HDS U60GTXR Upgraded to SSC U Bin P4

[acusifu]

Hi all without posting any beamshots. If you have a HDS that performs well good output and tint, etc. Don't hesitate to upgrade to a SSC P4.
I just got my U60GTXR back from Milky and WOW. About twice as bright, the tint is in my opinion better than the GT that was replaced. What a major improvement I recommend Milky's modding skills and the upgrade itself. Don't wait forever for the new HDS with increased lumens, you can have this now. I think Milky has a runtime graph he can add to this post. Let's see how the post modded light performs. Thanks

 

[milkyspit]

[size=+1]Acusifu HDS U60XRGT SEOULmated Initial Runtime on Max Output[/size]

Here's a runtime graph I did of Acusifu's HDS SEOULmation mod with the assistance of Luxlover. Essential equipment included a Fluke 189 data-logging DMM, Meterman LM631 lumen-normalized in conjunction with the recent CPF light meter calibration passaround, Luxlover's 'shampoo cap' apparatus, and a wiring harness I fashioned to pull outputs from the Meterman into the Fluke DMM. I'm no master of Excel graphing, so apologies if this is somewhat less than readable! oo:

Some offhand notes in no particular order...

Our original goal in measurement was to check the quality of emitter heatsinking for Luxlover's 'defective' and subsequently SEOULmated HDS U60... I put 'defective' in quotes because this particular U60 was allegedly within factory spec, and yet drawing an unusually high amount of current from the 123 cell... most likely there's some funny business going on with the power supply in this unit, possibly a substandard inductor... part of the graphing had to do with ensuring the emitter itself was performing as well as it could be, without output degradation due to heat pooling in the emitter slug itself. That's why these runtimes don't seek to capture the entire runtime of the unit, but rather the first 50 readings taken at 10 second intervals.

Acusifu's HDS U60XRGT happened to be here during these tests so I took advantage of the setup and measured his output as well. The results speak for themselves... at max output, Acusifu's light is capable of cruising along, pumping out a continuous 123 lumens or so, for the first 8.5+ minutes... it's interesting that at the end of the test interval output had actually INCREASED slightly over that first reading. Not sure how to explain that, but it's not something I'd consider a problem!

Luxlover's 'defective' unit seems to benefit from its high current draw in generating even higher outputs than Acusifu's for the first several minutes... roughly 140 lumens in this test... and the flatness of the output at that level suggests the emitter slug isn't heat-saturated, either, so the emitter heatsinking seems to be doing its job... the output falloff around the 6.5 minute mark indicates the HDS thermal limiting circuitry kicking-in, throttling back the output. My guess is the thermal limiting algorithm is calibrated to estimate die temperature for a Luxeon III emitter based on the temperature seen on the heatsink, taking into account the various thermal gradients... but the Seoul emitter wicks heat away from the die better than a Luxeon III, so the algorithm is triggering based on a 'false positive' thinking the die is overheating when it's really not. Unfortunately there seems no way to recalibrate the algorithm short of enlisting Henry to reprogram the units. Still, most folks would probably consider 140 lumens for 6.5 minutes followed by a steady 90 lumens or so to be a good upgrade from the constant 60 lumens of stock!

The stock U60 pretty much performed as expected. Note how its output gradually ticks down during the test interval, possibly suggesting the emitter slug is heat-saturated, though just barely.

I'll leave conclusions as an exercise for the reader!

 

[acusifu]

My conclusion is this light rocks now. Thanks Milky

 

[ggaarryy]

Just to add my comments, Milky recently did the mod for me on a u60 and the results are outstanding. The light although good before the mod is now more useable at all levels.

Thanks Milky.

 

[tebore]

It's not a false positive on the temp based step down. If it does draw a higher current the LED can could be heating faster. Or the power supply dumping more waste heat.

Remember Henry calibrated the light to step down to protect the LED and the user. I think it's calibrated to the range of ~60 degrees measured on the exterior of the light. If it's hitting 60 there's a reason for it besides the LED.

 

[milkyspit]

It's not a false positive on the temp based step down. If it does draw a higher current the LED can could be heating faster. Or the power supply dumping more waste heat.

Remember Henry calibrated the light to step down to protect the LED and the user. I think it's calibrated to the range of ~60 degrees measured on the exterior of the light. If it's hitting 60 there's a reason for it besides the LED.

I'm certainly not qualified to second-guess Henry or much of anyone else on this! In that particular light it would seem to be waste heat from the circuit, as the empirical evidence suggests the circuit isn't running as efficiently as it could be.

My whole point on the temperature thing was that we can't actually measure the temperature directly on the die itself for obvious reasons, so one would have to rely on the thermal gradient as heat flows from die to slug, to heatsink, to temperature sensor. If the temperature at the sensor were known and thermal resistance (conductivity?) at each stage from die to sensor were also known, one could calibrate the sensor to act at xxx degrees on the sensor itself because that would theoretically correspond to yyy degrees on the die... but a Seoul emitter will tend to pull heat off the die more effectively than a LuxIII, which invalidates the above calculation. If the calculation is no longer valid, then neither is the point at which the sensor triggers action.

I can tell you the exterior of the light was nowhere near 60 degrees (assuming Celsius) or Luxlover and I would both have burned our hands! Whatever was happening, happened as a result of something calibrated into the light (it was very sudden, and deterministic, as in repeatable at the same point in runtime again and again, and the cell wasn't anywhere near spent) and seemed to be something temperature related. It definitely wasn't the standard ratcheting-down a level or two that happens with a weak cell.
:shrug:

 

[tebore]

I'm certainly not qualified to second-guess Henry or much of anyone else on this! In that particular light it would seem to be waste heat from the circuit, as the empirical evidence suggests the circuit isn't running as efficiently as it could be.

My whole point on the temperature thing was that we can't actually measure the temperature directly on the die itself for obvious reasons, so one would have to rely on the thermal gradient as heat flows from die to slug, to heatsink, to temperature sensor. If the temperature at the sensor were known and thermal resistance (conductivity?) at each stage from die to sensor were also known, one could calibrate the sensor to act at xxx degrees on the sensor itself because that would theoretically correspond to yyy degrees on the die... but a Seoul emitter will tend to pull heat off the die more effectively than a LuxIII, which invalidates the above calculation. If the calculation is no longer valid, then neither is the point at which the sensor triggers action.

I can tell you the exterior of the light was nowhere near 60 degrees (assuming Celsius) or Luxlover and I would both have burned our hands! Whatever was happening, happened as a result of something calibrated into the light (it was very sudden, and deterministic, as in repeatable at the same point in runtime again and again, and the cell wasn't anywhere near spent) and seemed to be something temperature related. It definitely wasn't the standard ratcheting-down a level or two that happens with a weak cell.
:shrug:

How sure are you it's not power related because you'll see CLEARLY 2 blinks. If it's temp it's 1 blink. They are clear and definitive. It might not be thermal step down just for some reason this light is using more power. Perhaps this new LED is a higher vf. I know the Lux's vf are quite deceptive in the sense they rate them much higher than it is. The Seouls are pretty dead on.

If it is thermal:
Lets say that sensor is calibrated to X temp to step down. You're saying because this X temp is only for the module/ area around the sensor and the die temp is lower or not as hot as the Lux die, that it should be calibrated because of the new LED. It doesn't matter much what the die temp is because for some reason the sensor is still hitting X temp and thus stepping down. It could be that the module is heating up more because of the higher vf. What ever temp it's hitting it's doing it to protect the user now instead of just the LED. Do you want a light that can do 60 degrees (celsius) even with the die ONLY at 60? (assuming perfect heat conduction, impossible).

Or proper contact between the module and the body is not met because again Henry didn't just calibrate the sensor for the LED it's also for the user. If it's cool on the outside and still stepping down it might be a sign the module is not connecting to the body. It's VERY important the module touches the lip of the body. In fact in my mods, to promote heat transfer AS5 was applied to the lip (very thin) and the upper portion of the threads of the body.

My B42 can run somewhere between 20-30 mins without step down. If it's held it won't step down. Apple to oranges I know B42 - U60.

 

[milkyspit]

You know, I feel bad about getting into this discussion in this particular thread. It just occurred to me we're discussing LUXLOVER'S light and what may or may not be happening to it, whereas THIS thread is about ACUSIFU'S successful mod! And in fairness, Luxlover is actually quite happy with his light as well.

Sorry Acusifu! Guess I started us down this road.

Tebore, it's not that I don't consider the discussion valuable... actually it's quite interesting to me, and IMHO completely worthwhile... just that it probably shouldn't happen in THIS thread. Again, I'll take the blame for causing things to head in that direction.

Anyway, maybe getting things back on track... Tebore, think it's worth starting a new thread where we could get into the thermal stuff, maybe over in the HDS area?

Acusifu, very glad you're diggin' your new light! And if you ever decide it's expendable, may I have first right of refusal on it?

 

[acusifu]

The light is a keeper. I don't mind the discussion in this thread, the more the merrier.

 

[tebore]

I appologize as well, it's just that I love discussions. Well we have the permission to continue. We can continue here or where you want to.

I'd test my light to get some more numbers but my light is outta commission do to my clumsy butt falling on ice.

I'm glad you're enjoying your light acusifu and that it's over 120lm.

 

[Art Vandelay]

I can't decide which one I would pick if I had the choice. It's a fairly small difference to the eyes. However, I have been known to compare lights side by side to see which one is the brightest. Sweet mods!

 

[Led_Blind]

Milky,
Thanks for the graph mate, i did the mod myself and was absolutly certain of 120lm+ but had no way to test this. You however just proved me correct... yay!

My totaly unscientiffic method was the roof bounce test. For comparison i used a HD45 V's 2x U60's. The HD45 won this eaisly. After the swap the test was between the HD45 and a single Seulmated U60XRGT. This time the HD45 only just beat the HDS in output.....

Its one hell of an upgrade!