Today I received four Trustfire Flame Protected 26650 cells from Fasttech only to find out they will not fit in my Ultrafire T90. I sort of figured this already so I was prepared to remove the protection PCB if necessary. Once I got it removed my curiosity took hold and I tested the PCB using a bench power supply. My main goal was to document the voltage drop across the various components of the PCB and any tabs in the current path. What I found was measurable but not enough loss to be too concerned about.
Let's take a look at the PCB itself:
A brief description of the important bits is included in the picture. I'll refer to these points so it is clear where things are connected when testing.
The PCB is connected to the battery using metal tabs that are welded to the battery - and battery +. Only the Battery - tab needs to carry the load current. The + tab is a Voltage only connection that is used to monitor the battery for over/under charge/discharge. It also provides the miniscule amount of power needed to run the controller chip (I measured 3.2uA).
On the outside we see the Protected- or P- end of the PCB and the Battery+ or B+ end of the battery.
Using a traditional current flow approach, current flows from the B+ end of the battery, through the load (driver board, LED, laser, flashlight body, whatever; I'll call everything outside the battery and protection PCB the 'load'), back into the P- end of the PCB, through some PCB vias, through the FETs, through the B- tab, B- Weld points, and finally back to B-. Circuit complete.
Text schematic:
B+ --> Load --> P- --> P- vias --> FETs --> B- Tab --> B- Welds --> B-
The power supply will simulate the Battery, so it connects to B+ and B-. The resistor is our load so it connects between B+ and P-.
Here is a table of each component and the voltage recorded at ~3A of load current.
| Item | Description | Vdrop @ 3A |
| P- vias | Connects P- pad to rest of PCB |
18.0mV |
| FETs | Electronic switches that cutoff battery | 92.0mV |
| B- tab | 10mm long x 2.8mm wide x 0.14mm thick | 12.7mV |
| B- weld | Three tack welds | 2.5mV |
| Total | Sum of protection circuit Vdrop | 125.2mV |
125mV is not too bad. This represents about 0.375W of power lost in the protection circuit when running at 3A. If the load sees an average of 3.6V, or 10.8W at 3A, then the percentage of power lost to the protection circuit is about 3.5%. A bit of loss, but not a lot to worry about.
I will note that not all protection circuits are built the same. This is from a 26650 cell which is built for higher loads. An 18650 protection PCB will often have only one FET which doubles the drop for that item alone. It may also use thinner traces and tabs.
If you find yourself wondering why a flashlight is not giving the emitter current you expected, you may want to look at the host itself before thinking it's a cell protection circuit. I regularly measure 300-500mV of drop on clicky switches with springs. Usually it is the spring that is the culprit. Adding a piece of copper braid to the spring can help a lot. other areas to look are driver contact plates/springs thread contacts, and emitter wire size.
Thanks for reading!
P.S. In case you are beginning to think I'm a big fan of protected cells, I'm not. However, I will recommend them to anyone who is not willing or able to monitor the status of their cells regularly. They are in no way a fail-safe product simply by adding a bit of protection. Even 'protected' Li-Ion cells will not tolerate abuse like NiMH or alkaline.
