(answer to above question)
When the cumulative total of resistance is too great.
See thread for specfics
Ok to keep it short.
I built two generic S7 lights changing them over to 219C triple emitters lights with Qlite drivers.
A) is 4.56 amps
B) is 3.04 amps
Here is the issue, they work great except for the madding/niggling issue that after a very short while (say 15 seconds max) they both flash (completely off) three times and cut the power way back (I am assuming to 50%).
I have done all kinds of stuff mechanically to clean up the connections (removed, cleaned tightened, took apart put back together, etc, etc), and no it is not actual low battery voltage, I have charged three different sets of different 18650 cells no change.
I can’t remember, what is the actual low voltage warning signal? Either way, have you done spring bypasses? If you remove the tailcap and test the light for over 15 sec, does it still step down?
They have six modes (no blinkies).
The low voltage signal is (the three blinks then cut back on power).
It only happens in the highest mode (and as they are both floody 219C lights I need all the power I can get)
I removed the tail cap and no even in #6 it runs fine.
And yes I did a tail cap spring wire addition.
But boy it sure looks like it is in the tail cap (which was what I had suspected before, you idea of the cap removal test).
Are some switches just so bad they need to be replaced? (it seems to “work” well enough) but then this damn issue won’t go away.
7135 chips have a built it step down temperature threshold, which is when they hit 75C-80C (I can confirm it). Just in case maybe sometime you will encounter this issue (and it is not because of what program the MCU has).
Buy some of the cheap omten switches these are very cheap and also very low resistance. Simon has the five pack in his aliexpress store. Also djozz has a nice thread comparing switch resistance.
I have a variable power supply and I always adjust it to 5A and connect it to the switch, then I measure the voltage drop above the switch to see if it needs to be switched. Sometimes the pcbtraces are very thin than I drill a hole in the PCB to connect the spring bypass wire directly to the switch terminal.
You have to think of ohms law if you have a light which pulls 5A and has a bit resistance(all resistance added in series:battery, switch, cables, springs, button tops, protection circuits…)
Which gives a resulting resistance of 0.2ohm you get 5A*0.2V=1V voltage drop.
Considered that the battery voltage under load is maybe 3.7V the resulting voltage on the driver is 3.7V-1V=2.75V which lets the low voltage protection kick in.
that is why it is important to check all things the current needs to flow.
Ok here is the short check list of what fixed it (for now)
Took tail cap apart (again)
used brake cleaner to dissolve the remains of the rosin from the previous wire by pass
used Dremel tool to cut a groove in the top of each plunger.
put a dot of solder on the (brass plated aluminum plunger cap)
threaded the copper braid through the hole and soldered it to the top of solder on the plunger cap
pulled the spring apart for more tension
threaded tail of braid to be pinched in bottom of spring (and put a spot of solder on it also)
re assembled cap to test
At this point one or the lights worked and one did not (regardless of which light had which cap)
Checked battery and again the same light worked with either battery
after eliminating the cap, battery, driver and the threaded connections I checked the tension of each battery
It turns out that the two copper pennies I used as a shim (the S7 is 3xAAA sized so the cheap tube is to take up the slop on 18650) this also hold a penny perfectly.
So the light with remaining issues I soldered a third penny to the previous two pennies used as a shim between the plunger and the back of the battery (all pressed into the cheap 18650 rattle tube)
Bingo now both lights work.
Part of the issue was one of the lights used the closest 2+ heatsink/spacer at 7mm when it needed over 9mm but the next size up was 13mm which is too large. so the 7mm + three pennies on the back, tightened it up enough (but not too much) to resolve the remaining connection resistance point.
I am a bit confused…which light did you mod? Until now I thought we were Talking about a Tubestyle light like convoy s7 but the 3xAA size made me suspicious…
