DQG tiny 4 18650 triple mod

G'Day Dutcheee,

Great work.

Best Regards,

George

Great mod Dutchee!

Glad you found my DQG Tiny III mod helpful. :sunglasses:

The key to prevent the star from overheating is thermal grease. In a more recent mod I found I could actually get away with basically a hollow pill with nothing under the star… as long as I had plenty of thermal grease at the edges of the star. This is how I did it in my latest triple (Sunwayman C10R running a FET driver, triple XPL-HI and an Efest IMR purple 18500 cell). With the 18500 cell sitting where the stock driver used to sit there wasn’t room for a heatsink. With plenty of AS5 at the edges of the Noctigon triple the emitters don’t overheat and the leads don’t desolder themselves.

Thanks, I really was fulfilling to put all my gained knowledge from BLF to good work. This light would’ve never seen daylight without all the contribution on this great forum!

@ReManG

Some pictures of the intergrated driver/mcpcb of the stock DQG. I think a reflow will be quite a challenge. Besides that, the switch tends to “catch” between the slotted opening in the head, so preventing it from snapping off is difficult. Maybe just in my sample…


@Djozz I’m always awed looking at your mods. If you just follow Hoop’s post and all the liks he shows, it will be a piece of cake for you! I got less lucky reprogramming the original BLF firmware to erase the turbo stepdown for another light I’ve got, so not out of the woods myself!

Thanks about the info about the timing. I thought I was seeing some erratic behavior in the 15sec turbo stepdown. It seems not to step down when (over)heating.

great work!

thank you for sharing :slight_smile:

Dutchee,

Thanks for the driver pictures. I was going to give it a try with a 100W iron on the battery side contact while still in the head. Then attempt to put the new one on… One day, not anytime soon…. Thankfully the NW is a nice tint and not like a greenish one like some other brands have…

Do you think the switch could be desoldered before pushing the driver out? Looks to be a bear either way, but so glad it worked out for you on this mod… that looks incredible. Narsil and temp regulation would be the bees knees for a light like this…

Nice mod, it must have been quite a deal of work to put this together!

Wow, just wow,
nice work :slight_smile:

Great mod!

Great work!
I’m dreaming of build my own triple, dunno when. LOL.

What kind of solder did you use? Switching from lead based to unleaded solder should help both with elevated temperature and thermal conductivity(especially those heatsink plates soldered together). Here is the table of (almost?) all solder alloy types with melting points, and thermal conductivity for more common alloys:

Sn-Pb37 have thermal conductivity of 50.9 W/(mK) with melting point of 183C whereas Sn-Cu3 have 70.4 W/(mK) and melting point around 227C
(pure aluminium for comparison have 240 W/(mK) and copper about 400 W/(mK))

I know that there are some durability drawbacks in using unleaded solder but mainly in low temp environment so not exactly the flashlight related :wink:
I’m not familiar with any issues within flashlights. Anybody have any experience?

Wow, I was totally unaware. I thought lead based solder was the way to go and is what I used. Thanks for pointing this out!

I added thermal paste (sipik) and (more) properly resoldered the two wires. It didn’t happen again, but offcourse I was lucky it didn’t cause a short.

A 20% temperature increase in the solderbond and better conductivity is nice, I’ll scout for some Sn-Cu3 solder!

After 20secs on turbo (stepdown doesn’t seem to work for some reason, not even at 90secs :wink: ) the light get’s too hot to hold. But maybe the solder is nearing 180C right next to the LEDs…?

Does this type of solder need another solder technique, apart from a somewhat higher temperature?
Isn’t solder cracking an apparent issue with these lead free solders in flashlights, since they have a lot of warm/cold cycles?

I think most of us use lead based solder because it’s lower melting temperature and wetting properties is just easier to work with, there’s also easy availability factor and force of habit. No different soldering technique is required(or at least I do not use it) but higher temperature can require more power to work comfortably. With RoHS directive implementation in EU all current LEDs conform to lead free soldering requirements and major flashlight manufacturers should use lead free solder for that market. I am not aware of any problems with lead free solder in LED applications, but I do not reject the possibility. I had no need to reflow LEDs for high power flashlights yet, I did it for single AA lights so I used lead based solder I had on hand. It’s also worth noting that higher required temperature combined with makeshift temperature control (or lack thereof) poses greater risk of LED damage. But then again - pioneers are the ones making history :wink:

There’s also this thread here: Thermal conductivity of solder and pure indium is proposed as a solder because of superior properties (157C melting point, 81.8 W/(mK) thermal conductivity). But it seems it was not put to test ultimately.

I’ve got all the new components in, so it was time for the next mod on the modded DQG! :partying_face:

New parts:
20 AWG wire (18 was too thick for the center hole, too lazy to drill it out…)
BLF X6 Attiny25 driver (Banggood)
Atmel Attiny85 (ebay)

First I flashed Narsil to the Attiny85.

Then I desoldered the AT25 from the driver.
You can see, there is quite a size difference in the MCU’s

I added solderpaste to the driverboard

I bent the AT85 paws in a little and reflowed it to the driver, with my soldering iron.

Then I soldered the 20AWG LED wires to the driver.
Connected the microswitch to AT85 pin #2 and ground with thin silicone wire this time.
I’ve never been so happy with 2 blinks, the first sign of Narsil functioning. :+1:

After the first assembly, the switch came loose. So I glued it in with Sipik.

Then I soldered the LED wires in place. At first I got a short between the - Driver to LED wire and + on the noctigon. I added kapton tape to the top of the driverboard and solved the shortcut. - touched the copper heatsink and caused a short somehow :question:

Thereafter I reinstalled the GITD tape, originally installed in the DQG.
The light switches off! All works very nice. I love this firmware! Thanks Tom E, awesome job :beer:

The moon on level 1 I really like. It is so low! With the UT-210E I can allmost not even read the difference in moon vs off(parasitic drain).
Makes it a nice locator light at night. Lowest moon mode I’ve ever seen, love it.
In the picture below, moon is actually on!

I changed the temperature stepdowm delay in Narsil to 1 Sec. The light is way more managable now. However at 13.5A on a 30Q battery it still gets too hot to hold in about 10 seconds.
I’ve put a 3400mAh protected NCR18650B in it. Now its way more managable, safer and has higher capacity. And still stupidly bright :person_facepalming:

I use mode 4 with changed setting (moon+) 1-10-35-100.

@Firelight2, the thermal paste did solve the desolder problem. Now with 20AWG wire as well, I seems very oke.

So no need for other solder I guess.

Mission accomplished!
(for now :person_facepalming: )

Great :slight_smile: I will try changing to some of my fet+1 drivers to attiny85 and e-switch with narsil !
Do you have a EU source for the attiny85 ?

Dutchee, fantastic write up for this phase of the mod. Really appreciate the pics, and happy it is all working for you…

You can buy them at mouser. But due to the shipping cost, I bought a few from ebay.
These are the ones I used:
Ebay 5pcs Atmel Attiny85

Nice driver mod. I admire you guys modding and building these items.

Sorry, I think I missed this post…

Yes, with a tiny solder tip it should be doable.

I had to resolder the black wire to the switch while it was glued in the head. It came loose when I removed the driver to add the kapton tape. So I think desoldering should be relatively easy.

When heating the whole head as you mentioned, you might melt the rubber boot over the e-switch. Also the original glue might burn. I think its safer to remove the driver before the LED reflow.

Excellent job on this mod! Wish I had one of these lights, because this mod looks great for it.

Let me know f you have any flaky problems with the light - mode changes, going in/out of strobe or hi, etc. Most Tiny85 builds I had to tweak things to get it working reliably - add another 10 uF cap stacked on C1, or air wire a 0.1 uF cap from pin #4 to pin #8. Now could be more of a problem with the SIR800DP FET's I use, but probably still happens with regular FET's. Richard's new MTN-17DDm driver that supports a Tiny25 has several changes to properly support 25/45/85's.

Thanks for the nice replies, guys!

On the pannie cell it doesn’t overheat. When allmost burning my hand, it steps down at 30sec’s. On 35% it slowly loses the excessive heat. I don’t plan on testing this in a pocket though… I still have to do a tailcap reading on the pannie…
It does heat up real nicely all the way to the tailcap. But the cell stays quite cool, unlike in the stock DQG where it serves as a heatsink…
On a 30Q cell the strobe is so bright, it actually gives the same effect as a stroboscope in a discotheque!

@Tom E: Besides the somewhat high temperatures, I didn’t notice anything flaky, not even once. All modes work flawless so far. Really the nicest U.I. in a light I own. I usually EDC small twisty 10440 lights, so the lack of unexpected mode changes is something I’ll gladly get used to. Really nice with the e-switch.
But if I’ll notice something strange, I’ll give you a head’s up :+1:

Edit: Tom E just made clear there is not thermal stepdown in Narsil, so the heat has nothing to do with any flaw in the driver or Narsil. It actually performs 100% in a super hot light :+1: