I admit I didn’t read all the replies, but I chose not to because I want to give my own opinion before I’m skewed by those who are probably more experienced/biased, but I as a flashlight guy who’s been doing this stuff for a few years have come to the realisation that the “wow” factor wears off very quickly when the limitations and cost and other caveats come into play.
I like lights that work every time, for a long time, and provide a long term impression of the thought and foresight that went into their creation.
A big bang followed by a big frustration is fun for those of us that are always learning. Not so fun for everyone else.
That 3rd wire is meaningless - it can be removed. It's been posted about - think EMI related or something? I've never used any LED coating on de-domes and never had a problem - several lights for months now and I don't care for losing output and throw because all coatings WILL cause loss's, and not sure it really protects the LED surface from anything. Also, I haven't found it necessary to screw down the MCPCB's, but in theory it should help in heat sinking, but I'd be concerned of not getting the positioning right because you are very committed to getting the position perfect -- maybe beyond my ability...
I'm thinking you went down the more difficult path on this mod, but the screws could help, but for me would be hard/risky.
Thanks Tom. I do recall the EMI thing being mentioned before now, phew. I went the route of screwing them down because if I switched emitters I didn't want to have to remove the adhesive and have to polish the surface again. In order to get them aligned I flipped the reflector upside down, set the LEDs in them with a dab of thermal paste on the backs, then set the heatsink down. Lifted the reflector off, marked the spots and drilled/tapped holes. Lightly screwed down the LEDs, set the reflector down again to center them, took it off and tightened them down.
Hhmm, pretty nice method. I just roughly position the LED's, use AS5, wire them up, then when assembling with the reflector after tightening it down good with the reflector screw, I use a piece of tape on the side to match up to one of the LED's, then carefully monitor the reflector doesn't rotate much. It's been working out pretty well, but it makes the bezel tightening tricky. It's so much better when there are at least 2 screws holding down the reflector then there's no worries the reflector will twist.
I've seen other guys using shrink wrap sleeves over the LED wires near the ends, where they would rub against the drilled holes in the pill top (vinh) - really good idea to minimize chance of slicing the soft silicone wire coating (this has happened to me several times ).
I broke my one and only properly sized tap. Do you have details of what size tap, what size drill bit, and screws you use? Where to get the taps or tap set? I've done a couple of lights with taping hold down screws on MCPCB's, but really not sure what I'm doing...
I use a M3 x .50 tap (I think this one? http://www.mcmaster.com/#26475a34/=r5vox6) with screws found at my local hardware store. I can't read the drill bit anymore, but it should be 2.5mm. I really didn't need to go metric though, as it's rather spendy.
I might get around to working on this again tonight. The heatshrink is a good idea. Maybe I will redo my wiring.
Oh, side note, I actually got lucky drilling the holes. It was a good thing they were drilled where they were, because as I was drilling the first one, I looked down and noticed that a little bit further to the outside and I would have drilled right through to the cooling fins. That would have been a horribly stupid mistake. Where I drilled it didn't go through to the driver pocket or exit outside the light, phew!
Thanx for the info on the tapping. Yes, I use the stock centering rings and have not found any exact replacement - even ones from RIC which he said were replacements were not the correct size. I've got a few cracked centering rings -- very frustrating... I could get buy with using ones slightly smaller - those I have.
Tom, apparently it can. I don't know if I'd try it with 7135s, but with the FET it's working great. I've run it for a few hours on low mode along with 15 minutes straight on high.
Wow - the amps and lumen results are fantastic, but I thought an Atmel Tiny13A was rated for 6V max and pretty sure you got 12v feeding it. Not sure bout the other parts. There is of course a big drop off in the first 30 secs - probably good multiple reasons for that at these levels. Hhmmm..... All I can think of some applicable quotes, depending on how it goes:
Dr. Peter Venkman: Egon, this reminds me of the time you tried to drill a hole through your head. Remember that?
Dr. Egon Spengler: That would have worked if you hadn’t stopped me.
Dr Ray Stantz: Fire and brimstone coming down from the skies. Rivers and seas boiling.
Dr. Egon Spengler: Forty years of darkness. Earthquakes, volcanoes…
Winston Zeddemore: The dead rising from the grave.
Dr. Peter Venkman: Human sacrifice, dogs and cats living together – mass hysteria.
I think we can handle the 40 years of darkness thing pretty well...
It is, but that's what the zener mod does. It drops the voltage that the MCU sees, that's all. The MOSFET sees the full 12 volts but it doesn't care about that.
I'm not as old as you but I still got that quote! Classic.
Damn, you are making this mod'ing thing too easy... These are all very good signs. I'm probably a step behind on the full understanding of the Nanjg/Tiny13A design, but maybe worth checking voltage to the actual Tiny13A voltage - IN pin. We push this stuff past the limits/max of the ratings anyway, and sometimes it seems to work with no downside, other times - who knows. It's pretty impossible to do lifetime tests anyway. I suppose between checking voltage IN and checking temp would be the two things we can do, besdies run tests like you are already doing.
I just ran a modified TK61 for 20 mins to verify operation after a resistor mod, and it held up well, but I'm not setup/experienced to do component temp checks, so could be heating the hell out of something and don't know it. I should be getting my 10A bench supply any day now, so that will help, but also need better test fixture setups.
What's your lightbox by the way? Is it PVC based?
Ohhh - fyi, I work in an engineering dept. on the software/firmware and work closely with the EE's. A few times over the years we would get burned by operating a chip out of spec (more current, more voltage, etc.) and we'd get away with it, sometimes for years, then out of nowhere, units in production are not passing. So we find out a "new batch" of parts came in, or someone substituted a part wiht the same exact specs, and the circuit no longer works. Turns out what we used to get away with, being out of spec, no longer will work with either the new batch of the same part, or a substitute part. Sometimes it's caught earlier in the cycle - the design worked with a pre-production or couple of prototype builds, but not consistently for unit after unit. Funny, because in the proto development phase, we may discard a non-functioning part thinking it's a bad part, when in fact the part was 100% in spec, but that part just happened not to have the same tolerance to work out of spec. It's all sort of scary if this could happend with our CREE emitters one day. What of CREE starts making the LED's to enforce the max operating amps spec, and new batch's no longer work over 3A... Ouch for us, and a couple of manufacturers like ThruNite, SupBeam, etc.
I did verify the driver temperature...by feel. Barely gets warm, even the zener diode which is what I was concerned about. It's also why I ran it on low for about 20-30 minutes, because that's where the zener diode would be seeing the most voltage, with low sag numbers.
I also ran it against one of the IOS/LCK-LED/LIGHTMALLS 4.5A small toroid driver versions upgraded to a 15A rated large toroid with the sense resistors and output resistors bypassed. As expected, it only did about 6.3A, which is about all the cells will do in direct drive anyways. Before when I was seeing 7A with the small toroid on it I'm guessing that the extra amps were going into the driver (it smoked it really fast and was BURNING hot) and not into the LEDs.
My lightbox is PVC based, built and calibrated by manxbuggy so our results should be comparable.
Edit: Just tested with 20Rs. 6.7A at startup, which gives about 150 more lumens is all vs. the VTC4, at 60 seconds it was only 50 lumens more.
Nice results on your shocker. My Tom E Shocker does about 4260 at 60 seconds on 20R’s. I think he had it right at about 6A. So yours has a tiny bump over mine. Of course the focus is the fun part. Tom spent a ton of time getting mine focused. Lots of grinding on the back side of the reflector. Ended up with an outstanding 477kcd. So if you can get your focus perfect, you just might see 500k.
Ohh, k, we got the same lightbox . I did the calib myself, but cross-checked with same lights with the brothers, so we are pretty confident.
Yes - agree, could get 500 kcd. Ok, I'm convinced so far. This looks like the way to go. I assume you used a 5/8" washer from the the HarborFreight set -- I bought the same copper washer kit you recommended. The 5/8" looks really close to the desired inner and outer diameter.
Yes, I did use that washer. I could have used a contact PCB but i figured the least amount of joints and wires, the better! I think this is a bit uglier but it works great.
).
Classic.
. I did the calib myself, but cross-checked with same lights with the brothers, so we are pretty confident.