Fenix TK61 teardown and mod thread

186 posts / 0 new
Last post
dazed1
dazed1's picture
Offline
Last seen: 3 weeks 15 hours ago
Joined: 05/31/2013 - 11:16
Posts: 1557

@RaceR86, did you do all of this testing with stock transistor? (runtimes indoors)

Fenix TK61 MOD> resistor, transistors, copper heatsink, Artic Cooling MX4, dedome, 3A > 5.7A ~ 1700 lumens > (600.000cd)

RaceR86
RaceR86's picture
Offline
Last seen: 2 years 7 months ago
Joined: 02/22/2012 - 08:32
Posts: 3777
Location: Norway

dazed1 wrote:
@RaceR86, did you do all of this testing with stock transistor? (runtimes indoors)

No reasons to test before. Its built to handle 4 hours and 20 minutes on Turbo when stock. Or about twice that time if you have the battery extension kit of whatever its called. Basically its built to handle full output all the time. Like pretty much all lights...

I did two runtime tests after resistor mod.

I hope to give you some exact numbers regarding the modified TK75 in the distant future. Wink

 

RMM wrote:

If it can't hold up to 10 minutes turbo, it's no good.  This is a $200 (modded) super thrower and it should be able to run until it gets too hot to hold or until the cells die.  

Yupp.. Thats pretty much what I tried to say when I wrote that wall of text above. Big Smile


On a side note. I bought my light from WB.  TK61 is now back it stock from BG and slightly lower price than earlier. A tad cheaper than WB after coupon code are used in both stores it seems.

BLF LED database – collaboration spreadsheet and latest news about where to buy LEDs
http://budgetlightforum.com/node/19342

dazed1
dazed1's picture
Offline
Last seen: 3 weeks 15 hours ago
Joined: 05/31/2013 - 11:16
Posts: 1557

So let me get this straight.

You get the light to 4.8A from 3A at stock, with only resistor swap right, nothing else added?

And you did 2x runtimes ~ 30 mins each on turbo (one of them indoors you said?) without any issues at all? that’s totally impressive, knowing that it was suggested to me to not go over 4.2A for safety reasons (and even that can be dangerous)

Anyway, just to be on safe side i will wait for RMM and maybe your findings about new transistor (s) you guys have any idea of what can be used for reference? thanks.

Fenix TK61 MOD> resistor, transistors, copper heatsink, Artic Cooling MX4, dedome, 3A > 5.7A ~ 1700 lumens > (600.000cd)

RaceR86
RaceR86's picture
Offline
Last seen: 2 years 7 months ago
Joined: 02/22/2012 - 08:32
Posts: 3777
Location: Norway

dazed1 wrote:
So let me get this straight. You get the light to 4.8A from 3A at stock, with only resistor swap right, nothing else added? 

Yupp. As seen in OP.. More like 4,7x amps to be precise with one R030.

 

dazed1 wrote:

And you did 2x runtimes ~ 30 mins each on turbo (one of them indoors you said?) without any issues at all? that's totally impressive, knowing that it was suggested to me to not go over 4.2A for safety reasons (and even that can be dangerous)

Pretty impressive huh.. Not so useless info to know about afterall. Although, that does not mean its not dangerous for the driver. Silly I just means it went fine when I did it. Big Smile Its worth to mention again that I did it without any thermal pads or heat sinking done to the driver.  So with thermal pad + normal outdoor use you are probably much safer.

(btw, both tests were indoors. First were around 15-20 minutes or so. I did not time it.. Turned off the light when it felt "hot enough".)

 

dazed1 wrote:
Anyway, just to be on safe side i will wait for RMM and maybe your findings about new transistor (s) you guys have any idea of what can be used for reference? thanks.

I know some details about the transistor/MOSFETs mods that vinh is doing.

Ill let RMM continue with his MOSFET testing, and let him do the talking. Technically its beyond me...

BLF LED database – collaboration spreadsheet and latest news about where to buy LEDs
http://budgetlightforum.com/node/19342

dazed1
dazed1's picture
Offline
Last seen: 3 weeks 15 hours ago
Joined: 05/31/2013 - 11:16
Posts: 1557

Thanks for the great answer Smile

I understand that every part is different, and that even on stock can be burned, that’s why are electronics, but still you get very good results.

BTW, assuming on stock TK75 gets to 2850lm, (Fenix is spot on with ANSI) with this mod, 3.0A >4.6/4.7A, the 4400 lumens should be led lumens, and OTF should be around 4300? or driver efficiency comes into play additionally compared to stock?

Fenix TK61 MOD> resistor, transistors, copper heatsink, Artic Cooling MX4, dedome, 3A > 5.7A ~ 1700 lumens > (600.000cd)

RaceR86
RaceR86's picture
Offline
Last seen: 2 years 7 months ago
Joined: 02/22/2012 - 08:32
Posts: 3777
Location: Norway

Its probably around 4500 LED lumen at that current.

Differences in LED Lumen and OTF are the losses you get in reflector, lens, and possible things that blocks light (Bezel, locating ring and such). I don't know exactly, but I would probably expect at least 10-15% lower OTF numbers. Fenix does have very good design, reflectors and AR lenses. So maybe its as high as 4000 lumen OTF?... oh well. Lets try to keep to TK61 in this thread. Smile

BLF LED database – collaboration spreadsheet and latest news about where to buy LEDs
http://budgetlightforum.com/node/19342

dazed1
dazed1's picture
Offline
Last seen: 3 weeks 15 hours ago
Joined: 05/31/2013 - 11:16
Posts: 1557

Hmm 4000 is a bit on a low side, because Fenix get like 950 lumens per led on stock (it was measured)

The loses from lens/reflector should be the same going from ~ 2850 >4000+ lumens or even if a bit more, still it should be around 150 lumens per led loses compared to ~ 80lm at stock and this is big downgrade i estimate, it should be less.

It should be around 4200 imho…knowing that at 4.7 the leds should output around 1500 lumen x3 = 4500 – 200 we are around 4300/4200 at worst imho (with loses/efficiency)

Sorry for offtopic!

Fenix TK61 MOD> resistor, transistors, copper heatsink, Artic Cooling MX4, dedome, 3A > 5.7A ~ 1700 lumens > (600.000cd)

RaceR86
RaceR86's picture
Offline
Last seen: 2 years 7 months ago
Joined: 02/22/2012 - 08:32
Posts: 3777
Location: Norway

If you say stock emitter current is 3A that means about 1120 lumen at emitter based on graph above. I dont know if its 3A stock. But you said so. You said its measured at 950 OTF too right? That is about 15% lower than 1120 lumen.

Graph shows about 1500 lumen at 4,6A. Which means 4500 lumen. Minus 15% equals about 3825 according to the graph above.  I believe that match`s numbers (graph you used) were emitter numbers? I cant remember if it was T6 or U2 bin.

You have to take into account that these numbers are just different peoples numbers. They may not always be 100% comparable, or 100% accurate. Unless you measure it with a calibrated equipment we will always be doing guesstimates. People many not even have the same emitter temperature when doing these tests, which also changes things due to heat sag. What Im saying is. Some might say 4000. Some might say 4200. Some might say lower, some might say higher. If I use Djozz charts (which I believe similated OTF but without lens), we would probably be looking at around 4400 OTF with 3x XM-L2 U2 at 4,6A.

Also, on certain Fenix light, some here at BLF measure quite similar numbers, some measures higher with well above 10%.

Whatever you actually are seeing, you wont see the difference between 4000 OTF or 4400OTF anyway. It could be just the same, just like these guesstimates we are doing based on different peoples data. If you say 4200, then Im not going to argue with you against it. Your guesstimate based on available info seems to be as good as mine. When that is said. Relic38 had a giveaway where people guessed OTF numbers on a light based on given spec. You can see how large the differences were. I averaged the closest guesses I believe. This obviously makes me king of guesstimates. Silly

Oh well. Enough Off topic. Start your own thread if you want to discuss something not TK61 related. Big Smile

(Topic? "Modded TK75 lumen guesstimates based on assumed emitter current  - A non TK61 mod thread")

BLF LED database – collaboration spreadsheet and latest news about where to buy LEDs
http://budgetlightforum.com/node/19342

RMM
RMM's picture
Offline
Last seen: 3 days 20 hours ago
Joined: 07/23/2013 - 13:47
Posts: 4006
Location: USA

Mine showed about 2.85A to the emitter, stock, on my meter.  I have mine up to around 5.8A but haven't had time to put it back together yet, too busy building other stuff!

Mountain Electronics : batteries, Noctigon, and much more! What's new? 

DB Custom
DB Custom's picture
Online
Last seen: 15 min 18 sec ago
Joined: 01/13/2013 - 22:28
Posts: 18004
Location: Heart of Texas

Richard, just throw it in a box and send it to me, I’ll put it back together and do some, ahmmmm, testing. Smile Hey, no problem! That’s what friends are for! Wink

Dale

RMM
RMM's picture
Offline
Last seen: 3 days 20 hours ago
Joined: 07/23/2013 - 13:47
Posts: 4006
Location: USA

3 minutes on turbo and counting...fingers crossed.  

For maximum output I think the way to go is an FET driver, or four Knuckleheads running in unison. Cool

As a point of reference, my stock light measured 1200 lumens in my sphere, the light at 5.8A with the stock emitter shows 1900.

Mountain Electronics : batteries, Noctigon, and much more! What's new? 

dazed1
dazed1's picture
Offline
Last seen: 3 weeks 15 hours ago
Joined: 05/31/2013 - 11:16
Posts: 1557
RaceR86 wrote:

If you say stock emitter current is 3A that means about 1120 lumen at emitter based on graph above. I dont know if its 3A stock. But you said so. You said its measured at 950 OTF too right? That is about 15% lower than 1120 lumen.

Graph shows about 1500 lumen at 4,6A. Which means 4500 lumen. Minus 15% equals about 3825 according to the graph above.  I believe that match`s numbers (graph you used) were emitter numbers? I cant remember if it was T6 or U2 bin.

You have to take into account that these numbers are just different peoples numbers. They may not always be 100% comparable, or 100% accurate. Unless you measure it with a calibrated equipment we will always be doing guesstimates. People many not even have the same emitter temperature when doing these tests, which also changes things due to heat sag. What Im saying is. Some might say 4000. Some might say 4200. Some might say lower, some might say higher. If I use Djozz charts (which I believe similated OTF but without lens), we would probably be looking at around 4400 OTF with 3x XM-L2 U2 at 4,6A.

Also, on certain Fenix light, some here at BLF measure quite similar numbers, some measures higher with well above 10%.

Whatever you actually are seeing, you wont see the difference between 4000 OTF or 4400OTF anyway. It could be just the same, just like these guesstimates we are doing based on different peoples data. If you say 4200, then Im not going to argue with you against it. Your guesstimate based on available info seems to be as good as mine. When that is said. Relic38 had a giveaway where people guessed OTF numbers on a light based on given spec. You can see how large the differences were. I averaged the closest guesses I believe. This obviously makes me king of guesstimates. Silly

Oh well. Enough Off topic. Start your own thread if you want to discuss something not TK61 related. Big Smile

(Topic? “Modded TK75 lumen guesstimates based on assumed emitter current  - A non TK61 mod thread”)

Fair enough buddy, i will just addd that i’m not sure that the loses are proportional with current bump….imho you wont lose more lumens, if you use the same flashlight, if the loses are already 10% on stock, making the light ~ 40% brighter, wont produce additional 10% lose (but i could be wrong)

On,

WOW 1900 lumens @5.8A? thats sickkkk. I pray for your light lol Big Smile

Fenix TK61 MOD> resistor, transistors, copper heatsink, Artic Cooling MX4, dedome, 3A > 5.7A ~ 1700 lumens > (600.000cd)

RMM
RMM's picture
Offline
Last seen: 3 days 20 hours ago
Joined: 07/23/2013 - 13:47
Posts: 4006
Location: USA

Ok, it made it 11 minutes and is still working.  Body was very hot around the driver and pill, but it survived!  I fried four emitters playing with this driver, but it turns out that most of my tinkering was likely unnecessary.  I played around with every MOSFET on the board, but the only one I ended up leaving changed is the "C00C" MOSFET next to Q2 (I don't recall the number itself.)

I removed the stock "C00C" MOSFET and replaced it with four piggybacked IRLML2502 MOSFETs.  With only one or two it was too hot to touch in a matter of seconds, with four it will get too hot to touch within about 30 seconds, but it's a big improvement and it's holding for now.  

Here's what I have stacked on top of the stock sense resistors (I kept on adding until I got to 5.8A, so you can use different values to get the same effective resistance.)

(2) 100 ohm (1 on each bank for all)
(2) 75 ohm 
(2) 25 ohm

The flyback diode is also getting pretty warm, but not too hot to touch.  I will run it for a while longer tonight and see if it holds up.

I still think that (4) Knucklehead drivers running at 2A each would be pretty awesome...8A at the emitter, that's what I'm talking about.  Surprised  Will it work?  As soon as my coils come in I'm going to try some running in parallel with 1 MCU to rule them all!

Mountain Electronics : batteries, Noctigon, and much more! What's new? 

dazed1
dazed1's picture
Offline
Last seen: 3 weeks 15 hours ago
Joined: 05/31/2013 - 11:16
Posts: 1557

Oh my oh my, thats some HARDCORE PUSHING! impressive Party

Fenix TK61 MOD> resistor, transistors, copper heatsink, Artic Cooling MX4, dedome, 3A > 5.7A ~ 1700 lumens > (600.000cd)

18sixfifty
18sixfifty's picture
Offline
Last seen: 3 months 1 week ago
Joined: 12/25/2012 - 20:19
Posts: 4161
RMM wrote:

Ok, it made it 11 minutes and is still working.  Body was very hot around the driver and pill, but it survived!  I fried four emitters playing with this driver, but it turns out that most of my tinkering was likely unnecessary.  I played around with every MOSFET on the board, but the only one I ended up leaving changed is the “C00C” MOSFET next to Q2 (I don’t recall the number itself.)

I removed the stock “C00C” MOSFET and replaced it with four piggybacked IRLML2502 MOSFETs.  With only one or two it was too hot to touch in a matter of seconds, with four it will get too hot to touch within about 30 seconds, but it’s a big improvement and it’s holding for now.  

Here’s what I have stacked on top of the stock sense resistors (I kept on adding until I got to 5.8A, so you can use different values to get the same effective resistance.)

(2) 100 ohm (1 on each bank for all)
(2) 75 ohm 
(2) 25 ohm

The flyback diode is also getting pretty warm, but not too hot to touch.  I will run it for a while longer tonight and see if it holds up.

I still think that (4) Knucklehead drivers running at 2A each would be pretty awesome…8A at the emitter, that’s what I’m talking about.  Surprised  Will it work?  As soon as my coils come in I’m going to try some running in parallel with 1 MCU to rule them all!

Very nice info and thanks for sharing.

8amps would be simply incredible. I would love to see that happen! I have to think it’s possible as long as it’s really well regulated and there are Zero spikes.

I have to say I’m pretty happy with mine after adding the R025 but then again, who doesn’t want more?

I’m a junky, I mod lights so I can sell lights so I can buy more light to mod so I can sell lights to buy more lights to mod.

JohnnyMac
JohnnyMac's picture
Offline
Last seen: 10 months 2 days ago
Joined: 04/12/2011 - 16:03
Posts: 8864
Location: Eastern PA

RMM wrote:

I'll post full details tomorrow if all works out, but I've got mine cranked up to 5.8A to the emitter now.  There is no way the stock driver MOSFET(s) could handle this kind of current, I have a few burned fingers to prove it! (low tech thermometer! Yell)  If it survives a few 10 minute runs on high, I'll post details here.  I also have a stock light to compare it against and get some readings from (lumens, lux.)  I put the stock emitter back on the stock MCPCB, so the two lights will be equal in that regard.  

RMM, what other changes did you make to the driver besides the resistors?  What resistor did you add to get to 5.8A?


I should have finished reading the whole thread.  ROFL.  Don't mind me... Silly

RaceR86
RaceR86's picture
Offline
Last seen: 2 years 7 months ago
Joined: 02/22/2012 - 08:32
Posts: 3777
Location: Norway

Thanks for the update RMM and all your work!

Now process my latest order and take my money! Silly

My future plan is still currently HX-1175B.

I would check into your numbers btw. You measured 2,85A at the emitter and 1200 lumen OTF stock. Those numbers does not seem to go hand in hand to me. Even if the emitter saw 3A. Would you expect 1200 lumen OTF? Seems too high compared to  Cree`s own data, Match, djozz, Fenix, Selfbult, etc.. For all I know, Fenix could have gotten U3 and not mentioned it, but I doubt it. Still, would seem a bit high.

BLF LED database – collaboration spreadsheet and latest news about where to buy LEDs
http://budgetlightforum.com/node/19342

RMM
RMM's picture
Offline
Last seen: 3 days 20 hours ago
Joined: 07/23/2013 - 13:47
Posts: 4006
Location: USA

It does seem a bit high to me as well, but unfortunately the stock light I was measuring was a different copy.  I never measured the one I modded beforehand, so I don't know how high it really was.  I have a suspicion that the current is actually a bit higher at the emitter than what my meter says, due to the extra resistance of all the alligator clips, etc. I was using. 

Mountain Electronics : batteries, Noctigon, and much more! What's new? 

DB Custom
DB Custom's picture
Online
Last seen: 15 min 18 sec ago
Joined: 01/13/2013 - 22:28
Posts: 18004
Location: Heart of Texas

Solder a loop in (basically extend the negative lead) and clamp around it. Piece of cake and little to no extra resistance is added. On the bench, it’s easy to measure directly from the leads between the driver and emitter, no splices cuts or clips required.

Dale

antoninodattola
Offline
Last seen: 6 months 1 day ago
Joined: 11/22/2013 - 14:19
Posts: 102
Location: Italia
Congratulations RMM! 
good job! Wink 
 
These should be even better mosfet: 
 
 
if there is space in the torch, you could also change package mosfet ... Wink
RMM
RMM's picture
Offline
Last seen: 3 days 20 hours ago
Joined: 07/23/2013 - 13:47
Posts: 4006
Location: USA

Took this out last night.  All I can say is that my sight was limited only by the environmental haze / dust, not by the level of light.  I think I might like this one better with the dome left on but with the current cranked up, that way I get a bit wider beam.  The BTU Shocker cranked up has a better beam (MUCH wider, very similar intensity) than this IMO, but this is much lighter and the runtime is much longer!  

Mountain Electronics : batteries, Noctigon, and much more! What's new? 

RMM
RMM's picture
Offline
Last seen: 3 days 20 hours ago
Joined: 07/23/2013 - 13:47
Posts: 4006
Location: USA

antoninodattola wrote:
Congratulations RMM! 
good job! Wink 
 
These should be even better mosfet: 
 
 
if there is space in the torch, you could also change package mosfet ... Wink

If these didn't hold up, that was going to be my next step:  remote mount a 70N02.  I have run several 10-15 minute torture tests and it's holding strong so far.  

Mountain Electronics : batteries, Noctigon, and much more! What's new? 

dazed1
dazed1's picture
Offline
Last seen: 3 weeks 15 hours ago
Joined: 05/31/2013 - 11:16
Posts: 1557

RMM i hope you will stock some mosfets for this light, it will be very good to have an option to buy them from your store, instead of endless searching on mouser site.

Fenix TK61 MOD> resistor, transistors, copper heatsink, Artic Cooling MX4, dedome, 3A > 5.7A ~ 1700 lumens > (600.000cd)

RMM
RMM's picture
Offline
Last seen: 3 days 20 hours ago
Joined: 07/23/2013 - 13:47
Posts: 4006
Location: USA

I think I may try the 70N02 and try for 6.5A-7A with the stock driver.  If it blows I guess I'll have a $150 paperweight, huh?  Nah, that will just motivate me to try my 7A-8A option... Wink

Mountain Electronics : batteries, Noctigon, and much more! What's new? 

dazed1
dazed1's picture
Offline
Last seen: 3 weeks 15 hours ago
Joined: 05/31/2013 - 11:16
Posts: 1557

The driver should take it i think, with replaced mosfets at least? but i’m noob so don’t count on me lol Big Smile

Fenix TK61 MOD> resistor, transistors, copper heatsink, Artic Cooling MX4, dedome, 3A > 5.7A ~ 1700 lumens > (600.000cd)

RaceR86
RaceR86's picture
Offline
Last seen: 2 years 7 months ago
Joined: 02/22/2012 - 08:32
Posts: 3777
Location: Norway

RMM wrote:

I think I may try the 70N02 and try for 6.5A-7A with the stock driver.  If it blows I guess I'll have a $150 paperweight, huh?  Nah, that will just motivate me to try my 7A-8A option... Wink

If 5,8A is peak. Im already headed for my 7-8A option. 

But I could try 70N02 first. Just replace/remove "C001" from Q1 and connect the 70N02 in similar fashion + further resistor mod? I can try that if you want. But if you are going to do it anyway, I could just sit back and relax  get some progress done on my scratch build.

BLF LED database – collaboration spreadsheet and latest news about where to buy LEDs
http://budgetlightforum.com/node/19342

RMM
RMM's picture
Offline
Last seen: 3 days 20 hours ago
Joined: 07/23/2013 - 13:47
Posts: 4006
Location: USA

Racer, I'll let you try it first!  I'm pretty busy today.  I would keep the resistors the same to start with and see how the current compares.  I think that at a certain point we're really going to need a larger inductor, more input capacitance, etc. to make this reliable. 

Mountain Electronics : batteries, Noctigon, and much more! What's new? 

RaceR86
RaceR86's picture
Offline
Last seen: 2 years 7 months ago
Joined: 02/22/2012 - 08:32
Posts: 3777
Location: Norway

As a very quick and temporary test I connected the 70N02 like this:

Did not get over 1 amp output. Did I connect the legs in the wrong way or something?

Changed back to stock mosfet and saw 4,8A at the emitter. No harm done.

BLF LED database – collaboration spreadsheet and latest news about where to buy LEDs
http://budgetlightforum.com/node/19342

RMM
RMM's picture
Offline
Last seen: 3 days 20 hours ago
Joined: 07/23/2013 - 13:47
Posts: 4006
Location: USA

No, that's the correct way.  The gate charge, PWM frequency, etc. may not be suited for the 70N02.  I guess we'll have to stick with stacked "tiny" FETs for now.  BTW I have run this one quite a bit more and it's holding up well.  I feel confident with it around 6A.

Mountain Electronics : batteries, Noctigon, and much more! What's new? 

dazed1
dazed1's picture
Offline
Last seen: 3 weeks 15 hours ago
Joined: 05/31/2013 - 11:16
Posts: 1557

So with the parts swapped, 5.5/6A should not be a problem right?

Fenix TK61 MOD> resistor, transistors, copper heatsink, Artic Cooling MX4, dedome, 3A > 5.7A ~ 1700 lumens > (600.000cd)

Pages