I believe we have tried using a batter load tester. I’m not totally sure, but I think so. As Ally said, we are unable to get into the workshop until January to test the batteries any further.
Thanks everyone for all these ideas! We’ve been kinda lost about what is wrong and where to go from there, but these should really help us out
Each battery has a number (as well as the year we got them) and we are planning on keeping a record of the voltages of each battery/how they ran when we tested them.
In 2008, we did 2regionals and a scrimmage before atlanta, and had the battery mounted on it’s flat side in the robot. We then lost two matches at Atlanta because two of the six batteries we had been using would die during a match. They would read 13volts on a meter before a match, but under load it dropped to 6 volts or so. We eventually figured out that one or more of the glass plates within the battery must have cracked or have been compromised. We now use a 50 dollar load tester from napa auto parts to make sure the batteries are still ok when at competition.
New students see those batteries as industructible heavy objects and don’t realize how careful they need to be with them. I also recommend mounting them as vertical as possible to reduce the stress on the plates inside.
You have any of several issues:
You’ve read several suggestions to eliminate #3.
#1 is indeed unlikely, but possible. After a full charge (as indicated by the charger) measure the battery voltage with a multimeter. A normal reading for a battery just off the charger would be over 13.5 volts. Let the battery sit for a day, to eliminate the ‘surface charge’ and a normal reading should 12.6 volts.
#2 is easily determined with a load test. Al’s recommendation of a West Mountain Radio CBA is excellent, but they are not free.
Instead, take about 10 feet of #10 wire (or 8 feet of #12), which should be about 1 Ohm. Measure the length of wire with an Ohmmeter (be sure to comensate for the resistance of the meter’s test leads) to verify about 1.0 Ohms. (A high-wattage 1 Ohm resistor (25 Watts capacity) will do as well)
While measuring the voltage across the battery terminals, connect the 1 Ohm load (the wire or resistor) also across the battery’s terminals. Count exactly 15 seconds. Write down the battery voltage at the 15 second mark.
A Very Good battery will read over 10 volts. An OK battery will read over 8.5 volts. Bad battery will drop below 8.5 volts. (NOTE: These numbers are debatable, but that’s what I use).
If in 15 seconds with a 1 Ohm load your batteries all drop below 8.5 volts, and start above 12.5, they’re bad. Buy new ones.
What causes a Lead-Acid battery to go “Bad”? Other than usage, sitting in a discharged state (terminal voltage less than 12.4 volts) for any period of time. Like over the summer. This kind of battery loses 50% of its charge in 90 days. On our team, we make sure they don’t go more than 60 days without getting charged. Even then, we lose 3-4 batteries each year.
OK Sammy,
We will be waiting for a report in January. There is a possibility that you have one or more failed chargers and that in the normal course of events, all batteries have been touched by the bad charger. That is unfortunate but should not be ruled out. I am going to guess that the batteries have been used often and are just dying of old age.
Doug, the glass is a woven mat between the plates so there is really nothing to break. The lead plates can become dislodged with abuse and break off the common terminal inside a cell. In this case, the battery will test normal even with a momentary load tester. The CBA will show this damaged cell by a drop of 2 volts after a period of time. Black in the attached plot is a normal (2 year old) battery showing 16 amp hour capacity, the blue shows a two cell failure, green shows a single cell failure and the red shows the signs of intermittent plate connections within a cell followed by several cell failures.
Don, any advice on safe (ie sparks and heat) and effective (ie minimize extra resistance due to poor connection) ways to do this ?
Also, where did the “10 volts after 15 seconds at 12 amps” rule come from?
According to the datasheet, a new battery should still read 12 volts after 16 minutes at 18 amps.
**
Let me give a short thanks to Al and another plug for the CBA (computerized battery tester).
http://www.westmountainradio.com/product_info.php?products_id=cba3
We’ve used this at his suggestion now for 3 years, and even gone so far as to rate our batteries 1-10 based on their tested capacity. It takes a while, but is definitely worth it to weed out the batteries that are simply worn out.
He said to measure during discharge. The datasheet values are after discharge has stopped.
Joe,
As the datasheet does not give all the measurement specs, I would have interpreted this as load curves with calibrated load resistors. Generally, with the load removed the terminal voltage would rise. With a high impedance voltmeter, there is no voltage dropped across the internal impedance of the battery.
I interpreted Don’s #10 wire test as a go/no go test for his team, not a definitive battery test.
“He said to measure during discharge.” He did indeed. It would still be informative to know where the “10 volts after 15 seconds at 12 amps” rule came from. Is this a home-grown FRC metric based on experience? Or is it published somewhere?
“The datasheet values are after discharge has stopped.” The datasheet didn’t mention that fact anywhere. I guess this must be a common understanding among those who work with battery specs? It would be interesting to know how long the testing is paused at each datapoint to allow the voltage to rise before recording it. Perhaps that’s part of some battery-industry spec. Also, question for Al: in the CBA graph you posted, are the voltages open-circuit voltages or are they volts under load?
**
The graph you posted in this thread was for 1 amp load. Do you have a 12 amp graph (of the FRC battery) you could post?
Also, as I understand it the CBA is computer-controlled. Is it possible that it is momentarily opening the circuit before taking each voltage reading?
**
The curves are generated at 7 amps or about 0.5 C (see curve menu at right side of graphs). From the CBA webpage…
“Unlike a resistor load it has an electronic constant current load that is controlled both with software and electronics.”
I don’t think taking a reading after removing the load will give any valid information. I do not recommend that anyone use a voltmeter alone to evaluate the state of charge on an SLA battery. The input impedance of modern voltmeters is sufficiently high to prevent an accurate reading across the internal impedance of the battery. As this “resistance” changes with battery age, temperature and state of charge, the load test will tell exactly what the battery is capable of delivering in a curve that can be compared with both the published battery data and with past tests of the same battery. The one amp shown was generated when I overlayed the four previous tested battery curves onto a single page within the program.
Don suggested applying a 12 amp load for 15 seconds and measuring the voltage under load. Above 10 volts is Very Good. 29
Ether cited the battery discharge graph in the datasheet for the FRC battery which shows 12 volts after 16 minutes at 18 amps. 31
Joe stated that the voltages in the datasheet graph are taken open circuit. 33
Al stated that he believes the datasheet graph is showing load curves 34, and that to his knowledge the CBA graphs are load curves too. 36
Al’s 7 amp graphs show a healthy battery still reading over 12 volts (under load) after 43 minutes at 7 amps. Even a very sick battery (blue line) still reads over 12 volts (under load) after 17 minutes at 7 amps. 30
Can somebody help sort this out ?
**
I think you’re off by a factor of 100, Don. Isn’t ten feet of #10 wire actually more like 10 milliohms?
(A single ohm would only draw about 12 amps, which isn’t nearly enough to give useful information about the health of a battery after just fifteen seconds.)
I know that this can be confusing so I am going to make a stab at this. The MK battery charts are generalized, guaranteed specifications for a new VLRA battery. The company has set these standards for the entire line and the graphs are plotted for Amp Hour vs discharge current rated in C where C = the rated discharge rate in amp hours. For our battery C=18AH. The MK Company charts are published on their website for VLRA batteries. The curve we have been looking at is a converted chart where the C values have been converted to discharge amps. .1C=1.8 amps, .5C=9 amps, etc.
For Team 111, the CBA tester is then set to it’s maximum current (for this model) of 7 amps. Each battery, each year, is then tested at the beginning of the build season and the curves saved for each battery. If a battery tests out at 15 AH or above for the test, it is considered good enough for competition. Anything under is marked for “practice only”. If there appears to be cell failures as shown previously, the battery is recycled. Failures of this type can be frustrating to team members and I fear can cause damage to chargers. I might be persuaded to run higher current testing if I were to purchase the higher current model CBA but I would not test at anything above 18 AH rates to maximize battery life (charge/discharge cycles). This testing does duplicate published company specs and curves leading us to group like batteries as healthy enough for competition.
As noted, even bad batteries can show normal terminal voltage before a cell(s) fail. Don’s test, however, would also fail when testing a similar battery.
In Don’s test, as pointed out by Alan, the current is much more than 12 amps. It is however within the realm of current drawn during match play. By performing this test, Don’s team has been reasonably sure that the battery will survive a two minute match. I would agree, but point out that max battery life will suffer. The kind of current drawn during this test is beyond the current drawn by most FRC robots during a similar time period. An FRC robot might draw this current in a failed autonomous run where the robot is pushing against a wall or other barrier for the entire auto period. It is possible to draw this much current on an FRC robot as the 120 amp breaker will not trip. It is possible to draw up to 600% of the rated current for several seconds without a trip on this breaker.
Don and I are both amateur radio operators and a familiar term for us “key down”. In our hobby we relate “key down” time to the average battery usage of portable operation. It is usually calculated by a percentage of actual transmit time (high current) and receive time (low current) operation. Many of us have discussed using this analogy for FRC robots but unfortunately, there is no standard robot for which to make predictions. In either case, both of our testing, although possibly flawed, can predict with some certainty that a battery so tested will survive a two minute match. Simple voltmeter testing, very high current load testing (car battery handheld load testers) for seconds, and short driving tests will not guarantee a battery will last a two minute match plus auto.
It is also a problem when people quote sections of the manual out of context. For instance, the maximum current specification on this battery is 720 amps for several seconds. Note that terminal voltage is not specified here. 720 amps would drop nearly 8 volts across the internal impedance of the battery. It is why the designers of the PD made two boost buck regulators to feed the Crio and wireless access point. High current demands would shut down these two critical systems without the additional power supplies rendering the robot useless for most of the match.
Updating previous post with new info:
- Don suggested applying a 12 amp load for 15 seconds and measuring the voltage under load. Above 10 volts is Very Good. 29
It was apparently a typo. Don meant 0.01 ohm, not 1 ohm. [edit]Don did not mean 0.01 ohm. see his new post.[/edit]
But with 10 volts across a 0.01 external resistance, the load would be pulling 1000 amps… and that raises two problems:
Fifteen seconds of that could potentially damage the battery or lessen its life. It would also probably burn up or damage a 25 watt 0.01 ohm resistor.
The battery can’t supply 1000 amps at 10 volts. Its internal resistance is too high*.
Perhaps the resistance of the copper wire is increasing due to heating. Let’s say the wire gets boiling hot (100C). That’s a dT of 80C, which would increase the wire’s resistance from 0.01 ohm to 0.0132 ohm. At 10 volts, that would be 758 amps. Still too high.
Perhaps the test depends on additional external resistance, maybe from a switch that is used to complete the load circuit? (I hope someone is not just holding the wires and jabbing them into the battery terminals :eek:)
Don, do you guys have a clamp-on ammeter? Could you record the actual current draw next time you do this test and post the results?
Bottom line: Don’s test is certainly severe enough to challenge even a brand-new battery. I worry a bit about safety, though, and battery damage. Do not attempt this without competent training.
- Ether cited the battery discharge graph in the datasheet for the FRC battery which shows 12 volts after 16 minutes at 18 amps. 31
- Joe stated that the voltages in the datasheet graph are taken open circuit. 33
Still not 100% resolved, but I think this is one of those rare times that Joe is mistaken.
- Al’s 7 amp graphs show a healthy battery still reading over 12 volts (under load) after 43 minutes at 7 amps. Even a very sick battery (blue line) still reads over 12 volts (under load) after 17 minutes at 7 amps. 30
All true, I think.
*the internal resistance of the battery drops as the temperature increases, but not enough to allow 1000 amps without damaging the battery.
**
OK, Maybe this’ll help. Maybe not.
I meant 1 Ohm, not 0.01 Ohm. 1000 Amps would be nifty, but impractical.
Spark-free, safe & effective uses the 120A main breaker to bring the load online. We wired everything (including voltmeter leads) to a PowerPole battery connector for simplicity. I am using a 1 Ohm, 125 Watt resistor I ‘happened to have’.
Alan, of course you’re correct. if 10 feet of #10 had 1 Ohm, my house would’ve burned down years ago. And none of the power outlets would deliver more than, say, 20 or 40 volts AC. I just wrote what came to mind, and obviously I was off by at least 2 orders of magnitude. :o And, since 1000 feet of #10 isn’t usually just hanging around most shops…I’ll just suggest finding a 1 Ohm resistor somehow.
Ether, you asked where did those numbers come from? Empirical results over the years, specific to FRC batteries. These are based on recommendations from the Battery Council International (BCI), but they’re dealing with automotive batteries being tested at somewhat higher rates and so their data doesn’t directly apply here.
Lastly, a comment on the glass mats used in “Absorbed Glass Mat” technology batteries: Think of compressed fiberglass insulation, that’s what the mats look, feel and taste like.
Of course, you all know beyond a shadow of any doubt that** I am an idiot.**
I re-read Alan’s comment, and of course: 1 Ohm (12 A) shouldn’t bring any FRC battery down to 10 volts any time soon. I mean, how would we survive a 2 minute match?
I just wish I could delete all I wrote in this thread and just start over.
I promise to rely less on memory and focus more on reality in the future. I am mortified and embarrassed.
Don
Of course, you all know beyond a shadow of any doubt that** I am an idiot.**
I re-read Alan’s comment, and of course: 1 Ohm (12 A) shouldn’t bring any FRC battery down to 10 volts any time soon. I mean, how would we survive a 2 minute match?
I just wish I could delete all I wrote in this thread and just start over.
Ether, I’d best say that I just pulled those numbers out my (you know). They’re not even that good. What the #311 was I thinking?
I promise to rely less on memory and focus more on reality in the future. I am mortified and embarrassed.
Don
