Ultrafire JS-8054 Thrower - Driver Exposed - Heat Sink Finished...light is completed. Post #92

I’ve opened this new thread as a continuation of another open thread introducing the Ultrafire JS-8054 it includes more pics of initial breakdown and modifications to the tailcap.

I wanted to make sure I got the attention of modders that may not have looked in on the thread. I need advice on improving the stock driver. Not being very circuit savvy, I need someone to hold my hand and take me through some baby steps.

Stock emitter readings are:
High - 2.61A
Med - 0.77A
Low. - 0.24A

Lots of spots for more chips. :bigsmile: FT 10 pack amc7135 380ma current regulators
Were you looking to increase the current?

as said, there’s lots of room for more 7135 chips.

7 x 380ma = 2660ma.

pretty close to what your reading, why not whack a couple more 7135’s on there and see what it reads then?

With the top left group pf 6 * 7135 pads, looks like the ground isn’t connected to the others. Right in the center of the board there is a blob of solder & a small round pad to the left, that should connect them. You can also connect ” L3- ” with ” L2- ” or ” L1- ” if its easier.

It seems the pcb was designed to also be used for 3 * xml. 3 groups of 7135 can be separate or bridged together.

Yes, this light also comes in a triple emitter model. I’ve stacked chips on top of chips before, but am not sure how to go about soldering them on to a PCB.
Which pads would be the best choices?

The two directly inline with the existing 2 between L1+ and L3+ that already have solder on the contact points would be my first choice if only adding 2 more. As mentioned you’ll have to close the bridge in the middle of the board to make that bank part of the circuit, but with the others you’ll have to clean the contact, flux it and add solder, then put down the 7135 and solder it in.

This is an easy amps upgrade, just adding 5 more 7135's to get it up to 4.2A but I'd get it up to 5.6A either piggybacking or enabling the other bank. De-domed XM-L2 U2 1A at 5.6A with a reflector of that width should throw way up there, all depends mostly on the inner diameter of the reflector. The K40 is just a couple of mm's narrower than a TN31, but the kcd difference is huge.

Is this how you mean to bridge it?

Yep that is exactly it.

it’d be easier to bridge l1+ and l1- to l2+ and l2- then mount the chips straight onto the spare places on the board.

Just to be clear, don't bridge L1+ and L1- . No + with a - .

Minus with minus only. The positives don't need to be bridged at all.

L1- with L2- and / or L3-

obviously. ……bridge l1+ to l2+ and l1- tol2- although I was about to add, that solder blob in the middle of the boatd looks to already be achieving the same thing, so bang down 4 more 7135’s above the two lonely ones and go from there.

What is that little yellow line I drew bridging? Is it also connecting those L3 pads that already have solder on them?

Yep, that top right group (2 populated, 4 empty) is already connected. Those are good spots.
You have flux, yes? Solder wick?

Its bridging L3- and L1- (L2- is already bridged with L1- by the blob). It connects the the output ground of the 7135 regulators.
7135 chips regulate via the ground. LED positive goes straight to battery positive.

I have everything I need, although those chips always seem a lot smaller in real life. I’ll bridge that dot in the center and add some chips. I’ll take pics for you guys to inspect before I even test it. Does it matter whether I us 350 or 380 chips?

Do you think I can just remove those three wires for the charging port without any ill effects?

Only needs to be 2 wires for a charging port. The 3rd could be a switch (in the charging port) that connects only when there is no plug in the port.
Check for continuity between the two reds.

Mixing 350 & 380 chips should be ok. I believe others have. Your High of 2.61A seems to indicate they used 380 chips. Not sure why when 350 should be cheaper.

If you bridge, all the power from the other banks has to flow through the thin traces going to the L2- pad. The traces are probably adequate, but you can go another way if you want to reduce the risk of resistance in those traces. You could put the extra 7135's on the outer pads where they are better connected thermally to the external heat sink. Then connect all the L1-, L2-, and L3- together and then to the emitter. If go that route, I would not recommend bridge in the center so you don't create any feedback loops.

I'd do them all in one go, with solder paste on the pads, set the chips in place, then melt with hot air from a butane soldering iron (with no tip, so it blows hot air out the end). Solder paste will melt before the chip package is damaged.

Hmmm. I don’t have any solder paste…well, I have some dubious paste from FT, but I don’t think I’ll experiment with it on this board, or a butane iron. I think I’ll just have to try my hand with old fashioned reflowing by heating the solder pad and sliding the chip into place.

As far as bridging L1-L2-L3 externally, L1 and L2 already seem to be bridged by the center solder blob. Bridging the blob would only be connecting whatever chips I decide to add to the L3 sector.

Why do you think they left out that one lonely pad on the bottom of the board in the L2 sector?

It was slow at work today, so I had time to think. Would it be possible if I populated the entire board with 7135s, leaving the L1 and L2 sectors linked as they are now, but linking the L3 sector to a switch installed in the charging port. That way, by turning it on by the stock tail switch I would be powering 12 x 7135s. By clicking the extra switch in the charging port it would result in a kind of turbo mode engaging six more 7135s from the L3 sector, for a total of 18 x 7135s.
Does this sound feasible? I haven’t even looked to see if a switch will fit in there. Maybe I think too much.