Pitfalls on substituting larger Ah battery pack (with same voltage) on UPS?

[pqass]

I have an APC UPS (specifically a "Back-UPS XS 1300 LCD") that has started chirping which usually means battery replacement time. It takes two 12V8Ah lead-acid batteries in series (https://www.upsbatterycenter.ca/apc-back-ups-xs-1300va-lcd-bx1300lcd-compatible-battery-pack). 

However, I happen to have two surplus 12V18Ah batteries (from an APC Smartups 1500, https://www.upsbatterycenter.ca/apc-smart-ups-1500va-lcd-120v-w-network-card-smt1500nc-compatible-battery-pack).

Both packs are 2 x 12V (24V) nominal, but differ in Ah.  My load is about 120W.

I've extended the [short] internal red/black battery wires with additional 12in of 18AWG stranded wire (heater cord) to the larger 18Ah pack now sitting next to the UPS.

The only difference having the larger 18Ah pack I see is:
   a. tolerating a lower charging rate to the larger pack, and
   b. getting a longer on-line time bonus.

Was this wise?  Any pitfalls? 

FYI: The UPS seems to eat packs every 3.25 years.  This latest was my 4th in 13 years!
Also, the UPS seems have a case temp. of 45°C which I think may be the cause of their early demise.

[NiHaoMike]

18AWG sounds really thin for such high current. Should use wires at least as thick as the stock wires, preferably thicker. Other than that, you might want to add a large capacitor where the old battery was in order to keep the high frequency impedance low.

Also check the float voltage, some APC units have it too high or too low.

[pqass]

18AWG was the only stranded wire I had to hand. It should be good to 10A.
With a 120W load, 1A is drawn at 120VAC.  At 24VDC, it should be 5 times that, plus losses. 
So it should be good on a temp. basis.

The short internal wires seem to have more insulation so it's hard to tell if it's 16 or 14AWG; it's pretty stiff though.

The large pack is floating at 26.7V now; 13.4V per battery seems right.

How large a cap?

[elecdonia]

I doubt that changing the Ah of the batteries will make much difference as long as both batteries are the same Ah. Lead-acid batteries are charged to a specified terminal voltage, not by current. Larger Ah batteries can accept larger maximum charging currents than smaller batteries. But that isn't an issue here because you are increasing the Ah of the batteries. It might be an issue if you wished to use batteries with a smaller Ah than the originals.

I have a related question: Is there any need for a pair of resistors (one across each battery) to equalize their "float voltage" when fully charged?

I have two APC "Back-UPS BP1100" units that I recently restored. Both are from the early 2000's and both required replacement of several small electrolytic capacitors on their main PC boards. My BP1100 units contain a pair of 12V 12Ah SLA lead-acid batteries wired in series.

Both of my BP1100 units have  a "fully charged" DC voltage across both batteries of exactly 27.40 V. This corresponds to 13.7V per battery. This is within the specified range of 13.5-13.8V "fully charged for standby use," as labeled on the batteries themselves.

However I'm measuring an unequal division of this 27.4V between my 2 new batteries. One battery measures 13.16V while the other battery measures 14.24V.

Does this overcharge one battery while undercharging the other?
Or is  this difference  too small to  matter?
Will it gradually go away after operating the UPS for a week or two? (I haven't checked  this yet)

I'm requesting comments/recommendations from forum members who have worked with UPS units containing multiple 12V batteries connected in series.

-E

[duak]

A tangential caution. The UPS's charger may not be able to supply enough current to charge the larger battery after a major discharge.  It it's not current and/or temperature limited, could it overheat?  Bigger batteries have a lower ESR and the charger may not be able to handle it.

A friend of mine that runs a cabinet shop picked up a walk behind forklift for a song, but didn't get the right charger with it.  Even though it was rated for the voltage, it was too small and always struggled to supply even 1/20 C for that battery. (500 AH?)  The charger was of the ferroresonant type with intrinsic current limiting and is supposed to be as reliable as an anvil.  The transformer finally croaked so I cobbled up something with another transformer until he decides to get a proper charger.

[elecdonia]

A tangential caution. The UPS's charger may not be able to supply enough current to charge the larger battery after a major discharge.  It it's not current and/or temperature limited, could it overheat?  Bigger batteries have a lower ESR and the charger may not be able to handle it.

The majority of UPS units have current limiting in the battery charging circuit. The battery charges at a constant current up to a specified terminal voltage, and after that it is supplied with a constant voltage.

For most 12V SLA batteries of the sizes used in popular UPS units:
The maximum charging current is from 2A to 5A. This is often labeled on the case of the battery.
The "constant voltage" trickle charge for standby usage is 13.5-13.8V

-E

[james_s]

It works fine, I've seen someone connect a pair of 75Ah batteries in place of the original 9Ah battery in a UPS. The pitfall is that it will take much longer to recharge. Most small UPS's are limited to about 20W or less from the charger and will take close to 12 hours to fully charge the original batteries. If you double the Ah rating of the battery you will roughly double the time it takes to recharge. That said, the recharge time depends on how much energy you draw out of the batteries, if the outage is short enough that the original battery wouldn't have completely drained then the recharge time will be about the same.

One other consideration depending on your use case, most UPS's use heatsinks that are solid chunks of aluminum with no fins. They rely on thermal mass rather than convection and this works because the battery provides an upper limit on the amount of energy that can be supplied. At low load this shouldn't be an issue but if you are loading it up beyond about 50% of rated capacity you should consider adding better heatsinks or a fan.

18AWG wire is too small, if you're going to use that be sure to fuse the battery appropriately for the wire, otherwise the typically 40+ amp battery fuse in the UPS is likely to not blow in a fault condition and that 18AWG wire will be your fuse.

[TheMG]

I've extended the [short] internal red/black battery wires with additional 12in of 18AWG stranded wire (heater cord) to the larger 18Ah pack now sitting next to the UPS.

With batteries external to the UPS, you should install a fuse at the battery pack itself, suitably rated for the wire size in use. 10A for 18AWG. I would place the fuse at the series connection between the two batteries. Keep in mind that a fault internal to the UPS, such as a shorted inverter transistor, can draw considerable current and from my experience the fuses internal to the UPS don't always blow as the fault current may not be sufficient, especially as the batteries age, but would still be plenty enough to make 18 AWG wire go up in smoke in an instant!!!

Ideally, I would use larger wire. The internal wiring in these UPS is normally 10 AWG. Keep in mind as well that the added voltage drop due to the longer, smaller wire will cause the UPS to reach it end-of-discharge voltage sooner, resulting in reduced backup time.

FYI: The UPS seems to eat packs every 3.25 years.  This latest was my 4th in 13 years!
Also, the UPS seems have a case temp. of 45°C which I think may be the cause of their early demise.

Temperature would have a part to play, definitely. It is said that every 5 degrees above 25 (actual battery temperature, not the case temp of the UPS) reduces the life of the batteries by approx half.

However, the charging methodology also plays a significant role. Elevated float voltage especially if the float voltage is applied all the time, in combination with the batteries being at a relatively high temperature, will definitely lead to their early demise.

A lot of the better UPS apply a temperature compensation factor to the float voltage, and will also cease charging entirely after the batteries have been on float charge for a prescribed amount of time. UPS using such a charging method in my experience have no trouble getting 5 or more years out of a set of batteries.

The only difference having the larger 18Ah pack I see is:
   a. tolerating a lower charging rate to the larger pack, and

Charging rate will be A LOT slower! APC specifies a recharge time of 16 hours for that model of UPS. It only has a 9 watt charger. It's going to take more than 32 hours to fully recharge those larger batteries from a full discharge!!!

One other consideration depending on your use case, most UPS's use heatsinks that are solid chunks of aluminum with no fins. They rely on thermal mass rather than convection and this works because the battery provides an upper limit on the amount of energy that can be supplied. At low load this shouldn't be an issue but if you are loading it up beyond about 50% of rated capacity you should consider adding better heatsinks or a fan.

I do believe that is the case with this particular model, and most other smaller UPS in a plastic case without the built-in ability to connect external battery packs. Normally the batteries would run out of juice before the chunks of aluminum got excessively hot, but with larger batteries you might have long enough run time to actually heat up the inverter transistors to the point of failure.

Adding a fan would not be a bad idea. You could set it up so it only comes on when the UPS is on batteries.
 
Ultimately though, I would just look for a second-hand UPS that has proper support for external battery packs. There's a lot floating around on auction sites for cheap that just need a new set of batteries. For example, any of the APC SmartUPS models with "XL" in the model number have a connector for external battery packs. Eaton Powerware 9125 or 9130, those are very good UPS. You can score any of those for under $100 if you look around. I myself have a Powerware 9125, 1500VA, with a set of 4 77Ah batteries now 6 years old and still going strong, get HOURS of backup time even with the gaming PC going full tilt.

[TheMG]

I have a related question: Is there any need for a pair of resistors (one across each battery) to equalize their "float voltage" when fully charged?

However I'm measuring an unequal division of this 27.4V between my 2 new batteries. One battery measures 13.16V while the other battery measures 14.24V.

Does this overcharge one battery while undercharging the other?
Or is  this difference  too small to  matter?
Will it gradually go away after operating the UPS for a week or two? (I haven't checked  this yet)

I'm requesting comments/recommendations from forum members who have worked with UPS units containing multiple 12V batteries connected in series.

Normally the voltages will equalize after the batteries have been on float charge for long enough. I've worked with large UPS systems that have as many as 40 batteries in series, and on every maintenance check they're always within about 0.2V of one another or less. No resistors or anything fancy going on to balance the voltages. The UPS doesn't even know what the individual battery voltages are, only the total series string voltage.

That is, of course, provided that all batteries are identical and manufactured in the same batch, and that they are all healthy. Also, batteries of poor quality could also have manufacturing inconsistencies that lead to unequal voltages.

Give it a few days and see if they start to equalize. If they don't and there remains a big difference between the two, it may be an indicator of a faulty battery.

[sibeen]

FYI: The UPS seems to eat packs every 3.25 years.  This latest was my 4th in 13 years!
Also, the UPS seems have a case temp. of 45°C which I think may be the cause of their early demise.

Yes.  Basic rule of thumb for a VRLA battery (sealed) is that for every 10°C rise above 25°C you halve the life of the battery. Your batteries aren't going to be sitting at the UPS case temperature but they are in all likelihood going to be sitting well above 25°C. Many years ago I worked on a UPS system that was kept in a room that was constantly at around 50°C and the batteries wouldn't last a year.

The temperature rise rule of thumb comes from the Arrhenius equation ad relates not to just sealed lead acid batteries.

[TheMG]

Many years ago I worked on a UPS system that was kept in a room that was constantly at around 50°C and the batteries wouldn't last a year.

I've seen something like that before. Someone had the bright idea to put a network rack (with UPS of course) in a boiler room!!!

[duak]

Actually, I did the same thing.  I've got an older little APC unit (200 W maybe) to run a router and maybe a laptop.  The powerco scheduled a day long outage so I paralled a charged battery of about 3 times the AH to extend the time.  It worked so well I just left it.  After a few months, I noticed the UPS would randomly shut off without any power disturbance.  I also noticed the UPS was quite hot and so was the external battery.  Out of the unit, the internal battery was down to about 8 V and wouldn't accept a charge.  It was 3 years old and perhaps on the edge of old age (or maybe C19 got it).  Perhaps the heat killed it but good.  I haven't got a new battery or perform a post mortem to see if the UPS still works.  I can't remember what shape the external battery was in - it was over ten years old anyway.

Bottom line: check out the charger to see if it can safely supply the charging current and doesn't overheat.  Sometimes assumptions apply to every case, but sometimes they don't.

[sibeen]

One other caveat I'd add about adding extra battery capacity is that on some of the APC models, especially the cheaper ones, the inverter was only sized to run the load for the short time that the original battery was rated for; so the thermal capacity of the inverter was designed to run from 7 to 10 minutes at full load. The thermal capacity was crap in other words and the heat sinks on the switching devices were the cheapest possible. I'd give it a test and monitor the temperature on the output transistors and perhaps bulk up the heat sinks if approaching a problem temperature.

[Circlotron]

One other caveat I'd add about adding extra battery capacity is that on some of the APC models, especially the cheaper ones, the inverter was only sized to run the load for the short time that the original battery was rated for; so the thermal capacity of the inverter was designed to run from 7 to 10 minutes at full load. The thermal capacity was crap in other words and the heat sinks on the switching devices were the cheapest possible. I'd give it a test and monitor the temperature on the output transistors and perhaps bulk up the heat sinks if approaching a problem temperature.

Yep, that, also the thermal mass of the inverter transformer may only be enough to supply full load until the original battery quits. If that is the case, a big battery will kill it. Kinda like the tyres on a Bugatti Veyron will only last 15 minutes at full noise, but the fuel will only last 12 minutes. 

Another thing to consider is electrical safety. How good is the isolation between mains and the battery terminals? When the battery was completely inside the UPS nobody could touch it, so isolation could afford to be so-so. When outside it is a different set of requirements.

[james_s]

Yep, that, also the thermal mass of the inverter transformer may only be enough to supply full load until the original battery quits. If that is the case, a big battery will kill it. Kinda like the tyres on a Bugatti Veyron will only last 15 minutes at full noise, but the fuel will only last 12 minutes. 

Another thing to consider is electrical safety. How good is the isolation between mains and the battery terminals? When the battery was completely inside the UPS nobody could touch it, so isolation could afford to be so-so. When outside it is a different set of requirements.

As long as you run it well below the max rating it should be ok. I would recommend doing that anyway, the max rated capacity of consumer grade UPS's is absolutely brutal on the batteries. For example I have a little 550VA/300W unit that takes a 5Ah battery. I tested it with a 150W incandescent load and it pulled around 11A from the battery and that increased to 15A by the time the battery was nearly depleted. At the full 300W that's around 22-30A from a 5Ah battery, abusing a SLA battery with a continuous 4C-6C load will kill it after just a handful of cycles. I try to size a UPS for around a 30 minute run time unless there's a backup generator and the UPS only needs to carry the load long enough to get that running or just long enough to shut down the PC.

[elecdonia]

I'm measuring an unequal division of this 27.4V between my 2 new batteries. One battery measures 13.16V while the other battery measures 14.24V.

Normally the voltages will equalize after the batteries have been on float charge for long enough. I've worked with large UPS systems that have as many as 40 batteries in series, and on every maintenance check they're always within about 0.2V of one another or less. No resistors or anything fancy going on to balance the voltages. The UPS doesn't even know what the individual battery voltages are, only the total series string voltage.

That is, of course, provided that all batteries are identical and manufactured in the same batch, and that they are all healthy. Also, batteries of poor quality could also have manufacturing inconsistencies that lead to unequal voltages.

Give it a few days and see if they start to equalize. If they don't and there remains a big difference between the two, it may be an indicator of a faulty battery.

That's helpful information. Thank you!

After a couple of weeks I will retest the float voltage on my two batteries.  If a substantial difference in float voltage persists then I will update this thread.

BTW, these are brand-new Interstate Battery brand SLA batteries, 12V, 12Ah. They were purchased at the same time.

-E

[pqass]

Thank you everyone for your input.
I've summarized the issues identified so far and my response below.

18AWG WIRE TOO SMALL:
I've since upgrade the wiring to 10AWG solid; I don't have the equivalent stranded wire.
I will get another Anderson grey, gender-less 50A connector to match the one that came with the 18Ah pack to allow for easy separation rather than needing a screw driver.

See attached photo of current draw while on-battery. It actually draws more than my original estimate of 5A+losses.

ADD FUSE:
There is an existing 100A fuse between the pack batteries opposite where the wires are connected. I will lower it to 25A.

HARDER ON CHARGER:
Using a clamp meter on the positive battery wire, with the UPS on-line, I see it toggling between 160mA (5sec) current draw and 510mA (<1sec) charge. This should not abuse the 18Ah pack nor the charger.

Each battery in the 18Ah pack was charged individually with a smart automotive charger to full before I connected it to the UPS 10 days ago. Each battery is still virtually at the same voltage (13.26V and 13.27V, while connected to the UPS and measured during its 5s draw interval). This pack is not new and I haven't tested it until UPS cut-off but it will do at least 1 minute on battery which is all I need. When it does eventually go, I'd rather replace it with the larger 18Ah pack where currently the 10A on-battery draw will be less hard on it vs the originally spec'd 7Ah pack; thus should increase the time between replacements.

LONGER CHARGING TIME:
As per TheMG, the charge time for my UPS is normally 16hrs (on a dead 7Ah pack) but will be extended to 32hrs (on a dead 18Ah pack). This is not a concern as I don't intend to run on-battery for long; most of the load will be shutdown via apcupsd after about 1 minute; currently the front panel says I have 7 minutes while 20% loaded (150W). So, I should get at about 14 minutes on a new 18Ah pack.  My primary motivation is to save money in the long run; not extend on-battery time.

HEATSINK THERMAL MASS (OR THERMAL BUDGET, IN GENERAL) SIZED TO THE AMOUNT OF ENERGY IN THE PACK:
The UPS has a rear exhausting 30mm fan (it's off when on-line, it's on when on-battery), unattended load is 20W, attended load 120-150W (which will be shut down within a minute); the UPS is rated 780W max. I purposely bought more than my real need so I shouldn't be approaching the thermal budget even if I extended the on-battery time by double.

Last year when I disassembled the unit, the transformer core (4" square 2" thick beast) was quite hot to the touch. This appears to be steady state for this UPS. I don't remember the size of the heat sink for the output transistors.  But I have since mounted a 50mm fan sucking out air from the top vent.

The case temp of 45 degrees was measured with a infrared thermometer on the side of the case where the transformer is located. The case temp of 30 degrees was measured on the side of the case where the original pack would be. Keeping the pack external should extend its life.

SHOULD SHROUD WIRES; possible mains line voltage on battery terminals:
Battery negative is grounded; reads 0 ohms to back ground screw.

[james_s]

It shouldn't be harder on the charger, the charger is current limited, the battery cannot draw more current than the charger is willing to give it.

With fusing 25A for 10AWG ought to be fine, just remember that the fuse is to protect the wire, or whatever is weakest. The fuse should always be the weakest link in the chain so that if anything goes wrong, the fuse blows before anything else does.

[TheMG]

The fuse should always be the weakest link in the chain so that if anything goes wrong, the fuse blows before anything else does.

When it comes to UPS, the manufacturers normally choose the fuse ratings to protect the wiring and battery packs, nothing else. The components in the inverter circuit will almost always burn open before the fuses in the event of an actual UPS fault. I have yet to see a UPS actually blow its battery fuses. Usually one or more MOSFETs/IGBTs in the inverter will dump its magic smoke in a spectacular fashion.

Some of this is due to the fact that typical designs use many transistors connected in parallel to achieve the desired current capability, but usually not all of the transistors will fail short at precisely the same moment, and the fault current through one or two transistors is insufficient to blow the fuses.

[james_s]

Well, attempting to protect semiconductors from failure with a fuse is rarely successful, but what I really should have said is that the fuse is to fail before anything likely to start a fire. A short circuit that burns up the wiring can cause it to catch fire, melted wiring can short out and result in an even more direct short across the battery which can rupture. If the mosfet acts as a fuse that's sub-optimal but it probably won't result in a fire, just lots of smoke.