Short Review of Sofirn SP33 v2 with XHP50.2

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.

So you’re saying it’s not really a constant current driver?
Hmmm… I suppose i can live with that, unless the difference between 4.2 and 3.0 Volts is noticeably large…

I ordered the SP33 driver too, recently.
But i wonder if that’s the previous XP-L driver or this one…
We’ll see…

Mountain Electronics has 26mm boards for XHP emitters and they will reflow the new LED on the board, if they have the LED you want in stock.

I have not measured the output current. I measured the battery currents and the lumens in turbo mode with a fully charged battery. I also measured Turbo lumens and current at 3.75v. I think it’s safe to say the output current does drop a little bit.

I’ll measure the lower levels on a 3.75v battery to see if they are different as well.

He’s in Finland, so MTN is not a good choice. He can probably go through Kaidomain, etc… much cheaper.

I think Sofirn is limiting turbo power levels when below 3,8V/3,75V for their cell to keep on running happily.

The cutoff is 3.70v.

The regular Turbo still works at 3.71+v

If you get the original driver out please let is know how. So far I’ve tried heat and pulling on it and it has no effect on the glue. The glue does not even soften.

I’m hesitate to try pushing on it from above.

From comments on Aliexpress.

I wasn’t planning on taking out the driver, but i liked the idea of having a spare one, either the XHP50 version or the XP-L version.
It was only €4 + change so i though i’d order one.
I will report on it here though.