Yezl Y3 - a picture breakdown

Gotcha. In that case, you may be right about the performance problems with 1 partially depleted cell. Hard to say without some testing. As you said, 2s will knock the problem out for certain.

FWIW: While I know other people didn’t jump on the Y3 with the same intentions as me, I really didn’t/don’t intend on running the thing with less than 2s. It’s got a buck driver and extension tubes, so 2s+ was a no brainer for me :-). For 1s use I think I’d swap out the driver for something else?

Yeah I was initially one of those. After getting it in my hands I decided I liked the UI just the way it is and figured it’d be easier to look into modding the stock driver vs. trying to get another driver in there and rig up the momentary switch somehow? to make it function like it does when stock. So that’s where I’m at — modding the stock driver. I’m sorry for wasting y’alls’ time — looks like the additional R100 is exactly what I was looking for after all, and now with two cells everything looks ship-shape on the driver end. Time to reflow some emitters, shim up the mcpcb, braid some tailcaps and rock and roll.

I tried all kinds of resistor combos on this light. My 3rd one I have a U2 1A bin dedomed. Replaced the R150 with a R120 and then added a R100 on top (0.04186). 0.25 / 0.04186 is 5.97a? My meter doesn’t read after.

I was testing down quite a bit lower than that but it gets too hot and I lost one of the drivers when testing at 3 cells. At 0.04186 the sink can handle it fine with a 2 cell config and it’s about perfect throw on that sucker. Good balance. Noticed too you can get a better hotspot if you don’t go too far into the reflector. There’s a point where you get like a flower and that throws like crazy.

I wish I could keep this light but it’s going to my inlaw next week xd I’m doing two with a blfdd driver and retiring the mtg2. I didnt like this light for a flooder, my 15 led is good for that.

Some shots… It’s pretty awesome :stuck_out_tongue:


Ok - finished the Y3 mods. Here's some details.

20 mm square, 1.15 mm thick, MaxToch star, penny for size comparison, and originals:

MaxToch 26 shown positioned. I marked a spot for drilling a hole. Original holes are not aligned well:

After the hole was drilled. Used my cheap drill press, ala comfy:

I cut and sanded down the stock LED alignment piece. 1.23 mm thick originally:

Sanded down to 0.75 mm. Used a pad of 100 grade paper, and rubbed the part on it:

Shown here with original pieces:

I reflowed the emitter (XM-L2 U2 1A) and the 20 mm copper shim in one step. Used solder paste, and frying pan, monitoring the temp with a hand help IR temp probe. Surfaces sanded to 2000 GRIT. When installed, I used GC Extreme thermal grease between the shim and shelf:

Backside view. I did more sanding after this pic, before mounting it, to remove that oxidation from the heat:

Though I did drill and tap holes, they didn't come out aligned up well, so didn't use screws - I could have widened out the holes in the MCPCB and got them to work. No worries with any possible shorts to the reflector. Nice low profile, and beam looked well focused compared to a stock Y3 beam - tighter hot spot. Also, the black hole doesn't appear til about 2" from the wall - nice for a Y3.

This has the stock driver with one R100, measured 3.16A at the tail on one KK cell, and 2.60A at the tail with two KK cells. Hhmm - the 26700 did not work -- seems too long. I thought all the Y3's worked with the 26700 - apparently not.

Results:

On KK 4200, 3.16A tail: lumens - 1,170 @start, 1,139 @30 secs, throw: 79 kcd taken at 5m

On two KK's, 2.60A tail: lumens - 1,428 @start, 1,380 @30 secs, throw: 95 kcd taken at 5m

Note: all my lumens and kcd measurements are "my own" taken on an older LX1330B meter - kcd taken indoors and lumens in a PVC constructed lightbox, accuracy can be freely debated (CYA all the way...)

Conclusions:

  • note the 3.16A is sort of low, lower than I tested last night (3.6A) - can't explain, but used a different LED/MCPCB
  • 2.6A at the tail for 2 cells is maybe 4.6-5.0A to the LED I'm guessing (totally guessing).
  • this is with a UCLp installed, driver and tail spring copper wired
  • results using 2 cells is pretty awesome - lots of light, great throw, but the LD-1 driver in my XM-L2 U2 1D Y3 still does better (higher amps, 111 kcd)
  • total height of the MCPCB and copper shim is 2.63 mm, so about 1.15 mm increase. This was need to compensate for the reduction in the LED alignment piece, and the drop in height of the MaxToch compared to the stock MCPCB (about 0.5 mm less). I think the height/vertical dim turned out well - can full tighten the bezel and everything seems snug.
  • I'm impressed it holds output pretty well in the first 30 secs - not much drop there - I've seen far worse.
  • I like it, but don't like the stock driver UI - would rather have full control from the side switch, like the LD-1 driver or my (and others) custom drivers.
  • Better mod upgrade would be a piggyback'ed BLF17DD with custom firmware, but then you lose the multiple battery capability.

Hope this answers some questions and gives you a better idea of what a mod can do for this light.

Copper shim bought in this nice assortment pack: ebay-45pcs-Laptop-GPU-CPU-Heatsink-Copper-Shim-20mmx20mm. Comes to about 25 cents each.

Thanks for the detailed infos on your mod. I will just do the same.
I never used a shim until now, have you bought this piece as a shim or just cut it out from a bigger part?
I have a copper sheet but it’s not thick enough, maybe I have to do a little search in the scrap parts box…
Edit: just searched a bit and there are tons of shims on ebay.

Werner - post above updated with link for the source of copper shims I used.

Hey Tom, that’s what I’m doing next is the blf17dd driver on top of the switch like your ld-1 driver mod. :slight_smile: Been busy (putting up more solar)- but my drivers came in the other day.

I’ll piggy back it, but I’m not too sure where to put the eswitch, I was thinking run it as a temp on/off button or is there a specific place on the board to let that happen? I really need to get learned on flashing these things too but one step at a time.

The stock driver noticed on the Y3 is a lot more tanky than you’d think. I know wight was saying the diode is rated at 5a but I’ve ran that U2 1D as low as 0.036 with 2 cells successfully. It heats up like crazy though, not really practical. If you swap the R150 with a R120 and lay the R100 on top, keeping the R180 in place, it should give you 6a and still runs cool. I can’t measure it cause my meter doesn’t work at the tail after 0.45. :frowning:

You were right about the dedomed 1A. It is a bit more white than the 1D after dedome. The 1D is no slouch though, but I like the 1A better after the domes off. Playing with the depth in the reflector, I wasn’t being precise like you either. I should measure the centering ring after it’s sanded down. I’m just sanding and putting it in until it’s centered and I see that flower look. I noticed if it goes too far into the reflector you actually lose throw light. The beam gets tighter but not as bright? So I stick it back a little but not quite the thickness of the stock ring. It’s probably exactly where you have it sanded, wouldn’t doubt it anyway.

That mcpcb is perfect size… I can already see the cousins asking me to make them a few of these after my inlaw blinds a few of them. I’m not looking to my cousins kid though he always plays with the deer call in the stands, thinks it’s funny… And then there’s the hillbilly drama

Interesting fellfromtree... I was thinking of boosting the amps a little more - I'll do the resistor math and see. for the focusing, actually I didn't do any experimenting - should have. The throw is good, but hoping for a little better, but I'm not measuring ampds to the LED with 2 cells, so hard to tell. Are you having problems fitting 2 protected cells with these extensions?

I had to take off the button tops to fit protected. You can add o rings found too but that’s last resort sort of thing.

If you play with the depth on the dedome on a white surface about 20 feet away. You can see the light definitely loses throw at a certain point. Like it loses intensity- this is just with the naked eye though. The beam gets more floody if you go all the way in, but that’s what I noticed anyways.

Edit
If take a picture it looks different but you can see almost to the point of where it starts to get square… It looks perfectly round in rl. Maybe I need to play around with the camera some- but it has these 4 petals that come off the hotspot… Right at that point you have the richard simmons sizzler!

thanks tom e for the details, smaller quantity shims : 5 X Kupfer Kühlkörper Laptop GPU CPU Kühlkörper Kupfer Shim Sale - Banggood Deutschland sold out-arrival notice-arrival notice

Im watching this thread.

even a little cheaper at ebay:

sets of 10 pieces:

20x20x1mm = 3.62USD

15x15x1.2mm = 2.80USD

etc.

The eBay source though for qty 45 in 9 different sizes is still the best deal, if you want the variety of thickness's.

Instead of shims you can grind the base of the heAd down instead

I looked at doing that - won't the bare aluminum be exposed and seen? Specially if you have the light sitting head down? If so, does it look ok?

tome did you solder the shim with the led pcb ?

Yes - it's noted just above a pic.

I just took some measurements of LED current. The Stock current is between 2.4-2.6A over a wide voltage range. Power input is around 10W.

I will add a resistor and take measurements when I get the retaining ring out. It’s stuck and I used some oil and will let it cool down before I try again. The holes are a bit too small to use my good pliers and with tweezers it was unscrewable…
Edit:
I just played around with the driver. I soldered the driver to my test xml2 and played a bit. On stock there are one R180 and one R150. One added R100 boosts the led current to around 4.3A. Input power is around 23W

I also tried for a short period another R100 which gave around 6A to the LED at 35W input. After that I shorted something and needed to replace some parts so I couldn’t test the 6A for a longer period. After the repai the replaced FET wasn’t capable of the high current and I removed one r100….


I am now sure that the partly defective stock driver isn’t powerful enough for me in this light so I will probably also use a FET driver piggybaggedn one cell config.
Now I need to order a imr26650…

Thanks for the test results Werner. Did you record any output voltages?

Based on djozz’s Vf data I’m assuming >80% efficiency in stock form, down to <70% with one R100 added and <65% with two R100 added.

I didn’t took measurements of the led voltage(yeah I know that was lazy…) and you have to consider that I used a different led(noctigon on big heatsink) for the r100 test. The led voltage on the stock measurement was probably a bit higher than usual because the led was in the stock not ideal pressed down and got so a bit hotter. Also the input voltages are not super accurate so the data looks a bit worse than in reality I guess(that is why I didn’t post a lot datapoints). The regulation is not constant over the whole voltage range there are 10% fluctation like always on these buck drivers…but it is good enough to get an impression of the driver.
I also like a lot that this driver has voltage protection for one, two and three cells.

The efficiency goes down for sure a bit, I would also say efficiency is around 65-70%.

Think'n for my R100 moded light, add an ~R200 - think it would be a nice compromise. Also I have to play with the focus on that light a bit. Just wish I had time...