Group Actions
Engagement
  • Author Author: jw0752
  • Date Created: Date Created
  • Views 757 views
  • Likes 3 likes
  • Comments 8 comments
Related
Recommended

Nominal VS Actual Power of USB Power Pack

jw0752

The other day I was reading a blog by raspjoenoob   about trying to run a Raspberry Pi with a USB Power Pack.   USB POWER BANK // PORTABLE POWER SUPPLY  Later the same day I was in Walmart where I noticed they had 2200 mAh Power Packs on sale for $5.00. I bought 2 of these with the intention to test them to see just how much power they would actually deliver.

 

image

A while back I built a little battery duration tester that is used in conjunction with a Load Unit to test battery energy available under different loading conditions.

Battle of the Batteries - E*** vs D*** vs Bargain


image


These are better pictures of the USB Power Packs that will be evaluated. The specifications say that they are 5 volts in for charging and that they will supply 5 volts output at 1 Amp. The power rating was of course 2200 mAh.


    imageimage


I alternated between the two units in my tests. Each time the Power Pack was allowed to fully charge before testing. The units were tested under loads of 250 mA, 500 mA, 750 mA, 870 mA, and 1000 mA. In each test the unit was run until the internal circuitry of the Power Pack disconnected the unit from the load due to low battery condition. The time on the clock was then noted and a calculation of milli Amp hours was calculated. Here is how the test setup looked:


     image image

 

As I have seen before in battery tests the Power Packs were able to do slightly better under a low load than they did at the upper end of their rated load capability. Of course we are not dealing with a pure battery situation here as there is internal monitoring circuitry as well as a boost converter to deliver the nominal 5 volt output. Quite a neat little package all things considered. I will post a graph of the results of the tests below. The best power out that I was able to obtain was 1388 mAh under a continuous load of 250 mA. This level was very linear until the load got in excess of 750 mA at which point it dropped quite quickly until it was down to 750 mAh at it full rated 1000 mA output. Here is the Graph:


image


Obviously at $5.00 and clearly marked as Made in China this battery pack was not top of the line quality. It would be interesting however to see how well an expensive top brand would stack up against it.. When we tested the Alkaline AA batteries we were surprised to find little to no difference between the top brands and the dollar store cheapies. If I can get my hands on a better unit with the same specifications I will post an addendum to this blog.


John

.



.

  • in reply to jw0752

    Hi John,

     

    The battery analyzer is proving useful!! Seriously cool project.

    By the way, if you want a counter capability to count days, there is a project here that is already pre-built,

    no need for writing code, just needs plugging into a PC and it will look like a USB memory stick, and drag-and-drop

    the file at that blog post, to load it into the Flash memory. It has a counter function and a rotary encoder function, but just the counter can be used.

    Or a 4017 and some LEDs etc : )

  • in reply to jw0752

    That's not bad since I suspect there is some peaking of the current.

     

    It would be interesting to see if the performance changes over several cycles of the pack.

    In theory it shouldn't but I'm told that if you charge to the maximum they have a short life, while undercharging gives more cycles.

    http://batteryuniversity.com/

    http://batteryuniversity.com/learn/article/how_to_prolong_lithium_based_batteries

    http://batteryuniversity.com/learn/article/charging_lithium_ion_batteries

     

    I can see a modification coming if you need to do anything much longer than 48hrs....image

    mark

  • in reply to mcb1

    Thanks Mark,

     

    Of course one of the main reasons for carrying out such an experiment is to be able to make predictions about battery life in more practical situations. As a follow up test I decided to set up an Arduino running a simple double blink program as a load for the USB Power Pack. Using the data I had collected from the experiment and a measured reading of 30 to 32 mA for the draw of the Arduino I made a rough prediction of 48 hours. I then set up the experiment below and let it run.

     

    image

     

    The USB power pack in the back right corner is supplying all the power for the Arduino. The ATX is supplying power for the Battery Analyzer's circuitry. As long as there is voltage from the battery the clock will continue to tick off the minutes. When the test is going to run over 12 hours it is necessary to check the experiment once every 12 hours to log that the clock has rolled over and is still running. When I checked the experiment late in the 4th 12 hour cycle the clock had stopped at 10 hours and 8 minutes. Adding this to the previous 3 complete cycles gave me a total of 46 hours and 8 minutes. This is roughly 4% less than my prediction of 48 hours.

     

    John

  • Now that I actually have internet working I can see the whole test.

    This is a great way of presenting the findings, and clearly demonstrates the issue.

     

    Well done

    Mark

  • in reply to mcb1

    Hi Mark,

     

    Yes the unwary get fooled like this often.

    The term "Trust but verify" needs to be every engineers byword.

    Marketing brochures are aimed at selling, sometimes they lie.  No you should not be shocked.

     

    When possible, you test, just like @John Wiltrout did and get the real parameters.

     

    DAB