I think the KP 5200 mah came out in 2014. They have since been replaced by the 6000 mah version which is superior. The blue Liitokala came out, I think, in 2016 and is also superior. There have been a lot of better cells that have came out since the KP 5200 mah.
Do you think the chemistry is any safer with the 6000 mAh?
They could have used screws with a bit larger head to press down the MCPCB against the shelf and use thermal paste in between instead of using thermal glue. :person_facepalming:
They abandonned the driver retaining ring, it is now glued.
Someone at sofirn has a fetish with the glue.
The older chemistry has to do with performance, not safety.
I’ll try to explain why I think the protected KP 5200 mah is not a good choice for this light.
A boost driver tries to put out a set amount of voltage and amperage to the emitter on Turbo.
On the other side of the driver, it is pulling in roughly half the voltage and roughly double the amperage. I say “roughly” because driver efficiency and some other things come in to play.
As battery voltage drops, the driver has to pull more amperage out of it to compensate. This is Ohms law.
Now the driver has its limitation on how much amperage it can draw to keep from burning itself up. The driver engineers will decide on a safe limit. Once this amperage limit is reached, the light will either step down the the next lowest level or maybe do a half step down. This reduces the amperage being drawn by the driver, keeping it safe.
I don’t have the batteries to test it, but I’m sure that a fully charged protected KP 5200 mah will show a larger amp draw compared to an unprotected blue Liitokala 5000 mah, for instance, due to its voltage sagging a lot under load.
What this means is that you can test these two batteries by running the light on Turbo for 2 minutes, let it cool, run Turbo again for 2 minutes, repeat over and over.
The Liitokala may be able to run full Turbo for a combined 30 minutes (just an example as I haven’t ran this test). While the protected KP 5200 mah may only get 15 minutes.
So the KP 5200 is a perfectly safe battery to use, but you won’t get as much run time on turbo.
I would love to see Maukka or other more professional reviewers graph the output of this light with some different batteries to see how they compare. Other boost driven lights have shown to give more Turbo run time when used with unprotected high drain cells.
EDIT: For more on this boost driver design, see post #49 below.
Sofirns engineers have said straight up that they like to use glue because they believe it makes the lights more reliable.
I’ve make it clear to them (through Tracy) that no one on BLF likes glue. We hate it.
That was a great read, thanks Jason. It's probably the same scenario with the 18650's then as I tested my protected KP 18650 vs the 30Q and the 30Q was substantially better.
Here is a comparison at 10A. You can see the big voltage sag. This is the unprotected KP 5200. If you add the protection circuitry, the voltage sag will probably be even more.
Oh yeah, I edited that explanation post above to make it a bit clearer.
That's quite the drop and not even protected cells 
Looks like I've been missing out on the all the fun with my 26650 lights. Luckily I don't have many 26650's so I'm going to look around to start getting better ones this week. In the beginning I was so paranoid about unprotected batteries, but now nearly all my 18650's are unprotected.
KeepPower uses a lot of different 18650 models to rewrap with their protection circuitry, so you have to know what cell is under the wrapper.
Concerning older and newer chemistries, the same thing does apply to 18650’s as well. The Panasonic NCR18650B 3400 mah (aka Panny B) came out in 2012. It can not compare to a newer cell like the Samsung 35E. The test below is a protected Panny B (not a KeepPower version). This is only at 3A! Huge voltage sag.
.
The voltage sag is almost depressing in those graphs. I use my protected NCR18650B's only in my C8, SP32 v2.0 and my sofirn diving light and they seem ok in those ones. I'm pretty sure all the protected KP18650's I have are all using the NCR18650GA cells.
For linear drivers they seem fine. A Panny B in a Buck, Boost or FET driver tends to reduce performance, though.
Panny B in a Buck driver can make it fall out of regulation really quick. In a boost driver light, you lose Turbo quicker. In a FET driver light your max output is reduced. Overall, they are not so bad. Newer cells are definitely better, though.
thanks for the review jason. i ordered a couple as xmas gifts for the co-workers and 1 for myself. cant wait to get it.
Whenever you turn on the light there is a battery indicator feature in the switch which will light up green for about 5 Seconds when the battery is above 3.4 volts. Below 3.4 volts, it will light up red.
At 3.7 volts High mode (920 lumen) is only accessible with a double click. Another way to put it is to say that Turbo mode becomes the High mode. You have Eco, Low, Med and a 920 lumen Turbo mode.
At 3.4 volt, Med is only accessible with a double click.
I testing this with a low capacity junk 18650 which has a lot of voltage sag.
Candela is 18.6 kcd or 272 meters of throw.
Parasitic drain is 121 microamps/0.12 milliamps. Very nice!
Thanks for the new info.
Thanks for the review, great shots.
Why do you think the MCPCB is glued? It isn’t not in my light. As for why they didn’t use screws with a larger head, clamping the MCPCB down keeps things from sliding into alignment when the bezel is tightened down.
The driver is glued. I prefer a retaining ring with some non-permenant threadlock, but that’s going to push the price up. People returning their lights because they don’t work due to a loose retaining ring also pushes up prices.
Yeah, from a manufacturer’s point of view i can understand they want to use glue or thread lock.
Often it can be overcome with a little heat though, which is good enough for me.
…but sometimes it takes a lot of heat plus a lot of force…
I wanted to add an addendum to post #34 where I talked about how boost drivers work.
Not all of them work like that where the amperage goes up as the battery voltage goes down in order to maintain the constant lumen output. Some will run a lower amperage limit and reduce output as battery voltage drops.
When testing this SP33 on turbo at 3.75 volts I noticed the amperage is still at 5.80A. I checked the lumen output and it was down by roughly 200 lumen. Interesting.
I’ve been asking around, but not too many people are familiar with this style of boost driver. I’ve heard the Thrunite TC20 and Catapult V6 boost drivers may be similar to the SP33 design. I’m still learning about it myself.
I need to pull the driver out and look it over, but that is easier said than done. I can certainly add hot air to the driver cavity and slowly heat up the glue. Pushing it out from the top is a good plan, but you have to push at least one of the led wires all the way down into the driver cavity to get room for a pushing rod. If the driver is really stuck, you won’t be able to get the wire back through and the light will be non functional.
I might be able to tie some thread or dental floss to the wire so I can pull it back through. Maybe.
I’d have to also feel around in the dark with a flat tip punch to make sure I tap against the flat edge of the driver and not an important component. This is why I haven’t got it open yet. I like this light! Lol
I believe what I said earlier about using an unprotected high drain battery still holds true. Once this light gets to 3.7 volts you loose Turbo. So the less voltage sag, the better.
Great review! What size is the MCPCB, 20mm? I just got mine, and i am planning to change led, it is too cold for my taste.
It’s 26mm and 6v.
A 20mm might work, you just have to be careful the mcpcb doesn’t rotate and cut into the led wires. If the reflector is gripping the plastic centering ring tightly, just open the reflector hole a tiny bit so it spins freely. That should minimize the chances of the mcpcb spinning with any strong force.