Huh, I didn’t even know these kind of things existed.
Thank Bob_McBob.
Or get a digital caliper so you can custom make your own washers.
I do have calipers, it’s just that you seem to have done the research already, so it’d be nice to have a starting point for thickness 
No research from me. I just sliced my SST-20s to the closest to the wires as possible (from what I could see) in order to improve the DUV.
FBsLights has done some great research for slicing the 519As!
I use washers that have a hole just big enough for the emitter (5 to 6mm).
I prefer fender washers which are large relative to the hole size. I have a selection I’ve gathered with various thicknesses. I’ve sanded them smooth them out and remove any burs.
Take your calipers to the hardware store and pick up a few. There is usually a range of thicknesses to you can find the right size washers in bulk, so you can pick and choose.
If you want a complete selection, take a feeler gauge set, and drill an appropriate size hole in the thicknesses you want.
Something like this:
Overkill? Yes.
But you would have an almost unlimited selection of thicknesses.
Still playing with the thickness some. Adding a layer of silicone works just as you would expect. I’ve increased the silicone thickness to get higher CCTs. It is hard to get a nice smooth beam this way. It is better to slice to the preferred thickness the first time so you have a nice flat surface.
FB
1 Thank
Shim washers (DIN988), 3535 will fit into 5mm ones and with their 10mm diameter they usually fit onto most multiple LED MCPCBs without having to desolder other LEDs, if not it’s easy to remove a bit of material so they do. I have 0.1, 0.2, 0.5 and 1mm thickness and just stack a few to get any thickness in 0.1mm increments.
1 Thank
Don’t forget that the LED itself has to sit flush on the board. The slightest bit of “float” (ie, too much solder underneath) raises up the LED and you’ll be cutting unevenly at best, cutting too low at worst.
1 Thank
I’m surprised this is mentioned so rarely in dedoming discussions. When I do a traditional slice on an LED, I reflow it to a specific MCPCB with pressure from tweezers so it’s at exactly the same height every time. That’s the only way to be consistent about slicing height with a guide like a feeler gauge.
More emitters measured. All emitters behave about the same.
Emitter, CCT measured, duv measured
519A 2700k domed, 2660k, 0.0007
519A 2700k 1.4mm slice, 2564k, –0.0005
519A 2700k 1.2mm slice, 2445k, –0.0007
519A 2700k 1.0mm slice, 2273k, –0.0009
519A 2700k dedomed, 2138k, –0.0019
519A 3000k domed, 2993k, –0.0003
519A 3000k 1.4mm slice, 2907k, –0.0014
519A 3000k 1.2mm slice, 2770k, –0.0018
519A 3000k 1.0mm slice, 2607k, –0.0023
519A 3000k dedomed, 2388k, –0.0037
519A 3500k domed, 3423k, 0.0004
519A 3500k 1.4mm slice, 3319k, –0.0012
519A 3500k 1.2mm slice, 3196k, –0.0013
519A 3500k 1.0mm slice, 2999k, –0.0012
519A 3500k dedomed, 2709k, –0.0035
519A 4000k domed, 3779k, 0.0009
519A 4000k 1.4mm slice, 3663k, –0.0014
519A 4000k 1.2mm slice, 3478k, –0.0011
519A 4000k 1.0mm slice, 3329k, –0.0013
519A 4000k dedomed, 3027k, –0.0027
I can get a nice negative duv with CCTs from ~2150k to 5200k.
Happy modding!
FB
3 Thanks
First attempt at image, so I hope it works.
7 CCTs of 519A, all measurements with Opple (so just estimates, but numbers match what I see).
Each with: dome on, 1.4mm slice, 1.2mm slice, 1.0mm slice, dedome
Black = 5700k
Blue = 5000k
Red = 4500k
Orange = 4000k
Green = 3500k
Purple = 3000k
Olive = 2700k
Round = with dome
Square = 1.4mm
Diamond = 1.2mm
Trianlge = 1.0mm
Upside down Triangle = dedome
Also first time with ColorCalculator, so I didn’t know how to add a legend!
Lots of options. Pic your CCT and duv!
FB
2 Thanks
You don’t have to drill into the feeler gauge. Just take the bolt out and you have a hole. The Harbor freight one is 4.99. Thank PointyDogElbows for that one.
I bought a sm405 from Simon, dedomed I’m getting 3200K on 5mA, 3600K on 5A in S2+ OP reflector. This seems a bit high compared to what you are getting, do you think it’s sm405 or did Simon give me sm455 instead?
My sm403 measured ~3000k dedomed. This is at ~500mA.
With dome it measured ~3780k, so it started off a little low.
I’ve only measured (2) sm403 emitters and the other one measured 3860k with dome (never dedomed this one).
Your numbers are a little higher but I’d say it is within the normal tolerance.
Each emitter varies. You have a different emitter source (simon vs clemence)
Different bin (sm405 vs sm403). Different measuring equipment. Different drive currents.
I don’t know what you are using for testing, my I suspect my Opple measures a little low overall.
Happy modding!
FB
Pro-tip: deburring the edges of the hinge hole will give you better results with most cheap sets I’ve tried.
Been trying out sliced 5000K, 4500K and 4000K 519a leds in various lights.
Ran out of 5000K and very low on 4500K. So today was my first time trying out just dedomed 4000K in a light.
It’s …. too orange for me. Probably looks good outside with dark adapted eyes. But for general purpose EDC I often find myself using my lights briefly without dark adapted eyes and indoors. For this the color temp looks way too orange.
Guess it’s time to order more 4500K and 5000K leds. They look great dedomed.
I have 2x 519A 4500K, I installed one in a SC31pro, and sliced it first to 1.44mm. I saw hardly any change in colour temperature, so I went down to 1.26mm (these heights were not really chosen but came from what I had available in various washers). I used a razor blade to make these cuts, most of the dome is off but even at 1.26mm I still see very little change in colour temperature… I don’t have a spectrum measurement device yet so using other LEDs as comparison.
It is nowhere near the SST-20 4000K (IF25A, modified FC11)/Nichia 219C 4000K (Jaxman E2L).
Originally I estimate the temperature was at ca. 4200-4300K, a rough estimate from comparing against the 219b sw45k, SST-20 4000K from the IF25A and from a modified FC11 (using the SST-20 on 16mm PCB from Convoy) and 219C 4000K in the E2L. This conforms to measurements of the 519A 4500K which all appear to be in that same region as my estimate. The SST-20 from Convoy and Sofirn are slightly different to each other, but both are very different from the 519A sliced to 1.26mm (which should be somewhere near 4000K).
Comparing the sliced 519A 4500K 1.26mm to a standard 519A 4500K that I installed in a S2+, I see only a slight lowering of colour temperature.
Could it be that the lowering of the DUV makes it look far more ‘white’ (not as yellow), i.e. gives the impression of being more cool (higher colour temperature) when shining on a white wall or white furniture, even though in colour reproduction of most colours it is similar to a lower temperature?
I need to do some more tests. A quick test in the garden gives various impressions, such as that the shaved 519A may be better than the SST-20 4000K in the IF25A, although it is less easy on the eyes (being more white). It started raining so just a brief impression for now.
What was clear is that the beam is far narrower with the sliced LED in the TIRs that I tried, so that fits with what others say.
I think that is a sensible explanation. I observe the same myself and think there is a reason: if you look at isotherms (lines of constant CCT) on the CIE color chart, they are not orthogonal to the BBL; instead, they are inclined in such a way that lowering the duv corresponds to a shift in the direction of higher CCT.
Agreed, it’s hard to distinguish similar CCTs with different tints. At least for my eyes when comparing flashlights, the CCTs are supposed to be the same, but it doesn’t look that way on a white wall.
Without a measuring device, I think the only way to compare CCTs is with similar tints.
It works exactly that way.
My brain identifies a white wall under the rosy light of the 2000K E21A as white, not as yellow.
Great resource for 519A slicing. However, it would also be awesome if we had candela (or just relative lux readings) at each slice depth.
I’ve seen a detailed candela/throw for domed vs dedomed but not for sliced: https://www.reddit.com/r/flashlight/comments/twiik4/rough_comparison_measurements_of_output_and_throw/
What I’m looking for is minimal CCT drop (i.e. 1.4mm slice probably) but increased candela.