SLA Battery Failure Analysis

[klaxon]

I have a pile of small SLA batteries that failed in a backup application.

My tests indicated the problem is due to prolonged overcharging.

I am engaged in a long term argument with a non-electrical engineer who takes alternate viewpoint that the batteries were UNDERCHARGED. I want to get an external failure evaluation done to establish the fact chemically, just to drive home the point, beyond my own observations and tests.

My experiments were as follows:

Discharge the battery to an empty state based on terminal voltage at 25 degrees C.

Fully charge the battery using constant current + constant voltage profile, verifying the state of charge based on terminal voltage at 25 deg. C, 24 hours later.

Insert charged battery into circuit. Observe charging current that is 100x maintenance charge. (This is a 2300 mAhr battery and once fully charged, should require a mA or two of trickle charge. It is getting between 50 and 300 mA!)

This is a small camcorder battery (Interstate. 12V nom, 2300 mAhr) used as a 100 mSec backup for a computer stack on a vehicle during engine start. It is charged with the vehicle system, except that it is diode isolated. There is no charger to speak of, and there is no thermal compensation or charge monitoring in place. Pretty much a hack by a rank amateur kind of application.

So the question... does anyone know where to get independant failure analysis on a battery like this? See anything wrong with my testing methodology or approach? Any recommendations for articulating this better to manager types?

All replies appreciated.

 

[oldfieldguy]

Your non-electrical friend may be right.

" It is charged with the vehicle system, except that it is diode isolated."

The isolation diode introduces probably a 0.7 volt or so drop in charging voltage.

Measure your charging voltage in the circuit with the diode in series and you might find that you never get a good charge.

old field guy

 

[VEBill]

SLA = Sealed Lead Acid.

If the little battery is basically in parallel with the vehicle battery, except for the extra diode drop, then continuous slight undercharging (by about 0.7 volt) seems like a possibility.

 

[klaxon]

Thanks, guys..

Alternator voltage on this rig routinely goes to 14.5 V.

I considered the diode drop, though that's a good observation.

Again, even with an unambiguously fully charged battery, the measured charge current of 50+ mA that never goes away is what I am focused upon. <= .01C is usual self-discharge trickle current, which would be <= 23 mA.

Last item... on the dead batteries, weight inversely correlates to terminal voltage on the dead batteries... as if electrolyte had boiled off.

Also, this is exactly what I'm getting at... is there a chemical means of determining failure mode? Sulfation versus oxidation on specific electrodes? If so, any idea of where to get it verified chemically? (Our usage is kind of low for the vendor to help.)

 

[VEBill]

I can't help you directly to your question, but...

"...a 100 mSec backup for a computer..."

Would a Supercap be a better choice? Your local car audio shop can probably sell you a 1 Farad supercap - designed especially for automobiles (thumping audio amplifiers) - for about US$50. 1F should be able to supply several amps for 100mSec.

 

[klaxon]

I'm using a 58 F, 15V supercap in place of this for one application. Good idea, though. It works pretty damn well, too.

Powers the stack for 30 seconds at 8 Amps or so.

Two non-obvious issues... charge currents don't allow for a direct drop in. Fuses blow, wires melt. It takes re-engineering to use those things and for one used to dealing in a few amps, 200-300 is a lot. ESR of those things is in the milliOhm range... and low at that.

Second item is overvoltage. Stick a standard battery charger on the vehicle and set it for 200A start assist and you get 18 V or so on the cap. FYI, burning caps stink! Necessitates an overvoltage protection circuit.... simplest is relay-based.

I have solved my problem, technically. I'm just searching for external bolstering of my analysis.

Thanks for the input anyway!

 

[itsmoked]

klaxon; It sounds like you have a god handle on things. I have designed portable instruments that need to run in and be charged by vehicle systems and use 12V SLA batteries. They are truly a b@tch! Wish I could use super caps but I need too much energy.

Anyway, talk to the manufacturer! They have ways to tear their batteries apart and experts who will know at a glance what killed the battery. DO NOT go to some lab as you will be charged a fortune.

Just converse with the battery maker. Ask him to tell you what killed your battery. Don't 'suggest' what killed it just ask...

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[VEBill]

"...have a god handle on things."

That's quite the compliment. ;-)

 

[itsmoked]

Haha! So much for my syntactic spell checker.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[felixc]

In backup applications I have never let the voltage go over 13.5 volts. At 13.8 I have seen batteries starting to gas. The high trickle current indicates an overvoltage of the charging circuit.

At 12.0 volts, there is still 90% of the charge in a lead-acid battery.

Are these batteries tested when they are received? Depending on the manufacturer, and how they have been shelved in the store, I have seen a lot of bad batteries, batteries that do not meet their specs either. Enough to design a battery capacity tester for our receiving dept.

You opened a window about boosters and overcharge. How is this battery protected against overvoltages?

 

[itsmoked]

via "automatic gassing" LOL

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[ScottyUK]

At 12.0 volts there is 90% of the charge in a lead-acid battery IF the battery is discharging at a relatively low rate. A terminal voltage of 12.0V for a six cell battery during the charging phase equates to only a few percent of capacity.


----------------------------------

Sometimes I only open my mouth to swap feet...

 

[itsmoked]

I call 12V.. 'dead'!

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[felixc]

Oops, discharged at 90%, under load. My bad. Most of the action happens between 12 and 13.5 volts.

 

[klaxon]

hey felixc,

The battery of which I am speaking is NOT protected against anything, which is one of the reasons I think it's overcharged. It has no charger. It is simply connected in parallel (almost) to the vehicle power buss. (There is an isolation diode, of course.) Lame, huh? BTW, not my design. I have been advocating a legit charger since the first day I saw this.

Again, what I'm trying to find is a lab that will crack open several of these mothers, peek inside, and tell me, based on chemistry, if they died from overcharge or undercharge.

For whatever reason, my evidence does not seem to penetrate the knowledge-protection circuits of management. They see "14.7 V" printed on the side of the battery and think that's what you have to charge them to.

 

[VEBill]

How's the vehicle battery on the other (higher) side of the isolation diode? Must be a smoldering slag heap...

 

[klaxon]

Hey ve1bll...

They last several years.

I've never seen an automotive battery treated well....i.e., temp compensated charge, multi-stage charging, etc. They are basically a float application... they output power only when the engine is off or starting... the rest of the time, trickle charge. Since they are HUGE AHr capacities, a trickle charge of .01c is a pretty big number. My little camcorder battery, on the other hand... 2.3 AHr. A few hundred mA forever = doom.

 

[itsmoked]

Remember flooded batteries like starter batteries tolerate over charging quite well. Gelcells do not tolerate over charging at all! The flooded cells are able to boil and release heat. Gelcell's electrolyte is a paste that if over charged bulges the sheet of electrolyte away from the plates. This directly reduces the functioning plate area of the battery. This also directly reduces the capacity of the battery. Furthermore it means the ongoing overcharge is now subjecting smaller areas of the plates to excess current which exacerbates the cell's destruction.

I don't think you understand test labs klaxon, if you keep thinking of going to one. Go to the manufacturer!

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[klaxon]

Hey Keith...

I took your advice and contacted the manufacturer. It only took two calls to get to a guy who wondered how the hell I managed to find him, but who could not / would not / did not know how to deal with the paperwork portions. I've got a call into his boss, and it'll likely take a few more hours of explaining /cajoling to get them to work the problem. My collection of batteries is date coded, so I am going to use the most recent ones. Fortunately, they all seem to have SOME terminal voltage.

This is my second time trying this with the manufacturer. We buy 20-100 batteries a year.... $400 to $2000, and it's hard to get their attention except for warranty replacement, and the failures aren't their fault.

As far as an external lab, you are right... I came here looking for (hoping actually!) someone might have some leads to one. Sometimes I think my particular management would listen to some random SOB at the airport if he said something authoritative. I do have contacts at some rather large universities that I could probably cajole into doing my analysis, though, and your post kind of reminded me that I had overlooked them.

If I ever do get a formal analysis and this thread is still open, I'll post it. I appreciate your post, though, Keith!

 

[itsmoked]

LOL.

I do get your airport comment. Can be really frustrating. Enough to maybe work somewhere else where they actually care.

A university angle would not be bad.

The problem with a test lab is.... All they have is their "reputation". So to not botch things they will want to do a ton of tests and sub tests to back up the tests. They will want to take a bunch of good batteries and fry them to try for controlled failures. All this will take a bunch of time and money. That's my concern.

If you can convince the factory to tell you if it was over or under charged in an email via a couple of sentences you would have your proof. Inexpensively!

Tell them you're arguing that it's a charging problem not a battery quality problem. Be pleasant and politely insistent. This could help prod them a bit.

Well, good luck there.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com