BTU Shocker Triple MT-G2 with a twist -- Aiming for >100Watt ~9000Lumens -- With external 2S power pack, handle etc...

That sounds very fun! Make me a second one so I can strap it onto the top of my BTU. :wink: Crazy flood and throw in one. :bigsmile:

So… finally, finally after 39 days of not very patient waiting I have myself some MT-G2 emitters on noctigons!
These things are huge and gorgeous. I can’t wait to light them up.


I was a bit surprised to see that this one emitter is not particularly well soldered to the pcb.
It’s not sitting completely flush on the solder pads and slightly cocked up at one end. I wasn’t expecting that from Intl-outdoor but I guess it can happen, shouldn’t be too hard to reflow in any case.


As big as they are, the fit into the BTU reflector is actually a bit looser than I was expecting. Maybe I need to make some centering rings for them.
Just a quick assembly to see how things are looking.

MMmmm triple MT-G2 shocker, time to get cracking on the electricals to feed these bad boys! :bigsmile:


In the same order I also got the infamous 9A IOS buck driver, I won’t be using it to drive the emitters in this light but I plan to do some MT-G2 testing of it to see what’s what. Hopefully with some direct heatsinking it can still make a suitable option for driving a single mt-g2 at over 5A reliably.

My particular sample came with 3x R200 sense resistors on the top of the board and 6x (double stacked) on the bottom for a total of 9. I’m having trouble finding the thread where the sense resistor formula for this driver was discussed but I assume this combination should indeed deliver 9A output to an XML. I’ll see when I start testing anyway.

Cheers
Linus

That reflector with 3x mtg2 in it…

hehe :slight_smile:

So I tried fixing the reflow on that one wonky emitter but noticed that once the solder had melted I couldn’t get it to sit flat either. :~
It sorta rocked back and forth in a weird way like a see saw so I removed the emitter completely and found a chip of what looks like dried fujik stuck to the flux underneath it.
Very strange, must have gotten in there by accident on the assembly line.
I cleaned off the fujik chip and tried reflowing it once more. This time it sat down nice and snug with a hopefully solid thermal path to the pcb. :slight_smile:

Wow - I bought a ton of LED's now from Hank, and only thing I noticed is the annoying solder balls, but I trim them off because they usually interfere with the plastic alignment pieces. Ohh - found out best way to trim is an exacto flat on the surface of the star, and never cut down onto the star -- you can cut thru the trace -- not that I ever did something that stupid (Yell).

I had the same approach and thought process on this, I stopped using the knife because I was worried about damaging the traces or slicing through the dome if I slipped on the horizontal cut. :stuck_out_tongue:
I knew it was going to happen sooner or later so I went back to just taking the solder balls off with a fine tip on the soldering iron. A quick pass and they’re usually gone. Here I’ll just leave them since they wont interfere with anything.

Just playing around with the IOS “9A” buck driver and I’m not very impressed.

-First of all it’s only delivering max 6.7A to the LEDs (input 12v — output to 3 XMLs in parallel), I think I need 3 more of those R200 stacked to the top row to get the advertised output. I had a feeling it needed a double stack top and bottom to get the full output.

-Second the thing gets crazy hot on high even at this lower current. This is a well known problem with the driver, not very efficient and requires serious attention to heatsinking. But I was still surprised at how quickly this thing starts cooking! My guess is these are not really designed for more than about 5A sustained output so it’s no wonder they were burning up in MT-G2 builds running 9A…

-Third it whines louder than any driver I’ve ever heard on Med and Low, really irritating.

-Fourth, it takes about a week to switch modes, click and then you have time to make a cup of tea before the next mode comes up! :stuck_out_tongue:

-Finally it has visible PWM flicker on medium and low, not terrible since it’s at a reasonably fast Hz but still visible to the naked eye.

So am I glad I spent $15 on this driver? Nope
I was hoping to find a use for this driver but with the combination of all these issues I’ve lost all enthusiasm for it.

Edit: Saw that they were no longer available from IntlOutdoor, let’s hope he has a better alternative coming!

All those custom 5 LED, 7 LED, and the old IOS 3.5/4.5 and LCK-LED 5A drivers switch modes slowwwwlyyyy also. Think they also have the whining too. Think these drivers just don't like doin driver functions, but do it begrudgingly... Smile All I'm saying is don't be too surprised by seeing this more often...

Yeah I’m totally new to playing with multi cell buck/boost drivers. I guess I had my expectations set too high coming from the simplicity and downright UI perfection that you can get from a 7135 driver these days. These switching drivers are a whole other story and I’m not very impressed so far, good thing I don’t need to bother with them for this build. :wink:

I did notice this driver has an Atmel 268 MCU on it, maybe someone smarter than me can work out how to write some custom firmware to beat these things into shape! :slight_smile:

At first I was complaining about the slow mode switching, and got a response like that's how they are... Just weird the after market drivers are so different than the stock drivers, but we do want more amps and we do want only 3 modes instead of the stock 5 w/blinkies... But the high amps seems like the drivers aren't really designed to handle anyway, so not sure what we are gaining over resistor modding, if it's possible on all these.

This is all making the Task LED drivers look all the better, but I don't see a Task LED driver working for a single high amp LED - think they work well for multi-LED's of 5 or more. Even 3 LED's to max out at 3A each, I'm not that interested in.

Indeed, 3A is only interesting when talking about the 36v version of the MT-G2. :stuck_out_tongue:

Let the stacking commence!

3x KD 7135v2 boards stripped of everything but the 7135s and stacked. A total of 16x 0.380A chips on each board to deliver 6.08Amps regulated to each emitter.
Didn’t I say I would aim for 5A per emitter?…well… yeah, but sixteen chips is a much rounder number… :bigsmile:

These will be controlled by a single master board, a stripped Nanjg 105c with an Atmel MCU running some custom firmware and Zener modded for 2S input power.


This is the general layout of how I plan to mount the driver setup into the flashlight cavity.


Still need to work out the specifics but I may use this aluminium slug and drill some 17mm recesses into it to snuggly fit the slave drivers. Have it act as both a heatsink for the 7135s and a holder. The master will probably be mounted more loosely near the contact board, that way I can easily get it out and reprogram it if needed.

The bolt is a longer replacement for the one that usually holds the reflector down (standard M5 bolt from what I can tell). Just playing with ideas at the moment, there’s so much driver cavity space in this light that pretty much anything is possible. Trickiest part will be to route the 3 pairs of emitter wires through it all neatly and out to the emitters. Might be time to do some 3D mockups again.

Cheers
Linus

This is just… wow.

WOW alright.

Thanks guys, still so much to do :slight_smile:

That is not a microcontroller, that is a serial EEPROM.

Ah too bad, so I guess the other generic chip is the controller?

Yeap that must be it.

i hope you are taking notes on this build

because when you are done ill be sending you a valuable sum in dollars to build me one :stuck_out_tongue:

Hehe, I do have a spreadsheet of all the parts costs but I’m afraid to total it up for fear of what I’ll see! … :open_mouth: :money_mouth_face:
Just hoping the final light will be worth it :slight_smile: