1
\$\begingroup\$

We have a PCB with a DA14683 that is powered by a 90mAh Li-ion single cell. For testing the software and hardware of this PCB, we want to emulate the battery. For example, we want to test how the PCB behaves when the battery voltage drops below 2.8V or the battery has very low internal resistance. Another example is to verify charging current and charging voltage in the different charging phases. What's a schematic that emulates such a battery and can be controlled by another microcontroller?

There are basically two tasks: a) battery being charged and b) battery providing energy. For b), we can simply use an adjustable LDO like the one used in DA14683-00A9DEVKT-P and ajust the voltage with a microcontoller, here's the schematic: enter image description here But we also want to test the switch over to charging the battery, so we need both tasks integrated into a single solution.

There seem to be commercial devices such as the Keithley 2281S for USD 3000? or qoitech.com for USD 1600 with 1 year of software use. For one, I'd like to understand how they work and second, I prefer to have the emulator integrated in our development environment and production testers. ADALM1000 provides a 2 quadrant source, could it work as an example for an battery emulator? Are there any application notes or open-source projects for Li-ion battery emulators?

\$\endgroup\$
5
  • \$\begingroup\$ I can't see a sensible reason to integrate your own spin of the 2281S into your production testers from a QA point of view. \$\endgroup\$
    – Andy aka
    Commented Jan 14, 2023 at 11:11
  • \$\begingroup\$ If it was only for production testing and money is not an issue, I would agree. But we also want to have (preferably the same) solution for our development process. So every developer needed a Keithley and integrating external devices into our continuous integration test setup is also quite a task. Finally, I suspect that I'm just not aware of an existing, simple solution of battery emulators, that's why I ask :-) but I might be wrong? \$\endgroup\$
    – mpa
    Commented Jan 14, 2023 at 11:37
  • \$\begingroup\$ Presumably R127 is a digipot controlled by your external MCU. How about integrate a TLV431 with a PNP power BJT and a digipot to make something like a programmable load, and put the two solutions in parallel? \$\endgroup\$
    – Neil_UK
    Commented Jan 14, 2023 at 11:45
  • \$\begingroup\$ @Neil_UK great suggestion with the TLV431. I pushed it a bit further but I'm not sure about the behaviour when transitioning between current sink/current source/voltage source. While learning more about source measure units (SMU), I found the LT1970. All the hard stuff is integrated in the chip and current/voltage source/sink limits can be programmed with 3 voltages. Though not sure yet how low the currents can go.. will post an answer once I come around to implement it. \$\endgroup\$
    – mpa
    Commented Jan 28, 2023 at 9:34
  • \$\begingroup\$ @mpa LT1970 sure packs a lot of functionality into one package, just had a quick drool over it. Now where could I use something like that? The bahevaiour when transistioning between sinking and sourcing? It's all about the accuracy and stability needed to maintain a deadzone between them, or limiting the unwanted current that flows when the deadzone vanishes, perhaps using a finite gain rather than comparators. I'm sure it's easily handled. \$\endgroup\$
    – Neil_UK
    Commented Jan 28, 2023 at 10:01

0