That's why I always copy & paste too 
The ‘R’ in X7R signifies that its specified uF value varies +/- 15% over its specified temperature range. Worst case will be at 125 C, where it will be ~15% less than at 25 C. This should be good enough for what we are doing.
Some better curves and info here:
http://product.tdk.com/capacitor/mlcc/en/faq/faq00035.html
Going with a capacitor that is 1/20 uF (to accommodate C0G type), the voltage value read at boot up will be 1/20 of what you would read with a 1 uF.
If we assume that most of this effect is due to the temperature (in)stability of the OTC, this indicates that we have an xxU capacitor. Not a good choice for this application. ‘U’ would be –56% capacity at its upper temperature rating. If it is a x5V in addition, this temperature is only 85 C.
If I have time this weekend, I will remove an OTC and put it on a meter while heating it. Or just swap out with an X7R and try to compare with an unmodified light.
Are they checking QC? Banggood keeps moving restock date back It’s now october 2
Finally got my 25R cell to work with my 3D. Turbo is impressively bright, UI is great, the light works flawlessly. Thanks to all who make this GB happen, really appreciate all your hard work.
If so, I’d count that as a good thing! 8)
Wonder if that also mean further delays in sending out replacements for the defective parts.
This is the one I’ve used for all my builds and I notice mode change problems when hot. It’s also the part specified in the original driver build thread.
I've been using only X7R 1 uF caps for all my OTC driver build/mods and I still see the temperature related problems in most of the lights I modded using it. Now I did buy the caps off of eBay at the time, so no guarantees I got caps as labeled. I suspect though they won't make much a difference, but it would be interesting to try. In my next DigiKey or Mouser order, I'll definitely will be ordering some 0603's and 0805's - cheap enough.
The inconsistencies with the OTC's is some of the reason why I much prefer e-switch's now -- you have true control of timing under the MCU. I've also used the NOINIT (brown out) method as well, but still seems to have some timing inconsistencies.
The circuit is the cap and the resistance it bleeds down through, if the resistance goes down with heat the timing will get shorter also.
What was not working with 25R and what did you do to fix it?
.
I actually think that they are!
Obviously the first batch was rushed, and now we are seeing the spotty QC as a result of it.
I am down for four, but have thus far only ordered one. - Received (3D), and I am very happy with it - But some it seems are not so lucky…
I just thought that one would be enough to play with to start, and wanted to see how things would go before I order the other three. - I think that was a good decision.
Many, Many thanks to all involved for Continuously working to make this GB Right! - It can’t be all that much fun 
Thank you,
-Chuck
I ordered several aditional 3d binned ones:
Well, this aint a XLP 3D but smth like 3C, kinda disappointing- the left is IOS V6 3D
1A ones were just fine though
One of them dont have a moon mode, just a blank space, i will reflash the driver but for someone it could be an issue
Many thanks to the TK for that lovely all arround firmware! And the OP ofc
Mitko - If a reflash brings back the moon...please let us know. Actually if you could directly flash TK's UI and tell us that would be a big help to all.
Will do that- its actualy a black space instead of the moon mode, only on one of 3 lights, maybe a glitch or smth- i usualy get glitches like that when i overhear the a13 while soldering
FWIW, the color temperature will change depending on what mode it’s in. This is a side effect of having two power channels at very different amps.
So, it can be tricky to compare the tint against other lights unless they run at similar power levels and have a similar type of AR coating on their lenses.
I’ve had a time to look at my 3 lights I’ve ordered. One of the lights I have to tighten as tight as I can to get it to work. If I put this tail cap on the other lights they do not work at all unless I put a shim the switch. I’m not sure if I received 2 short body tubes or 1 short tail cap. The length of my body tubes are 2.732, 2.723, and 2.730. I also measured the tail caps from the thumb indent to the end and tail cap that is giving me problems is a little shorter. The tail cap measurements are .998, 1.000, and 1.000. The tail cap that measures .998 only works when the 2.732 body tube, but i have to crank the tail cap as tight as I can to get the light to work and have it not flicker.
That really sounds like a tube issue to me - Too short.
Thanks,
-Chuck
That’s a perfect illustration why having a production that respects and understands specified tight tolerances matters.
There you have one that is less than .01 too short on one part, and less than .01 too long on the other piece, and — put them together, and hey, this one sample works fine.
Checking one product fails to detect the problem — too much variation around the spec — because some combinations of parts will average out to working properly.
Too much on one part, too little on the other, result — it lights up if you happen to get those pieces put together.
That’s why taking a real sample of the products — not just one, not just one they know works, definitely not one prepared to show as a “sample” — is what you need to check them.
Measure, write down the result, find out how variable the work is.
Given sloppy work with overmuch variability, the workaround — assemblers fishing through the bins looking for a slightly longer tailcap to go on a slightly shorter body tube, trying to make the damn thing light up — isn’t the right fix.
Want to build a modern nation that doesn’t melt down and fall apart?
Here’s how: production tolerance variability specification interaction - Google Search
What are other people’s tubes measured at? My issue seems to be that one of my tail caps. Unless all 3 of my body tubes are too short. If someone just ordered 1 light then this issue may be missed. The only reason I noticed it was because I ordered more then 1 light. I was wondering why one of my lights I had to crank the tail cap as tight as I could to make the light work. When I put the tail cap in the other lights the only way I could get it to work was by putting in a paperclip shim in the tail cap. If people order the 18350 tube the light may not work, but it works with their 18650 tube.