UPS Battery Failure

[bertramh]

Hi

I have a 40KVA ups and it's been three years since installation,the batteries are ten year design but they have failed instantly. The specs. on the are 90AH 12V in a serise combination giving 420VDC. I measured the ripple current and voltage and they are as follows 1.9A ac and 3.9V ac, IS this too high and if so do you have any documentation or could you point me to some useful liturature on ripple current?

B.F.H.

 

[advidana]

I would contact the people that makes the UPS as some as possible. have them check out your ups and batteries.

 

[Laplacian]

C&D has some excellent information on this topic:

http://www.dynastybattery.com/cd_dyn/contact/tech_support/techsupport.htm

I've done a lot of research on this problem myself and the values you stated for AC ripple seem very low. C&D's recommendation is no more than 5A (AC rms) ripple current per 100A-hr of battery capacity. Ripple voltage should be less than 0.5%rms of total DC float voltage.

Here is C&D's documentation on ripple:

http://www.dynastybattery.com/cd_dyn/contact/tech_support/2131.htm

How did your batteries fail? Did they fail a capacity test or fail to provide backup? VRLA (sealed) cells are notorious for opening without being detected until called upon to provide power. That is a major disadvantage for such a high DC bus voltage - you need many more cells which means that much more opportunity for that failure mode.

Have you verified that your float and equalize voltages are calibrated correctly (per battery manufacturer's recommendations)? I'm assuming you have sealed type batteries - equalizing is not good for them (it is beneficial with flooded or wet type cells) since the electrolyte is in a suspended medium and water can't be added.

High temperature is also a killer for batteries, though I'm sure this UPS is in a climate controlled room. There are many other factors that contribute to poor sealed battery performance in addition to their inherently inadequate design.

Please provide more details about your specific application and battery operating conditions such as float voltage, nominal battery charging volts per cell, ambient temperature, etc.

 

[bertramh]

No the batteries failed while there was a "brown out". The float voltage is calibrated as per spec. and they are C&D Dynasty VRLA batteries. The room temperature is 77deg.F and all other conditions are acceptable, I was told by one vendor that this is a good battery but judging from what happened I beg to differ.

 

[alehman]

Have you done cell internal impedance tests? This test will frequently turn up bad cells before they are called upon. It can be done "on-line" and is good to do quarterly or semi-annually. Also consider an annual discharge test. In my experience it is common for VRLA cells to fail well before their stated warranty period. In spite of Laplacian's statement and Dynasty's paper, the ripple current seems high for this size system during float condition. Your ripple voltage is 0.9%, which seems a little high.

Were you able to find the failed battery(s)? How many failed?

 

[hmchi]

Almost all static UPS failures result from battery failures. The static UPS has the battery always 'on-line' thus cannot be equalized. They may have seen many charge and discharge operations which are not known or recorded. And yes, when you need it most, you then found out it is unceremoniously dead !

This is why the UPS industry to looking to go 'off-line', where the loads are fed from the utility directly during normal operation without the UPS being involved, yet within 4 to 6 milliseconds of a power disturbance, the batteries come in through inverters to pick up the loads.

This way, the batteries can be better 'taken care of' when 'off-line', and has a chance to really contribute when they are needed.

A lot of the UPS for PCs are of the off-line type. They are not always of good design, but they could be. The important thing is for the manufacturer to pay attention to taking care of the batteries [such as equalize charging them during off-line times]. Most UPS people pay a lot of attention in the 'boost mode' [when they take power from batteries to serve the emergency loads], yet precious little efforts in looking after batteries.

 

[stevenal]

Fives years seems about right for a Dynasty in my experience. Might want to set up a program to catch the bad ones before the outage, like the impedance testing mentioned. The failure did not occur at the time of the outage, it just wasn't noticed until then.

 

[alehman]

hmchi - your statement regarding on-line UPS's being unable to equalize batteries is incorrect. Most UPS's from the major manufacturers provide a battery equalization function. However, most VRLA battery manufacturers do not recommend routine equalization. Equalizing VRLA's can lead to thermal runaway if not carefully controlled.

 

[bertramh]

Thanks much for all the assistance and guide lines is has been an enlightining one. You know what amazes me, the fact the vendor used his method of determining the good batteries and putting them in service and conducting a load test the 440V instantly disappeared from the fluke. It appeares the entire 2 strings of 35 battery is "dead".
What is one of the better battery testing analyzer on the market?

 

[jbartos]

Suggestion: The original posting focuses on the UPS internals that are manufacturer's responsibility and may be proprietary. Visit

http://www.mesta.com/PDF/UPS3500 Specs.PDF

for typical UPS data offered by a UPS manufacturer.

 

[alehman]

This is only one of several good battery testers available.

http://www.albercorp.com/ProdPageCellcorderCLC200.htm