connecting batteries in parallel or series, myth and theory

Good advice. Thanks. I just have to figure out how to do it -- shouldn't be hard.

--ron

In Agent, simply be on an article authored by the target, hit ^k and make your selections. Agent and that stuff from microsloth both call the process "filtering". Check your help file for details.

This (I'm posting from AEH) is a great group once a few card-carrying *ssholes are filtered out.

John

-- John De Armond See my website for my current email address

| Here is another way to look at it. Series strings have enough problems | of their own. No two series strings can be 100 percent identical. When | we parallel two or more strings, the chance for problems is multiplied.

So I can build a string of 24 2V cells, or 4 12V batteries, with a given capacity, for about X dollars. I'd be inclined to go with the 24 2V cells in one big string. But, suppose I could build a string of 4 12V batteries that are smaller with only 1/4 capacity, for much less than X/4 dollars. In that case, 4 such strings in parallel gets me the same capacity, but it costs less than X. How much less than X does the 4 parallel strings option begin to look attractive, balancing the economic gains against the technical issues?

FYI, I'm looking at having a 20A loading capacity at 100 hour rate and 48V. The Surrette 2KS33PS (x24) certainly looks like it would do the job well. But I'd rather find something less pricey. It seems multiple strings of

6V or 12V batteries would do the same in capacity at half the price. But is it worth it?

One factor in favor of multiple strings is the ability to take a string out of service when there is a cell/battery issue. If something goes bad in a single string, then it's all down at once. I'd rather be able to replace a bad cell/battery in a derated system than in a shutdown system.

Other options: separate strings to separate syncronized inverters that are then paralleled on the AC side.

Ron, how hard is it?

Kill file is for wimps. But is dead easy. Funny you should find it hard, I assume that it is easier than telling us how you designed your system.

Still glad to be rid of another talker, non walker.

Like wayne you never designed your system.

Bye (I hope)

No gains, just more problems.The economics will cost more in the long run.

Okay, Say you have 4 parallel strings. If you lose a cell in one string you are left with 3 parallel strings. It's no good replacing one cell in the affected string and just as bad replacing that string with a new one. One bad cell in a string and you are up for a new set of 4 parallel strings, or, forever chasing problems related to the different ages of batteries/cells.

So you save money by using 4 parallel strings so you can afford 4 inverters that can be synchronized. Shock, horror, 4 of this type of inverter will probably cost more than a single series string of cells of the correct Ah capacity.

Thanks, John.

Done.

--ron

I struggled with the same problem. My conscience finally let me connect three groups of two batteries. I didn't like doing it and I'm not proud of it. It's never mentioned around the dinner table.

I'm using 225 amp/hr 6 volt golf cart batteries. Interstate model 2200.

63 lbs. They come in different company names. It was the most cost effective solution as the huge two volt cells were going for almost 500 dollars each. These batteries were a bit over a hundred each. (cheaper in the States) It was a compromise, but if I can get 5 to 8 years use out of them I'll be happy. The Surrettes and Trojans were, I thought, overpriced. We will see if these last.

These are what I'm using:

Looks like even Costco sells them I use a 12 volt inverter as anything higher was either not available or more money than it was worth. The banks are connected to 2 inch copper water pipe buss bars, which are fed/tapped at opposite ends in an attempt to balance out the charge/discharge. So far each battery has taken the same amount of water when refill time comes around, which is every three months. Roughly two litres/half a gallon of water are used in total.

The advantage of twelve volts is in the ease of charging and really common automotive accessories. I have a 200 watt 12 volt stereo wired into the shop. I use 12V wherever I can, like in the lighting. No sense turning on the inverter if it isn't needed. The disadvantage is the need for monster wiring and paralleling. I have 2/0 cable feeding the inverter. Voltage drop becomes an issue if your use is any distance from the source.

That makes sense. Some large knife switches could isolate each bank.

These look very nice. I'm sure there are others out there. I don't know if 86 percent efficiency is good or not. As far as I know, the waste may be less in normal stand alone inverters.

For the rest of us:

--=20 __ __ __ __ __ __ __ __ / /\ / /\ / /\ / /\ / /\ / /\ / /\ / / / /\ \/ /\ \/This space for rent/\ \/ /\ \/ / /_/ \/_/ \/_/ \/_/ \/_/ \/_/ \/_/ \/_/

Densa International=C2=A9 'Think tanks cleaned cheap'

Due to the insane amount of spam and garbage, I block all postings with a Gmail, Google Mail, Google Groups or HOTMAIL address. I also filter everything from a .cn server.

If you stick with Surrette, the price is about the same for similar Ah capacities. For example, one vendor will sell equivalent strings of 2KS33P for about 7.3% more than for the 6CS25PS's. But the former has about 5.7% more capacity.

If you start comparing different brands, and I'll use Trojan as an example since it seems to be a popular "cheaper" brand in the US, you need to look at more than just initial purchase price. Given the same DOD, the life span of a Trojan L16 will be less than a 2KS33P. You'd need to check with mfg, but the data I have indicates that at 80% DOD cycles, the L16 is rated at 698 cycles to failure, whereas the Surrette is rated at 2000. At 50% DOD, the Surrette should do 3170 cycles to failure and the L16 1036.

If your system is mostly in standby mode, and you only cycle to 20%DOD, the Surrette should go 5100 cycles to failure, and the L16 only 2322.

So you will either replace the Trojans more frequently, or have to purchase significantly more Trojan capacity in order to obtain the same number of cycles to failure.

You also need to check about connectors and shipping. When I purchased my Surrette's in 2001, shipping was included as were the connectors, and I requested and they sent 4-0 copper for the connections.

If you have a battery bank with a 15 year life span, how often would this kind of event have to occur, in order to "pay for" the extra maintenance involved in having multiple strings.

I've not had to do that with my Surrette's since they were put into service in 2001. My plan, if I ever have to replace a cell, would be to switch the house to my backup generator while doing the work. That involves throwing a grand total of three switches.

That's likely to be your most expensive option.

--ron

The cells are separated because if they were not, being series connected, they would short internally. I've had old rubber-case batteries with cracks between the cells, and they did not work.

You need 3 or more plates (one double sided -) (making 2 "cells" ) before you can parallel anything.

Sandwiched plates can have significantly lower resistance than a long plate and being double sided, can have double the active material area per unit. They can also be significantly more durable, mechanically.

Long plate cells do exist, such as the Cyclon and theOptima (coiled plate)

** Posted from **

If the 2 parallel string solution costs 35 to 50% less, take a chance. If a 3 or 4 parallel string solution is being considered, savings would need to be 50 to 80% to make it worth while to me. The possibility of failure, and therefore the projected cost/hassle per AH goes WAY up with multiple strings

Use 2 SEPARATE strings - one backing up the other, if redundancy is required.

Or just split your loads. Unless a single load excedes the capacity of your "half system" it is much simpler to just run 2 separate systems.

** Posted from **

address

Tellico Plains, Occupied TN

Yeah, run away from the hard questions.

In alt.engineering.electrical snipped-for-privacy@gmail.com wrote: | On Aug 10, 5:09 am, snipped-for-privacy@ipal.net wrote: |> In alt.engineering.electrical snipped-for-privacy@gmail.com wrote: |>

|> | On Aug 9, 6:09 am, snipped-for-privacy@ipal.net wrote: |> |> In alt.engineering.electrical Ron Rosenfeld wrote: |> |> | On 8 Aug 2008 18:32:00 GMT, snipped-for-privacy@ipal.net wrote: |> |> | |> |> |>If the effect is that the _older_ strings gets _older_ faster, |> |> | |> |> | And if it is the newer string that gets older faster? ... Well, maybe you |> |> | won't get the longevity you thought you'd paid for. |> |>

|> |> Right. So we need to know which it is ... which string gets older faster. |> |>

|> |> | In any event, the "best" setup does depend to some extent on the |> |> | application and goals. |> |>

|> |> How about a power backup system (charged by the grid) that gradually shifts |> |> to a renewable power system (charged by solar, wind, etc) and then eventually |> |> to a completely off-grid system (or at least a sell-only grid system). |> |>

|> |> -- |> |> |WARNING: Due to extreme spam, googlegroups.com is blocked. Due to ignorance | |> |> | by the abuse department, bellsouth.net is blocked. If you post to | |> |> | Usenet from these places, find another Usenet provider ASAP. | |> |> | Phil Howard KA9WGN (email for humans: first name in lower case at ipal.net) | |> | |> | Phil |> | |> | People have explained several reasons for "NOT" paralleling batteries. |> | It should be avoided wherever possible. |>

|> If you'd like, feel free to summarize the thread. I plan to look it all over |> again once it seems all the responses are done. |>

|> | You response tells us that you have/want parallel batteries and you |> | are looking for someone to tell you it's ok. |>

|> No. If there is an alternative, I'm willing to consider that. So what would |> you suggest as an alternative? More in series and increasing the voltage? | | A single string of 2V cells of the correct Ah rating. | If there is no other option then limit the parallel strings to 2.

Let me summarize/rephrase, since my intentions didn't pass through the layers of conversation. What is the alternative for paralleling cells/batteries/strings when the desired capacity is NOT available in single cells or small batteries?

If a 24V system needs to be doubled in capacity, and larger cells is not an option, maybe you'd suggest a 48V system, especially if no investment has yet been made in the 24V ssytem. But what if the system is already at 48V? Then what? Even higher in voltage?

|> | So, Yes it is ok. As long as you also accept that it is the second |> | best option and are prepared to take all responsibility for your |> | choice of battery bank. |>

|> It's 2nd best compared to what is 1st best? | | A single string of 2V cells of the correct Ah rating.

As you go up in single cell capacity, other issues come up. These can be things like cost (double capacity single cell might be 3x the cost) and handling (you can't lift the double capacity cell).

What I want to know is not so what _the_ best system design is, but information about the advantages and disadvantages so that I could weigh one design over another. ONE way to explain this might be in terms of the cost of everything. For example, just how much of a savings on smaller cells/batteries makes it worth going that route, in your opinion/experience. If I can build a system of one string of single cells for $30000, and would like to reduce the price, at what price level would _you_ decide it's worth going with parallel strings? $10000?

For me, once I can get a good handle on the _actual_ issues of parallel strings vs. parallel cells vs. parallel batteries, then I could answer the above for myself. Right now I'm getting answers like "batteries obey the laws of physics" without any explanation of what laws apply (there might be more laws that apply beyond the obvious ... and that is crucial to know).

Right now I'm not getting much better answers here than I got from Googling.

But I did get one useful answer that the measurement of a single cell can be masked when another is in parallel to it ... but that just tells me it is better to parallel the strings rather than the cells (which also happens to be a lower cost option).

And maybe some big rectifiers to isolate the strings from cross-charging might be called for. It would then seem to me the only way to keep the strings charged with the rectifiers in place is to separately charge each string. But that might be an economic benefit from smaller chargers.

In alt.engineering.electrical snipped-for-privacy@gmail.com wrote: | On Aug 10, 11:40 am, snipped-for-privacy@ipal.net wrote: |> In alt.engineering.electrical m II wrote: |>

|> | Here is another way to look at it. Series strings have enough problems |> | of their own. No two series strings can be 100 percent identical. When |> | we parallel two or more strings, the chance for problems is multiplied. |>

|> So I can build a string of 24 2V cells, or 4 12V batteries, with a given |> capacity, for about X dollars. I'd be inclined to go with the 24 2V cells |> in one big string. But, suppose I could build a string of 4 12V batteries |> that are smaller with only 1/4 capacity, for much less than X/4 dollars. |> In that case, 4 such strings in parallel gets me the same capacity, but it |> costs less than X. How much less than X does the 4 parallel strings option |> begin to look attractive, balancing the economic gains against the technical |> issues? | | No gains, just more problems.The economics will cost more in the long | run. |>

|> FYI, I'm looking at having a 20A loading capacity at 100 hour rate and 48V. |> The Surrette 2KS33PS (x24) certainly looks like it would do the job well. |> But I'd rather find something less pricey. It seems multiple strings of |> 6V or 12V batteries would do the same in capacity at half the price. But |> is it worth it? |>

|> One factor in favor of multiple strings is the ability to take a string out |> of service when there is a cell/battery issue. If something goes bad in a |> single string, then it's all down at once. I'd rather be able to replace a |> bad cell/battery in a derated system than in a shutdown system. | | Okay, Say you have 4 parallel strings. If you lose a cell in one | string you are left with 3 parallel strings. It's no good replacing | one cell in the affected string and just as bad replacing that string | with a new one. One bad cell in a string and you are up for a new set | of 4 parallel strings, or, forever chasing problems related to the | different ages of batteries/cells. |>

|> Other options: separate strings to separate syncronized inverters that are |> then paralleled on the AC side. | | So you save money by using 4 parallel strings so you can afford 4 | inverters that can be synchronized. Shock, horror, 4 of this type of | inverter will probably cost more than a single series string of cells | of the correct Ah capacity.

Many many years ago I worked at computer services company (as a programmer and system administrator) that had a battery based "UPS" system sized for

5 mainframes. I never found out how the batteries/cells were actually wired. I didn't spend a lot of time in the UPS room and when I did need to go in there to get some readings, the place was HOT! But what I do know is there were 216 "batteries" (which may have been cells). They were in a black case and each was significantly larger than a typical car battery by about 8 to 12 times in volume. The system would drain down in about 25 minutes if we left all 5 mainframes on. If we lost power, the usual procedure was to shutdown 3 of them if the power was not back in 10 minutes, giving us about 45 minutes on the remaining 2. Management was unhappy with the time frames but did not have the room to install a larger system.

Would it be your guess that this system, installed by the vendor that sold it, was a 432VDC system? Sorry, I didn't get any specs on the true capacity and I can't tell you what the mainframes used (but I can say there were 600+ disk drives of mostly IBM 3330 size). I wish now that I had enough interest in it back then (1970s) to have checked up on the power capacities and requirements.

I thought I would allow everyone to get their two bits in before I opened my mouth. So it's time for my view of things.

Most of the myths, or bad advice isn't really bad advice. It is just misapplied advice. Most of the solar energy system installers are at a disadvantage right now. Solar power is a growing field and most of the know how, designs and regulations are not crafted for that type of application.

Battery bank design has been crafted around the emergency backup systems. Like for telecom systems, mainframe backup system, and other mission critical intermittent high load systems. Something to keep things running until the diesels start. The things that are important in that system are of no concern to a off grid PV system.

Lets look at the factors.

Classic application emergency backup. It is monitored by a live person, or multi thousand dollar monitoring system

24/7 The value of the battery bank and it's enclosure is immaterial. The batteries may cost $250,000 but if the system shuts down, it will cost you $2,000,000. Because of that fact, batteries will get changed out as a group every two or three years, irrelevant of condition. Any batteries that show any problems in the mean time, will get pulled as fast as you can spin your head. When they are called on to perform, they don't have to run a load for a day or two, or even three or four hours. The load they are called on to run will drain the entire bank in an hour or two. But all they need is for it to run it long enough for the big MW diesel generators to come on line in about 30 minutes. Once the generators take over, then the battery bank starts it's recharge, and maintenance is dispatched to do a physical check of the bank after the outage event. If a section of the battery bank burst into flames before the generators come on line, it is no concern for the company, as long as they maintain the load long enough for the generators to pick it up. The fire suppression systems, will put out the fire, and the maintenance crew will replace the melted parts.

Lets look at solar power systems. There is no 24/7 multi thousand dollar monitoring system, or live person looking over the bank all day. It just sits in the basement and minds it's own business. Neglected so to say. The battery bank usually cost more than most of the things running off of it. If it fails to supply the load that day, then the only damage is you miss the last part of "three's company" The battery bank is not changed every two to three years. It is changed when it can no longer supply load over a day to day basses. The only time individual batteries are changed out is if they misbehave so badly that it is impossible for the owner to ignore it. It is preferable for the owner to get as much usage out of a set of batteries as he can because they are a big investment. They have to be in the system 24/7 They have to supply a light load for days, or even weeks between a significant recharge. A generator only kicks in to supply the load and recharge them when they hit the bottom of the barrel. No one will be dispatched to inspect them after each usage. If part of the battery bank burst into flames, it is a real problem. No automatic fire suppression system will put out the fire. And the most important thing to the owner will be burnt down by that fire (that being his house). If the load is maintained is the least concern to that owner at that time.

As you can see. Very different factors affecting the design decisions. The classic deign is a set of big batteries of the same vintage connected in series parallel with large cables in a big battery bank to supply mega current to a device for a short period of time. Internal battery problems is of little concern to them because the batteries are always pretty much new. Cables have to be big because of the heavy loads that can drain them in an hour.

In an emergency backup system

Small batteries in parallel with large ones is a no no. Large currents won't drain all the batteries equally.

Fuses in the battery system wiring is a no no. The batteries are not of concern.

Long wires between batteries is a no no. Uneven resistance will cause the batteries to discharge unevenly under high load.

Old batteries in parallel with new ones is a no no. The new batteries will support most of the load, and the old ones will do nothing under those heavy loads.

All those things will reduce the chance of a system maintaining the load. In a solar power system the situation is a bit different.

In a solar power system

Slow discharges/charges will discharge/charge old/new and small/big batteries more evenly than heavy intermittent loads.

Small wiring is a less of a problem with small continuous loads, as log as it is properly fused so the small wiring won't melt if there is a short or malfunction.

Fuses in the battery bank are helpful because it will stop the battery bank from going south and burning everything down while no one is at home. And if they do blow and shut the load down, then it is just a minor inconvenience.

Hooking large/small and old/new batteries in parallel is a perfectly workable battery bank for a solar power system. As long as you properly fuse between parallel strings to stop the new big batteries from melting down the small old batteries if the small old batteries have an internal problem resulting from their age.

The only time you have to worry about mixing old/new and large/small batteries is in a series battery string. They have to be matched in age and size or there is a possibility of one of them reversing polarity in the string during a load cycle, or being overcharged during a charge cycle.

So it should be perfectly fine for you to have a big new matched series string in parallel with a small old matched series string of the same voltage. It will allow you to get the last bit of life out of the old string.

Worst case scenario is the old small string won't contribute to the load, but you won't be any worse off for them being in the system. You just won't be any better off. Real worst case scenario is the old string shorts out internally and blows the fuse that connects it to the rest of the battery bank, and then it's life has come to an end.

Right now, the prevailing logic is a result of people dealing with emergency power systems (the main use for large battery banks up to this point), and they are trying to apply that logic to solar power systems (a relatively new application for large battery banks) with mixed results.

My No1 rule. Design your system around it's application, not around prevailing logic. You will always be better off in the end.

A 2000Ah battery bank supplying a 20A load for days on end is far outside the traditional applications of large battery banks. The factors that go into designing, and the design of that large battery bank will be irrevocably different from traditional battery bank design.

| Use 2 SEPARATE strings - one backing up the other, if redundancy is | required.

[...]

| Or just split your loads. Unless a single load excedes the capacity of | your "half system" it is much simpler to just run 2 separate systems.

That might make more sense.

Assuming all the DC systems are the same voltage, is it safe to do a closed transition switch when changing over, where the transition is several seconds (as in manually switching one on then switching another off)?

Based on what I see about inverters (Xantrex specifically), they are 120V only and 120/240V is achieved by syncronizing two of them and running them in series. So what about _feeding_ these inverters with separate DC systems? Unless they are self-syncronized by some signal on the DC feeds, I would think there should be no issue with this (but I've been surprised by other things before).

Two 12V in parallel for occasional surge like starting is one thing, but a large number of banks in parallel and having frequent charge / discharge is another.

You've heard from folks that parallel two batteries for starting (with/without diodes) and not had any problems.

But for large storage applications, parallel batteries are the second best option, always. If you want to keep all the parallel strings 'up to snuff', you will need a way to separate them and charge/test them individually. When in parallel, it becomes more and more difficult to have them 'share the load' as they age or as you add more parallel strings.

Series connections do result in much higher voltages and you *can* have a problem with reversing a cell. This can be minimized by performing good equalizer at manufacturer's recommendations.

Testing individual sp. gr. and ICV (individual cell voltage) during charging is also a good way to find 'weak cells', regardless of connection.

FWIW, when commercial applications such as large UPS and even larger (submarine batteries are about the largest in the world), the engineers have always opted for special 2V cells that have the right A-H rating and as many such cells in series as needed to get the desired voltage. I've never seen a commercial UPS system (or submarine battery) that was connected in parallel. (maybe they know more about this than alt.energy.homepower posters :-)

Depending on the voltage needed, I've seen UPS systems with 36 12V batteries connected in series (about 430V). Submarines used about 128 2V cells (about 270V).

daestrom

"N9WOS"

Great post!!

You hit on exactly the correct points, and did it in a sensitive way. Still, if my experience is any guide, you may expect some nit-picking negativism, just don't be discouraged by any of it.

Luck; Ken

I see no particular advantages, and I see one disadvantage...unequal discharge on your strings. (unless your 120/240 system is perfectly balanced.)

Vaughn

"N9WOS"