Thrunite TC20 - XHP70 - 3800lm

Unfortunately, I think all three of those spheres were calibrated a bit too high. Maybe –20% to 30- too high (might be closer to 10 to 15%). They were never tested against a NIST certified reference light in a commercial sphere. I think their calibration was based on guess work in relation to factory rated flashlight specs.

Texas_Ace is trying to get some good accurate data to calibrate the lumen tubes he recently built. He’s working with Maukka, who has a pretty sweet $10k integrating sphere, to figure out a more accurate correction factor.

The amp draw between the emitter and battery are quite interesting. Due to some loss in the driver circuit, I would just assume the emitter would read a bit less than half the battery draw. Instead, it reads a tiny bit more.

The discrepancy is probably due to the emitter voltage not being exactly half the battery voltage.

No big whoop, I’m just thinking out loud here. :partying_face:

This is why I only posted the amount of lumen drop and not the output readings from the TA tube. At this point any output figures from these tubes cannot be considered accurate.

Richard also uses the same tube, made by the same people, as mine, Robo’s, and TomE’s, and they’ve built themselves another one as well that compares to mine and Robo’s directly. These are not guesswork, but based on the ANSI ratings of a great many expensive factory lights that are reputable.

Tossing opinion’s out about 20-30% high readings is just insulting.

I was thinking about that 20 to 30% thing this morning also…….

I know the factory averages a given amount of readings to get their results , from a certain run of lights and the 4 meters mentioned here are usually pretty darn close to the factory listed specs.
I mean they are not exact , but I am usually within around 50 to 70 lumens or so on the 900 to 1200 lumen lights and within 100 to 150 of the lumen ratings on the higher output lights and many times I am within 10 to 20 lumens of the factory specs.

That is the biggest part of the reason I trust these light tubes to be fairly accurate , although they may not be pin point accurate.

I’m not talking about emotions, just facts and figures.

From what I understand, Texas_Ace’s lumen tube was calibrated to the same level as you three. Maybe it wasnt? I thought it was pretty darn close to you guys.

Since his lumen tube is reading 20% to 30% high, then that means you 3 are also reading high. Right? Maybe I’m wrong here?

Also, it does seem fundamentaly wrong to base an integrating sphere calibration on “ANSI ratings of a great many expensive factory lights that are reputable”. Doesn’t it? I would not do that unless it was my only choice. Maybe there was no better alternatives at that time?

Now we have access to some very precise measuring equipment with certified light sources.

Hmmmm well this TC20 is a good example of how mine reads at least………I get 3791 lumens at start……Thrunite list 3800……pretty close to me really.

If each of us with a PVC light tube measured a light then sent that light, and it’s power source, to Maukka for further testing, we’d then know exactly how our measurements compare to his. But at the end of the day, don’t Fenix and NiteCore and Thrunite ALSO use that massive green integrating sphere we’ve seen pictures of in their laboratories? Aren’t their ANSI ratings determined with expensive actual testing equipment?

And it hardly matters, for what it’s worth, as it’s all relevant to actual use… after depleting cells and recharging you can’t get the same numbers again in the testing equipment, so do the “facts and figures” actually mean anything to us in real world use?

What I said was insulting was meant on the terms that a great many of us have gone to great lengths to show results, while you simply sit and speculate, that’s the insulting part, throwing all the efforts of quite a few forum members in the trash with your 20-30% high comments while you yourself have no testing equipment to prove and verify what you speculate. Put a light on the testing box and watch the numbers change, constantly. Recharge it and test it again, and again, and again, for years, you’ll find that “facts and figures” are speculative as nothing is carved in stone. Variables are the only constants.

Remember too, that ANSI ratings from manufacturers are an average of lights tested, not a specific lights actual abilities. I don’t know if they test 10 out of 1000 or what, but they average out what they see and they use whatever cells they have on hand, usually their own brand and not what we use, so once again, the variables come into play.

Does 4.9A at the emmiter seem like enough amperage to support your lumen readings?

I was trying to find a 70.2 review from someone who has good measuring equipment. I could be wrong, but I think Köf3 from TLF has a good setup. At 4.9A the raw emitter (a P2 bin 4000K) was measured at 3800 lumen at room temp 25°C, before it got heated up. Once you subtract for the reflector and lens losses, DB knows more about this than me, it’s going to be well under 3800 lumen. So your 3790 cold lumen reading does seem high to me. I don’t know how much higher, though. Also, Cree emitter binning allows for plus or minus 7%. So that’s another wrench in the works.

I think TA’s idea of sending some stable output lights to get tested in Maukka’s sphere and then using those lights as a calibration tool is a smart move. It eliminates all the factory ANSI-FL1 rating variables and gives an exact output for that exact light.

Anyway, sorry if I upset anyone.

The Convoy L6 makes similar output readings using a buck driver with 2 cells. Larger head, different reflector size and shape, different glass, so again the variables and probably emitter choices that add to it. And XHP-70 vs 70.2.

The K70 with 70.2 XHP-70 emitter and buck driver has been tested at over 10A tail current and 4600+ lumens out the front, so for the TC20 to get ~3800 at 4.9 emitter amps, yeah, probably about right. Emitter binning and size of the reflector, shape and finish of the reflector, glass lens used, all comes into play of course.

That’s what I’m trying to get at though, we use the numbers to show gains when we mod a light, so we know if we’re on the right track or if something else needs to be done, getting too caught up in the numbers is counterproductive.

I am not really “upset” I am just saying that with the factory numbers that are given for their parameters I am usually pretty close on the numbers with my readings on the lumens with my meter and tube setup.

I think Dale may have mentioned that you can test the same light with the same cell 20 times in a row , with that same cell fully topped off and each and every time you take a reading it will be a little different. Even if you make sure the light and cell are at the exact same charge and the exact same temperature each time. There is always a little bit of a variation in there. I am sure you know also as the cell and emitter get used more , then they will lose performance over time.

Even with the high dollar professional equipment the factories use , I don’t know for a fact , but I feel you will still get at least a slight variation using the same light and same cell through a series of 20 test.

The light tubes that DBC , TomE , Richard and myself use , were calibrated using I think it was 25 different type and sized lights , that were from reputable manufacturers such as Fenix , Nitecore , Thrunite , Surfire and such. I wasn’t there but the way I understand it , they took the averages from all those lights that they were using and the ANSI factory listed specs from each and then got an absolute average from all those numbers to get the calibration multiplier to the correct value on these meters / light tubes.

I just feel with my meter being so close to at least 85% or more of the lights I have tested , that are straight from the factory using their listed specs , that I am getting pretty accurate results , unless the factory integrating spheres are not giving correct numbers ?

Just a little update, Newlumen did his measurements on the TA tube using a 0.7 correction factor. Texas_Ace got his measured lights back from Maukka and determined his lumen tubes need to use a 0.68 correction factor. These tubes still are only plus or minus 5% between all 30 or so of them so his TC20 measurements are still pretty accurate. Maybe they can be rounded down just a tiny bit.

Is the general consensus that the TC20 will hold turbo for about 3 minutes before doing a big stepdown? Maybe 2.5 minutes is more accurate?

Does it get burning hot to hold?

Thrunite tc20 with the shockli imr…

High mode, 1554 lumen @. 30 seconds…

This light maintain 1554 lumen pretty well… it doesn’t drop fast…

Isn’t it supposed to be making 3800 lumens? And isn’t high supposed to be 1800?

From Thrunite…

“Output mode/Runtime:

-Turbo (3800 lumen; 130 minutes)
-High (1800 lumen; 145 minutes)
-Medium (320 lumen; 10.2 hrs)
-Low (38 lumen; 46 hrs)
-Firefly (0.5 lumen; 37 days)
-Strobe (2280 lumen; 180 minutes)”

Yes… it make 3800 lumen with their way of measurement… we dont know how they take the measurement… turn on lumen ?

Either way, i am not concerned… i am here to play with the TA tube and have fun…

Robo tested his at 1792 lumens in high, 3806 in Turbo, he found it maintained output pretty well up into 80 or 90 seconds, if I recall correctly.

Robo has a review on it with run time charts, for what that’s worth. Tested various cells in it as well.

I was looking hard at this one earlier today, decided on spending the money for components and driver boards instead, reckon I’ll build my own as there’s more challenge there. lol

I did get 1750 lumen on my thrunite catapult v6 ( 1700 lumen )…

My TC20 (NW) gets 2960lm on Turbo @ 30 seconds on the TA tube, using the .7 correction factor
I’m using the included Thrunite cell (fully charged) and have meticulously cleaned all contacts.

Using the same correction factor, all of the other lights I’ve tested besides this one pretty much live up to their factory specs, including some ANSI rated lights.

TBH I’ve been wondering why I’m not getting anywhere close to the advertised spec of 3800lm.
My thought was the 3800lm rating is LED lumens, Thrunite’s website does not show anything about ANSI ratings that I could find.

Figures for some popular (non ANSI) lights on the same tube using the same correction factor:

BLF Q8 (stock)…………4865lm
Sofirn C8F (stock)……2338lm
(Both tested @ 30 seconds with fully charged 30Q cells) These readings appear to be pretty much in line with the factory ratings.

At first when testing the TC20 I wondered if the tube was reading low, however the 4 ANSI rated lights I tested (Eagletac, Jetbeam, Nitecore) were nearly spot on to the factory ANSI specs.(within 4lm to 50lm)

My TC20 NW does 3530lm after 30s with a liitokala 26650 cell. Other lights give similar results according to their specs so I’d say accuracy is OK