[TriEmbed] Battery Rundown Tester - Is this Interesting?

Charles McClelland chip at mcclellands.org
Tue Jul 8 15:13:24 CDT 2014


To all, 

Since I can’t make it to the meeting on Monday, I wanted to share my most recent creation and see if it sparks any interest.  I have been working on some switching power supply designs with Pete and I got all excited about the Texas Instruments TPS63031 Buck-Boost 3.3V DC DC converter which could handle input voltages from 5.5 down to 1.8 volts.  I got it in my head that it would be neat to build a test rig which would allow me to test different converter designs and different battery chemistries, sizes and packages.  Here is what I cam up with:

The setup uses a relay to connect a battery pack to a constant 165mA load.  Every ten seconds, I measure the input and output voltage and current and log it to the terminal in a table I can then analyze to compute the efficiency as the voltage from the battery drops and the converter moves from buck - to pass through - to boost modes.  

I have documented the components and their connections in the sketch: 




In order to keep this email small, I have put all the files for this project in my public Dropbox folder - here   I have included the output for two runs - three AAA NiMH (~700mAH) batteries and a single cell LiPo (2000mAH) pack. In each case, the measured capacity from the test rig matched the expected capacity of the cells.  

Here is what I learned:
	1) Buck is where it is at - at least under a 165mA load.  95% of the run down was in the buck or pass through modes.  Adding the ability to “boost” added only about 5% to the rundown test.
        2) It seems the biggest advantage of these buck boost converters is that they have no “drop out” voltage.  Neither of these packs would have worked at all with a converter with a 1V drop out like the 7803
	3) The current demand on the battery supply spikes dramatically once the converter switches over to “boost” and the efficiency of the converter drops like a rock.
	4) That SparkFun LiPo pack delivers a lot of juice in a small package - almost three times longer on the rundown in a package that is much smaller than three AA cells.  

So, my question is whether this is an interesting project or nothing more than a noobie - me -  “discovering” the information that is already in the data sheet.  My plan was to try some different battery packs and some different DC DC converter designs in order to better understand what the best combination may be for my projects but can I get this information using already published information?  

If this is something that could be of general interest, I was thinking about building a simple PCB board so I am not tying up one of my breadboards  What do you think?

My next step is to take Jim’s advice and test my actual circuit with amore realistic load and a big capacitor instead of a battery to see the relative performance under a more realistic and modest load.  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.triembed.org/pipermail/triembed_triembed.org/attachments/20140708/9c0ca778/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Battery_Rundown_Tester.ino
Type: application/octet-stream
Size: 11113 bytes
Desc: not available
URL: <http://mail.triembed.org/pipermail/triembed_triembed.org/attachments/20140708/9c0ca778/attachment.obj>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.triembed.org/pipermail/triembed_triembed.org/attachments/20140708/9c0ca778/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Screen Shot 2014-07-08 at 4.11.05 PM.png
Type: image/png
Size: 50542 bytes
Desc: not available
URL: <http://mail.triembed.org/pipermail/triembed_triembed.org/attachments/20140708/9c0ca778/attachment.png>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.triembed.org/pipermail/triembed_triembed.org/attachments/20140708/9c0ca778/attachment-0002.htm>


More information about the TriEmbed mailing list