Performance of VEX Batteries Under Load

This thread is to discuss the results of a series of VEX battery load tests. It was first discussed in these two threads 1] 2], but they are both in the “New VEX Product Ideas” forum, so I thought it would be better to start a new thread here.

To catch everybody up: In order to answer some questions about Vex battery performance under load, I decided to run some tests.

I’m using a borrowed integrating power meter that takes very accurate readings of V, A, W, mAh, & mWh and reports them to the computer (4 samples per second).

I’ve built a dummy load that can be set to consume a constant current (Amps), and will shut off when the battery voltage drops below a safe range.

I’ll be running deep discharge tests on the entire range of Vex batteries at various loading levels. I’m testing two of each battery type since there is likely to be some variation. For comparison, I’ve also tested a AA Battery Holder with 6 Eneloop NiMH batteries.

Before each test, the battery is fully charged using the “Safe” mode of the new Smart Charger. The battery is then allowed to rest till it comes down to room temperature (gauged by hand).

I’ve tested them a 4A, and will be testing them at 8A and perhaps 12A (if the dummy load survives). I’m having the load shutoff at around 5V, since the Power Expander battery LED turns red around 5.1V and the Cortex battery LED turns red around 5.5V.

I’ll post results here as they become available, and I’ll post full data and graphs over on the wiki once I’m done collecting everything.


  • Dean

Here are some numbers from my sweep of batteries discharged at 4 Amps.

I ran two sets of batteries, so each statistic has one value for each battery. (I didn’t run a 2nd set of Eneloops because I didn’t have enough spare ones to form two complete sets.).

Larger values are better For all these statistics, except for Internal Resistance.

*]Vex NiMH 7.2V 2000mAh Battery
*]Runtime (down to 5.1V): 26m 19s, 22m 31s
*]Total mAh delivered: 1753.1, 1499.2
*]Voltage at end of 5 min: 6.409, 6.61
*]Voltage at end of 10 min: 6.372, 6.575
*]W•hr delivered in first 5 min: 2.1823, 2.2447
*]W•hr delivered in first 10 min: 4.308, 4.439
*]Internal Resistance: 0.248Ω, 0.224Ω

*]Vex NiMH 7.2V 3000mAh Battery
*]Runtime (down to 5.1V): 40m 45s, 41m 27s
*]Total mAh delivered: 2711.1, 2762.5
*]Voltage at end of 5 min: 7.372, 7.305
*]Voltage at end of 10 min: 7.326, 7.276
*]W•hr delivered in first 5 min: 2.4953, 2.5041
*]W•hr delivered in first 10 min: 4.9399, 4.9284
*]Internal Resistance: 0.080Ω, 0.078Ω

*]Vex NiCD 7.2V 2000mAh Battery (discontinued)
*]Runtime (down to 5.1V): 25m 13s, 23m 25s
*]Total mAh delivered: 1678.4, 1561.1
*]Voltage at end of 5 min: 7.176, 7.068
*]Voltage at end of 10 min: 7.117, 7.005
*]W•hr delivered in first 5 min: 2.4558, 2.4370
*]W•hr delivered in first 10 min: 4.8316, 4.7786
*]Internal Resistance: 0.118Ω, 0.115Ω

*]Vex NiCD 9.6V 1000mAh Battery (transmitter)
*]Runtime (down to 7.5V): 5m 38s, 6m 27s
*]Total mAh delivered: 426.5, 460.3
*]Voltage at end of 5 min: 7.726, 8.327
*]Voltage at end of 10 min: n/a, n/a
*]W•hr delivered in first 5 min: 2.7262, 2.9283
*]W•hr delivered in first 10 min: 3.4181, 3.9171
*]Internal Resistance: 0.354Ω, 0.298Ω

*]Six Eneloop AA cells in a Vex AA Battery Holder
*]Runtime (down to 5.1V): 0m 12s (yes, that’s correct!)
*]Total mAh delivered: 13.5
*]Voltage at end of 5 min: n/a
*]Voltage at end of 10 min: n/a
*]W•hr delivered in first 5 min: 0.0734
*]W•hr delivered in first 10 min: 0.0734
*]Internal Resistance: 0.675Ω


It takes a while to crunch all those numbers from the raw data (almost fifty thousand samples). It’ll take a while longer to get graphs up, but I wanted to post these results. They are not very surprising:

The new 3000mAh NiMH batteries are the best in every category when discharged at 4A. The 2000mAh batteries (both NiCd and NiMH) are close behind.

However, looking at the Watt•hours provided by the first five minutes of a transmitter battery, you will see it delivers more power because of its higher voltage. They were really struggling at 4A, though, and these transmitter batteries won’t be moving on to the 8A round.

The Eneloop batteries look like a complete failure from the data, but that was simply because they have too-high of an internal resistance to be useful at 4A. They were not fully discharging this quickly, but they simply couldn’t keep the voltage above 5V with a 4A load. I discharged this same pack at 0.5A and it provided over 2000mAh. They are very useful for Wii remotes and the like, but not much good for robots. They, too, will not be moving on to the 8A round.


  • Dean

This is terrific data.

Thank you for sharing, maybe one day VEX can release a formal Datasheet for their products. Eventually, students need to appreciate the value of comparing detailed specifications of batteries, motors, sensors and mechanical systems.

What circuit did you use to stop draining the batteries when they reached the desired voltage? Would you be able to post a diagram of this part of your device?

Thanks for the data.

Certainly - I’ll post full details after the data all goes up. I was even thinking about running off a few boards since it is mostly scary spaghetti wiring. (PM me if you are interested in a group buy of boards.)

As for the cutoff circuit: I used an LM339 quad comparator to detect when the battery voltage goes below a reference voltage (set on the potentiometer). I used the spare comparators as (1) a latch to prevent cycling between on and off at the end of discharge, (2) as an LED driver, and (3) as a driver to shutdown the pass transistor.

Honestly, I think this would have been easier with an ATMega microprocessor, but I was having fun getting my analog geek on, so I just went with it.

I’m running the 8A test now…


  • Dean

Dean, et al,

A couple of years ago, my VEX students thought that they had a bunch of dead batteries. I gathered some data (see attached) and made a PowerPoint presentation for them which, if refined, could cure insomnia…

At any rate, you might find the data interesting. My conclusions were:

[INDENT]Most of the data “scatter” is likely attributable to temperature variations.

All of the batteries underperformed from their rated specifications. However, they did so in such a uniform way as to suggest that the rated specifications might have been reported in error.

While a couple of the batteries appeared to benefit from exercise and/or reconditioning, all batteries were considered acceptable for use in a VEX competition.[/INDENT]

I used a CBA II battery analyzer from West Mountain Radio to perform these tests…



Postscript: Most of these batteries had at least 3 years of “heavy” use and abuse at the time of the analysis…
VEX 6-cell 1C and 3C Tests.jpg

Yep, I think all batteries are sensitive to temperature to some degree (actually, so are electronics, but digital mostly hides the issue). I’ve noticed some thermal effects with my battery tests too.

Age and abuse will derate the capacity somewhat (I’ve seen specs that indicate you should derate to 80% after 1000 cycles).

Also, most NiCd and NiMH batteries are rated at 0.2C, and the available capacity decreases as the load increases. This is at least partly due to internal resistance, which is particularly visible in the 3C runs on your graph. Their voltage is depressed across the entire run, and crossed the cutoff voltage before the curve even gets to the steep part. I suspect you could have gotten several hundred mAh more from them if you continue the discharge from that point with a lower load.

After I’m done with the high-amp discharge tests, I’m going to run all the batteries through at 0.2C to see if they still their rated capacity.


  • Dean

Here are some numbers from my sweep of batteries discharged at 8 Amps.

As before, I ran two sets of batteries, so each statistic has one value for each battery.

Larger values are better For all these statistics, except for Internal Resistance.

*]Vex NiMH 7.2V 2000mAh Battery
*]Runtime (down to 5.1V): 11m 22s, 10m 3s
*]Total mAh delivered: 1512.2, 1337.0
*]Voltage at end of 5 min: 5.447, 5.619
*]Voltage at end of 10 min: 5.314, 5.135
*]W•hr delivered in first 5 min: 3.5468, 3.6708
*]W•hr delivered in first 10 min: 7.1538, 7.3447
*]Internal Resistance: 0.286Ω, 0.263Ω

*]Vex NiMH 7.2V 3000mAh Battery
*]Runtime (down to 5.1V): 20m 29s, 20m 35s
*]Total mAh delivered: 2721.0, 2734.5
*]Voltage at end of 5 min: 6.939, 6.958
*]Voltage at end of 10 min: 6.852, 6.881
*]W•hr delivered in first 5 min: 4.6565, 4.6567
*]W•hr delivered in first 10 min: 9.2362, 9.2531
*]Internal Resistance: 0.083Ω, 0.082Ω

*]Vex NiCD 7.2V 2000mAh Battery (discontinued)
*]Runtime (down to 5.1V): 10m 59s, 8m 29s
*]Total mAh delivered: 1461.2, 1129.2
*]Voltage at end of 5 min: 6.587, 6.030
*]Voltage at end of 10 min: 5.933, n/a
*]W•hr delivered in first 5 min: 4.4699, 4.3258
*]W•hr delivered in first 10 min: 8.7127, 7.0589
*]Internal Resistance: 0.110Ω, 0.107Ω


Again, the new 3000mAh NiMH batteries are the best across the board when discharged at 8A. They had no problem holding up to the load, and in fact seemed to be a just as efficient at 8A than 4A (total delivered mAh just over 2700). It is striking how identical the discharge graphs are for these two batteries - it almost looks like the same line plotted twice. I don’t think these batteries will have much trouble at 12A.

The NiCd batteries also did OK, though one of them stumbled a bit - it may have a marginal cell or perhaps it didn’t charge completely. These will move on the the 12A round as well. I expect they will do OK, but I’m just hoping they both can make it to the 5 minute mark.

The small NiMH batteries (2000mAh) had a very interesting discharge curve. The both dipped down to near 5.1V (the test cutoff) and then increased for a while before settling into the normal discharge curve. The only thing I can figure is that these NiMH batteries are slightly more efficient when they are warm - I need to do a bit more research on this. By the end of the test, these small batteries were pretty hot, so I won’t be advancing them to the 12A round.


  • Dean

Keep up the good work Dean.

An 8A discharge represents a fully loaded Cortex,
or a fully loaded PIC+powerextender.
Since the 2000mAH new small NiMH battery runs for 10minutes at this load,
the idea of a power Y cable for PIC+powerextender is viable,
and a fully charged battery may last two matches (but I’d swap if I could).

A 12A discharge represents a fully loaded Cortex + power extender.
I’m still interested in seeing the 2000mAH results at > 8A, but it is not my risk…

For lightweight robots that intend to hang, a 2000mAH new small NiMH battery may be a bettery tradeoff of power to weight than the lower resistance
3000mAH batteries.

I previously suggested that NiCD may have lower internal resistance than NiMH.
There are no apples to apples comparisons here (for same size, formfactor, capacity, age) to prove or disprove that, but for the apples to oranges comparison of available products that Dean provided, its not true.
The 3000mAH NiMH batteries are awesome!

Agreed - the batteries have generally done better than I assumed they would. Going into these tests, I thought the older NiCd’s would have retained less of their capacity, and I assumed the NiMH’s would have a higher internal resistance (and thus more voltage depression under load).

Note that the small (2000mAh) NiMH pack runs about a volt and a half lower than the large (3000mAh) pack under an 8A load. This means it will deliver less power to your robot during that time and the motors will run slower.

Keep in mind, though, these tests are somewhat artificial. A real robot will not present a constant-current load - it will spike well above the 4A limit briefly when motors stall, and it will rest at well below 4A when fewer motors are running (and not stalled).

If at any time, the voltage dips below 5.5V, the Cortex battery LED will go red, and I assume it will be at risk of rebooting (the 9V battery just backs up the VEXnet link, not the CPU, right?).

There really isn’t much point. The small NiMH packs hovered just above the Cortex cutoff voltage for almost the entire 10m. A 12A load would have them fall below the cutoff voltage in just a matter of seconds. I’ll give it a try tonight, but I expect it’ll look much like the eneloop batteries in the 4A test.

I agree. The NiMH 2000mAh batteries are perfect for cases where weight and/or size are critical - they offer options that the larger packs can’t. I am particularly impressed how much better they are than the eneloop batteries I tested. They are both 6xAA packs with about 2000mAh capacity. The eneloop batteries are designed for long shelf life, which translates to higher internal resistance and a useful load limit of about 0.5C (~1A). The Vex pack is solid up to 2C (~4A), but will probably self-discharge faster than the eneloops.

That being said, I would always use the larger 3000mAh packs if weight and size are not an issue. The higher voltage under load will translate to slightly faster motor speeds.

The general rule of thumb is that the internal resistance on NiCds is lower than NiMHs, so I also assumed that would be the case here. Lucky for us, IFI made sure to select cells that are optimized for high current discharge.

Agreed. I’ll post the 12A data shortly, but the 3000mAh batteries handled it easily. These new NiMH batteries are very much an upgrade over the older NiCd ones.


  • Dean

So, here are the numbers from my sweep of batteries discharged at 12 Amps.

As before, I ran the test with two batteries of each type.

And as before, larger values are better except for Internal Resistance.

*]Vex NiMH 7.2V 3000mAh Battery
*]Runtime (down to 5.1V): 13m 33s, 14m 56s
*]Total mAh delivered: 2711.9, 2957.8
*]Voltage at end of 5 min: 6.559, 6.664
*]Voltage at end of 10 min: 6.310, 6.473
*]W•hr delivered in first 5 min: 6.6256, 6.7681
*]W•hr delivered in first 10 min: 13.083, 13.350
*]Internal Resistance: 0.080Ω, 0.080Ω

*]Vex NiCD 7.2V 2000mAh Battery (discontinued)
*]Runtime (down to 5.1V): 5m 46s, 4m 23s
*]Total mAh delivered: 1153.9, 877.4
*]Voltage at end of 5 min: 5.667, n/a
*]Voltage at end of 10 min: n/a, n/a
*]W•hr delivered in first 5 min: 6.179, 5.0749
*]W•hr delivered in first 10 min: 6.9957, 5.0749
*]Internal Resistance: 0.101Ω, 0.097Ω


The new 3000mAh NiMH batteries seem to live for this kind of load. Amazingly, the 2nd battery nearly delivered its full 3000mAh under this heavy load! The dummy load had to dissipate up to 90W to keep up with these batteries (got very hot, but no drama). I think these batteries could handle 15A in a pinch; not sure if my dummy load could though…

Both of the the NiCd batteries gave out around the 5 min mark. The combination of age and higher internal resistance proved to be just a bit too much for them to handle a 12A load for very long.

The bottom line is that the new 3000mAh NiMH batteries can handle pretty much any load a Vex robot can put on them. They are clearly an upgrade from the previous generation of Vex batteries.


  • Dean

Since you asked, here are the numbers from the small NiMH (2000mAh) batteries discharged at 12 Amps.

The first test was run on a single small NiMH.

The second test was run on a pair of them wired in parallel (4000mAh combined capacity). I noticed that the small NiMH batteries are half the weight of the large NiMH, and I wondered how a pair of them in parallel might compare to a single 3000mAh battery.

*]A single Vex NiMH 7.2V 2000mAh Battery
*]Runtime (down to 5.1V): 0m 20s
*]Total mAh delivered: 66.2
*]Voltage at end of 5 min: n/a
*]Voltage at end of 10 min: n/a
*]W•hr delivered in first 5 min: 0.3504
*]W•hr delivered in first 10 min: 0.3504
*]Internal Resistance: 0.213Ω

*]A parallel pair of Vex NiMH 7.2V 2000mAh Batteries
*]Runtime (down to 5.1V): 14m 35s
*]Total mAh delivered: 2920.4
*]Voltage at end of 5 min: 5.871
*]Voltage at end of 10 min: 5.793
*]W•hr delivered in first 5 min: 5.9321
*]W•hr delivered in first 10 min: 11.7800
*]Internal Resistance: 0.139Ω


Not surprisingly, the single battery could not keep the voltage above 5.1V for long under a 12A load. This battery would be fine handling brief 12A peaks, but it can’t sustain that kind of load.

As for a pair of 2000mAh NiMH’s in parallel -vs- the 3000mAh battery: The answer is that it performed almost as well, but not better than the large pack. It had a similar run time and delivered a similar number of mAh to the load. However, the parallel internal resistance was still just a bit higher than the IR of the larger battery. This resulted in more voltage depression and therefore less Watt•hours being delivered .

Again, the new 3000mAh NiMH battery pack seems to be king of the hill.


  • Dean

Thanks Dean! I appreciate it.


Do you intend to post this data to the Wiki? It might be nice to test some new batteries also. I can get you some to test while your test fixture is up and running.

Yep - I’ve been generating the graphs and I’ll be posting them in the next day or so. Excel was simply choking on the volume of data (10K samples for some of the runs), so I’ve been teaching myself R and I’m finally happy with the results.

That would be outstanding! Of course, all the NiMH batteries I tested were brand new, but all my NiCds are years old. If you still have stock of any of the discontinued 7.2V NiCd packs, I would appreciate a pair of them so I can do “New NiMH -vs- New NiCd” comparisons.


  • Dean

This writeup has been posted to the wiki for a while now, but I just went back through and cleaned things up a bit. Have a look here if you are interested.

Also, I added some graphs to the wiki pages for the three 7.2V batteries.
*]3000mAh NiMH Battery
*]2000mAh NiMH Battery
*]2000mAh NiCd Battery


  • Dean

The wiki description of the 12A load doesn’t match my memory. If I recall correctly, the Cortex has dual 4A limits, “twice the power of the PIC”, so a 12Amp load is not hopelessly unrealistic for a (poorly engineered, overloaded) competition robot running a Cortex plus a power expander.
Your data shows that my proposed “power Y cable” would work at the limits of (theoretical, zero safety margin added) adverse conditions for the 3000mAH battery, but not for the lighter 2000mAH battery.
Thanks for posting to the wiki.

You are correct. I caught this on the 3000mAh battery page, but didn’t fix it in the big writeup. I’ll adjust the wording to match reality, and update the Cortex wiki page to describe the dual-PTC configuration.

Yep - that seems to be my conclusion too,

  • Dean