Re-visiting the Q5 recoil thrower: mod with a a White Flat

While discussing the good old Ultafire WF-008 XR-E recoil thrower here, Agro found a aliexpress source for a clone of this old-school thrower and so I bought one a while ago. Looks cheap but so was the Ultrafire, and the (glass! :+1: ) mirror looks like it had better times but it is good enough to me.

It has been mentioned many times over the years but I never saw it happening: modding it with a different led. And to be honest, still the XR-E is hard to beat: very throwy and relatively small beam angle, the WF-008 seems tailor-made for this led. I gave it a go anyway, with a White Flat, knowing that the beam angle is wider and light is wasted outside the mirror, but hoping for enough output overkill to still beat the XR-E.

The XR-E is huge, and in this flashlight it is not mounted on a board but directly with its central thermal pad on a small brass pillar, with the ledwires hand-soldered to the pads on the led itself. The White Flat is much smaller and although the brass pillar can be adapted (with some very fine soldering :weary: ) to direct mounting, I chose for adding a ledboard, an adapted 10mm DTP board (bought it some time ago from ASF which is vestureofblood’s store), the total size is almost the size of the original XR-E. The led was reflowed and the board soldered on the pillar in one go with the blowtorch.

Then the pillar was seated in place in the hole in the aluminium frame, the lens was put in place too, the wires connected to a ledtester and then the focus could be checked on the wall. It appeared that with the pillar as far in the hole as it would go, the led was still not in focus. So the ledboard was unsoldered from the pillar, the head of the pillar filed mostly off and the board soldered back on top of the pillar. With the pillar fully down in its hole, now the focus was good.

The aluminium cap that fixes the mirror is press-fit, and so is the driver :expressionless: . I chose a 6x7135 (2.1A) Jaxman driver that was in my junkbox. The heatsinking of this setup is minimal (the mirror-led-assembly is held in place with a spring, P60 style) so I reckoned I should moderately drive the led.

Finished:

The beam:

The mirror does really good imaging btw, here an overexposed pic of the beam in which you can even judge my soldering :person_facepalming: :

I indeed measured 2.13A with a clamp meter on the tail, 230 lumen makes its way out of the light which is probably not much better than the XR-E, and the output is not dropping fast so the heatpath is good enough. As you can see, a lot of side-emitted light is not hitting the mirror, still I measured a throw of 51 kcd. But stupid me forgot to measure the stock light so I do not know if it is improved :person_facepalming: . And guess what, this recoil type light is so old that no-one on CPF or BLF ever measured actual throw numbers :weary: (luxmeters were not common 8 years ago)

So I just ordered an extra copy from aliexpress to make the comparison, to be continued.

I’ve been wanting to do something similar but have been stuck on choosing the best LED for it. My top choice would have been the SSL-80 which you had tested.

Another nice thing about using a modern LED in this setup is the greater efficiency. That means less heat and the recoil configuration can perform better.

Beamshot 80 meter

:+1: nice job. It’s cool to see something a bit different.

That is interesting that the recoil reflector creates a nice image of the LED. Whereas with a traditionally oriented reflector the beam is made of a superposition of rotated images making the beam round. :question:

I’ve got an UltraFire WF-008 here, and this topic makes clear again why i have no idea what to do with it… :smiley:

The projection is cute though. :slight_smile:

Maybe some precollimator would be a good idea?

A good idea in theory but a very bad idea for me because this will then become a mod that is several levels more difficult, too much hassle for me and this host is not worth more than what I did today.

Using a SSL-80 instead of a White Flat would count as a pre-collimator mod, but the SSL-80 is not as optimised for luminance as the White Flat.

Yeah, but that changes the focal distance, which is not easily adjusted…
It won’t increase throw either, but make a larger projection.
But also the big fat boom and center piece ought to be a lot smaller, so that would mean removing the stock boom and put in a thinner copper one, which means you can place a new one at the right distance for when you use a tiny pre-collimator lens.

The led mount in the light path covers 13% of the surface of the mirror, so there is some output gain in reducing its size but not something you would notice.

The focal distance has quite some adjustment actually, by sliding the pillar in and out its hole you have about 5mm play, by filing down the pillar I added another 2mm play. Given that this light is made for a domed led, adding a pre-collimator to a domeless led may get you even close to the stock position of the pillar.

True, but it just annoys me… :wink:
It didn’t matter for the XR-E, of course.

Oh, okay. In mine, the pillar is stuck in its fixed position.
Probably glued.

It’s a cute old school light though.

It is glued. I first unsoldered the XR-E from the pillar (small blowtorch), then placed the ledmount thing on top of a small piece of pipe that just fitted around the pillar, then with a pin hammered the pillar out of its hole.

I intentionally said “some collimator” because I didn’t mean aspheric specifically - I wanted to use a term wide enough to include TIRs.

How do they affect focal length? I suppose it depends on the TIR but is there some general rule?

Thought I do understand how this host is not worth an involved mod….

Thing is, I got 51 kcd, which is not impressive for this size flashlight, a classic Uniquefire T20 is about the same size and with a White Flat almost 150kcd is reached. The reason is not the concept of the recoil thrower, in theory a mirror of a given surface area should get you the same throw as an aspheric lens of the same surface area (this light has a 30mm diameter mirror, the T20 a 32mm aspheric lens) .
But then comes the 13% blockage of the mirror by the mount, the sub-optimal heatpath that does not alow you to run the maximum current through the led plus the led still running hotter than usual , the dirt on the mirror, and likely the reflectivity of this cheap mirror is less than the transmissivity of an aspheric lens.
So the concept is fun but it needs to be very well done to get the same performance as even a very cheap aspheric lens light. The real advantage comes when you go big, like Enderman did, and aspheric lenses become too heavy (and fresnel lenses too low quality).