NXT battery test

Discussion specific to projects ideas and support.
kvols
Posts: 29
Joined: 14 Oct 2010, 22:09

NXT battery test

Post by kvols »

I've thought about this for quite some time, and I really believe it would be quite interesting: Using the NXT brick and motors to test batteries!

Battery testing is a bit more complex than one would expect: A battery is based on chemical reactions, and the battery may be optimized for a specific use. Some batteries a very good at delivering low power for long periods of time, whereas other batteries can deliver more power over a short period of time, but will self-discharge over relatively short time.

Using the NXT brick and motors, a standardized NXT battery test platform could be built. The set has it all: A rather precise volt meter, a method to load the batteries, timing and a computer for logging, ...

To ensure usable and independent test results, a standardized platform needs to be built, and it needs to test batteries in a way that to some degree resembles the way a Mindstorms NXT robot uses power: A relatively high load over short periods of time.

To get repeatable test results, that can be compared between users, the following should be decided and standardized:
  • A mechanical platform.
    The platform must me stable, and it must be able to give a repeatable load. Another person building the device, must give the same results. The platform should be built from a standard NXT 1.0 or 2.0 set.
  • A program performing the test.
    The program should ideally be able to run in both standard and non-standard firmware with the same results.
  • Environmental parameters.
    Batteries are quite sensitive to temperature, so this should be taken into account.
I have a couple of ideas, but I would like some input from others...

Anyone up for the idea? Wanna know which batteries are the best value for money? Which are the best when money is no option? Are Duracells really better or worse than NiMH? How good are these Energizers really? (for LEGO Mindstorms, that is!)
doc222
Posts: 117
Joined: 29 Sep 2010, 03:02
Contact:

Re: NXT battery test

Post by doc222 »

I think in the end all will work well enough. Though at times i like the heavy Nh if you need weight. If you need lighter batteries then the Lego one is almost half the weight, though it takes up a bit of space. If money is no issue and you want the top dog very lightest toss away then nothing has the edge over E2 energizers( peak new about 10.3-10.5 V @ (ive seen a set hit 10.7 new selected)(6cell) , they are very nice will make a bot faster. You can hear it. that extra 1.2-1.5 voltage at peak. It holds that for longer than you would think. But they are expensive and toss away.

One thing the bot could do is rate them and give the best of a group? Say you have a gang of 24 AA's and want the ones that are holding the highest peak charged value. Then it could give you six matched cells to run your NXT bot. even deal them out at six at time in decending peaks of 6.

doc
"Anyone who has never made a mistake has never tried anything new."
Albert Einstein
muntoo
Posts: 834
Joined: 01 Oct 2010, 02:54
Location: Your Worst Nightmare
Contact:

Re: NXT battery test

Post by muntoo »

kvols wrote:Anyone up for the idea? Wanna know which batteries are the best value for money? Which are the best when money is no option? Are Duracells really better or worse than NiMH? How good are these Energizers really? (for LEGO Mindstorms, that is!)
When there's something robotic? Who you gonna call? MythBusters! (Specifically The Imahara.)
Image

Commit to LEGO Mindstorms Robotics Stack Exchange:
bit.ly/MindstormsSE


Commit to LEGO Stack Exchange: bit.ly/Area51LEGOcommit
h-g-t
Posts: 552
Joined: 07 Jan 2011, 08:59
Location: Albania

Re: NXT battery test

Post by h-g-t »

You mean something like this? http://www.youtube.com/watch?v=6DAVCAvNlx4

Or this? http://www.youtube.com/watch?v=oLql-JD6x5s

Or even this? http://www.youtube.com/watch?v=s4Nmn6LC1A0&NR=1

Looks like any of these could be adapted to automate the process.
A sophistical rhetorician, inebriated with the exuberance of his own verbosity, and gifted with an egotistical imagination that can at all times command an interminable and inconsistent series of arguments to malign an opponent and to glorify himself.
kvols
Posts: 29
Joined: 14 Oct 2010, 22:09

Re: NXT battery test

Post by kvols »

Yes Imahara is a pretty cool robot builder. I love the Mythbuster show!!

But no, that was not what I meant by a battery tester. Let me try again:

I want to test the batteries inside the NXT brick itself. Build a robot, load the NXT brick with the batteries to be tested, and run the program. The program will then use its motors to spend battery power in a well-defined pattern (alternating loads, letting the batteries rest etc.), until the voltage get to a sufficiently low level, eg. 6V in total. During the test, the motor speed and the battery voltage and voltage drops will be monitored, so we can give the batteries some sort of score (speed, stamina, stability, ...).

There are battery tests to be found on the net, such as:
http://www.jeffsreviews.com/Articles/Ho ... 0comp.html
http://gizmodo.com/#!5152116/battlemodo ... -batteries
And many more...

But there are a few problems with these tests:
  • The battery tests does not mimic the typical NXT Mindstorms load.
  • The tests use different loads and equipment and are difficult to reproduce and tests cannot be compared.
  • The batteries tested may not be the ones you have available.
  • You might want to measure your current batteries with other batteries.
By building a standardized way of testing, more users can submit and compare battery tests. Doing a large battery test is both quite time consuming and costly. Especially is you want to have some level of statistically well-founded base.

I would really like some input to a battery tester that can be built with both a standard NXT Mindstorms 1.0 and 2.0 set. It must be "fool proof", so that tests will vary as little as possible (simple mechanics that can produce a well-defined load on the batteries). The test must also include a way of doing a reasonably realistic battery load, in order to give some useful results (a good performing battery should produce a high score in the battery test)

I believe there are quite big differences in batteries and how well suited they are for a LEGO Robots. Have a look at this test:
http://www.powerstream.com/AA-tests.htm

It is actually quite interesting to see that despite the fact that rechargeable batteries start at 1.2V and alkalines start at 1.5V, the alkaline batteries goes under 1.1V pretty fast. Some of the NiMH maintains a higher voltage during heavy loads than the alkaline batteries!

I think it would be really cool, if we could build a battery tester out of LEGO, and test a set of six batteries (or the rechargeable battery packs from LEGO), compare the results.
h-g-t
Posts: 552
Joined: 07 Jan 2011, 08:59
Location: Albania

Re: NXT battery test

Post by h-g-t »

Ok, sorry if I got it wrong. One point though, you said rechargeable batteries start at 1.2V. I regularly see my nimh batteries at just over 1.4v when charged and rarely see them at 1.2v or less.
A sophistical rhetorician, inebriated with the exuberance of his own verbosity, and gifted with an egotistical imagination that can at all times command an interminable and inconsistent series of arguments to malign an opponent and to glorify himself.
kvols
Posts: 29
Joined: 14 Oct 2010, 22:09

Re: NXT battery test

Post by kvols »

Yes, the odd thing is that the alkaline batteries are rated at 1.5V, but very quickly drops to 1.1-1.3V, whereas rechargeable NiMH batteries actually start at 1.4V but stays well over the alkaline batteries for quite some time - especially at a high current. This means that you NXT motors might start out running faster on alkaline batteries, but might be taken over by NiMH batteries shortly after...!

Did you notice the "E2" and the "DC" batteries in the last test...? Quite a difference! ;-)
h-g-t
Posts: 552
Joined: 07 Jan 2011, 08:59
Location: Albania

Re: NXT battery test

Post by h-g-t »

Came across this on Flickr. The NiZn batteries look very good but they are not widely available and expensive.

http://www.flickr.com/photos/imager/489 ... hotostream
A sophistical rhetorician, inebriated with the exuberance of his own verbosity, and gifted with an egotistical imagination that can at all times command an interminable and inconsistent series of arguments to malign an opponent and to glorify himself.
jwiger
Posts: 125
Joined: 24 May 2011, 20:41
Location: Where ever the military sends me (currently Central Texas)
Contact:

Re: NXT battery test

Post by jwiger »

You've got my interest! I wonder if any of the data from Philo would help here:
http://www.philohome.com/motors/motorcomp.htm
I've been wanting to get my hands on one of the new Energy Meters to try out just such a test
http://www.legoeducation.us/eng/product ... ku=W779688

I think the simplest way to load the motors is to build a basic winch and lift a mass, say 1Kg lifted 1m over and over again. the only downside i could see is the motors may act as a generator while lowering the weight, what would be the best way to prevent that? Maybe have it stop and hold in a brake fashion a few times on the way down?

For the endurance: possibly two motors working in an alternating fashion, each with their own weight at a moderate power setting (~60%) I would hunt for a speed that puts the motors in their highest efficiency.

For changing load: find a let down speed that moves slowly and puts a minimum amp draw on the way down, then raise the weights at the same time at 100% then repeat.

For high load: Weights might not be the best (or safest) way to do it. I like the idea of a fan/propeller type load. It works for exercise bikes, it might work here too. It may just be more difficult to calculate the work done.

I would love to calculate this in Joules too, since many of the battery companies have proprietary methods of measuring mAh ratings. It shouldn't be too hard to convert the work done into Joules, if you use the weights method.
JimmyJam
"The more you know, the more you know, the less you know."
kvols
Posts: 29
Joined: 14 Oct 2010, 22:09

Re: NXT battery test

Post by kvols »

Sorry - been away for some weeks now! :) Great to hear that I caught your interest!

Regarding the fan method for putting load on the motors, the wind resistance will depend on the air density, which again depends on the local air pressure, temperature and humidity, and then it becomes rather complex.

Using the motors for lifting a heavy object is interesting. It will not depend on the temperature, But this solution also has some variable factors: The internal resistance of the construction (especially if some gear is involved), and type of string being used (a thick string may make the circumference of the reel vary from start to end).

Simply running the motors at full speed might put quite some load on the batteries. The motors may (probably will) vary quite a bit, but using an average of all three and measure the RPM might give a good indicator of the power being used. I think I'll have a go at measuring how many amps. the motors actually use at 7-10V input, note the variance between the motors...

If you have an amp. meter, perhaps you could do the same with your motors?

Thank you for your inspiration!
Post Reply

Who is online

Users browsing this forum: No registered users and 0 guests