7135 drivers with higher voltages revisited

Uses a voltage regulator chip (same size and package as 7135) to protect processor from higher voltage battery.
Uses the forward voltage of the LEDs to protect the 7135 chips from the higher voltage battery.
Good for modding lights that have series batteries( 3C,3D Maglights).
Good for lowering current from remote packs like bike lights or lower current switches
A battery pack with less sag will likely still need a resistor but I did everything in my power to push it. I don’t have any more cells to throw at it.

nice job! If you still have any energy left, give yourself a pat on the back :slight_smile:

Current eta on the new LDO chip for the 8.4V 2s2p xte mod is Monday. I think I will use my 380mA chips for this one now that I know things work. There will be 8 chips in two sets of four(1.5A) with each set pushing two series Leds so each led will be safely driven at max. The heat sink is ready and the 7612 board is stripped with the traces cut and the excess board cut off. It will look similar to this but with two output wires(black one).
A next logical step might be to see if improved heat sinking would allow this mod with the 7135 chips remaining on the board. I have seen various attempts with potting but none that involve direct soldering to the board.

The new chips came today. These are very low dropout voltage regulator chips. They do the same thing as the 78L05 in lowering the input voltage to the Atmel Tiny13A to 5V but will operate down to a 2s battery voltage of 5.5V. The max input voltage is 16V so these are not an option for a 4s battery pack. The 78L05 has a dropout voltage of 2V and would only work down to 7V, good enough for >3s but not 2s. I should be able to play with one on Sunday. One disadvantage of the new chips is that they require input and output capacitors so the Mouser order also has some insanely small 0805 caps. More carrot juice please.

So I open up the box from Mouser and everything’s there, the new low dropout voltage regulator chips, the capacitors…(omg, ever try and solder a piece of coarse salt?). I actually knew how small they were because I sized them to sit end to end across the three pins of the regulator. Anyway, no need to worry about board space.
Skipping over all if the pita soldering the set up is similar to the previous one with these differences: Instead of three XML’s in series there are four XTE chips on a 25 mm quad pcb wired 2s2p. As before there are eight 7135 chips slaved to the modified mcu board but the outputs are split into two groups of 4 chips each. The battery pack is two Keygos 26650 4800mAH(right!) wired 2s. I kept the same set of resistors as before to allow me to ramp up the test to full Vbatt.
I first used a 1 ohm 4watt resistor(effective) and everything was fine. I measured the voltage drop across the resistor, each string of emitters, the driver, and the battery under load. I did this at low, medium, and high. I ran this series of tests twice more with a 1/2 ohm resistor and no resistor and as one might expect the emitter voltage increased in each series as the voltage across the resistor dropped. On the last test with no resistor and the driver on high the drop across the 7135’s measured 87mV(no load Vbatt was 7.95V by then) and the only thing getting hot was the LEDs. The highest drop across the driver(mcu+7135’s) was 3.25V on low with 1 ohm resistor. I’d have to conclude that this mod passed with flying colours. The next step will be to take one of my 3A Illumination Supply boards and run this test with the new regulator and the chips left on board. It’s almost 2am so pics of the setup will have to wait.

It works just fine leaving the 7135 chips on the board. Or , I should say, the second one works just fine. The first one died a fiery death under my iron so I hit the poor defenseless bugger with a hammer. Made ME feel better anyway. As I said, it works just fine using a voltage regulator chip for the mcu and up to 3 LEDs in series with an equal number of Li-ions(4 LEDs if 2s2p). I suppose various combinations like 3s2p or 2s4p using slaves shoul work as well. The main thing I’ve found is that virtually no heat is generated by the driver on low or medium and on high the heat generated by the LEDs far exceeds that generated by the 7135 chips.

Sweet! You've really come a long way since you started out with this Rufusbduck

Congratulations on the progress :-)

I cringed a bit seeing this. I don't even like soldering the 7135s.

Thanks for the great read.

I solder the two caps end to end across the three legs. That puts one across the input and one across the output of the VR chip. Getting them to stay put till I solder them is the biggest pita. I’m thinking of making a cradle to pre install the caps on the pins before the chip goes on the board. The rest of the mod is otherwise pretty easy. I haven’t posted in this thread recently as since this part works, I’m working on the next part which is a battery fuel guage similar in size to a 16mm pcb. Maybe even use it to replace the tailcap switch pcb and show the charge level. I think I understand the circuit well enough. The question is do I A) have boards made and populated(expensive), B) have boards made but populated by me(less expensive, more pita but I can make them for different voltage at no added cost), C) make the boards myself using photoresist on either presensitized(better for small traces) or regular pcb material. Feel free to chuckle if you think I’m in over my head.

you might be in over your head, but you’re a strong swimmer!

I mistakenly posted the following questions in a different thread:

I meant to post here because this is were the original idea was posted.

I would like to try the mod in Post 22 above. I want to use it to drive an MG-T2 with 2 li-ions or 6 Ni-Mh’s. I’m not following the bit about adding a resister with a switch to cut it (the resistor) out at lower battery voltages. Would the resister part be necessary for my app which has fairly limited difference between supply voltage and Vf?

Secondly, wouldn’t the 2 volt dropout from the 78L05 serve as a low voltage cut off? 7 volts comes out to 3.5v for each cell.

Not needed. Remove & replace the original diode with a ~100 ohm resistor, and put a zener diode across (in parallel with) the top of the capacitor on the other side of the board. That's it, works fine.

I have pictures of what to do posted here somewhere, but with the forum's method of only listing threads started instead of all posts in all threads, it's a serious pain to go back and find stuff later, like now...

Click "My Account"

Click "My Posts"

Click "General Settings"

Change "Content" from "Node" to "Both"

Click "Apply"

Thanks comfy. I’ve seen your mod and it is great. Haven’t tried it yet, but will eventually. I already have some of these regulators and want to give this approach a shot.

The 78L05 is only regulating the supply voltage of the mcu and is not otherwise involved in the led power circuit. Since the mcu will operate down to under 3V(1.5V/cell), there is a risk of over discharging the cells unless you pay attention to the led and recharge the cells when the light begins to dim on high(<6V or 3V/cell). With an overdriven led this should happen at a higher voltage but at lower output there’s still a risk of over discharging the cells. The switch and resistor was a means of burning excess voltage over what the 7135 chips could handle and when the battery voltage drops to something manageable, the resistor would be bypassed, allowing the 7135 chips to handle the remaining load. Probably not necessary with 2cells/2xmls or 2cells/1mtg but may be needed for 3cells/3leds and certainly for any more than that.

Thank you Rufus. So the regulator will just quit regulating below 7v and just let the current pass. Makes sense.

Thank you for your opinion on necessity of the resistor/switch. It’s worth risking inexpensive 7135’s to not have to develop a separate resister circuit with a mechanical or automated switch of some sort.

I’m not an EE so I don’t know what happens to the 78L05 below minimum but there’s another chip, the LD2981 that works down to 5V and is the same package size and pin layout. It would save a boatload of effort if a nanjg driver were layed out with this chip located next to the mcu. With more mtg mods going on it might make more sense now.

Check out the LM2936, it has a VERY low quiescent current and a 0.2V dropout.

The 7805 chips typically look like a 2V dropping diode at input voltages below their dropout voltage.

Rufusbduck, I'm finally going to try your idea in Post 45. I'm not going to surface mount, but I am going connect the L78L05 to the same point as in the below pic.

Question. Will the 5v regulator consume power all the time (i.e. even when light is off)? I can test this myself, but was wondering if you know right off the bat.

No, only when the circuit is complete. It does the same thing as the Zener mod but has slightly wider input voltage range so it could work with 2s liion - 2s led, 3s liion - 3s led, or 2s liion - MT-G2. The LD2981 will work for any of those. The L78L05 has a slightly higher voltage drop and works better on 3S mods. DrJones was the one who pointed me towards them way back when.