I did the train test again to fill out the missing fields in the spreadsheet, especially the extra time from dead battery, without the sleeve. So I started with a fresh battery, same brand as my first test, but this time it ran for 169 minutes, instead of 127 minutes, similar to my last test with Duracell instead of Energizer. I guess the reason might be that the train was new and it needed a few hours burn-in time. After the test, I let the battery recover for one hour, exactly as with the Batteroo test. Then it ran for another 31 minutes! Extra time with the sleeve was 4 minutes. Even if we take into account that the train might run longer after the burn-in, it is still a lot longer, albeit at a much slower speed, with Batteroo it was 2 to 5 times faster. This is expected from a boost converter.
As usual, I recorded the test run (only the second run, didn't expect the high difference with the first run). With my OpenCV script it was no effort to measure the individual lap times. I updated the
data and the diagram:
![](http://www.frank-buss.de/batteroo/batteroo-train2.png)
So if you want 4 minutes additional full speed instead of 31 minutes slow speed, after 2 hours without the sleeve, you should buy the Batteroo sleeve
![Smiley :)](https://www.eevblog.com/forum/Smileys/default/smiley.gif)
The cutout voltage for the train is difficult to measure, because it is very low and at this voltage the interaction with the high internal resistance of the battery is a problem. So I used a power supply: the train motor stops at 0.6 V and starts again when increasing to 0.8 V. It might stop when on track a bit earlier, so I used the 0.8 V limit for the spreadsheet, which is what I could measure at the battery shortly after the train stopped, too. This would explain the bad performance of the Batteroo sleeve, because efficiency at this voltage is very bad.