Hi,
How do you guys maintain the 12V batteries? Do you charge them until they’re over 13V…use them in practice until they drop down below 12V or the motors have no more uummph!? Do you put a load on them and measure voltage over time to determine if they can ‘hold’ a charge? Do you wait until they begin leaking and toss them?
Have you built a load tester? If so, what did it consist of and conceptually how did you use it?
In short, how do you guys ensure your battery has what it takes to make it through the 2 minute match?
I have a lot of batteries (over 20) that have been collected over the years in robotics. Like to use them if we can…toss them if they’re junk, etc.
I can use a Dashboard viewer program I wrote to grab data from the battery along with any other variable that can be monitored. Have any of you utilized the Dashboard for battery load testing or monitoring?
First and foremost, don’t leave a lead-acid battery sitting around with little to no charge. They’ll sulfate and then be mostly useless. Not instantly, but over a few hours.
After that, I have no idea. We generally run through batteries on the practice bot and see which ones seem good. A more objective testing methog would be a good thing.
We are currently in the process of redoing every part of our team side projects such as, cart & crate, but we are also building a new battery charging station for the pits, and also to put on the cart.
I know that it will have a built in load tester, but I’m not sure how that will work, or what it will look like when it is finished.
I know that it is going to be wired up to just be a plug and charge box, and have a status indicator for every battery that is in there normal, and under load.
When we finish all these projects, I am probably going to upload some white papers on the cart, crate and battery box.
The sealed lead acid gel type batteries that we use do not die instantaneously like NiCads. You will find that a particular battery doesn’t seem to last as long as it once did. Load testing is better than a voltmeter for testing but will not tell you the overall capacity of the battery. In these batteries the electrolyte is a gelled sulphuric acid that allows them to be used in any configuration. As they are charged and discharged a small amount of moisture is lost with each cycle. As moisture is lost, the amount of electrolyte that is in contact with the plates also diminishes. Eventually, there just isn’t enough electrolyte for our demand and as the robot draws current heat builds up inside the battery causing further destruction. If a battery feels much lighter than your others, it is near the end of it’s life. Use it for practice and then recycle.
You will find people that tell you to leave them on trickle charge, or charge every month to keep them topped off. Our batteries are stored at our playing field that is away from the mentor’s place of work. We charge them when we need them. They have been charged maybe once since IRI in July for homecoming. We held a mini invitational this past weekend and all of them only needed a few minutes on charge to bring them up to full. Store them in a heated area, protect the terminals and don’t leave them on the floor. If any have cracks in the case or the terminals move around also recycle.
If you want to get moderately serious about sorting out your batteries you can build a reasonable “capacity” tester using the robot controller, a spike, and a high power load resistor from Digikey. Last years RC can be used for this, or you can do it with the EDU using a suitable relay wired to a solenoid output.
You can use a one ohm, 225 watt resistor, part number FVT200-1.0-ND costing $12.36. The mounting bracket for this resistor appears to be B1001-ND, you would need two of these per resistor.
At 13 volts, and the battery won’t be there very long, the resistor will draw 13 amps and dissapate 169 watts, within spec. It is also within the spec of the current limit for the spike. Pointing a small muffin fan at the resistor will be a good idea!
If you want to load at higher currents, you use more than one spike controlled resistor in parallel. The concept can actually be taken to a rather insane level using high power bipolar transistors to load the battery with a constant current independent of battery voltage, and programming the load current in binary using small relays to drive the transistors. A serious senior project, indeed. Some of us are quite crazy…
You sense the battery voltage through a voltage divider hooked to an analog input. You measure the time that a fully charged battery takes to drop to a specified voltage, drop the load, and display the time in suitable units. A reasonable choice for voltage limit, examining the discharge curves for the battery in question, would be 11 volts.
The task of programming the RC and calibrating the voltage sense for the analog input is not difficult. It is a nice pre-season project that will “get used” to sort out your batteries. A setup using one load resistor and a voltage divider is cheaper than a battery, assuming that you have last year’s RC and a spike laying about.
You can calibrate the measured capacity, the time to reach 11 volts, by checking the new batteries that come in this years kit (after charging them over night). A good rule of thumb is that any batteries that are lower than %75 of the capacity measured for the new ones should be used only for practice, but you can come up with your own rule of thumb based on your own experience with your batteries.
These batteries don’t take much maintainance as their self discharge rate is low, but you should fully charge them before storing them, store them in a cool dry place, and give them a charge when taking them out of storage, before using them.
Al,
Just read the white paper. Great reminders about wire resistance and bad connection. Thinking through the routing of wires from the battery disconnect to various breakers is important as you mentioned.
Didn’t know the details about the battery. In the past it was just a black (grey) box with two terminals…use…recharge…use…recharge.
It was designed for R/C battery applications, but I think it may be useful with our 12V sealed batteries also. I’m going to evaluate it’s potential for testing our batteries, and if it works out I’ll add it to our team’s electrical test bench.
Bruce,
We are also planning on this as a purchase. Have you received yours yet? I would love to discuss your review when you get a chance. PM me, please?
Team SPAM got a “battery cycling” setup a couple years ago from someone that one of our team members knows at Exide. It’s basically a lightbulb with a timing circuit that drains at specific intervals and rates; it’s used when the battery is new - we put all of our batteries through it - and we’ve noticed a measurable difference in how long our batteries will hold a strong (>12V) charge. Warren Boudreaux can give a more detailed description of it - I wasn’t there the night they brought it in and explained it.
I recently took an online course from RedVector.com on batteries for continuing education credits required for my PE license in Florida (I needed the knowledge for work as well so they paid for it) - it was very informative especially in regards to how different types of battery chargers work.
If you were going to load test the FIRST batteries for capacity comparisons, what amp draw would you use? C/10 or 1.8A seem reasonable? Or should it be a 20 hr rate?
C/10 is a reasonable value to test the capacity of a battery at, but your goal should be to evaluate the batteries as close to the actual robot loading as is possible with your tester. I would test right below the continuous power limit of your tester, with enough margin that you are sure that you won’t damage the tester. Assuming 13 volts, the 100 watt limit is 7.7 amps. You are looking at close to 2 hours for a fresh battery.
The highly variable current loading in a competition is actually much higher than this (on average), and most would call it abuse of the battery, but that is what we are doing in the robot…
One added note on battery testing: If the battery gets warm during testing, turn down the loading, or intermittently load the battery in a manner that allows for cooling between current draws. One wants to simulate the robot loading as much as possible, but one does not want to test destructively.
All,
There are two types of tests I would perform when given the chance. The first is at c/10 to insure that the battery meets specs +/- 10%. Then I would perform an average match draw to see what you leave in a battery for the next match. My conclusion, rule of thumb actually, is that if you are unable to run three matches from the same battery, you have a problem with mechanical design that puts too much load on the battery. With these new computer controlled testers it would be great to simulate a match and run a test under those conditions. There is a formula for determining approximate battery life for portable radio operations. (QRP for you hams out there.) It is based on how much a typical radio contact will be in transmit and how much the receiver runs in between.
Match simulation can be a more valid test now that the RC has a backup battery. The deciding factor had always been to not exceed the current draw that would pull the output below 8 volts. All things being equal that was about 350 amps on a fully charged battery. Modeling an actual match would be a great help. Has anyone compiled valid current data (sampled throughout a match) for their robot or others?
As a side note, do not attempt to draw this much current from our battery without safety precautions, knowing the risks and using a circuit cutout in case all else fails.
