[Teardown] SEEKNITE S11 GITD

I got my SEEKNITE S11 GITD a few days ago. Bought it for $12.99 from Banggood using Fin17’s code:

I also bought some identical looking ones (not GITD though) from AliExpress but they have not arrived yet. I suspect that this flashlight is sold under many brands, such as Alonefire S11, BORUIT S11, etc. but I couldn’t find any teardowns for them so I can’t be sure.

Generally speaking, I mostly like this flashlight. It is very similar to the NITECORE TIKI (UI is almost the same) and not too much bigger:

I have drawn a chart for the UI:

I don’t have any way to measure lumens output but it is brighter than the TIKI on turbo.
I don’t have a way to measure color temperature either, but it doesn’t look like 6500K to me. My guess would be around 5000K-5500K. It is warmer than the TIKI and I think the TIKI is lower than 6500K.
The GITD works well, although I don’t find this feature useful at all:

The Auxiliary white LED is CW, unlike the TIKI. The UV light IS UV, but it’s UV component is weak and it’s not as “clean” as the TIKI. My guess is that it has purple wavelengths in addition to the UV (just like the sunlight has all the colors, including UV). I tried to capture the difference with my camera but While the XANES KT and the S11 light looks about the same to my eyes as the photo, the TIKI came out very different, in real life it is not only much brighter, but more things light up in the money bill. You can, however, see in the S11 photo that there is visible purple:

The tail magnets do hold the flashlight horizontally, but any small bump or shake will make it drop.

Teardown:
Disassembly is very easy, just unscrew the metal head (mine was partially undone already :open_mouth: ), remove the heatsink+TIR lens and then push from the USB port and the whole innards will slide out:

The thread is lubricated, I lack the experience to say if enough, though. The SST20 LED looks genuine to me:

Only the sides of the the bare copper have thermal compound on them (not sure why):

Some of the battery markings are faded but most of them are intact. The battery is marked “701040 3.7V 260mAh (0.9)62Wh DD10”:

And with the gunk removed:

The circuit board looks pretty simple:

I have not measured anything but I’ll make some assumption as to how it works just by looking at the photo:
The big unmarked SOP-14 IC is the (unknown) MCU. It runs the 5 LEDs with PWM via the five A2sHB MOSFETs (the SOT-23-3 packages), the primary LED has an additinal 0.5Ω current limiting resistor (two 1Ω in parallel). The SOT-23-5 IC marked LTH9 is the charge controller (probably TP4054). The battery protection is on a small PCB in the battery, I have not opened it but my guess would be a DW01+8205S duo.

I measured 228mAh on the battery, discharged from full to 2.7V at a constant resistance of 71Ω:

I charged the battery with an external charger at 128mA (read the last paragraph to see why). The capacity discrepancy is probably because the manufacturer rates the battery with a slower discharge and/or to a lower voltage (E.G. C/10 to 2.5V).
Funny thing though, the battery measures only 38.5mm x 10mm x 6mm, not 40mm x 10mm x 7mm as the marking implies.

Charge current is about 500mA:

That seems too much to me. I don’t know the specs of the battery but I have doubts as to it’s longevity with this much current. I barely charge my Li-Ion batteries with 1C, let alone 2C like this flashlight.
If I could be assed, I would change the 2K resistor (marked 202) next to the TP4054 with a 4.7K one to get a 198mA charge current, but I can’t, so I won’t.
Also, since the TP4054’s termination current is one tenth of the charge current, it means it will terminate the charge at 50mA, so the battery will never be fully charged to 260mAh (my guess would be around 240mAh).

[update: Aug. 23rd 2021]
I have received the non-GITD S11 that I bought from an AliExpress seller for $10.97 (free shipping), but before I begin I would like to correct a couple of things: first, the UI chart has a slight omission: you can switch to each mode from the other, I.E. double-press while the side lights are on will turn on the main LED, and triple-press while the main LED is on, will turn on the white side light.
Also, I forgot to mention the S11 has a low battery warning, the charging LED will turn green for a few seconds in any of the modes if battery voltage is higher than 3V, and red (blinking) if it is below. The red will also blink if the battery < 3V while the flashlight is on.

The S11 from AliExpress is slightly different. The package claims a different main LED, the Samsung LH351B:

And it does indeed looks like it is:

Judging by djozz’s photo:

The color temperature is also different. It is cooler. I would guess about 6000K to 6500K.
I couldn’t see or feel any lubrication on the threads or the O-ring, so the waterproofing is questionable (might still comply with the stated IP65 though).

Ultra-Violet is about the same, strong light, but most of it is in the visible range:

Once again, I tried to capture the TIKI’s UV but no matter what I did I couldn’t take a photo that shows what my eyes see. I did, however, manage to take one photo where the TIKI’s light shows more UV activated things light up:

(These glowy lines light up all over the note easily with the TIKI but not at all with the S11)

The board is slightly different, but pretty much the same:

The battery is the same. I didn’t see any reason to test the capacity again, it’ll just be the same:

I can say for a fact that the protection circuit works well, because I accidentally shorted the battery wires (oops) and the battery output was turned off and did not reset until I connected it to a charger.

That sums up the differences. In addition, I have made some further tests:
Quiescent current: The flashlight draws about 1uA from a fully charged battery when off.
Main LED current draw (from the battery):
Economic: 7mA.
Low: 38mA
Medium: Starts at 150mA, but within a few seconds starts ramping down slowly. This doesn’t seem to have anything to do with heat, it is completely time based.
High: Starts at 500mA , but within a few seconds starts ramping down quickly. This doesn’t seem to have anything to do with heat, as using cooling (fan) doesn’t change the time it stays on 500mA and also turning the flashlight off and on immediately as it starts throttling down doesn’t change anything as well.
Turbo: Starts at 750mA and after a while, starts throttling down. I didn’t measure the time but I would estimate it’s about 30s to 1min.
UV: 260mA.
I also measured the white and red side LEDs but after re-assembling the flashlight I discovered that I forgot to write the values down ( :person_facepalming: ). Sorry.

Please note: These values were measured with a Ruideng DPS3005. It is an amateur power supply and I don’t know how well it handles the PWM measurement, so take these numbers with a grain of salt.

This time I actually took the time to change the 0603 charging resistor:

(Don’t judge my soldering… It’s very small and I didn’t use any magnification)

And now I like the charging current much better:

A weird thing happened though, before I re-assembled the flashlight, I did another test with the PSU and noticed that all the currents were lower than what I measured earlier, E.G. the turbo was about 600mA and the high was about 400mA.
Maybe I damaged something when soldering the new resistor? I don’t believe the actual change to the resistor could have had any such effect.

I have a question to the experts though: How is the brass(?) heatsink thermally coupled to the stainless steel(?) head? The only contact I see is through the top, but, the silicone washer goes there. Is the silicone washer thermally conductive?

Thanks Flail for your highly informative review above.

I found on Alibaba, the cheapest supplier is selling S11 keychain light for only US6.20 each, with a minimum order of 10 pieces.

https://www.alibaba.com/trade/search?fsb=y&IndexArea=product_en&CatId=&SearchText=keychain+flashlights

Thanks TermsakC, but this is not a review. A review would have beamshots, light output graphs, etc.
$6.2 is very low but what am I going to do with 10 of them? :open_mouth:

I found one on Aliexpress for $8.94 but the seller didn’t ship it and then raised the price to $15.

BTW, all of the cheap ones don’t have GITD plastic. I didn’t find a cheaper GITD one on Aliexpress.

Thanks Flail, nice post!
Can you measure the quiescent current/standby current draw?

Unfortunately, I have already reassembled the flashlight.
I wouldn’t mind taking it apart again, but in order to desolder the battery you have to remove it by undoing the new glue, which I really don’t feel like doing. Sorry.

I’ll tell you what though, eventually I will find a seller on AliExpress that will actually ship those flashlights to me, and when it arrives I will also measure the quiescent current (heck, maybe I’ll even measure the PWM frequency), and if we find that the other one is completely identical to this S11 (other than the GITD shell, of course), then I think it’s a good bet to assume that they have the same (or very similar ) quiescent current too.

Thanks for the teardown, this seems like a mixed version of Rovyvon A5 and newer A5X series, with the upgraded battery capacity but old PCB design. I paid a little bit more for mine on aliexpress boruit store.

I have now received my S11, and I observe the opposite. The TIKI GITD main LED seems to be warmer for me, maybe 5500K. The S11 looks like 6000K or 6500K.
Edit: Actually, my light has a completely different main LED than the one you received. The packaging says Samsung LH351B, but I’m not familiar with this type and can’t visually confirm this. The white markings on the PCB that holds the main LED are also slightly different.

Again, I observe the opposite. The Nitecore TIKI’s UV light is surprisingly strong for such a tiny thing, and the S11’s UV is even noticeably stronger.
I think the unclean UV could be because the UV output goes through the GITD shell and because some UV is reflected onto the phosphor layers of the neighboring white and red LEDs.

I have had this only for a day, but so far I quite like it. The cool white LEDs are a bummer.
With NW, high CRI side and front LEDs (side is more important, and probably harder to mod) this light would be extremely good.

I measured battery voltage after the built-in charger terminated at 4.17 V, that seems a good, healthy value.
Terminating the CV phase a bit early also helps to achieve the incredibly short charging time.

Is the footprint oft the side LEDs known? Maybe blue could be replaced with amber to make the police strobe more like traffic strobe.

master-of-disaster, I received my S11 from AliExpress today and I can confirm your observations about the main LED: It is indeed LH351B, and it is cold white, 6000k-6500k looks spot on.

I like cold white better than natural and warm, but if you like the warmer ones, why didn’t you buy the GITD version from Banggood? (it’s still $12.99, I believe)

It’ll take me a while to finish the tests, maybe a day or two, but in the meanwhile I can tell you that the quiescent current is about 1.0uA with a full battery. I also measured all the current consumptions. I will add everything to the OP once I’m done.

Is there any way to tell without desoldering them?

Yes, by knowing them from looking at them :smiley: .

More seriously, I wonder why no one thus far tried this mod. That thing is or was cheap enough to buy it out of curiosity.

I didn’t know there were Versions with different LEDs. The aliexpress seller I bought from shows the exact same specs as banggood (SST-20 LED).

Regarding battery health I’m getting concerned about just how hot this thing may run - with the light disassembled and UV on, the main PCB gets painfully hot within ten seconds or so.
I will attempt some battery temperature measurements in the assembled state tomorrow if I can shove a thermocouple through the charge port.

Oh yeah, mine actually shows SST20 too. Huh, I didn’t even notice.

You can probably get a refund, or at least a partial one if you care to go to all the trouble of a dispute. If you do, take a photo of the S11 box with the package (envelope) next to it and in the description you can link to this image:
Imgur
and tell them that this is what you should’ve received.

I ordered mind from Fasttech and it have parasitic drain. The battery last about a week with full charge.

I contacted Fasttech and they ask for video proof but i find it hard to video.

Do others have same issue?

I have updated the original post with the new flashlight tests.

As I mentioned earlier, mine has a 1uA parasitic drain, so no, I don’t have the same problem.
I can’t think of any good proof other than a time lapse with a clock next to the flashlight, but that is a lot of work.

If you want to try to fix it, maybe you should try to desolder one battery wire and after a few minutes reconnect it. Might reset the MCU and help. Also, while you’re in there, maybe put some Kapton tape on the wires’ solder points on the board, because maybe the metallic battery case is shorting them.

Thermal measurements
Because this type of flashlight does not have an easily replaceable battery, and temperature is the most important factor for battery health (excepting misuse), I measured battery temperatures during operation.

The light is an S11 from aliexpress that seems to be the same as the one Flail has added to his original post update. I added a bit of thermal paste between the brass heatsink and the PCB that holds the main LED, because there was only a tiny amount (see Flail’s pictures).

The thermocouple was placed inside the plastic shell of the fully assembled light, next to the middle of the battery and lightly touching it. The actual battery temperature next to the thermocouple could have been higher due to poor thermal contact. The battery temperature will not be uniform, so this measurement likely does not represent the hottest part of the battery (which is the part that matters). It also does not represent the hottest part of the light (not at all) which will be at different locations depending on the mode you are running.
The battery seems to be glued to the PCB with some type of foam, this is good and should help protect the battery from the hot PCB.

Front LED, medium setting (highest continuous setting without stepdown), 90% battery charge
The measurement was stopped after 25 minutes because the temperature didn’t increase any further.
Highest measured temperature was 7 °C above ambient. This is absolutely fine.

Front LED, momentary high (highest possible output), 100% battery charge
The measurement was stopped after 200 s. Temperature was still increasing at this point, but not very quickly. I doubt anybody would use momentary for much longer.
Highest measured temperature was 20°C above ambient. This is high, but acceptable and to be expected with such a high performance in a tiny package.
The stainless steel bezel was painfully hot after this run, presumably much hotter than the battery. This indicates that the heatsinking through the brass piece to the bezel works.

White side LED (“lantern mode”), 100% battery charge
The measurement was stopped after 30 minutes because the temperature didn’t increase any further.
Highest measured temperature was 11 °C above ambient. This is fine.

Red side LED, 70% battery charge
The measurement was stopped after 25 minutes because the temperature started decreasing (battery voltage dropping).
Highest measured temperature was 16 °C above ambient. This is higher than the white side LED and would have been even higher with a full battery. If you are worried about battery health, don’t run this mode continuously on a full battery, especially in a hot environment.

UV side LED, 100% battery charge
The measurement was stopped after 8 minutes because I didn’t want to cook my battery. Temperature was still rising at this point.
Highest measured temperature was 28 °C above ambient. If I had not turned the light off, Temperature at the measurement location would probably have reached 33 °C above ambient and stayed there. This is not ok and could be dangerous. Don’t run this mode continuously, even if you think it’s a nice party trick. It will ruin your battery, unless you place the light in a very cold environment. Running the UV for 2 minutes is ok.
If a manufacturer puts in a side LED with this much power, there must be a time-controlled stepdown or a temperature regulation.

Charging from ~ 50% state of charge
Flail’s measurements show that charging in this light is unhealthily fast, so I measured battery temperature during charging. Charging from empty would have been a more useful measurement (probably with a bit higher temperature), but I didn’t have enough time for that.
Highest battery temperature during charging from 50% was 13 °C above ambient. This is ok.

Thanks very much for these terrific tests and analysis. I see $8.39-$10.97 with 50 piece MOQ at that Alibaba link now. At 10 pieces a group buy might have made some sense. Oh well. Current BangGood price is $20.99 and it looks like Fin17’s coupon is still active (edited), so I’m tempted.

There is sure a lot of stuff on the board! I don’t really see the usefulness of all those different side leds. I think I’d change them all to warm white. It would be interesting to identify the MCU: it could conceivably be reprogrammed.

I’ve been a bit intrigued by the RovyVon A5X but didn’t want to pay $30+ for a disposable light. At $12.99 this similar Seeknite is more attractive, especially if I can locate replacement batteries.

I like the low weight and small size of the light, don’t care for the complexity of the side leds, don’t care for the scorching max output of the main led, but can live with those things. At least this light really gets some size and weight reduction from the internal lipo and plastic shell, compared to traditional metal cylindrical lights.

About a month back, I found a seller on Alibaba that had a price of about $7 per piece with a minimum order of 10. I actually thought about trying Alibaba for the 1st time, but eventually I decided it wasn’t worth the hassle. Not when the price on AliExpress is $10.97:
https://www.aliexpress.com/item/1005002393158222.html
(for the non-GITD cold white version)

As for the MCU, IMHO, the chance of it being reprogrammable is 0.0000001%. There is simply no reason whatsoever to use an MCU with a flash memory when OTP MCUs are at least an order of magnitude cheaper! For example, take the PADAUK PMS152 - a 14 pin MCU with 7 outputs which were specifically designed to drive LEDs and LED flashlights. It costs only 5 cents a piece:

As for the battery, I couldn’t find the exact same battery for sale, but you could probably put a slightly smaller battery without a problem.