Please step inside to help me mod my 6 x XM-L SkyRay XY-600

Thinking it would be a fairly descent flooder, I bought a SkyRay XY-600 during the Wallbuys Christmas sale.

First, the condensed version. THIS LIGHT SUCKS IN A HUGE WAY! The small deep poorly polished reflectors are worthless and the light is extremely under driven. I would like to change that. I was hoping that some of you could offer some suggestions to help make this light worth modding, rather than throwing it into the river (where it belongs in its present form)… it really is that bad, especially when compared to an SRK with 3 terroid driver.

The long version:

Out of the box, this light is a total fail – but might make for an excellent mod host.

The SRK (the good one with the original 3 terroid driver) blows away my SR XY-600, both in brightness, spot and spill. The problem with the XY-600 is that it doesnt know what it wants to be when it grows up. Combined with an anemic driver, the tiny reflectors are far to small and deep to be efficient, so they throw a small (but not distant) concentrated hotspot, all while wasting a huge amount of lumens in potential spill that may have been realized with shallow reflectors. Also the reflective coating isnt as shinny as it should be, which absorbs even more lumens. The tint is the coolest white that I have yet to see in any XML… absolutely horrible.

Tests taken @4.18V

SR XY-600 high mode before losses – 0.76A per emitter
H – 4.60A
M – 2.30A
L – 1.37A

For comparison

SRK high mode before losses – 2.1A per emitter
H – 6.20A
L – 0.55A

On the bright side (uh, well it isnt that bright at all… lol) my example is very well constructed and free of any obvious imperfections. If driven to 3A per emitter with 3 or 4 XM-L’s, the large heavy head should probably be able to wick away heat about as efficiently as the SRK (which is driven @2A per emitter x 3). The switch is a clicky contact type, not electronic. Since this light is 4 x 18650’s in parallel (4.2V), a 7135 based driver solution seems like the easiest choice.

Goals:

  • Mod for a wide angle flood light with minimal beam artifacts.
  • Not exceed 12A total draw. 9A max would be preferred.
  • Attempt to maintain a balance in efficiencies (lumens per watt, driver efficiencies, reflector, battery life, heat sinking)

Considerations:

  • Im thinking about removing all the components from the stock driver and using it as a contact plate. Then attaching 1 x 3.04A 7135 driver per emitter. Would this be a good one? Im considering the 3rd Star option : Low (5) - Medium (30) - High (100%). Can I ground the memory on the drivers to prevent synchronization problems between the drivers? (so it would always come on in low first and have no memory). That, or is there a way to sync the drivers modes and maintain memory without having to go to a master/slave configuration (as described by Hill in post 32)
  • I want to get rid of the stock reflectors and mod for 3 or 4 much larger ones. Preferably 3 if I can find the correct sizes.
  • I would consider 6 replacement reflectors (and drive each emitter @1.5A - 2 emitters per driver) if they would produce a better flood beam (more on that below).
  • I could probably remove the emitters and put a round aluminum plate under them to shim them closer to the lens and swap in larger diameter/shallower reflectors.
  • Can anyone recommend some quality small shallow reflectors that play well with XM-L? (prefer to avoid TIR because they arent very efficient, unless you know of some that are)

Possible emitter configurations:

  • 3 x XM-L U3 at 3A each = 3120 lumens nominal - I prefer this option.
  • 4 x XM-L U3 at 3A each = 4160 lumens nominal
  • 6 x XM-L U3 at 1.5A each = 3744 lumens nominal
  • 6 x XM-L T6 at 1.5A each = 3276 lumens nominal
  • 6 x XP-G2 R5 at 1.5A each = 3108 lumens nominal

I will add photos and parts measurements to post #1.

Thanks to all that stayed with me so far. :slight_smile:

Parts pics and measurements:


This is the reflector array removed from the light. Each reflector cup sits within the 6-hole reflector retaining plate.


Each 14mm emitter board sits flush in a recessed hole cut into the head. I removed the electrical isolation plastic thing from the center emitter to have a better look. Emitters are parallel wired.


Closeup of one of the reflectors.


Driver.


Closer look at the driver. I dont think anyone knows how to squeeze 9A out of this with a resistor mod, do they?

Measurements:

  • Reflector retaining plate: 75.02mm x 6.07mm
  • Reflector: 17.67mm long x 20.79mm wide
  • Head housing ID (where the reflectors sit and where replacements would go) 68.17mm
  • Driver: 45.96mm x 2.18mm (same as SRK)

I wonder if you could use the DRY driver and wire the emitters 3s2p

If it were mine (and I was very close to buying this at christmas sale), I would probably find all the reflectors and/or optics that were the right height, then decide the rest, including the number of emitters.

If I found a good 7up optic that would fit, I’d probably strip the components from the stock driver, take this driver
http://www.kaidomain.com/product/details.S020166
separate the components and mount them to the stock driver board.

Is this driver regulated?
Theses resistors look like it is, this makes hope….

Thanks dthrckt. Depending on optics, Id like to run 3 or 4 emitters at 3A each or 6 at 1.5A each. Can you recommend some good efficient optics for a flood beam?

The driver you linked looks great, but: Working voltage: 6-12V. I need to stay at 4.2V

I think so, but going from 4.6A to 9A might be a stretch. I need others to take a look and comment.

If it were my light. I would stick a small Aluminium heatsink to the back of that mosfet driver.

Then I would place one of these http://www.fasttech.com/products/0/10002999/1224400-10-ohm-potentiometer-trimpots-5-pack in parallel across any one of those sense resistors.

Then ‘Very’ carefully dial in the amps you want while simultaneously watching your amp meter.

Have you measured tailcap current on lower voltage? If it is higher than its regulated and if it is you could do a resistor mod. Worst thing could happen is to burn the useless stock driver.

Are all the LEDs parallel? Are the resistors in parallel?

Spasmod & Werner, great ideas! I’ll have to rig an axillary battery pack with lower voltage and take some readings. All emitters are wired in parallel.

I'm in the process of doing the same (or some kind) of mod to this light as well. I currently have it dissembled and posted a possible mod here (post #32) using Techjunkie's multi-AMC7132 circuit. He responded it could feasibly work provided I reconfigure the circuit with a bypass since 12A would likely fry the stock switch. He also suggested to use IMR cells for this application due to the high current necessary, but was really not keen on having parallel batteries due to the potential of uncontrolled cross charging. A better config would be to use 6 separate boards, but that would require some pretty skilled soldering and cramming them into the driver cavity.

I am open to any ideas here...

Thanks Hill, I had already linked your comment in my original post above. I know there is a lot there to read, but Im trying to avoid a master/slave setup if possible. If you can get it to work, have you considered how much heat 52W will generate within your host? (especially with the constant current IMR will deliver till the cells drops off). IMO, its probably going to be way to much heat to deal with for anything over a 10 minute run. Im considering 3 x 3.04A 7135 drivers, one per emitter for a total of 38W. Remove the memory from the drivers and run it in 3 mode without memory. 1.3A per cell should give descent run time without IMR.

Ive had far better luck while using 4 x 18650 matched cells in parallel rather then using them in serial (especially while driven hard). In parallel, if initial cell voltages are close to equal at the time of discharge, they will all discharge at the same voltage… they dont have a choice. Ive had some startling and often not repeatable results while measuring serial cell voltages after discharge… even while using brand new expensive tested and matched cells.

Are you going to change your optics or run stock?

The driver in mine is the same except R4, R5, R6, R7 and Q1 were all resistors marked 0

duh, what was I thinking, same problem w/ dry driver - there’s really no way to change battery config to serial.

unfortunately, I know very little about optics. I have done quite a bit of searching for small reflectors though, and CNQ has the best selection of reflectors near the size of what you require. kaidomain and dealextreme also have some.

spasmod’s suggestion sounds like a great way to start, and if it gets fried, you could pursue the amc mod.

Hi,
It’s a long shot but maybe RIC can help you , I bought an APEX 5t6 from him to make my project (APEX 5t6 bike light) and it was under driven 5.5A H and he send me a new replacement driver 8.5 A.
The driver alone is not listed in his site but maybe you can arrange with him the price and send you one - still you will have under driven t6 with 1.6 - 1.7 A but the amount of light will be good enough like an APEX
Jim

jlogres, thanks for the tip. An 8.5A driver split 3 ways would deliver around 2.8A per emitter (or 1.42A if driving 6 emitters) . Can you please measure yours and tell me the diameter?

And then the pot would burn up… they are only rated for a few milliamps of current, not the manyamps that can flow through a sense resistor.

I remember reading somewhere that the sense resistors only saw a few milliamps, I'm guessing that is incorrect?

yeah, i remember someone saying that too

No! Sense resistors see the full battery (or LED) current.

The sense resistor is direct in series with the led, so it sees all current. Of course it’s splited to all parallel resistors. That’s the way this works, the mcu measures the voltage drop and if there is a lower voltage because of a more empty battery than the mcu regulate the duty cycle up so that the current stays constant over the whole voltage range.

These pots are 0.25W max. So if these resistors are parallel we have

1÷(1÷0.082+1÷0.082+1÷0.082+1÷0.082)=0.0205 ohm
If we want double the current we have to parallel another 0.02 ohm resistor, that’s a too low value for the pot either.
P=R*I^2……so Pmax @3.5A. But real current 4.5A…Preal=0.41

Not too bad for such a little pot, worth a try perhaps to see what we can get from this driver and that is what spasmod said. I would try this too.

But first he has to find out if it is a regulated driver….