I haven’t tested my Q8, but it’s obvious that it’s massively brighter on high/turbo than my ~1000 lumen lights. So fwiw, I believe you have a problem if visually it only slightly outperforms the TK-15.
Oh thank you ValuseekeR.
Empirically having an error of over 100%, other tests with fine batteries, also less lux, and I did not want to throw in the 14 minutes to reach 60º C, I may have a faulty Q8. Just wanted to inquire if others had done some true tests but got a lot of hearsay.
There have been true tests, my Q8’s perform as they should and put out well over 4000 lumen (measured) at 16A (measured).
You will always get questions like you have had if you ask something on this forum. That is simply because very often it is the people and not the flashlight. So that option must be ruled out first before diving into the technical stuff. And your picture showing a way of measuring the current that is known to give large errors does not help convincing people that the problem is the flashlight, you can expect questions about that.
FYI, I got my Uni-T UT210E (DC clamp) last year from Amazon for around $30. Not sure what they cost in Canada.
If you use fully-charged 30Q’s, your light should get fairly hot after 3 minutes on maximum. It will only output 5000 lumens when fully-charged high-drain cells are used. Output (and heat) will be much lower by the time the cells are half-drained.
So, if your light is not getting hot with fully-charged cells, then that should also tell you there’s something wrong.
Have you made sure all 4 cells are making contact with the positive brass ring? You are using button-top cells, right?
Cells are new and fully charged. Button top Samsung 30Q. My set-up (not that I showed all details) have the cells spring loaded to a clean brass ring. The wiring is all via 16 AWG mechanically connected leads (which the people don’t seem to understand an error of 100% is clearly beyond 10% due to contact/gage resistance). Setting the thermal step-down was timed at 14 minutes, using a Fluke laser thermometer onto head body near the switch. Set at 70º C (its cold here in Canada).
I can understand the electrical properties of low voltage pushing high current. The smallest of resistance makes for leveraged effect. My set-up tried to offset these details but still have an overwhelming error. Djozz is ponderant about a huge difference yet his own results have an error of less than 10 %.
All in all, if there is no visual fault as connections go, I can only infer a faulty component. This is not to be taken personally as BLF has designed this light, more of misgiving in assembly/solder/component.
Sidney, you try to perfect an inherently flawed measurement method.
There are still some reasons why your results might be much worse than djozz’s:
- much higher currents involved
- different meter
Now I actually don’t think they explain the different result, but they are enough to leave doubt.
Good meter is far more reliable, but we surely can’t expect you to buy one.
Maybe you could thermal measurements without the multimeter in the loop?
Not sure what ‘ponderant’ means, but I never said that what you see is not real, just that you must be patient with the obvious questions. and so you answered them, and ruled out most trivial causes of measuring too low current and output 
The resistance of your ammeter certainly can cause the low measurement you are observing. Have you measured the total resistance of your ammeter? It could easily be 50mOhms. This more than doubles the circuit resistance (cells, wires, FET, etc) and reduces the resulting current that flows.
It seems like you are reasoning that the leads will not dissipate much power and therefore they are not the source of the measurement error. Indeed the ammeter and leads will not be dissipating much heat, but they will decrease the total current that flows in the circuit.
But, if you have also done a reliable comparative lux reading (with flashlight assembled with tailcap), and are getting low readings then that could indicate something wrong.
If I understand what you’re saying, it takes 14 minutes to reach 70C.
1. That’s taking too long to heat up. It should be that hot in about half that time.
2. If it’s reaching 70C, but only producing a bit over 1200 lumens, that seems odd. It’s getting too hot for that low of an output.
Maybe there’s some internal components/resistance that is causing it to heat up, without directing that energy into the LEDs?
Whats a good meter? Amprobe made in the USA (if I recall correctly). Been faithful over 25 years. How are you gonna trust your measurements if you doubt the instrument? Meaning I used this many times for DC readings and Ohms law do concur (within an acceptable margin). And you still instill an error of over 100% is due to leads/connections? This meaning that the people all say that the measurement method is hugely wrong but the values that are presented differ by only 10%.
I just took the tools at hand to verify a problem and it begs for heresy.
Had already looked at a Fluke with DC capabilities - too expensive!
Thermal measurements were made prior to any testing, some two months ago. May have to redo.
I see that my methods aren’t accepted, yet they are to give a ballpark value. I may leave this alone or perhaps bring it up with BangGood.
I’ve put too much effort in this. Drawing too much contentious debating. I’ll explore the anomally on my own. Your insight as thermal problem gives me a lead - verifying dissasembled light under load.
Thank you for your interest.
This suggestion may be approaching Captain Obvious levels and I’m a relative noob with these lights, so feel free to disregard, but have you let it run for a while and then checked to see if the voltage is dropping from all 4 cells proportionally?
I’m sure the meter is accurate, but it’s presence in the circuit is affecting the measurement. Your circuit resistance with the ammeter is likely twice or three times that of the circuit resistance in the properly assembled light. This really can cause a large difference and is not in the ballpark of the real value (in assembled flashlight).
There may be a problem with your light, but you have not presented adequate evidence. This method of measuring current, especially in a light like the Q8, is very inaccurate.
Oh, the shunt resistance is leveraging the power drop. Explained in an academic manner it makes sense.
Thanks!
Why don’t you power the driver and the light with your bench power supply or just without the multimeter in it’s path. If you get roughly the same amount of light as you did with the fully charged 30q’s then there is likely a problem with the flashlight itself. If the output is a lot brighter you know your light uses more amps 
I seem to have jumped a few of your replies (hadn’t seen them);
1- Yes the internal shunt (meter) must be high as compared to overall supply/load circuit.
2- If I had a quantitative way to compare lux - but as stated my subjective comparison with a 1200 lumen is about 1 1/2 bright. I am aware the square law of illumination and it is an opinion.
3- I will be re-testing thermal step-down. This is the best tool with a laser thermometer. Testing indoors this time.
4- And to answer TUNE4IT - my bench supply only delivers 5 amps. But testing light output with/without multimeter will provide a comparison. May not be a problem. As now I better understand the multimeter’s shunt is the root of the low current value.
There are apparently various light measuring apps for cell phones, which, although likely fairly inaccurate, would still be perfect for comparative testing.
When used with ceiling bounce it could give you at least an objective measure of comparison to your other lights, and a way to check how much your meter setup is affecting output, and whether any changes you make are having a beneficial effect.
This thread shows how to build an external shunt to that you can measure high levels of current using the voltage reading on you DMM. Unfortunately the pics in this thread are temporarily down.
https://budgetlightforum.com/t/-/26092
Charge the cells and place on turbo on your table and let us know. Outdoors in Canada 2 months ago the light will take a 15 mins to heat up easily probably what 2p 30 degrees outside. With my triple ee x6 making around 3500 lumens in the winter it can go about that long before its to hot to touch even with calloused hands. In the spring and summer 2 3 mins and it’s scortching