Guys, end of the road for my Sky Ray 3 x T6 SR3800

53 posts / 0 new
Last post

Pages

2100
Offline
Last seen: 2 years 7 months ago
Joined: 05/28/2011 - 10:28
Posts: 4512
Location: SINGAPORE
Guys, end of the road for my Sky Ray 3 x T6 SR3800

Was just using my SR3800 when it suddenly shut off.  Something is wrong, coz it is just a couple of warm 2-3 min runs testing the lux and the batteries are XTAR 2600/Solarforce/TF Flames/Panny.

Did a tailcap measurement and noticed the amps ramp up and then down very fast, probably tripping the PCB.  Fearing the worst, i took out the retaining ring for the driver and pulled out the driver.

Lots of transistor burn smell, the thing wasn't really hot even.  So i guess it is toast.  Pulled out the -ve and +ve wire and tried to use 2 x AA to see if the thing lights, well not even a flash.

 

It has been a fun run of about 20hrs.    Careful guys, use good cells with good PCB. 

Edited by: 2100 on 07/23/2011 - 02:49
2100
Offline
Last seen: 2 years 7 months ago
Joined: 05/28/2011 - 10:28
Posts: 4512
Location: SINGAPORE

Now what, i can't take out the drop-in and replace with a 5 x R2.

agenthex
agenthex's picture
Offline
Last seen: 7 years 7 months ago
Joined: 07/14/2010 - 04:49
Posts: 3070
Location: Merica

You'll have to find a new driver which isn't trivial for this kind of light. You can perhaps find out if it's parallel and maybe get a sst driver for it.. The host isn't worth all that much since the similar tr-1200 is like $10.

Reading this makes you smarter: http://lesswrong.com/

2100
Offline
Last seen: 2 years 7 months ago
Joined: 05/28/2011 - 10:28
Posts: 4512
Location: SINGAPORE

I found that the driver took out the LEDs too.  Took the wires and wired to 3xAA, no flash at all. Then tried a Li-ion, also no flash.

agenthex
agenthex's picture
Offline
Last seen: 7 years 7 months ago
Joined: 07/14/2010 - 04:49
Posts: 3070
Location: Merica

If you only connected to the big wires to the driver, it's possible that they're actually series inside. 

Reading this makes you smarter: http://lesswrong.com/

fishmaniac
fishmaniac's picture
Offline
Last seen: 6 years 7 months ago
Joined: 05/03/2011 - 07:21
Posts: 1019
Location: Peoples Republic of California

My condolences. Hope your time together was filled with nice memories and ear to ear smiles.

On the bright side, you got a nice looking $50 paperweight/nutcracker/doorstop. =)

 

DasFriek
Offline
Last seen: 9 years 3 months ago
Joined: 06/26/2011 - 16:26
Posts: 294
Location: Ohio

Man that sux! My condolences.

What you think killed it? I just got my Panasonic 2900's from HKE today and put them in my 3800 even tho they are unprotected. Think i should switch to the HiMax 2600 protected cells?

BTW check out the stickied 3800 review at the top of the page in the 18650 cell section, I think you will find my last post most interesting. 

If you weren't so far away and not a modder id be trying to buy that paper weight from you as im just looking for a good excuse to mod something and i cant decide where to start yet. 

(Edit) I guess my post in the 3800 review isnt the big news i thought it was, But i ended up being lucky none  the less. 

2100
Offline
Last seen: 2 years 7 months ago
Joined: 05/28/2011 - 10:28
Posts: 4512
Location: SINGAPORE

Dasfriek.....well, actually with this turn of events that mitro's and my drivers went toast suddenly, I think better change to protected batteries.  What happens is that within 0.1s of the driver failing, the PCB would cut the current and prevent a thermal runaway in the cell.   No problem, not really heartpain, in fact it is good that I had the Xtar 2600 inside.

You can still use the NCR18650s, they are very safe.  Just that you need to be aware that when a light suddenly cuts, you need to immediately cut the power at the tailcap.  Also do not use magnets with them as they might shift and cause a short.  A blob of solder and filed down flat is much safer.

DasFriek
Offline
Last seen: 9 years 3 months ago
Joined: 06/26/2011 - 16:26
Posts: 294
Location: Ohio

Dang, What a bummer as i bought those cells just for that light.

BTW which revision was yours? Solid head, drop-in, or screw in pill? 

agenthex
agenthex's picture
Offline
Last seen: 7 years 7 months ago
Joined: 07/14/2010 - 04:49
Posts: 3070
Location: Merica

But the protected cells didn't protect your light??? I still don't think you necessarily fried the xml's. To kill them in para requries >15 amps, surely some trace on the driver would go short before the emitters do.

Reading this makes you smarter: http://lesswrong.com/

2100
Offline
Last seen: 2 years 7 months ago
Joined: 05/28/2011 - 10:28
Posts: 4512
Location: SINGAPORE

fishmaniac wrote:

My condolences. Hope your time together was filled with nice memories and ear to ear smiles.

On the bright side, you got a nice looking $50 paperweight/nutcracker/doorstop. =) 

 Yeah this light was sure fun and it is easily the longest running and best tested in my whole range.. 

And the best thing is, Agenthex is right, it is run in series.  I am able to get the head to light again with 2 x 18650. So I'll just ask MF's Audrey for a replacement driver, well at least i did testing and advertising for them.  LOL!

agenthex
agenthex's picture
Offline
Last seen: 7 years 7 months ago
Joined: 07/14/2010 - 04:49
Posts: 3070
Location: Merica

:bigsmile:

Reading this makes you smarter: http://lesswrong.com/

DasFriek
Offline
Last seen: 9 years 3 months ago
Joined: 06/26/2011 - 16:26
Posts: 294
Location: Ohio

Id look for a SST 5A driver and see if that dont push the emitters to their maximum power output.

2100
Offline
Last seen: 2 years 7 months ago
Joined: 05/28/2011 - 10:28
Posts: 4512
Location: SINGAPORE

DasFriek wrote:

Id look for a SST 5A driver and see if that dont push the emitters to their maximum power output.

12.6V @ 1.9A DD, that's 24w. The crocodile clips and 22awg wires are my driver.

I'll try my 3S3P Li-Co batt pack from my HID and see what it can do in DD. And measure the lux at my 8.17m test distance.  If it fries, it fries...

agenthex
agenthex's picture
Offline
Last seen: 7 years 7 months ago
Joined: 07/14/2010 - 04:49
Posts: 3070
Location: Merica

Brave man. I would run them off a bit first, though.

Reading this makes you smarter: http://lesswrong.com/

Match
Match's picture
Offline
Last seen: 5 years 2 months ago
Joined: 11/14/2010 - 11:57
Posts: 1488
Location: South Carolina, USA

2100 wrote:

I'll try my 3S3P Li-Co batt pack from my HID and see what it can do in DD. And measure the lux at my 8.17m test distance.  If it fries, it fries...

You should be fine.  During testing, I pushed 8 amps through 3 xm-l's in series from 4 26650 LifeP04's DD.  Granted, it was only for a second until my fingers could react to what my eyes just saw, but all the emitters lived. 

2100
Offline
Last seen: 2 years 7 months ago
Joined: 05/28/2011 - 10:28
Posts: 4512
Location: SINGAPORE

Match wrote:

 

You should be fine.  During testing, I pushed 8 amps through 3 xm-l's in series from 4 26650 LifeP04's DD.  Granted, it was only for a second until my fingers could react to what my eyes just saw, but all the emitters lived. 

Ok no issue then. Thinking of leaving this "as it is" to play and getting the Trustfire 3xT6. Lots of threads + lots of external fins, 7 of them vs 3 on Sky Ray/Hi-Max/Fandyfire.

Actually I am using SLA to play now, no issue. 4S1P is definitely too high.  I guess in stock form, it is not a matter of LEDs surviving but rather the driver electronics surviving. From what we can see only the lower PCB is somewhat heatsinked.

2100
Offline
Last seen: 2 years 7 months ago
Joined: 05/28/2011 - 10:28
Posts: 4512
Location: SINGAPORE

Direct drive with a fully charged SLA measured 2.2A to the emitters in series, i am getting 324 lux @ 8.17m.  21626 lux @ 1m, or candlepower. 

Using slightly larger diameter wires on one of the them i could do 2.8A, giving me 360 lux @ 8.17m. 24029 lux @ 1m.  This thing comfortably outthrows stock C8/UF-980L/U80.

Update, i pushed 3.3A and got 380 lux, that's the max. Head gets hot really fast.

Stock SR3800 is 260 lux @ 8.17m. 

Engrotelis measured 1580 OTF lumens in his lightbox (he said his is pretty accurate). If that's the case 380 lux / 260 lux = 2309 lumens OTF.

 

mitro
Offline
Last seen: 1 year 8 months ago
Joined: 01/25/2011 - 11:58
Posts: 707
Location: Montgomery, IL USA

Sorry to hear yours is dead too, 2100. I took mine all apart and the LEDs appear to be fine. I guess I need to find a suitable driver but I'm not sure I want to bother.

2100
Offline
Last seen: 2 years 7 months ago
Joined: 05/28/2011 - 10:28
Posts: 4512
Location: SINGAPORE

mitro wrote:

Sorry to hear yours is dead too, 2100. I took mine all apart and the LEDs appear to be fine. I guess I need to find a suitable driver but I'm not sure I want to bother.

Run it in DD with a SLA, quite fun. Laughing3S would drive it to approx 1.9-2.2A depending on charge level (12.4V-12.6V unloaded).

wiljen
Offline
Last seen: 1 year 6 months ago
Joined: 04/28/2011 - 16:04
Posts: 399
Location: East TN, SW VA

 How about a taskLED.com H6flex or H6CC?   Both are capable of 6.6 or 6.7A delivered and would provide a driver circuit designed for series wired Xmls.    Granted at $30-40 you have almost as much in the driver as you do the light, but I'm betting you have a much harder time hurting one of those drivers.

bushytails
Offline
Last seen: 8 years 1 month ago
Joined: 06/11/2011 - 21:53
Posts: 167

Both of my skyray 1x14500 lights failed as a dead short too.  TF Flames instantly shut off.  I don't know why anyone insists it's OK to use unprotected cells, but people still do...

I'll give you $5 for your broken one.  Smile

--Bushytails

2100
Offline
Last seen: 2 years 7 months ago
Joined: 05/28/2011 - 10:28
Posts: 4512
Location: SINGAPORE

bushytails wrote:

Both of my skyray 1x14500 lights failed as a dead short too.  TF Flames instantly shut off.  I don't know why anyone insists it's OK to use unprotected cells, but people still do...

I'll give you $5 for your broken one.  Smile

--Bushytails

$5, no way.....I am getting comfortably above 2000 lumens OTF from the broken Sky Ray, form a good basis for comparison with say the TK70.  Laughing Probably will order the Trustfire as replacement since it has so many threads.  Sheesh why do I need 4 x 3xT6 for?  (actually supposed to spend on HIDs but plans foiled again and again)

mitro
Offline
Last seen: 1 year 8 months ago
Joined: 01/25/2011 - 11:58
Posts: 707
Location: Montgomery, IL USA

wiljen wrote:

 How about a taskLED.com H6flex or H6CC?   Both are capable of 6.6 or 6.7A delivered and would provide a driver circuit designed for series wired Xmls.    Granted at $30-40 you have almost as much in the driver as you do the light, but I'm betting you have a much harder time hurting one of those drivers.

I've always wanted to try one of their drivers but that a big investment for this thing. But from the looks of it (and I'm an idiot when it comes to electronics), they need a higher Vin than Vout and with two cells that's not going to work with the LEDs in series.
wiljen
Offline
Last seen: 1 year 6 months ago
Joined: 04/28/2011 - 16:04
Posts: 399
Location: East TN, SW VA

based on the tech info from the TaskLed site, I dont think that is a as big a problem as you might suspect.

 

From the site:  H6CC is capable of running from 8V up to 22V as shipped (guaranteed operating range). H6CC may operate as low as 5V, but operation between 5V and 8V is not guaranteed at the full 6.7A output current over the full operating temperature range (0C to 70C).  

 

mitro wrote:

I've always wanted to try one of their drivers but that a big investment for this thing. But from the looks of it (and I'm an idiot when it comes to electronics), they need a higher Vin than Vout and with two cells that's not going to work with the LEDs in series.

mitro
Offline
Last seen: 1 year 8 months ago
Joined: 01/25/2011 - 11:58
Posts: 707
Location: Montgomery, IL USA

But it still needs Vin higher than Vout, correct? Its a buck circuit and unless the 3 XM-Ls are run in parallel, its not going to fly.

bushytails
Offline
Last seen: 8 years 1 month ago
Joined: 06/11/2011 - 21:53
Posts: 167

With 6.7A, you'd want them in parallel - an xm-l doesn't do well at those currents.  Someone recently posted some very nice graphs here showing how poorly they do when you go over the rated power.

--Bushytails

2100
Offline
Last seen: 2 years 7 months ago
Joined: 05/28/2011 - 10:28
Posts: 4512
Location: SINGAPORE

bushytails wrote:

With 6.7A, you'd want them in parallel - an xm-l doesn't do well at those currents.  Someone recently posted some very nice graphs here showing how poorly they do when you go over the rated power.

--Bushytails

That'd would be match's excellent work....   For my "gut feel" last night on a real production light, there is some meat left from 2.8A range to say below 3.3A.  But it got hot real fast, already starting to get uncomfortable to hold by the 1 min mark so that's fast.  I suspect even a big thermal mass like the TK70 / Ebay HID aluminum torches it's just enough mass for a couple of mins run.  Unless if you are in sub-zero temps so that host to air heat interface would be high.

2100
Offline
Last seen: 2 years 7 months ago
Joined: 05/28/2011 - 10:28
Posts: 4512
Location: SINGAPORE

Ok some updates :

Direct drive with a fully charged SLA measured 2.2A to the emitters in series, i am getting 324 lux @ 8.17m. 21626 lux @ 1m, or candlepower.

Using slightly larger diameter wires on one of the them i could do 2.8A, giving me 360 lux @ 8.17m. 24029 lux @ 1m. This thing comfortably outthrows stock C8/UF-980L/U80.

Update, i pushed 3.3A and got 380 lux, that's the max. Head gets hot really fast. 25364 lux @ 1m.

 

Again pushed it by cooling the head with ice-cubes, got 451 lux at don't know what current but killed it quite fast within 60 sec. The LEDs were just glowing faintly. Gave it a hard knock and it came back on (prob internal wires made contact again). I think the emitter internal wires melted or something.  End of tests....  Smile   

The last one was a cool 30103 lux @ 1m.  Using engertolis test measurements in his light box, lumens of 1580L OTF @ ~ 260 lux, that means i was pushing in the 2700L OTF range.

 

Glow_Worm
Offline
Last seen: 10 hours 19 min ago
Joined: 04/02/2011 - 23:20
Posts: 22

Awesome.  Just awesome...

2100
Offline
Last seen: 2 years 7 months ago
Joined: 05/28/2011 - 10:28
Posts: 4512
Location: SINGAPORE

Glow_Worm wrote:

Awesome.  Just awesome...

As powerful as the 3xT6 is, it is still "not bright" comparatively.  When i gave it a hard knock to make it come back , i was looking at the emitters which were glowing faintly. and figured it could come back to life with a tough man's repair method. Wah la.....dumb idea eh?  Sealed

The 3 emitters + reflectors after-image was burnt into my right eye's vision, but it went off quite quickly....like within 15 minutes.

The last time i was looking at the faint bluish and reddish glow of the arc chamber of a 75W HID bulb overdriven to 100W, i accidentally hit the switch and it hot-restriked. That "impact" was BIG enough to render me somewhat nauseous and nearly caused me to lose balance, yep its that bad. That one took like 24hrs to subside to about 20%  and my eyes went puffy, one was especially puffy.  I am on transition lens, so that took care of ~ 405nm plus a big part of UV-A (my transition lens attenuates a big portion of my 405nm laser defocused).  Eye drops took care of it, a trip to the optemetris gave me a clean bill.  With a laser you get a very small blind spot as long as you don't hit the optic nerve which ironically itself is already a blind spot (hit it and there goes your whole vision).....but this one left a huge burn after-image for like nearly 2 days before it really went 99% away.   The lucky thing is, i probably got hit by the corona + at such close range the beam was not fully converged yet.

Pages