Romisen RC-A8 -- Beautiful, Unusul and Well Made

I received two of these relatively large 1 x CR123A / 16340 lights on the same day. One is from (new to me) for $15.90, and the other is from (known to be unreliable) for $16.59. It is also listed here, but I don’t know whether they actually have them. And here, but they don’t take PayPal. I ordered one silver and one black and received two silver. The gamble payed off nicely! They are the most smoothly curved flashlights I have seen. They really look as good as they do in the pictures on the Web sites. The lanyards didn’t break the first time I tugged on them.

They have XR-E LEDs, one mode “linear” regulating drivers and forward clickie switches. Current draw with a new Energizer CR123A cell and a cheap multi-meter is 0.25 A. The star is 16 mm. and the driver and contact board are each about 15 mm.

The driver sits in the bottom of the pill, and there is a contact board at the top of the battery tube and a single wire between the two. The one mode driver consists of a single IC.

It appears that it regulates current, so it can work on a 16340 without overheating. There is no inductor or large capacitors, so it is not a boost driver. With thin leads, I measure 0.1 A at a CR123A cell and about the same at the original XR-E LED. With a 18650, I measure 0.7 A. With much thicker leads, the current is about the same, so the driver does appear to regulate, at about 0.7 A, similar to 2 x 7135. Maybe it either was designed mainly for lithium ion cells or else before boost drivers were available. Current regulation classifies it as not a typical budget light, a resistor would have been cheaper.

With the XR-Es, they must be old, but I had not seen them until recently, so they must be uncommon. describes them as designer flashlights. They are large and heavy for the battery size, but the big head throws a small spot and the waist makes them easy to hold. I first ordered one from and had my money refunded as it was not in stock. Other stores list them as sold out. lists it as Cree P4-WC, with Q3 out of stock, so it must be really old. That also means that almost any LED swap will make it brighter.

Tekna Mono-lith, TrustFire Z1, RoMiSen RC-A8, TrustFire Mini-01, Yupard

Same order:

An odd detail: There is a stainless steel insert at the bottom of the pill where it holds the driver.


One of the six examples I have (a black one), as of 4/7/2016, has a defective switch. The switch and battery spring are integral, and a CR123A cell puts too much pressure on the switch for it to work. Switching to a CR2 cell fixed the problem, but with the reduced capacity I am only using it in stock form. The switch is not easy to fix, because it is at the end of the long tail piece and it is crimped in, rather than held with a threaded ring or a plastic insert.
I have a third RC-A8 with a switch problem. It won’t reliably stay on when the switch is released.
Now a 4th bad switch, about half of them!


One might say POP lites are the Zagato bodied flashlights, at least in an enthusiastic moment, visions of things to come. Or maybe they are more like Citroën or even Panhard who weren’t always right.

Romisen RC-A8 and RC-A6 are the Figoni and Falaschi of flashlights. They are flamboyant, but not the most light from the smallest package.

—- Modification —-

A 17 mm. driver will fit under the pill, as sometimes done with SK-68s. Shortening the pill seems easier to me than building a 15 mm. driver. The pill will need to be further shortened to allow for the thicker star.

For lithium ion 16340 one can use the usual combinations which currently are Cree and BLF A6 driver or Nichia 219C and 7135 based regulating driver. For CR123A, to get much current increase, it needs a low forward voltage LED such as the new Nichia and/or a boost driver. It is not practical to make a high output light to work on either one, because for 16340 it needs low voltage protection and for CR123A it needs the low voltage protection to be disabled. My first choice is a Nichia and the 3 amp. Qlite version of the Nanjg 105C, star 4 jumped for moonlight mode. I didn’t have a 16340 at this point.

Nichia 219C and Qlite 8 x 7135, Modification # 1

I filed the corners of the regulator chips to fit in the neck section and in the pill, to make good contact and minimize the extra space needed.
I hope they still work and don’t short.

There was a stainless steel ring forced into the bottom of the pill. Without that, the bottom of the pill fits over trimmed components on the Nanjg 105c. I managed to pry it out with a screw driver after filing a bit off the bottom. I shortened the pill by about 2 mm.
I used thicker wire between the two boards and beefed up the connection on the contact board by slipping a thin wire through one of the two vias and covering both sides of the board with solder.

On a CR123A it is not much brighter than the stock one, and has less throw. I expected the lower forward voltage and brighter LED and reduced stray resistance to increase the output, but they don’t by much. I now read 0.37 A, up from 0.25 A stock. Does that single IC have voltage boost? Does it have less resistance than 8 x 7135? I can’t imagine how. Maybe the old XR-Es have less forward voltage than I thought. To get good output, I will need a 16340, rather than removing a resistor to disable low voltage protection.

Efest 16340 Lithium Ion Cell

Charged at 3.7 V, I read 1.4 A in high mode.

With the same two TrustFire CR123A lights:

With an UltraOK SK-68 on an alkaline AA:

None of these are fair comparisons, but they show the modification is successful. For significantly improved performance on CR123A, I am looking at boost drivers.

Boost driver from SK-68 clone for CR123A with Nichia 219C, Modification # 2

For Mod. # 2, I planned to use this 17 mm. 5 mode boost driver and this same Nichia 219C on 16mm Noctigon - 80+ CRI 5000K LED
The stainless steel insert holding the driver gradually came out with penetrating oil, my grandfather’s hand vise, warm water (Aluminum has a higher coefficient of thermal expansion than stainless steel.) and a small screw driver twisted at various places around the crack.

The plan was, as in the first mod., to file the pill to make space under it for the 17 mm. driver, SK-68 mod. style. But the components on the back, bottom, spring side of this two sided driver did not fit in the narrow waist below the driver. That could be gotten around by filing the pill even shorter and using the stainless steel driver insert as a spacer. The back of the driver will fit in the ring, and the front fits in the wide part of the pill. But that seemed too much filing, and I gave up on using the DX driver. I may do one more like that some other time.
That left a Nanjg 210 or a driver taken out of a red SK-68 clone to make room for a lithium ion conversion. In a crude test the two read about equal LED current, but the 68 clone had less input current. It also has three modes. I had already selected it after similar tests as my best leftover boost driver, with the highest output and reasonable input current.
So the modification became almost a routine LED and driver swap. I tapped the insert back in, connected and placed the LED, shoved the driver into the insert and connected the driver to the contact board wire. I soldered the 22 gauge wire to both vias in the center of the contact board.
The resulting light is about as bright (on high mode) with a CR123A as the stock A8 is with a 16340, for short bursts. From a data sheet, lithium primaries are like alkalines in only being able to provide high current for short bursts.

Here it is next to the cool white Ultra OK, zoomed in and zoomed out:

The cell in the SK is closer to full charge than that in the RoMiSen is.
I measure 1.8 A on high and 0.4 A on low at the CR123.

BangGood BLF A6 Driver and XP-L HI U6-7A LED, Modification # 3

Both vias of the connection board are now in the circuit.

The driver fit ring came out with penetrating oil and, screwing it part way into the light body, putting a screw driver through each wire hole and tapping it with the back of a knife.

I filed the pill down from 15.3 mm. to 13.5 on one side and 13.6 on the other to make room for the driver and thicker star.
I enlarged the wire holes in pill and on the Noctigon star. The reflow worked on the second try.
The Toykeeper user interface is a bit surprising in a small forward clicky that used to be a one mode. I had to think about how to change mode groups with the forward clicky switch. But maybe my sister in law can just use it as a four mode.

With fully charged Efest 16350s, a stock RC-A8 and the new one:

I read 2.25 A (direct drive) with my cheap multi-meter. The stock one reads 0.75 A (regulated).

Nichia NVSL219C 5000K - 80 + CRI and boost driver for CR123A— written off, Modification # 4, (black)

These will make nice presents, especially for women, so I continue.

I have one of these LD30 drivers:

I put it all together and it sort of works, but is not really usable. It is about as bright as # 2. Medium mode is maybe half as bright or more and draws only 0.4 A which is encouraging about driver efficiency, but, with a used CR123A reading 2.9 V, it drops out of high mode into low. With one AA cell it drops out of high much faster and also drops out of medium, so it is probably programmed to act that way, though I can’t think why. Why would a boost driver have low voltage protection?
I had great hopes for these drivers, because they are from a good source and they are clearly more expensive to build than the common budget boost drivers. They have more than twice the component count, with components on both sides. So they aught to work much better.
Simon Mao responded quickly to my AliExpress message that he thinks he sent me the wrong type.
This sort of behavior has been observed for drivers of this type before:

Someone, who apparently is very skillful, seems to be so confused that he puts low voltage protect on drivers that are not for lithium ion cells.

Added 7/18/2016: I have a replacement driver from Simon I don’t presently have a light to use it in, but it tests better. It appears to hunt for the optimal current to draw from the cell or cells, but does not seem to try to signal low voltage.

Next I tried a 16 component two sided driver from DX. It draws 2.5 A on high compared to 1.8 A into #2, and it isn’t nearly as bright. The switch is now not staying on, so the whole thing goes in the junk bag. I don’t know if it is a defect in the switch or whether it couldn’t take 2.5 A.

It is possible that the driver is shorted or damaged, as it is probably the same one that wouldn’t fit in #2. It is also possible that the driver and/or switch was damaged by a short the first time I put it together and it didn’t work.

Nichia NVSL219C 5000K - 80 + CRI and Nanjg 110 boost driver for CR123A, Modification # 5, (silver with orange o-ring)

The only tricky parts were getting the driver and contact boards and the star in place.
It behaves similarly to mod. #2.

Tail current: Mod. 2: 1.85 A at 3.1V; Mod. 5 2.5 A at 3 V

Mod. 2 on left, mod. 5 on right:

The Nanjg 110 and this SK-58 clone drivers behave similarly under these conditions. Being 15 mm., the salvaged driver was the easier mod.

Nichia NVSL219B and LD30 driver from Simon, for CR123A, Modification # 6

Simon sent a replacement driver that does not have low voltage protect. The driver works well, starting in the low of three modes: 0.06, 0.13, 0.8 A. About as bright as one drawing 1.5 A. The switch is broken. Yet another switch is intermittent only, most of the time.

Nice. I’d like to see the pill. These would make nice 14500 lights.

man those lights are smooth! I’m gonna have to snatch me one & designer lights for sure. Great share these Fritz

They do look nice, I haven’t seen them in a while. Let’s send one to Dale, see how much he can reliably wring out of it, then start a group buy. :wink:

That would be 16340. I added the size to the OP.

I remember seeing those years ago. Google Image Search might find their earlier incarnations.

Yeah, these are definitely from back in the day. Way back when there was more discussion about the merits of various stock flashlights. 16340s are definitely limiting for today’s modded burners.

It is a big light for that size cell too.

Odd but kinda sexy

I line that little monolith too

I used the Mono-lith for quite a few years backpacking. When I looked, there was no LED conversion available. It is an incandescent reflector zoomie.

I think what I paid is a reduced price for lights with out of date LEDs, so a group buy price would probably be higher. But if others see the beauty of it, Romisen might be able to dig up the old drawings and make some updated ones.

I corrected the title.

I ordered a Romisen RC-A6 from
Again, it is also listed here:

This is a 2 x AAA light that looks very similar to the RC-A8. Again, it is out of stock at the larger stores.

Added: both are also available here:

This flashlight is pretty high on the list of lights that I would never loan to anybody except for my girlfriend.

This shows that it is not the manufacturers’ fault that flashlight shapes are so simple minded. These shapes have been available, while people have preferred SK-68s and Sure Fires.

Romisen makes good lights. Seven or eight years ago they were one of the top budget brand flashlights. They have kind of been left behind the last several years.

Yeah, I picked up 3 different Romisen focusing flashlights from a few years back. They were pretty good bang-for-the buck lights for the $15-20 price range. They seem to have disappeared from the flashlight world since then.

I know what you mean, but I’m not sure any of the brands were, or are, real brands, with real differences.

I remember a few years ago when SanRenMu knives became popular on BLF.

SanRenMu was described as the "Romisen of the knife world."

That was a real compliment back then.

Yeah, it's old.These are from my 2008 'office' collection. All are single AA lights except the Romisen RC-A8.