Acebeam x70

Wut.

Thank you. I always thought it was the opposite, since in my LED builds, using my constant current supply, as my LEDs heated up, so the voltage went up.

Is there a reason for this?

Never seen that happen. Did you measure with a DMM from the LED wires instead of trusting the PSU?

me to I thought the same thing

Another test with another LED where I turn the heatsink fan on at 10 minutes.

Well this one makes sense…

You can even use diodes (which LEDs are of course) as temperature sensors, because the relationship between temperature and forward voltage is pretty linear (for Si diodes at least, don’t know about LEDs).

That’s true, for example the Nichia 144A is very linear from 20°C upwards. Use a very small current not to heat up the LED itself much and it can be used to estimate ambient temperature.

The exact reason “why” is pretty much a entire undergrad class in silicon junctions. Needless to say it involves quantum effects with the bandgap. I guess an ELI5 answer would be that the increased thermal energy can give lower voltage electrons a push across the bandgap? That is probably wrong on several levels though.

The effects of a falling Vf are huge though. It is why LEDs need current regulated supplies. Otherwise you get a thermal runaway and the LED destroys itself.

Anything with a negative temperature coefficient will need regulated current. Very few things have negative coefficients and voltage regulation is sufficient.

On a related note I’m always kind of amazed that direct drive FET lights work at all. The wires and contacts and the battery itself form a positive coefficient that magically balances out against the LED that is trying to self destruct.

Thanks to you all for the answers.

The reason I thought forward voltage of an LED would go up is because of my experience in CPU and GPU over clocking.

See, when you up the voltage and frequency beyond its normal parameters, as you may already know, the additional power creates more resistance, and therefore more heat.

Then, what happens of this effect is that it accentuates electromigration since electrons have less incentive to stay inside of the silicon transistors, and if there is too much electromigration, boom! a crash even before speed ans temperature were a limit for the chip itself.

I had noticed this on my GPU, when I kept temperatures below 40°C, I could achieve massive frequencies and push voltage to its limit without any crashing.

Once I reached even 1°C above that threshold, it crashed. And observing the voltage and current behavior, I noticed that voltage slowly upped as the temperature went up, as well as current. Lowering the fan speed did this and if the voltage was not locked, this is what would happen.

Thanks for the explanation everyone again!

Oh it is trial and error for sure. Many emitters get burned up before finding the limits. You have to get everything matched just right. The xhp50.2 has a tendency to burn up on a pair of weak 18350 cells while the 70.2 can handle a pair of the best 26650 cells. You can also help control things by emitter wire size and length to add that extra bit of resistance. Sometimes you want to bypass the battery springs and sometimes you don’t. Too hot of a combo can be tamed by using a weaker battery, etc…

interesting stuff… Now where’s that Acebeam X70 ?

Still testing it etc last i heard… they can take their sweet time too, not like anyone else have something at 40k lumens coming anytime soon unless Imalent pulls an 50k haha…

true but there are modded light that are much higher… than 40K

yes true, but talking production lights nothing ive seen that comes close to the x70 yet…

Not true DX80 32K lumen…

Not sustainable, and quality control has been even worse than Haikelite lights sometimes.

I own one… quite sustainable…. and quality is just fine… at this point in the cycle… most kinks are out of the light…

well yes but imalent has alot of quality issues imo its worse then olights x9 even tho more lumens the body cant handle that lumens from what ive seen. I wouldnt call that brand quality exactly acebeam and olight are on a whole different level.

You can get 2 DX80s for the price of one of these…Was expecting at least 50k at this price. I think they can push out more lumens but they’ll do it after a competitor releases something brighter.

Umm, that would be extremely hard to do even with 8 cells.

With with 8 30Qs, you would need to push above 500W! At a max of 70W per cell, that would mean a max of 560W! At 50k lumens, we are at the limits of 8 cell packs unless we go bigger.

Also, even with active cooling, how are you going to get rid of 500W in an LED flashlight? You would need to push an absolutely massive heatsink with a vapor chamber and some monster fans, or watercooling, which is not doable in a flashlight.