very much agree. It's often said to use the manufacturer specs. But these specs are usually for batteries being charged by grid power. Solar charging has limited time and non-steady currents. So it's necessary to use higher voltages and currents to get the battery fully charged. Or do a weekly generator run.

For example, adding up the times below (3 stage, flooded) is 9+ hours even with moderate discharge.

http://usbattery.com/wp-content/uploads/2014/05/usb-charging-recommendations.pdf

Maybe (I'm no expert) the one universal truth is to verify (using SG or at-rest voltage) that full charge is being reached at least weekly. If not, you need more aggressive charging (even if it violates manufacturer recommendations).

edit:
This being said, Roll's seems to say that their absorb time formula is applicable to solar situations (see page 32):
http://rollsbattery.com/public/docs/user_manual/Rolls_Battery_Manual.pdf

Ie, if you bulk charge at 10-20%, you can achieve a full charge with only 4.2 to 2.1 hours of absorb. Sounds optimistic - I'd like to see this verified with measurements.

Ako · July 2018

Read again Rolls method of calculating Asorb times and yes Vic and Jonr your both definitely right , they are talking about a method other than solar , as you say solar has far to many variants for a one size fits all formula .

Jonr , i wonder if page 32 they are assuming everything is always equal and theres a constant known current from the PV system as with a grid tie system in which case , solar or grid tie its a formula for constant current irrespective of its source .

Unfortunately i have come across this before with Rolls where they assumed a grid Tie system , they really need to update their manuals to make these things clearer .

On my initial post i was concerned about abnormally high SG readings , im pleased to say that from advice and recommendations given here i have not only learned an awful lot more than i knew before my problems now sorted . Rolls said Lower voltage to less than they say in the manual and reduce the absorption time , that did very little , SGs dropped by only .01 after reducing 0.6v and absorption by half . Taking everything int account it seems fluid levels although at the lower end of Rolls specks so correct they were the reason for the problem , needed to be higher than Rolls levels and as Mcgiver also pointed out going right up to the tab didnt cause any problems even through bubbling through equalisation at 38c for 2 1/2 hours The last piece of advice from Mcgivor said to add a little water to a single high cell of 275 , i added about 3 reasons full only and that did the trick . My SG readings have now been stable for the last few days , 4 batteries of 3 cells @ 34c Rolls range for fully charged 255 - 275 . Below are readings a few minutes ago , temperature compensation will make them slightly higher but still within specs .

265 265 265 265
260 260 265 260
265 255 265 265

Feeling chuffed , think im about there now . Thank you everyone .

Estragon · July 2018

Great to hear it's getting to where you want it. Sometimes takes a bit of experimentation to get things dialled in for your specific application.

I've found having input from others who've gone through the same process extremely helpful as well.

<rant> That said, it's a bit surprising to me that there's apparently (at least to me) a real lack of hard data on a technology that's been around for ~150years with relatively minor charges, and which is one of the most recycled products in the world. I've read papers from the likes of Sandia, but they tend to rely on a fairly tightly defined lab condition. That's what you need to test a hypothesis, but may not be applicable in real world conditions. You'd think battery makers or someone might have an interest in doing a failure analysis on batteries at point of recycling to see if there's a pattern. Or not. Maybe it's better we don't know, and just keep buying batteries.</rant>

t00ls · February 2019

things i have noticed about batteries since I have been off grid for several years now with new trojans

new batteries have been sitting for a while and the plates inside wiil be grey in color....means a bit of sulfation exist...it is always recommended to do aequalization charge right from the start ....this is to re-mix the sulfation on the plates back into the fluid
it may take a few times back to back to do this....but you dont want to equalize the first time with hard hitting amps and voltage due to overheating quickly....you just want the voltage to stay at a higher than normal level while keeping the batteries cool

looking at the color of the plates, good batteries in good health will have a chocolate brown color, meaning all of the sulfur crystals have mixed into the fluid

if your batteries reach a complete charge fast, usually means low resistance between the cells,power will also make them discharge faster, denotes a short, possibly from sufuric crystals...when the batteries take a long time to charge thats when they are in good health and power will discharge them more slowly

the only way to get a true specific gravity is to weigh the fluid...there are formulas to tell you X amount of fluid should weigh X amount, thats how the manufacturers do it

refraction I believe is more accurate than a hydrometer due to as one said bubbles may be causing the floats to rise higher

and as someone else said, you really want to deep discharge the batteries from time to time...it wont kill your batteries as long as it isnt cold and the batteries could freeze, and you do a good equalize charge afterwards

deep discharging really does make a difference. I have noticed an increase in length of time of discharge and recharge after I have done it...I usually do it every 3-4 months, sometime down to 35% and 40% of charge, then most of the time I only discharge down to 50 %

but as I am off grid, thats what works for me....someone who is on grid that doesnt use the batteries down to 60%-70% may not have the same results

I dont do SG readings as much as I used to ...maybe once a year now, I think if batteries are being used that the readings wont make much difference in a short time period.....but, if they show a difference through a year, something is probably wrong

these are my opinions and observations, others may disagree with me

High Specific Gravity Unexplained

Comments

Categories