Solar series battery charging problem

[jim hardy]

You have to distinguish between the meanings of the terms Current, Charge, Potential difference and Power and follow 'the rules'- even when attempting a 'friendly' and chatty explanation of the way a circuit operates.

Amen,
and i was too lax above . it was in interest of brevity.

my old standby
Laviosier, quoting de Condillac
https://web.lemoyne.edu/giunta/EA/LAVPREFann.HTML

"We think only through the medium of words. --Languages are true analytical methods. --Algebra, which is adapted to its purpose in every species of expression, in the most simple, most exact, and best manner possible, is at the same time a language and an analytical method. --The art of reasoning is nothing more than a language well arranged. ....
... the sciences have made progress, because philosophers have applied themselves with more attention to observe, and have communicated to their language that precision and accuracy which they have employed in their observations: In correcting their language they reason better."

not bad thinking for 1780, eh ?
Though i'd make it read

We think exchange our thoughts only through the medium of words.

 

[BernieM]

Well I am not going to belabor this any longer. It's possible I have gotten things mixed up in this discussion, as to which is fixed, i.e., current or voltage. But don't forget a battery cell is an electrochemical reaction and a cell is not a carbon resistor. One has to capture 2 electrons to reverse the process to charge a cell as I recall. The earlier comment made in one of the posts was that you have X coulombs of electrons flowing in the circuit, a coulomb of course being a fixed quantity of electrons. When you have electron capture, you have to be subtracting from the total number of electrons flowing once capture occurs.

Regardless, this is a real problem evidenced by the dead batteries.
However, I think this could be a power density issue. And that's why I went to stratification. If the electrolyte is stratified, the current density flowing through the electrolyte will be higher where the electrolyte is stronger, less current flowing through the weaker electrolyte. So there would be a gradient from top to bottom of the battery in each cell. Any lead plates conducting current will see the lower portion of the plate with a much higher current density compared to the upper portion of that plate. If I put say 10000amps charge on a battery, the post would evaporate. At lower amperages it would melt, and so forth and at zero amps of course no damage. By reducing the area over which the current is entering the electrolyte, or increasing the the current densities in one area, one would effectively have a higher current density and may be degrading the plate starting from the bottom and as it falls to the bottom of the case, works on up. At least that's the theory I had come up with. Wrong or right it seems to me to make the most sense. This is what I was tending toward in the conclusion that I made earlier. I guess it's time to cut a few batteries open.

 

[jim hardy]

But don't forget a battery cell is an electrochemical reaction and a cell is not a carbon resistor. One has to capture 2 electrons to reverse the process to charge a cell as I recall. The earlier comment made in one of the posts was that you have X coulombs of electrons flowing in the circuit, a coulomb of course being a fixed quantity of electrons. When you have electron capture, you have to be subtracting from the total number of electrons flowing once capture occurs.

nope. That's the source of your confusion.
See
http://ecee.colorado.edu/ecen4517/materials/Battery.pdf

However, I think this could be a power density issue.

Could be.

Might also be a ripple current issue. I don't know if it is, but good troubleshooting rules out the possibilities one at a time.
Industry papers on effect of ripple on batteries confirm it heats the electrolyte and to an extent the plates too. It shortens battery life.
The heating is easily measured. Do your batteries get warmer when running that inverter from them?
http://www.cdtechno.com/pdf/ref/41_2131_0212.pdf

http://www.battcon.com/PapersScanned1999/GernerPaper1999.pdf

http://www.ecmweb.com/ops-amp-maintenance/get-know-your-battery

Ripple current — Excessive ripple will cause internal heating of batteries. Ideally, ripple current should be less than 5A for every 100Ah. However, ripple current says more about the state of the charger than the battery. Ripple current does not provide any information about SOC, capacity, or SOH.

You ought to look at your inverter's current draw with your oscilloscope if only to rule put the ripple issue .

However - all that said

Yours is not the only inverter of its type out there. Were this an inverter issue the brand would have a reputation as a "Battery Killer".
Which leads me to think it's a battery issue.

Where did you get your batteries?

They are not recycled batteries. They came out of a 12v solar power system that someone was upgrading and were 1 year old when I got them. I tested them all and they all tested about the same on a load tester. They are two different brands but have the same specifications.

http://www.cres.gr/kape/publications/photovol/batt-bar.pdf

A techno-economical analysis showed that the use of vehicle batteries should be avoided in PV applications.
Daily cycling of two SLI batteries resulted in more than 70% capacity loss after 160 cycles.

Are yours deep cycle or plain old starting(SLI) batteries?

I don't know what's going on. That's all i can think of to check.

Let us know what your autopsy on a battery finds ?
Photos always help. Might be a battery guru in the house.

old jim

 

[OmCheeto]

...
I guess it's time to cut a few batteries open.

You might want to google the hazards of working with sulfuric acid first.

 

[jim hardy]

At 50%, a 2 year lifespan is standard for flooded lead acid.

If one keeps depth of discharge less than 50%, one should be able to nearly double the life span. The life span of a battery is considered to be the number of full discharge cycles the battery can tolerate without failing. So a 50% discharge would take 2 discharges to that depth to equal one complete discharge. This is how I run my setup, I do not allow it ever to get below 50% (24v.)

That's good...
Never below 70% charged (30%discharged) might be better ?

http://pvcdrom.pveducation.org/BATTERY/charlead.htm

Relationship between battery capacity, depth of discharge and cycle life for a shallow-cycle battery.

Have you said whether yours are shallow cycle or deep ?
How deep did previous owner cycle them ?

 

[BernieM]

As I stated in an earlier response, there is one system that is at 12v using a Powerjack 8/32K inverter too. That system never has these failures, only the 24v systems (there are three systems I set up, one for myself (24v), and two for friends nearby (one 24v and one 12v.)

Batteries are two different brands (the batteries here, however the batteries at the other 24v system are all the same brand and model) They are 8A8D with about 265Ah storage each. The batteries as I said, were from a large commercial solar installation and had been in use for about 9 months when they decided their banks were too small and they upgraded their entire array and I picked these batteries up from the company that swapped them out for that company. The brands are Deka and Interstate but both have the same exact specs are far as Amp Hour capacities, etc. Both types have failed, so it isn't just say the Interstate failing. One battery so far out of 8 on the other system, and 4 out of 24 here. My use on the system is far greater than the other system sees. I can be running up to 4 a/c's at times (small a/c's) and a microwave oven as well (for short periods of up to 5 minutes.) If you have been watching the temps here in Arizona recently you will realize we hit 121F the other day (where I live,) so power to run my a/c's is a matter of life or death, truly. A generator running 24 hrs a day at these temps won't last long. As a result, at times (like when it's hot like it is these days) I push the inverter to the limits pretty much. I have had it whining at me (beeping due to overload) more than once. But that's why I love this inverter, not as fancy as say an old Trace, but this thing takes a lot of abuse without failure. It has a huge torroid in it (it's a low frequency inverter.) Unlike the one you presented in your post that shows 2 of them, they later went to a larger single torroid for even more robustness.

You might want to google the hazards of working with sulfuric acid first.

Dangerous? Wow didn't know that. Thanks for the heads up. Will buy some hazmat gear ;)

 

[jim hardy]

Unlike the one you presented in your post that shows 2 of them, they later went to a larger single torroid for even more robustness.

You've not offered any information about your inverter besides its brand name and 8kw.
Google pointed me yesterday to this manual but it's a different brand so i ignored it.

. http://pdf.donrowe.com/wag8000_owners_guide.pdf

Got a link to yours ??

It has a huge torroid in it (it's a low frequency inverter.)

Low frequency inverter ? Not pwm synthesized sinewave? That's curious, this day and age. Does it have input filter?

They are 8A8D with about 265Ah storage each. The brands are Deka and Interstate

East penn's site is hard to navigate . I don't know what the LTP means.

http://www.mkbattery.com/images/8A8DLTP-DEKA.pdf

Here's the life vs discharge depth curve from lower right

at 50% discharge maybe 700 cycles ?
C/20 = 245 AH / 20 hours = 12.25 amps, X 10 sets = 122.5 amps, X 24 volts = 2940 watts at 20 hour rate.
So you're not overworking it a lot at 8kw.
At 5KW, 21 amps per battery pair, i get a little less than 6 hours to 50% discharge? Does that agree with your calcs?

background

http://www.eastpennmanufacturing.com/wp-content/uploads/Guide-to-VRLA-Batteries-1927.pdf
http://batteryuniversity.com/learn/article/absorbent_glass_mat_agm

Everybody stresses temperature and you have a lot of it out there. Try taping your temperature sensor right to a battery case ?

Still, the preponderance of failures you report in + side is curious.
Is negative side of 24V input tied to chassis as in that Wagan inverter? If so it'd be easy to 'scope.
A section of the interconnecting wire would useable as a primitive shunt to observe current waveform , DC coupled of course so you could use a DC ammeter to determine how many amps/division..

Good Luck

old jim

 

[BernieM]

The PowerJack is not made by Wagan. Also what I can find on the internet today regarding the inverters I use are not up to date as to what they currently are. It seems that most of the posts, videos or reviews are several years old. Most inverters jack the voltage up pretty high and high frequency as well, then drop it back down to the output voltage. Typical. By using high frequency of course they can get away with small inductors cutting costs. The LF (low frequency) does not jack the frequency up and so a larger inductor is needed. But the advantages of that is that it is much more robust and harder to burn up as it heats up more slowly than a small inductor at high frequency. Also LF inverter is not susceptible to aircraft radio traffic interference. Where I live that's actually a concern as I am right in the flight path (several miles out) of the F35's flying into and out of Luke AFB in Phoenix.

Here is a link to eBay's ad that gives specs about the inverter. This is the inverter I have (mine is a few years older but specs about the same.) Please note that I have never found a useful manual other than a user manual that's included in the box. If you scroll down the page they give quite a few of the technical details and a description there that might be somewhat helpful.

http://www.ebay.com/itm/8000W-SP-LF...177748?hash=item4af2a064d4:g:U5oAAOSwDNdVihY5

I'm not sure what the LTP is either. My battery AmpHour rating is 265 not 250 as the one pictured (on the label of the battery.) But it's close so I don't think there is a lot of difference, or at least enough to be concerned about, so it's about right for my batteries.

I don't usually run loads of 5kw. Generally my loads typically run 1 to 2kw with intermittent 3 or 4kw loads for a few minutes at a time. The only loads I have that are large here other than a skill saw or table saw (which are used infrequently) is a microwave that pulls 1kw or so. But I can run an A/C (about 1kw) for 12 hours, and the battery bank will be discharged to about 50% (23.6v) by morning (with the load on it, of course the actual voltage recovers if I were to turn off the a/c which drops the voltage about .4v when on, so 24v.)

Since day 1, ALL battery failures have been the battery (of a pair) connect to the + of the system. Never the battery that's connected to the -. Now as I recall, heat moves in the direction of current flow from positive to negative. One of the principles of of a peltier diode. It is true it gets very hot here, and where the batteries are currently in this system, they see a fair amount of heat, (yet never as high as outside temps. To cure this heat problem I almost have finished an underground battery cellar 9 feet deep.) So if it is a heat issue, then the + terminal would be the hottest in addition to temperatures externally (which max out at around 98F.) Current would be related to post/wire heating as well, which then may be a cause for this strange phenomenon.

The other locations where the batteries are stored are similar to mine. In shaded locations where it does not get as hot as outside but not as cool as inside. The other 24v system is in a garage attached to the house, mine is on the front covered and enclosed, insulated porch.)

 

[Tom.G]

Since we are about down to brain-storming for a solution, check the temperature of the connections to the battery posts during a high current situation. Gross problems would be detectable with your finger.

 

[jim hardy]

Since day 1, ALL battery failures have been the battery (of a pair) connect to the + of the system. Never the battery that's connected to the -.

That's four failures ?

I still think you should 'scope the 24 volt voltage across and current through a couple of battery pairs.

Secondly - it's always something simple... How accurate is your voltmeter ? Have you checked it against another one? Got any friends at Palo Verde who'd do a calibration check it for you ?
If it reports low edit- oops,, too high you'll undercharge your batteries...

 

[Tom.G]

It looks like you have the wrong battery chemistry for your usage.

Also, is the charge controller in the same thermal environment as the batteries? That way it can adjust charge voltage to track the battery temperature.

20 Flooded Cell batteries...

But the datasheet clearly states Lead Calcium plates, Valve Regulated Sealed, AGM (Absorbed Glass Mat.)

For an overview of battery types see: http://www.battcon.com/PapersFinal2009/ClarkPaper2009FINAL_12.pdf
Where it states that Lead Calcium is good for float usage as for an occassionally used UPS but has a relatively low cycle life. A Lead Selenium construction is suggested for long cycle life... or better yet a battery marked for photo-voltaic use if available. (pg.9)

And also indicates that there is something non-obvious about where a cell is located in the battery string. Unfortunately, no explanation is given.
Pg.4 states:
"In extreme cases ICV (Inter-Cell Voltage) variation can result in a gradual loss of capacity in low voltage cells. To achieve acceptable ICV variation, some manufacturers add platinum salts to the high voltage cells in a string to depolarize the negative plate."

From the Mfg of your batteries: http://www.eastpennmanufacturing.com/wp-content/uploads/Guide-to-VRLA-Batteries-1927.pdf

pg.3
"Immobilized gel or separator-absorbed electrolyte will not “float”
or “sink” within itself when a non-uniform concentration exists
so it cannot stratify. Therefore, no high-voltage equalizing charge
is necessary. Simply recharge at the recommended recharge
settings. This means longer life and consistent performance."

pg.4
"Any VRLA battery will dry out and fail prematurely if it
experiences excessive overcharging.
Note: It is too much voltage that initiates this problem, not too
much charge — a battery can be “over-charged” (damaged
by too much voltage) even though it is not fully 'charged.'"

p.s. (@jim hardy I think you'll also find the above articles informative.)

 

[jim hardy]

Thanks TomEast Penn's site is confusing me
their 'complete line brochure" at http://www.eastpennmanufacturing.com/wp-content/uploads/Complete-Line-Brochure-0083-.pdf
shows only one 8A8D on the gel/agm page

but it should be marked "Intimidator".

He said his were flooded
but East Penn claims they're a replacement for flooded.
http://www.eastpennmanufacturing.com/wp-content/uploads/Intimidator-Deep-Cycle-Flyer-1740.pdf

i'll study up on those links.

 

[jim hardy]

pg.4
"Any VRLA battery will dry out and fail prematurely if it
experiences excessive overcharging.
Note: It is too much voltage that initiates this problem, not too
much charge — a battery can be “over-charged” (damaged
by too much voltage) even though it is not fully 'charged.'"

That's interesting.. So if a local voltmeter is consistently reporting LOW it'd make one overcharge his batteries.

Feel those batteries when inverter is working hard.

Do gel batteries dissipate internal heat as well as flooded? Seems there'd be no convection in the electrolyte.

 

[BernieM]

Don't know why you guys don't just take me at my word. I know the difference between flooded cell and AGM. These are NOT AGM. I got photos of them but this editor won't allow me putting them in without uploading them to the internet somewhere.

I found the datasheet for them. Here it is:
http://www.dekabatteries.com/assets/base/0010.pdf
The battery numbers for the two case styles are 908DFT and 908D. Most of the batteries are Dekka. I have about 6 Interstates. They are just 8D-MHD.
Here is the datasheet for them (2nd page.)
http://www.interstatedealers.com/IBCatalog/common/commercial_spec.pdf
As you can see they are well matched.

I have 3 meters. One is built into the controller. One is built into the remote battery monitor I have attached, and I have a hand meter. All of them agree within a reasonable row of corn as to the voltage. There is no meter that's mysteriously reading low.

I will see if I can get the time to move the scope out there and check it for a/c ripple superimposed on the dc.

Regarding post temps: I just went out and checked, and though my optical thermometer was recently borrowed and not yet returned, I put the contact of the outside temperature probe right on the positive post. Air temp is 95F and post temp is 91F. It's about 1pm and the reason the post is lower than air temp is that overnight it all cooled down and takes a long time to heat up all that mass again the next day. But I have never found posts to be hot. By the way, the charge controllers advise connection of the temperature probe that comes with them to be attached to the + post. Which it is.

Load is about 2.5-3kw right now, as I am running 2 a/c's, a tv, a desktop pc with a 500w power supply in it, a chest freezer and a refrigerator and some lights. And that's about as much load as I typically run. Though there may be occasion where I run a heavy load like table saw, it's rare and usually isn't for very long.

No, it is not 4 failures. It is 4 failures in the last year with this array of batteries which have been in a year and a half, or thereabouts. Prior to that I had some smaller batteries which I had the same problem with, and had two of those go bad in about 1 year. And my friend's installation is now in place about 1 year and has had 1 failure. ALL OF THESE FAILED WHILE BEING THE BATTEY CONNECTED TO + IN THE PARALLEL BATTERY ARRAY ... and that's why I came here to see if there was something I was missing on what could cause battery failures like that. So far I have not found anything, even the a/c ripple to explain it. Why? Because that ripple would be spread out over 10 pairs of batteries. I think that would be a very noticeable thing.

 

[jim hardy]

Don't know why you guys don't just take me at my word.

Which word ?

Batteries are 8D's

They are 8A8D with about 265Ah storage each.

The battery numbers for the two case styles are 908DFT and 908D.

looks like a truck battery to me.
http://www.eastpennmanufacturing.com/wp-content/uploads/Deka-Commercial-Batteries-Flyer-0010.pdf

 

[BernieM]

If you read the sentence just after that I made clear it was regarding whether the batteries were AGM or flooded cell. I have said that they are flooded cell several times in this thread.

They use them in commercial vehicles I guess. I don't know where I saw it, (possibly on a label) the term 'deep cycle.' Will look again. But even if these were not true deep cycle batteries, I still wouldn't understand the failure on just the array + connected battery.
One thing I read at one time or another in the past few years was that the batteries have 9 negative plates and 8 positive (per cell??) Not sure how that might affect this kind of a failure, or if it would have a role.

Regarding the earlier statement by Tom G. re: Failure of AGM's and high voltage even if battery is discharged. I had a good example of that on a small array with four 4D AGM's and a friend of mine got to playing with my controller settings while I was gone. 1/2 hour later when I came back and discovered it, he had fried all 4 AGM batteries. He set the controller to go 'on hi' instead of 'on low' which meant that when the array voltage got to the high voltage setting, to turn the charge controller on. He exposed the array to about 42 volts @ 100 amps (across 4 batteries, 2 ea in series!) He went on the internet trying to find a manual for the controller and let them cook, rather than disconnect them, and THEN look for a manual!

If random failure was the cause, what are the odds of 7 failures on a positive connected battery only? Well it would be like flipping a coin 7 times, always coming up with heads (or always tails) and the odds of that is 1:128. Fairly high odds, but actually reasonably possible. So there is a remote chance it is just a statistical fluke which I can't ignore. 2 or 3 more failures would push it out of the range of acceptable odds however I would think. Or perhaps the odds are weighted a bit by something like heat beginning at the positive terminal and heading toward the negative. Would be nice to see an expert on battery physics render some sort of speculation on this.

 

[anorlunda]

But even if these were not true deep cycle batteries, I still wouldn't understand the failure on just the array + connected battery.

You seem to have a knack for misunderstandings.

Deep discharge batteries AKA marine batteries, are critical to your application. They have thicker plates than so called starting batteries. They are meant for a smaller number of cycles, but deeper discharge and slower recovery. They typically cost more than starting batteries.
For maximum life, they should remain over 70% charged all the time. Their most common cause of failure is undercharging leading to sulphation. But you rejected sulphation because of you profound misunderstandings about basic electricity.

If you ignored deep discharge and used truck starting batteries, bad outcomes are expected. Compare it to putting diesel fuel in a gasoline engine.

From what I read in this thread, I put no credence in your idea about the +terminal being factual or relevant. You brushed aside advice to not mix batteries of different brands or ages. You have no records of the brands or ages connected to the + terminal and how they were paired.

You ignored my advice to ask the manufacturer for help and instead proposed cutting the failed batteries open yourself.

At ay ay ay ay.

 

[jim hardy]

One thing I read at one time or another in the past few years was that the batteries have 9 negative plates and 8 positive (per cell??) Not sure how that might affect this kind of a failure, or if it would have a role.

Physical position in a battery string will affect both cooling airflow and radiative cooling. Middle ones have the tough spot. We haven't seen a picture...

Here's the only esoteric brainstorm i can dream up for electrical position in the string to cause causing unequal heating. It's why i suggested a look at ripple.
It requires that ripple be high enough frequency for the area of those outside negative plates to make enough stray capacitance with nearby grounded structures that a few amps can flow through it.

Easy to rule out - 'scope both terminals of both batteries to ground. Look for ripple in the kilohz range at various inverter loads. Non sinewave load (like modern appliances) can affect ripple current on the DC side. .

If you can scope voltage drop along a few feet of wire, or make yourself a shunt, you can look directly at current. #10 is one milliohm per foot.

Direct current HAS TO flow through all the cells of both batteries.. HF ripple doesn't.
One cell of a battery: (Did i get all seventeen plates?)

i said it was esoteric...

Test of 'heat following electron flow' hypothesis:
In failed car batteries-
is it always the cell adjacent positive post that fails?Good Luck,old jim

 

[BernieM]

Anorlunda

Regarding the myths around mixing battery types and ages, most of what people say regarding the maximum number of strings or the ages of batteries in a string, capacities, etc., are mostly not true or conditionally true.

Apparently you believe some of those myths. Here is a paper that should address at least the ones you have expressed here in this thread:

http://battcon.com/PapersFinal2002/McDowallPaper2002.pdfAgain, at some point, when I got these batteries, I seem to recall that I saw or was told by someone that they were in fact deep cycle. I will have to clean all the labels off now and see if there are some that are deep cycle. Regardless, even starting batteries can be used, I just have to be careful not to cycle them as deep, and of course they won't have the same life span a deep cycle will. They have plenty of capacity for what I am doing for now. When they die I will replace them with better batteries more suited to the purpose.

JIM, My time in a day for this is very limited unfortunately, but I did set up the scope out by the batteries, but haven't had a chance yet to do everything you asked. I see only a 60 hz waveform, approx. 2v p-p. Plates would have to be pretty large for 60hz to be the issue wouldn't it? Nothing else. It's super clean up to 1 MHZ. I am using a Hitachi V1150 scope with input probes rated up to 100mhz. Scope is good to 150mhz. So it's overkill for khz. Again, nothing in the khz range at all. zero. zip. nada.

I am familiar with the problems with solid state devices and the power factor problems they cause. This inverter actually displays PF (power factor) on it's display, and the occasions I have checked and noticed the PF, it has been in the .9x range. However, I do have a line conditioner between the inverter and most of the house (except for the lines that feed the a/c's,) so it should eliminate most problems coming from my devices attached to the inverter. The a/c's are not on the conditioned line due to max amperage limits of the conditioner. The conditioner is a Tripplite.

 

[jim hardy]

I see only a 60 hz waveform, approx. 2v p-p. Plates would have to be pretty large for 60hz to be the issue wouldn't it? Nothing else.

Good. Looks like that one's nearly ruled out. Thanks !

60 hz not 120 ? Curious. Sine or spikes ? Narrow spikes have high frequency content. But i expected them at frequency of a pwm synthesizer, ten khz or so.

So, maybe that thing is a step wave with only low frequency switching(as i think you said earlier) .

Troubleshooter's last stand: When nothing will show itself bad you have to make everything prove itself good.

one by one.

2 volts p-p is about 0.7 rms,
Seems curious the DMM didn't pick that up.
It's roughly 3% of your nominal 24 volts.

Ripple usually comes from the battery charger but yours has to be coming from the load, obviously.

Given that typical recommended ripple is ½% of nominal voltage and yours is 3%, 6X that

In the spirit of troubleshooter's last stand,
it's something to look into and make sure you're okay.
And resolve that difference between meter and 'scope.

When all those little mysteries are understood you're getting to know your system.
I lived thirty years with a power plant control system. Believe me, Mother Nature loved to humble me but i learned well from her regular floggings.
Pride is the enemy of learning.
One of those lessons was battery ripple from inverter loads. I had over 75 volts p-p on a 130 volt battery. So i might be oversensitive to the subject.

anyhow, to get back on topic

Here's a snippet from a sort of relevant article on the phenomenon
of course it's speaking to charger ripple not load induced ripple

article is at http://www.cdtechno.com/pdf/ref/41_2131_0212.pdf

another article that's long on pictures and short on words is here
https://www.victronenergy.com/live/_media/ve.bus:4._ripple_in_a_ac_battery_system.pdf
Between the two i think it's comprehensible

I sent East Penn an inquiry asking about ripple current rating of their 8D's, got no answer. Used their "contact us" button on the page. Maybe if you call up their sales guys and explain you have an array of twenty big ones that are dropping like flies* maybe it'll get their attention.
Given your high ambient temperature coincides with your airconditioning load time of day, you don't need any extra heat in those batteries.

Good Luck

old jim

* mild exaggeration is sometimes expedient

 

[BernieM]

I have been working with solar in one form or another for over 40+ years. A lot of that high temperature steam power generation using concentrators, so that doesn't apply to this of course. But I moved into a shack (literally) about 3 years ago, where this is no power (and nowhere near here to get it if I did want grid power.) I started out here with about 400 watts in 40 watt solar panels, and 4 deep cycle RV type batteries (group 27.) Later I increased that to 1KW+ and barely survived a summer (using a generator a significant portion of every day during the heat of the day. Went through 3 of them!)

The system I have now isn't very pretty as it has had to be installed on a budget and in a hurry at times (getting up at the break of day and trying to get panels installed, wiring done, supports built, etc., before the sun made it impossible to work.) Of course all of it needs to be redone actually, because it was done in stages and done cheaply or whatever, and once up, hard to justify tearing it all down and redoing it all.

But it does work and most of the time works very well. Other than this battery issue from time to time, I am never without power, or too low on power to do what I want or need to do.

The voltage on the system when I checked the a/c output and read 2 volts a/c was 28.8v.

Didn't I mention that this inverter is low frequency? They use high frequency in inverters so that they can use smaller inductors. That makes them cheap. This one uses a huge torroid, so they don't need to jack up the frequency. They can work at the output line frequency. However, the ACTUAL frequency that this inverter works with, I do not know for sure. It is possible it is 120Hz. I will check again on the frequency, but I am pretty sure it was 60hz. Regarding the waveform, guess in my editing of the last post, I accidentally deleted the part where I said sine wave. It is a sine wave. A very clean, well shaped sine wave.

There is one thing that just thought of. One of the indications I get just prior to a failure in a battery, is that my inverter will start whining about being overvoltaged. It's overvoltage point is 30.5v. Even in cases where I set the max. charge voltage to be say 27.2 volts, even this low, when a battery is in the throes of dying, the inverter sounds off and I have to blow off some power fast from the array (run microwave a minute usually) or the inverter will fully shut down after 1 minute in this state.

The no load voltage of my panels is 21.8v (42.6v in series.)

When this happened, I had it figured that when the batteries are fully charged, with nowhere for power to be stored, that the voltage would rise faster (near instantaneous) than the PWM charge controller I was using would poll the voltage and be able to turn off power from the array, and this was the cause. If this is the problem or related to it, now I guess the question would be: is the inverter seeing higher than 30.5v caused by the failing battery, or are intermittent short peaks of voltages higher than 30.5v causing the battery to fail? Every time I have ever checked battery voltage with a meter however, unless I am equalizing the batteries, I never see a voltage more than the charge controller is set at. But then the meter does not respond instantly either. Perhaps I should look for it on the scope or get the old analog meter out! I actually figured that it might even be an idiosyncrasy of the inverter iteslf (haven't ruled this out either.)

Is there a way I can post a picture directly up here without having to post it somewhere else on the internet first? If so I can post some pictures.

 

[jim hardy]

They can work at the output line frequency. However, the ACTUAL frequency that this inverter works with, I do not know for sure. It is possible it is 120Hz.

It probably uses an H bridge to switch that big toroid. (Yes you did say it was low frequency - just that's hard to believe this day&age)
The H bridge switches at twice line frequency, once each half cycle to provide AC to primary. So it takes gulps of current at 120 hz , every time it reverses polarity and that's why i expected 120 hz not 60. That's sort of minor though , significance is it's line frequency ripple not high frequency chopper ripple.

Is there a way I can post a picture directly up here without having to post it somewhere else on the internet first? If so I can post some pictures.

no "Upload" button? Image icon and Photobucket(it's free) works pretty well just you have to log in once a month or your pictures won't link anymore.

DC meters usually average over period of at least a line cycle. SO when set to DC they hide any AC that's present.

If ripple gets really bad your inverter might sense instantaneous overvoltage at a ripple peak? - i don't know. With so little info about it , who knows what those computers are doing ?
Need a readout of the AC that's present on your battery bus to see if this is related to high ripple voltage.
Try a capacitor (about a microfarad) in series with your DMM meter's test lead. Some analog meters half wave rectify so won't work. An ancient Simpson 260 with "Output " jack would be great.

Got a really HUGE electrolytic, like tens of thousands of microfarads ? Hook it across your inverter DC terminals and see if ripple goes down. I used one microfarad per milliamp on my 130 volt inverters you'll need more at only 24 volts. Could be the ones inside your inverter are losing capacity . That you see sine wave ripple suggests there is filtering just not enough of it.Yesterday i stumbled across a CDE article on selecting aluminum caps for this application, taming ripple. Will look for it again later.

Good Luck

old jim

 

[jim hardy]

more than anybody wants to know about aluminum electrolytic capacitors

http://www.cde.com/resources/catalogs/AEappGUIDE.pdf

http://www.cde.com/resources/technical-papers/selectinvcap.pdf

http://www.cde.com/capacitors/aluminum-electrolytic/screw-terminal
i like the CCE 550 series but they're expensive
36D from chemicon is cheaper
Look at ripple current ratings of both in about 100,000 uf size

http://www.cde.com/resources/catalogs/550.pdf
http://www.mouser.com/ds/2/420/1687836876505b6669848e0-1100623.pdf

if you find out you need them.

Might look inside your inverter and see what input filtering they included, for a starting point..

 

[BernieM]

Biggest caps I have here are 60μfd @ 400v. Motor 'run' caps. But I can build one, just have to do some calculations on charge/area and make one from sheet aluminum that I have and plastic sheets to separate the layers. Saw a nice on in a laser setup at a university once. Simple, cheap and easy to build. Will have to dig up the formula to calculate out the square area of the aluminum sheet needed though.

FYI the email for powerjack is powerjack.albee@gmail.com
I have contacted him for scematics for the inverter, or schematic or specifications of at least the input filtering of the inverter. Will let you know if I get anything useful in response.

As I recall, when I opened up the inverter to blow it out of dust (very dusty here so I have to clean it regularly) it seems that I recall some hefty caps toward the rear of the unit where the battery connections come in. Problem is these days in my old age, I am finding out that memory isn't what it used to be.

One additional note here: I did check all the labels and none of them have any indication of being 'deep cycle' batteries on them marked anywhere. But it does bring up a point, what is a deep cycle battery really? Basically it's a battery that uses one of a few different possible lead alloys that make a battery last longer, with thicker plates. But in the terms of application, which is a deep cycle battery? A very robust commercial battery that has a 7 year warranty on it? Or a really cheap RV style battery with a 1 year warranty (like a type 27)? And I know that a lot of people believe that a marine battery is deep cycle battery, but technically it's not a true deep cycle battery either. It will list cold cranking amps, and generally deep cycles do not list cranking amps, just discharge rates at different loads. So it's kind of fuzzy gray area if you get what I mean. The batteries I have are essentially like a marine battery that is kind of a middle ground between a true deep cycle and a starting battery.

 

[BernieM]

I found the reference to the deep cycle application of the Deka 908D batteries. It is considered 'dual purpose' meaning a cranking battery and a deep cycle battery. So it's the equivalent of a marine type battery in that respect. Here are a couple links. The second link shows it to be a 200ah @ 20 amp drain rate battery. This would only apply to a deep cycle application.

http://www.islandwaterworld.com/bro...y-8d-deep-cycle-908d-1730-mca-12v/4,7152.html
http://islandwaterworld.net/iww-files/pdf-description/LeadAcidBatteries.pdf

 

[jim hardy]

Will have to dig up the formula to calculate out the square area of the aluminum sheet needed though.

http://hyperphysics.phy-astr.gsu.edu/hbase/electric/pplate.html

use meters
k typically 3 for everyday plastic http://hyperphysics.phy-astr.gsu.edu/hbase/Tables/diel.html

That's why aluminum electrolytics have such high capacitance. Their dielectric is an aluminum oxide layer not very many atoms thick so it's very very thin giving a tiny denominator.

 

[BernieM]

OK I noticed yesterday that there are two pairs of batteries exhibiting an odd difference. Here is a picture of one, the other isn't quite as bad as this yet. Both batteries are dusty, and neither has been cleaned. The one on the right which is coincidentally connected to the positive lead that connects in parallel with the rest of the array, is dampish. There has not been large amounts of acid or water coming out of the battery, because there is nothing dripping down the side, it's only on the plane of the top. I disconnected the batteries from each other and measured the voltage. They are both fully charged. The voltages are exactly the same. However, I think this is an early stage prior to failure in the manner that I have been talking about. That left uncorrected (whatever the problem is) that in the near future this battery will fail.

I checked the water level on it and it is not significantly different, if any really. So not much acid or water has come out of the battery. I am betting it is water vapor that has vented out across the battery while equalizing.

I have been doing a lot of research lately, and I have found some references to differences in charge rate of batteries in series, caused by internal temperatures. The positive connected battery may be heating up just a bit more than the second one, due to the effect of heat flowing from positive to negative as mentioned in the peltier effect. So heat generated by current will start at the positive and migrate toward the negative post. This small difference in temperature when temps are in the 90F temperature range, makes a large enough difference between the two batteries that while charging one will charge a little less than the other, amplifying any small difference to begin with. Over time the difference is cumulative, and eventually gets into a condition possibly of runaway. The a/c ripple may be another factor involved as to why one battery is heating possibly more than the other. However since I haven't seen any high frequency, I'm not totally sold on that yet. Will be checking into that further in the coming days.

Here is a post that describes the same issue I have experienced:
http://www.continuouswave.com/ubb/Forum6/HTML/003774.html
http://www.continuouswave.com/ubb/Forum6/HTML/003774.html
So, I think the bottom line is that there is nothing perfect, so the cells all have minor differences in impedance. This minor difference becomes amplified with time as the battery is cycled. At one point it becomes large enough to become a problem and cause the failure, and that heat is the major factor, whatever it's source.

 

[jim hardy]

Well now you're sounding like an old troubleshooter.

Conduction across the top of a battery , through acid slime between its terminals, is a load on that battery just as surely as a would be light bulb. It'd steal charge from that battery and not its series partner. One would expect though to see bubbling if it got as high as an amp ?

Try your dmm on current scale, one test lead to a post and other into that dusty stuff . Wave it around to see if there's conductive spots. That's ruling out long shots, part of troubleshooter's last stand.

However since I haven't seen any high frequency, I'm not totally sold on that yet. Will be checking into that further in the coming days.

we never resolved why your 'scope reports different than your DMM. Some DMM's won't read small AC in presence of substantial DC , that's why i suggest try a small capacitor in series with test lead.

Keep after it...

 

[sophiecentaur]

But I can build one,

(Big Capacitor)
I would be interested to know how well you manage to do that. I guess if volume is not a problem, you could hope to do what the manufacturers manage to do in reasonable sized packages. It's quite a problem to roll two foil strips and two dielectric films with many tens of turns and to get them even and undamaged. Why not just buy Capacitors, anyway? The film and the foil are pretty expensive to get hold of, in any case.

Conduction across the top of a battery , through acid slime between its terminals, is a load on that battery

I visited a Post Office Telephone Exchange in the late 60's and the batteries in the battery room were all sparkling clean! The terminals had clean grease on them too.

 

[BernieM]

Well now you're sounding like an old troubleshooter.

Conduction across the top of a battery , through acid slime between its terminals, is a load on that battery just as surely as a would be light bulb. It'd steal charge from that battery and not its series partner. One would expect though to see bubbling if it got as high as an amp ?

Try your dmm on current scale, one test lead to a post and other into that dusty stuff . Wave it around to see if there's conductive spots. That's ruling out long shots, part of troubleshooter's last stand.we never resolved why your 'scope reports different than your DMM. Some DMM's won't read small AC in presence of substantial DC , that's why i suggest try a small capacitor in series with test lead.

Keep after it...

No current flow either to negative or positive post in the wet dust.
Checked DMM again. Still the same result. I'm betting the current is too small for the meter to detect the frequency. When I check a/c current, it shows zero.
So it can be seen on the scope but not a meter. I also tried my clamp on ammeter but it showed no a/c as well.
I'll disconnect the + lead and put the meter in line to check a/c amps and see if it reads that way. That's in my mind the best way to check it anyhow. It will show current flow regeardless of freq if there's a/c on the line, as long as it's significant enough. Though I don't know what the frequency response will be on the meter in the a/c current mode, but I bet it's higher than 3000hz. And the current can be tested into the μa with the meter I have.

(Big Capacitor)
I would be interested to know how well you manage to do that. I guess if volume is not a problem, you could hope to do what the manufacturers manage to do in reasonable sized packages. It's quite a problem to roll two foil strips and two dielectric films with many tens of turns and to get them even and undamaged. Why not just buy Capacitors, anyway? The film and the foil are pretty expensive to get hold of, in any case.

I visited a Post Office Telephone Exchange in the late 60's and the batteries in the battery room were all sparkling clean! The terminals had clean grease on them too.

A flat capacitor is easy to make. Just make two bar clamps made of copper or aluminum and layers of aluminum foil or sheet between the two clamps, alternating which clamp they are clamped by, each sheet coming just short of touching the contact/clamp on the opposite side. Place a sheet of thin insulating material between each conductive layer. For more capacitance add more layers. Attach wires to the two clamps. Instant non-polarized cap. Pretty simple really.

That was a Telephone Exchange. This is a shack in the desert with batteries on a screen enclosed porch. Lots of dust. I don't use grease I use a battery spray that is conductive and inhibits corrosion. If you have OCD and dusty batteries bother you, you are welcome to come here and clean these daily! I hope you can handle 110F+ temps though!