I have been playing with arrays of 18650 batteries for the last 10 months, charging and discharging them all the time with my LM338K based power supply, that, until tonight, where I inadvertently connected (backwards) a semi charged array (pack) of 2P3S delivering 10v and who knows how many amperes in shortcircuit, and it looks like I fried the LM338K, I know I can avoid that with just a single beefy diode in series, my question is: is there any other way where I don't lose the 0.7V voltage drop in the diode? Should the regulator have survived? The circuit for the regulator is the one in the datasheet and surely it does have overload protection and overcurrent/thermal and whatnot, but I think hooking my battery pack in reverse was just too much and unexpected.

  • \$\begingroup\$ Schottky diodes have lower dropout; a simple part swap. also, the lm338 does not have OCP, I exploded two last week hooking up a shunt to the wrong wire. twice. i hate hearing "POP"... \$\endgroup\$
    – dandavis
    Oct 26, 2017 at 5:12
  • \$\begingroup\$ A small suggestion - use mechanical polarity connection. If you (semi) permanently connect a polarized connector to the battery pack, with the mating connector connected to the power supply (as well as your load), it’s simply a matter of plugging and unplugging the connector. Be careful to select a connector which is rated for the current and avoid plugging and unplugging while much current is flowing. \$\endgroup\$
    – DoxyLover
    Oct 26, 2017 at 5:36
  • \$\begingroup\$ @dandavis Schottky diodes do have greater leakage currents. This could mean that you have higher errors if you are drawing very low currents. \$\endgroup\$
    – Joren Vaes
    Oct 26, 2017 at 6:41

1 Answer 1


You will need two sets of protection and separate the charge input from the discharge output. If you just put a diode in series with the pack, it will be reverse biased when you try to charge the pack.

Depending on the level of performance needed, you can use a simple PFET circuit (http://www.ti.com/lit/an/slva139/slva139.pdf) or an ideal diode controller with an external FET. Some of the ideal diode controllers use NFETs which (apples for apples) have lower Rds-on resulting in lower losses (and less heat).


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.