Transparent washer is just for protection during transport! ILC-1 should be installed without any washer.
That explains a lot. Thanks 
led4power, sorry to bother you again. I’ve noticed some annoying oscillation in brightness. It is visible only in 1.5% and 2% modes (checked for 7A and 9A configuration, with and without moonlight enabled). Measurements with multi meter shows oscillation between 135.3 and 136.4 mA (for 7A). Is my driver defective? Can I fix it somehow?
"Oscillation" between 135.3 and 136.4mA is pretty normal because it comes from thermal noise superimposed on useful signal(we're talking about 1-2uV of noise), the real question is how you can see 1mA fluctuation at 135mA current? Anyway,one way to reduce it is to increase signal level from shunt at a cost of increased driver internal resistance.LD-2 to LD-4 drivers have very low internal resistance because at that time there was demand for driver with DD-like resistance on high mode,which was because of high Vf snd generation CREE LEDs. It's possible to reduce that effect by 10times with shunt and one 0402 resistor change,at a cost of increasing driver's internal resistance from ~5mOhm to 14mOhm (still pretty low, difference would be noticeable only if you drive 2nd generation of CREEs close to max. current).
I’m using the driver with triple nichia D240. Those measurements (around 135mA) was taken with cheap multi meter (not high resolution oscilloscope), so maybe fluctuations are bigger and that’s why they’re visible. Anyway - I can live with that 
Hmm, <1%, imho nice result.
Idea: switch that would glow blue when in off position and red when flashlight is turned on (or any other color combination) 
That is not possible with tailcap switch, because when switch is on, there is no voltage over switch.
Added big gold plated bronze springs:
Gold plated phosphor bronze spring 12x9x1mm
Bottom diameter is 9mm,so they can fit on quite many drivers, switch pcbs etc.
I will try to measure resistance soon, but I expect something like 5mOhm under realistic conditions (~30-40% compression).
Also added new ILC-0 listing with fixed default currents at 1€ lower price:
You still can change current set resistor as you like. Default values save me some time for build, that's why price is lower.
Also, one note related to LD-x4 drivers listings: you can now choose do you want 1mA bleeding or not. If you don't plan to install illuminated tailcaps, it's
better to not have bleeder for two reasons: lower active state driver consumption (1.6mA vs 2.6mA), and 1mA lower sleep current (~40uA vs. ~1mA).
Did you have those springs made? They are about perfect dimension for most tail springs
(if you had them specially made, that is seriously cool, and having had springs made myself, I’m curious what is the manufacturer. But I’m aware that that may be a trade secret of yours )
Edit: what is the height of this spring after a couple of compressions?
They are not custom made. They get shorter by ~1mm after 10+ compressions.
Good find then :+1:
And they keep their length well too
I just finished a build with a LD-A4 and ILC-0 combo, there are some images and details on the mod thread , but overall I can say I’m really satisfied with bot products 
. And as soon as the optics arrive I’ll build an e2l with ld-b4 too.
Guess what these are (what makes them different compared to other PCBs), and win pack of 5 (any type)!
Maybe they aren’t DTP boards but still provide good thermal conductivity? For LEDs with thermal pads connected to the cathode, so no need to isolate the mcpcb any more…
The boards are aluminium and central thermal pads are not separate but are connected to led+, so these must be non-DTP boards. And that only makes sense nowadays if the dielectric layer has an extreme low thermal resistance, perhaps the ano-layer type that Clemence also used for his new VirEnce boards?
They look 
Do I spot triple and quad 4040 boards? :+1:
That’s obvious. They’re white.
That they are DTP but have isolated the conexions of the FET, temps sensor etc
Negative DTP for the LED and mosfet.
Except copper Noctigon that have direct thermal path (DTP) trough central pad, they are aluminium metal core printed circuit boards (MCPCB) designed and developed from scratch named MOSLED EXTREME or short MOSX. They have pad for external Metal-Oxide-Semiconductor-Field-Effect-Transistor (MOSFET) combined with pad for Negative-Temperature-Coefficient (NTC) thermistor (thermal sensor). This makes them different from any other MCPCBs from market.
There are huge benefits to this design because of true temperature regulation, better dissipation of heat coming from MOSFET thus leading to more reliable operation and cooler running Driver board thus longer runtimes on highest modes.
Also they are 3 times lighter than Copper comeptition because aluminium is used as core material with only 2.3% higher thermal resistance compared with copper DTP competition and 1.5% lower thermal resistance compared to non-DTP copper MCBCBs.
Another thing to note is that currently there is no other MCPCBs on market that support 4040 LED footprint in form of single, triple or quad configuration, especially combined with pads for MOSFET and temperate sensor!
Also there is support for Osram Oslon Black Flat that is known for not having electrically negative central pad and support for Lumileds Luxeon Z ES!