What is smarter choice. Number of 7135s

BlueSwordM · 2018-11-18T03:06:53.000Z

Well, you can’t really driver any harder, or any less a 7135. You can just add less/more CC regulators until the light/driver can’t heatsink the chips anymore.

What you want is many 7135 regulators as possible for the light’s size while putting the LED at its most efficient/power ratio is.

For a single LED light, a 1x7135+3x7135+FET arrangement would be your best bet, as it would give you nice regulated runtime at around 300-400 lumens depending on the emitter, being perfectly sustainable even on a light like the H03/Convoy S2.

For a triple/quad, the 1x7135+6x7135+FET would be good, as the current is spread across the LED array, and is more efficient than a single LED, so doubling the power to about 950-1050 lumens would give you decently sustainable runtime as long as it has thermal management.

For a biiig light like the BLF Q8, I would go with 1x7135+16x7135+FET if possible. Now that would be nice

chadvone · 2018-11-18T03:59:19.000Z

thanks blueswordM. The Skilhunt has 1 led. Can you dumb it down a bit for me.

BlueSwordM · 2018-11-18T04:02:09.000Z

Just ask for 1x7135 + 3x7135 + FET driver from Lexel then.

Would be perfect

DavidEF · 2018-11-18T04:03:46.000Z

BlueSwordM:

Well, you can’t really driver any harder, or any less a 7135. You can just add less/more CC regulators until the light/driver can’t heatsink the chips anymore.

What you want is many 7135 regulators as possible for the light’s size while putting the LED at its most efficient/power ratio is.

For a single LED light, a 1x7135+3x7135+FET arrangement would be your best bet, as it would give you nice regulated runtime at around 300-400 lumens depending on the emitter, being perfectly sustainable even on a light like the H03/Convoy S2.

For a triple/quad, the 1x7135+6x7135+FET would be good, as the current is spread across the LED array, and is more efficient than a single LED, so doubling the power to about 950-1050 lumens would give you decently sustainable runtime as long as it has thermal management.

For a biiig light like the BLF Q8, I would go with 1x7135+16x7135+FET if possible. Now that would be nice

No need to overthink it. He already said:

chadvone:

…I have been thinking what would be the most efficient set up for um, lets say 500 lumens.

So, just go to CREE PCT and pick the emitter you would use, and find where you get ~500lm output. Then put the number of 7135’s on your driver needed to reach that.

Example:

The CREE XP-L HI V3 reaches (just over) 500lm at 1.15A according to the PCT.
The AMC7135 comes in 350mA and 380mA increments.
So, using the 350mA variety, you could choose to use three chips, for 1.05A, and have ~476lm or four chips, for 1.4A and ~608lm.
But, if you go with the 380mA chip, you could hit closer to your goal, 1.14A with three chips, giving you just over 500lm output.

You ultimately want to be as close as possible because of efficiency. Too many chips, and you’ll be using PWM heavily, losing efficiency of the emitter.

chadvone · 2018-11-18T04:07:00.000Z

BlueSwordM:

Just ask for 1x7135 + 3x7135 + FET driver from Lexel then.

Would be perfect

LOL, thanks.

So your saying 3 7135s will give longer run time at 400-500 lumens then 6 7135s at 400-500 lumens. My next step would be all she’s got captain.

chadvone · 2018-11-18T04:11:39.000Z

BlueSwordM and DavidEF. Thank you both for the replies, I think I have it now.

Lightbringer · 2018-11-18T05:33:20.000Z

DavidEF:

No need to overthink it. He already said:

chadvone:

…I have been thinking what would be the most efficient set up for um, lets say 500 lumens.

So, just go to CREE PCT and pick the emitter you would use, and find where you get ~500lm output. Then put the number of 7135’s on your driver needed to reach that.

[wak wak wak]

Even that’s overthinking it. 500 lemons could be ±100 and still not be visually perceptible. Right off, about 1A would be right smack in the middle of that range, so 3×7135, either the 350mA or 380mA variety, would give about that.

Use 350s or 380s, and you’d be hard-pressed to tell the difference, even side-by-side. Probably even 3 vs 4 chips.

So, I’d stick with 3×7135 and be happy. :laughing:

chadvone · 2018-11-18T15:31:53.000Z

how much then???

BlueSwordM · 2018-11-18T15:46:27.000Z

3x7135 then.

Pete7874 · 2018-11-18T15:56:03.000Z

My S2+ 4x7135 (1.4A) is rated at around 500 lm.

chadvone · 2018-11-18T15:59:09.000Z

now wondering how much more efficient 3x7135 at 400 lumens compared to 6x7135 at 400 lumens.

noticeable run time improvement ?

I will take you guys word for it.

Quadrupel · 2018-11-18T16:12:21.000Z

I’m using Lexers driver with 5x7135 so it’s about 1.7amp and no FET. H03 stock driver runs maximum at 1amp and 2amps on turbo for few minutes. 350ma is more then enough for every day use. But without FET NarsilM and Anduril ramping behaves strangely. Looks like after 350ma there is an empty space or pause for about 1sec. But I like it. I programmed Anduril to ramp up only to 350ma and if i need more, always can double click and get 1,7A on turbo. Here is a stock driver power usage :

Т1 - 2,05 А
T2 - 1,12 A
Н1 - 0,78 А
Н2 - 0,28 А
М1 - 0,12 А
М2 - 0,04 А
L1 - 7,5 мА
L2 - 3,9 мА

EasyB · 2018-11-18T16:29:09.000Z

chadvone:

now wondering how much more efficient 3x7135 at 400 lumens compared to 6x7135 at 400 lumens.

noticeable run time improvement ?

I will take you guys word for it.

It will not be a large difference in efficiency with those two situations. Based on some measurements that I and Texas_ace did (separately) regarding this I would estimate 5-7% difference in efficiency.

BlueSwordM · 2018-11-18T16:31:32.000Z

3x7135 at 400 lumens will be more efficient than 6x7135 at 400 lumens by about 20%, not counting parasitic losses by PWMing the 7135s.

EasyB · 2018-11-18T17:02:34.000Z

EasyB:

chadvone:

now wondering how much more efficient 3x7135 at 400 lumens compared to 6x7135 at 400 lumens.

noticeable run time improvement ?

I will take you guys word for it.

It will not be a large difference in efficiency with those two situations. Based on some measurements that I and Texas_ace did (separately) regarding this I would estimate 5-7% difference in efficiency.

My measurements are at the bottom of my post here. I compared the output of a light using two dimming methods, PWMing a FET and adjustable linear constant current. The difference in efficiency is not as high as one might expect from looking at LED test curves. I.e. if you PWM a FET driven LED that gets 6A with direct drive to 3A average current, you get more than half the output at 6A. Texas_ace did a similar measurement in his test of the XHP35.

The reasoning here is that LED efficiency suffers from two main effects at high current. One is the increased temperature at high current and the other is the high current itself. When you PWM the current to the LED you get reduced efficiency due to the higher pulsed current, but the temperature won’t be as high as it is at 6A so you don’t get a big hit to the efficiency.

The exact numbers will depend on the LED and how hard it’s driven at the peak currents. 6x7135 is only 2.3A which is not high to begin with, so that’s why I would estimate the efficiency loss would not be high compared to 3x7135.

chadvone · 2018-11-18T23:04:21.000Z

Thanks everyone. Looks like 3 is what I am gonna order for H03.

chadvone · 2019-01-21T04:54:57.000Z

UPDATE. I ended up with a couple Lexel TA 1-6+FET drivers. After having one in a Wowtac A2 for some time. I tore it down for a led swap. While it was down I removed 3-7135 chips.

I tweaked the UI to allow for 3-7135s and 4-7135s and put in a SST-20 4000k 95cri.

With 1 chip I get 150 lemons.
With 3 chips I get 390 lemons.
With 4 chips I get 450 lemons.
With max FET I get 1000+ lemons.

Have been running the light a few minutes at 450 lemons. I think it will be able to sustain that output just fine.

I don’t know if I have 350 or 380 chips.
Need to take some current readings.
Anyone want to guess the run times of the above set up?

I am hoping for 3 hours at 390 lemons.

Barkuti · 2019-01-21T05:08:39.000Z

Suffice to say, in my honest opinion the ideal number of 7135 chips is…

Source: https://avaloncakesschool.com/shop/digital-tutorials/donut-cake/

O:)

Barkuti · 2019-01-21T06:07:40.000Z

chadvone, are you 0K? I don't think so.

It's just a pretty innocent joke, and you overreacted. There are thousands far more senseless posts in this forum. If you're high, inebriated and/or sick I hope for you to get better soon. I won't come back to brawl.

There are linear drivers which let you tune the desired current, by the way.

^:)

e1000 · 2019-01-21T06:19:38.000Z

that’s a lot of lemons!