Haikelite HK90

1lumen selects and reviews products personally. We may earn affiliate commissions through our links, which help support our testing.

Haikelite HK90 Specifications

Brand/model	Haikelite HK90
LED	3* Luminus SBT90.2
Lumens	18,000 lm
Beam intensity	750,000 cd
Battery config.	3*21700 (included)
Material	Aluminum
Modes	Many
Blinkies	Many
Reflector	Smooth
Waterproof	IPX
Review date	October 2020

Please don’t look for the Haikelite’s website, since you may end up on a less than ideal website that is NSFW. Anyways, the Haikelite HK90 is something that not many manufacturers have done. Namely, using 3 Liminus SBT90.2 LEDs into 1 flashlight powered by 3×21700 batteries. As most flashoholics know, these LEDs require an immense amount of energy, so you won’t cut it with your average batteries. My box included 3 high quality Samsung 40T batteries, and that is really necessary.
I’m very curious at how this beast will handle the power and if it steps down quickly or not. All of that will be tested in this review.

Package

The packaging is adequate including some interesting accessories.

• The Haiktelite HK90
• Lanyard?
• Carry strap
• USB-C charging cable
• O-rings
• Manual

Do you ever train with dumbbells? No, then get the HK90. Just kidding.

The HK90 is heavy though, so I wouldn’t be surprised that people will get muscle pain actively using the HK90 for an hour or so. If you’re a tough buff, don’t worry about it.

Haikelite included 2 kinds of accessories. But I don’t know what 1 is actually for without the other. One is a carry strap, which makes sense, but I can’t find a way to attach it to the light, without using the paracord strap. This doesn’t matter too much to me, since I have enough ways to attache other straps etc.

The side switch is your typical electronic switch with a click and press-and-hold functionality. A little fun fact is that it has a secret empty space in the back of the flashlight behind the batteries. Not 100% sure why that is done, but I assume it’s just to have enough ‘body’ to hold the flashlight onto. Without that extra dead space, the battery tube would be too short to carry the light conveniently.

Tailstanding: yes, it can stand on its tail without a problem

I chose for the silver color on purpose. The coating looks good on all part of the light. But there is 1 thing to it, that I don’t like so much and that is the shininess. I’m not sure if that is just my bias, but it looks a little cheap. There might be people out there, that think it looks great of course, but I’m not 1 of them. Again, this is just my first impression, and it might change over time.

Unscrewing the body feels smooth as silk. The threads are generously lubed and there is no squeaking sound or anything. The tailcap is much harder to unscrew, but there is no need to do that anyway. When you unscrew the tailcap, you’ll see a secret space. This space is big enough to hide your money.

For the rest, all accessories look fine in terms of quality. Not so sure why 1 end on the paracord has those 2 strings, but maybe you can make a knot so it can be carried like a normal lanyard? Anyways, nor the manual, nor the box shows how to use these accessories.

In terms of LEDs, the Haikelite has 3 high power Luminus SBT90 gen2 LEDs. Each LED is able to push more than 5000 lumens. This means, you can expect to reach an output of 15,000 lumens or more. Haikelite’s specifications show a max output of 18,000 lumens, so we’ll see if that can actually be achieved. These LEDs have a CCT (correlated color temperature) of 5700K

One thing that really stands out is the stainless steel bezel. Most light with this size bezel don’t use stainless steel, but keep it aluminum. Since it’s so heavy, I’m not 100% sure how much it can protect the light when dropping it onto concrete. The bezel isn’t completely flat, so when you put it down with the bezel on a flat surface, you will still see some light.

In the lowest outputs, the beam is a little on the greenish side, but when you increase the brightness, you can’t really see that green anymore. The beam also has a shape like a flower, close to a white wall, but out in the field, it’s just a very bright hot spot with enough spill.

• Length: 20.3 cm / 7.99 ”
• Bezel diameter: 99.2  mm /  3.9 inches
• Body diameter:  53.2 mm /  2.1inches
• Tailcap diameter: 58.8 mm /2.3 inches

Weight: 

• Empty:  1129 grams / 39.81 oz
• With batteries: 1336 grams / 47.13 oz

Flashlight comparison

Size compared to other throwers and high power flashlights

From left to right: Astrolux MF04S, Haikelite HK90, Acebeam X45 II, Acebeam X80-GT.

The driver is using a Anduril UI. This is currently one of the most advanced firmwares for flashlights at the moment.

This is the image of the UI, but forget about the bottom part, because this particular light has no AUX LEDs.

From OFF:

• Single-click: ON
• Double click: High (Top of ramp, to get to Turbo, you need to do another double click)
• 3 clicks: Access the blinky/utility modes.

Enter Special/Fun modes from OFF:

• 2 clicks + hold: Strobe modes
• 4 clicks: Lock Out mode
• 5 clicks: Momentary on
• 6 clicks: Muggle mode

From ON:

• Single-click: Off
• Double click: Turbo
• 3 clicks: change ramping mode.. Instead of a smooth increase, it has 6 little steps between Lowest and Max.
• 4 clicks: change to ramping configuration mode
• Press and hold: brightness ramps up.. release and press and hold again to ramp down.

INTERESTING, SPECIAL AND FUN MODES:

Please read the manual carefully to know how to access or customize these modes. Also see the firmware picture, above.

• Blinky Utility mode:
• Battery check
• Sunset Mode
• Beacon mode
• Temperature check

Strobe / Mood modes:

When in the strobes mode, double-click to rotate between them – all modes except party strobe are brightness, changed the same way as a normal ramp. In party strobe mode, ramping up and down will increase and decrease the frequency of the strobe, for incredible effects.

• Candle
• Bike flasher
• Party strobe
• Tactical Strobe
• Lightning mode
• Lockout mode (can’t use the light)
  • Lockout mode is available by clicking four times from off – while in lockout mode, click-and-hold will light up at bottom of the ramp, and a double-click-and-hold will light up somewhat brighter. This is to give you a way to quickly use the light if necessary, but if it activates in your pocket, it will only be at very low modes, and only for as long as the button is pressed. No more holes in pants!
• Momentary mode (signaling/ morse coding)
• Muggle mode: (safer for children)
• Configuration mode
• Ramp config mode

PWM

• You will be able to detect some PWM at low modes via phone camera, but as soon as it kicked in higher, it disappears. None of it was able to be detected by eye, though.

Protection Features (Low Voltage Protection)

• LVP makes the light step down to a lower level when the battery is low, and if the light is already at the lowest level, it shuts itself off. This activates at 2.8V. This won’t happen often, since the light will run for many hours before it shuts off. I couldn’t get it to shut off during my testing. I had to do the runtime tests, so I just couldn’t wait a few more hours. I waited a few more hours for each runtime test anyway.. but not long enough for the light to switch off.
• Thermal regulation attempts to keep the light from overheating, and otherwise adjusts output to stay as close as possible to the user-configured temperature limit. By default, it’s 45 degrees. For my testing I increased it to 60 degrees.

The HK90 was shipped with 3 Samsung INR21700 40T batteries. (this is optional) At first sight, the light didn’t turn on, and I didn’t see any plastic washer used during shipment. But when I looked closer, all 3 batteries had a sticker on the negative terminal. I don’t see this so often. Also, these batteries have button tops, so you can’t use flat tops with it.

The built-in charger uses a USB-C port for fast charging. But charging 3 batteries in parallel, this means 12,000 mAh. This takes some time. Therefore, I prefer to use my SkyRC MC3000 to charge them. The charge current is 2A according to my measurement. Maybe it’s just my charger tho. During charge there is a dim red light glowing behind the switch.

Lumen measurements:

All output numbers are relative for my home-made Integrating Sphere. It is set up with an Extech SDL400 Lux Meter for measurements including a Kenko PRO1D ND-16 filter. The base measurement is done with a Convoy S2+ that has been tested at 255 lumens.

All of my readings were taken from fully-charged Samsung 40T

I used the standard modes instead of the ramping menu. The 30 seconds readings are done by me, counting. lol. So it could be a few seconds off. And looking at the runtime test, I believe I was off by a couple seconds.

Tested with 3*21700 Samsung 40T

Mode	@30 seconds	@ startup
1/7	13.08	13.08
2/7	40.54	40.54
3/7	108	108
4/7	237	240
5/7	817	821
6/7	2517	2557
7/7	5257	5486
Turbo	3923 lm	16281 lm

The specifications show a lumen output of 18,000 lumens. I was able to measure 16280 for a split-second because the output dropped like a rock. You can see that Turbo was dropped to about 3923 lumens within the first 30 seconds. You should look at the runtime graphs to better understand its behavior. I don’t really like this.

Runtime:

The runtime test was done with the 50cm integrating sphere, including the Kenko Pro1D ND-16 filter and Extech SDL400 data logging Lux Meter. The room temperature setting was calibrated before the runtime tests.

I totally agree; this is a lame runtime graph. The 5th mode actually continues for many more hours, but the graph would look even more ridiculous. The next image will show the first 10 minutes of the graphs.

From the runtime graph we can see that Turbo drops dramatically within the first minute. It then starts to ramp up again till it reaches the maximum temperature again. The temperature configuration was still used in the default settings.

I didn’t test mode 1-4 since these tests already took me a week to take. Not every day, but during the times when I was home. In most cases I just stopped the runtime, when the output was just 10-20 lumens. This output stayed for many more hours. Not really interesting to see in a runtime graph.

The following runtimes I measured till the last largest drop in output. It doesn’t take into consideration the hours or low-lumen output. Based upon the standard Max. temperature configuration.

• Turbo: 3 hours and 24 minutes
• 7/7: 3 hours and 12 minutes
• 6/7: 3 hours and 16 minutes
• 5/7: 8 hours and 43 minutes

Next, I set the Max Temperature configuration to 60 degrees. I hoped this would give a much higher output in Turbo after the initial drop, and sure it did. But still not very drastically different. You wouldn’t really be able to notice the difference IMHO. But the runtime is completely different. The higher temperature setting pushed the amps a bit harder and the batteries ran out of juice within 1 hour and 40 minutes. In standard configuration, it was about 3 hours and 30 minutes. If you want to be more conservative, don’t touch the Max. temperature configuration. If you want the highest output possible, set it higher. I believe the max setting is 75 degrees. You will burn your fingers if it did that.

Throw Measurement

Measurements were taken outdoors at 20 meters, with a professional Hagner E4-X Lux Meter. As you may imagine Turbo dropped the output quickly, so I had my daughter helping me out by turning the flashlight on, while I measured the beam intensity. Tip. If you do that, start with a lower output and find the brightest spot, and then turn to Turbo.

• Turbo:  688000 cd = 1659 meters / 1814 yards of throw (only at start, the output drops fast)
• Mode 7/7: 233600 cd =  967 meters / 1057 yards of throw
• Mode 6/7: 106800 cd = 654 meters / 715 yards of throw

For the following beamshots I used a Canon EOS 5D Mk2 and a 100mm lens. manual settings: ISO1600, .5sec , F4, 5000K

The tower is about 450 meters / 492 yards away.

Disclaimer: This flashlight was sent to me for review at no cost, by Haikelite, through the hands of Neal at Nealsgadgets. I have not been paid to review, nor have I been holding back on problems or defects.