Odd battery discharge?

Philldapill

I recently finshed my battery voltage logger. The battery under test is a 12V, 7AH Lead-Acid battery. I've attached a graph to show the anomally. The Y axis is the battery voltage in mV(minus 8V to better show the anomally). The X axis is in seconds. This is a fairly old battery, so that is why all this happened in 1 hour. I believe I was drawing about 1.5A's at the beginning.

Anyway, please not the normal discharge slope. The voltage tends to drop faster and faster as it discharges, which I believe is normal, but note the last minute or so as the voltage rises all of a sudden, then starts back down. I wish I could have logged more, but I programmed the propeller to only record 1 hour worth of discharge.

Any ideas what this last part means?

Bean

What was the load ? Resistive ? Inductive ? Constant Current ?

I suspect the load caused the glitch.

Bean.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Uhhhm, that was on fire when I got here...

www.iElectronicDesigns.com

·

Philldapill

Purely resistive... a small automotive light bulb. I just charged the battery up, then hooked up my logger, connected the bulb to it and let it go for an hour. At the end, the logger stopped, and the propeller turned off. That glitch is probably about 3 minutes in length...

Anymore done on your solar panels, Bean?

Philldapill

I'm actually logging a 29AH deep cycle battery charging from a 15W panel right now... I'll see if there is anything odd about the charging on it, too.

Phil Pilgrim (PhiPi)

Philldapill said...
Purely resistive... a small automotive light bulb.

Well, sorta. An incandescent bulb is resistive but doesn't behave like a normal resistor. Its resistance is not constant but increases as the filament begins to incandesce, like a PTC thermistor. In your case, the resistance was dropping as the battery voltage decreased. (That doesn't explain the blip at the end, though. Maybe the filament became disconnected briefly, allwoing the battery to bounce back.)

-Phil

Post Edited (Phil Pilgrim (PhiPi)) : 6/24/2008 7:30:35 PM GMT

Philldapill

Phil,
Yeah, I spent a good bit of time a year or so back, trying to figure out why no measurements of voltage vs. current worked on a light bulb. I then found that out that the resistance changes, which makes sense considering how HOT that filament is, and that resistance does typically go up with temp(semiconductors have the opposite behavior).

I considered that exact thing about the blip at the time I pulled in the data to excel. That was the first thing that came to mind, but then again, I'm pretty sure my connections were solid. I soldered wires to the bulb and made sure everything was stout.

I've read up on batteries today and haven't came across anything like this... Maybe I DID have a bad connection... Anyway, here is a graph of what happened to the voltage on the battery RIGHT after I disconnected the charger(solar panel) when it was fully charged. Each x axis value is a 5 second interval.

Phil Pilgrim (PhiPi)

I was thinking more of a problem inside the bulb. I've seen bulbs with loose filaments that work intermittently.

-Phil

Paul Baker

The I-V curve for a filament as Phil has mentioned is a non-linear load. The anomoly is happening at 9.5V which I believe is considered deep cycle (11.89 V open circuit is considered "discharged" for a lead acid battery, did you happen to measure the voltage of the battery without a load at the region of interest?). A batteries behaviour in deep cycle is also non-linear. When you combine non-linear systems in a feedback topology (the behavior of one system affects another which in turn affects the original system), you can sometimes achieve unexpected behavior (but this is conjecture).

Another question, was the blub still illuminated in this region?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Paul Baker
Propeller Applications Engineer

Parallax, Inc.

Philldapill

I wasn't watching the entire time, but I'm pretty sure the bulb was illuminated the whole time. This is a new bulb from Autozone, so I'm doubting it was a bad filament. I have a bank of 11, 29AH batteries that were actually from a backup telecom piece of equipment. I want to use these as a rechargeable power source for a shed that I have about half a mile from my house. Right now, I am trying to run tests on these batteries to see what kind of shape they are in. I am charghing one hooked up to a battery charger with a 10 ohm resistor in series. This is to simulate a fairly constant voltage source being applied to the battery.

With this setup, I am logging the battery voltage every 5 seconds from the propeller, onto an SD card. From there, I pull the data into Excel and graph it. Once the battery is charged, I will do the same in reverse; discharge the battery through a 10 ohm resistor and log the voltage. Ideally, I suppose I should be logging the current as well, but I know the battery charger voltage is fairly constant upstream of the resistor during charge, and I will know the discharge current since I know the resistance and the battery voltage.

Anyway, this is something very simple but I'm just amazed at how easy and sophisticated the things I can do with the propeller are.

Fred Hawkins

"fairly old battery" = "a couple punk cells?"

I wonder if the glitch was from strong cells capable of retaining a decent charge.

Philldapill

Well, I've got another pickle for you... This is a graph of my 29AH deep cycle that I charged up fully, then discharged through a precision 10.5 ohm resistor. The resistor is designed to keep it's value even when hot. Disregard the discontinuity at about 7:12hrs. I had to stop logging and take the card out, save it to my computer, and put it back in.· Usually, I do this in about 1 minute so you can barely(if at all) tell that I stopped logging. This time, I got tied up inside too long. While I take the card out, I keep the discharge going.

Anyway, notice the behavior at the end. It goes down quickly, levels off, then goes down again... Like I said, this battery was used in some telecom equipment as backup power. I don't know if it was used extensively, but I only got about 20AH our of it.

Phil Pilgrim (PhiPi)

It could be that the cells aren't all discharging at the same rate. Hence, the double knee on your graph.

-Phil

Paul Baker

@ Philldapill, theres something you need to know about stated battery capacity. The capacity stated on a battery is only applicable to a specific method of testing capacity, iow that capacity only applies if you drain it at the same rate as the manufacturer did to determine the capacity rating.

This discharge rate is refered to as C, so standard discharge is 1C, 1/2 discharge rate is 0.5C, 2 x discharge rate is 2C and so on. When discharge rate is < 1, the observed capacity of the cell is greater than what the battery is rated at. When the discharge rate > 1, the observed capacity of the cell is less than what the battery is rated at.

Another factor which can skew the observed capacity from the rated capacity is the temperature of the cell. Nearly all chemical reactions occur at a faster rate when the temperature is raised, this is also true for batteries (which are really just electrochemical devices, converting chemical energy into electrical energy). Lowering the temperature will reduce the effective capacity, and if you lower it far enough no chemical reaction will take place. This is why radioisotope batteries are the gold standard for batteries used in space especially for deep space, radioactivity isn't effected by temperature nearly as much as chemical processes are, and the radioactivity is self-heating which keeps a floor temperature for the unit even in the coldest parts of space.

Here's a Maxim ap note discussing the issues: http://www.maxim-ic.com/appnotes.cfm/appnote_number/121


▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Paul Baker
Propeller Applications Engineer

Parallax, Inc.

Post Edited (Paul Baker (Parallax)) : 6/26/2008 6:17:30 PM GMT

Philldapill

@Paul - The battery states "29AH @ 8hr discharge rate". My load for the battery was 10.75 ohms, and the average current is between 1.1A and 1A. The rated discharge rate @ 8 hrs would be 3.625A. If my understanding is correct, reducing the discharge rate should actually increase the capacity, not decrease it. Again, this is a used battery, so I'm not sure what kind of life it's had.

@Phil - I bet you're right. This isn't a new battery, so I bet some cells are in better shape than others. The knee occured at less than 10.5V, which is below the "discharged" voltage anyway. I assume that one cell finally went kapoot, while the others still had a little life. Anyway, I'm recharging the battery at a constant rate, and logging the voltage.