Osram Oslon Boost HX

204 posts / 0 new
Last post
Jerommel
Jerommel's picture
Offline
Last seen: 5 days 7 hours ago
Joined: 01/04/2014 - 13:18
Posts: 6376
Location: the Hague, Netherlands

luminarium iaculator wrote:
Why?
It has perfect sense for me. You don’t have to have flashlight pressed on turbo all the time Smile
For example:
Turbo mode 5A, Middle mode 2,5A, low is 1A. You can run it on middle or low mode(saving the battery) and turn turbo for a minute or two when needed.
A minute or two is probably what it takes to discharge to 4.0 Volts, and so your current will drop.
Jerommel
Jerommel's picture
Offline
Last seen: 5 days 7 hours ago
Joined: 01/04/2014 - 13:18
Posts: 6376
Location: the Hague, Netherlands

Jerommel wrote:
luminarium iaculator wrote:
Why?
It has perfect sense for me. You don’t have to have flashlight pressed on turbo all the time Smile
For example:
Turbo mode 5A, Middle mode 2,5A, low is 1A. You can run it on middle or low mode(saving the battery) and turn turbo for a minute or two when needed.
A minute or two is probably what it takes to discharge to 4.0 Volts, and so your current will drop.

It may be better to start with a battery that is already discharged to 3.8 Volts, because that’s where the voltage starts to stay the same for a longer time, just slowly dropping.

But hey, do as you see fit.

luminarium iaculator
luminarium iaculator's picture
Offline
Last seen: 1 day 2 hours ago
Joined: 07/07/2013 - 09:09
Posts: 2332
Location: X
luminarium iaculator wrote:
Since we know that mentioned Osram emitters loose very little on performance and we know how fet driver behaves we could approximately calculate that 4.2-3.3V of battery cell = 5-3.5A current draw on turbo mode which is approximately 30 minutes of turbo mode on 1×18650 Samsung INR 30Q cell by looking HKJ minutes/runtime diagram for Samsung INR30Q cell. Even if it is not in regulation it will still run strongly and one that uses it will not visually notice significant candela/lux intensity change.

I done hundreds of diy FET drivers and some Texas Ace one so I know how they behave. I am looking for permanent solution of restricting FET driver current, and since Djozz mentioned that some kind of resistor can be added to driver circuit I would really like to know what type of resistor and where should I place it in a circuit?

If someone wants to run Osram WF with FET driver just use low current 18650 cell and don’t bypass springs. But however that is not proper and safe solution. Proper solution for FET drivers would be current restriction on FET driver. It should be doable but I don’t know how.

djozz
djozz's picture
Offline
Last seen: 16 min 13 sec ago
Joined: 09/07/2012 - 17:04
Posts: 16549
Location: Amsterdam

I’m sorry l.i. but although I produced the AK-47 hack back in the days, which was purely an extension of comfychair’s work, I’m not very knowledgable in drivers, not like Texas_Ace or Lexel or DEL or wight.

The idea of putting an extra resistor in this hacked driver is not a really practical one because it would have to be a big one that can handle 5A current.

Maybe if you can guarantee buying a large amount of them, you could ask Simon from Convoy Flashlight to produce a batch of his 5A drivers with high-med-low without memory?

luminarium iaculator
luminarium iaculator's picture
Offline
Last seen: 1 day 2 hours ago
Joined: 07/07/2013 - 09:09
Posts: 2332
Location: X

Yes I could. But what should I do with hundreds of AK47C1 drivers(also ordered from Simon) ? LOL I would try to place big resistor on driver cause I pot drivers anyway. Can you suggest me proper resistor for this potential current restriction FET driver mod?

EasyB
Online
Last seen: 5 min 46 sec ago
Joined: 03/09/2016 - 15:24
Posts: 1817
Location: Ohio
luminarium iaculator wrote:
Yes I could. But what should I do with hundreds of AK47C1 drivers(also ordered from Simon) ? LOL I would try to place big resistor on driver cause I pot drivers anyway. Can you suggest me proper resistor for this potential current restriction FET driver mod?

The added resistance is going to be in the 0.05 to 0.1 ohm range I think.

http://budgetlightforum.com/comment/1458705#comment-1458705
http://budgetlightforum.com/comment/1443460#comment-1443460
These are two resistor mods I did for the 2mm WF. Added resistance was about 0.1ohm. For the 1mm WF it should be a bit less.

Agro
Agro's picture
Offline
Last seen: 3 hours 15 min ago
Joined: 05/14/2017 - 11:16
Posts: 5579
Location: Ślōnsk

I wonder…would it be possible to adjust PWM as battery discharges to maintain something closer to constant current?

EasyB
Online
Last seen: 5 min 46 sec ago
Joined: 03/09/2016 - 15:24
Posts: 1817
Location: Ohio
Agro wrote:
I wonder…would it be possible to adjust PWM as battery discharges to maintain something closer to constant current?

I think it’s possible. Some discussion here.
http://budgetlightforum.com/node/49230

luminarium iaculator
luminarium iaculator's picture
Offline
Last seen: 1 day 2 hours ago
Joined: 07/07/2013 - 09:09
Posts: 2332
Location: X
EasyB wrote:
luminarium iaculator wrote:
Yes I could. But what should I do with hundreds of AK47C1 drivers(also ordered from Simon) ? LOL I would try to place big resistor on driver cause I pot drivers anyway. Can you suggest me proper resistor for this potential current restriction FET driver mod?

The added resistance is going to be in the 0.05 to 0.1 ohm range I think.

http://budgetlightforum.com/comment/1458705#comment-1458705
http://budgetlightforum.com/comment/1443460#comment-1443460
These are two resistor mods I did for the 2mm WF. Added resistance was about 0.1ohm. For the 1mm WF it should be a bit less.

Finally someone with a will for FET solutions Beer

Is there any way to put some kind of resistor on a top of a driver and where would you put it? I pot drivers so it would be nice if I could merge resistor with a driver for permanent bond Thumbs Up

EasyB
Online
Last seen: 5 min 46 sec ago
Joined: 03/09/2016 - 15:24
Posts: 1817
Location: Ohio
luminarium iaculator wrote:
EasyB wrote:
luminarium iaculator wrote:
Yes I could. But what should I do with hundreds of AK47C1 drivers(also ordered from Simon) ? LOL I would try to place big resistor on driver cause I pot drivers anyway. Can you suggest me proper resistor for this potential current restriction FET driver mod?

The added resistance is going to be in the 0.05 to 0.1 ohm range I think.

http://budgetlightforum.com/comment/1458705#comment-1458705
http://budgetlightforum.com/comment/1443460#comment-1443460
These are two resistor mods I did for the 2mm WF. Added resistance was about 0.1ohm. For the 1mm WF it should be a bit less.

Finally someone with a will for FET solutions Beer

Is there any way to put some kind of resistor on a top of a driver and where would you put it? I pot drivers so it would be nice if I could merge resistor with a driver for permanent bond Thumbs Up


You would put the resistor anywhere in the main current path. The challenge is the heat dissipation. It will be producing between 2 and 5W depending on what current you choose (5 to 7A), and that is a lot for a small resistor so it will get very hot.
The_Fat_Controller
The_Fat_Controller's picture
Offline
Last seen: 5 days 6 hours ago
Joined: 09/13/2019 - 07:34
Posts: 331
Location: England

has anywhere got stock of these yet? Also does convoy do a 4040 gasket or are we stuck with reaming out the 3030 ?

luminarium iaculator
luminarium iaculator's picture
Offline
Last seen: 1 day 2 hours ago
Joined: 07/07/2013 - 09:09
Posts: 2332
Location: X

EasyB,
Can it fit to 17mm diameter? Can I cut it to fit if needed? I have aluminum and copper plates as potential heat sink for resistor and when I merge everything with potting compound it will look like it is made from one piece and whole driver should act as a heatsink.

But if it generates serious heat at 5A (more than 60 celzius) that could be problem…

EasyB
Online
Last seen: 5 min 46 sec ago
Joined: 03/09/2016 - 15:24
Posts: 1817
Location: Ohio

luminarium iaculator wrote:
EasyB,
Can it fit to 17mm diameter? Can I cut it to fit if needed? I have aluminum and copper plates as potential heat sink for resistor and when I merge everything with potting compound it will look like it is made from one piece and whole driver should act as a heatsink.

But if it generates serious heat at 5A (more than 60 celzius) that could be problem…


The extra heat is not large compared to the total heat from the LED. The risk is that if the resistor you choose is too small or not heat sinked enough it will burn up. You will have to look for resistors that will fit in your lights.
luminarium iaculator
luminarium iaculator's picture
Offline
Last seen: 1 day 2 hours ago
Joined: 07/07/2013 - 09:09
Posts: 2332
Location: X

Can you link me some kind of big but no to big resistor that could be bent into 17mm diameter please. I really don’t know what to order? Pretty please… Innocent

Jerommel
Jerommel's picture
Offline
Last seen: 5 days 7 hours ago
Joined: 01/04/2014 - 13:18
Posts: 6376
Location: the Hague, Netherlands

The_Fat_Controller wrote:
has anywhere got stock of these yet? Also does convoy do a 4040 gasket or are we stuck with reaming out the 3030 ?

Use a square file, only cuts out the necessary corners.
chemnitz
Offline
Last seen: 17 hours 31 min ago
Joined: 01/02/2020 - 17:24
Posts: 5
Location: Detroit MI

Anyone have any leads on this chip yet? Seems impossible to find, but some people have gotten their hands on it.

JaredM
JaredM's picture
Online
Last seen: 3 min 2 sec ago
Joined: 10/31/2011 - 13:33
Posts: 768
Location: Pittsburgh, Pennsylvania

I haven’t heard anything about it, but apparently they have another boost family emitter 0.5mm² named HM I think.

The_Fat_Controller
The_Fat_Controller's picture
Offline
Last seen: 5 days 6 hours ago
Joined: 09/13/2019 - 07:34
Posts: 331
Location: England
Jerommel
Jerommel's picture
Offline
Last seen: 5 days 7 hours ago
Joined: 01/04/2014 - 13:18
Posts: 6376
Location: the Hague, Netherlands

Okay, so the ‘boost HX’ is basically the 2mm² white flat with a 4040 footprint.
So the 1mm² white flat with 4040 footprint will probably beat it in throw (like with the 3030 sizes)?

JaredM
JaredM's picture
Online
Last seen: 3 min 2 sec ago
Joined: 10/31/2011 - 13:33
Posts: 768
Location: Pittsburgh, Pennsylvania

@Jerommel
Yes. Boost line seems to take a given chip and move it up one package size/class. This probably didn’t mean much for the 1mm² chip but should for the 2mm² . Where 1mm² saw a ~10% “boost” in current handling, I would expect 15-20% for the 2mm² . This might close the gap entirely between these two. The choice between would then be based on your host/driver/runtime restrictions.

@Fat_Controller

Interestingly the If vs Vf curve maxes out at 8A (despite 6A rating)! That should mean we have a bit of overhead there I would imagine…

EDIT: Looking at the datasheet for the 1mm² (Boost HL) it appears they did the same thing. If vs Vf goes up to 4A and max rating is 3A. If overhead is proportional then the HX should be good for 12A. We’ll still just have to see if the HX ever shows up.

PS: Interesting to note as well, the HL product page notes it’s “not for new designs”. I wonder what why? Is there a successor already?

contactcr
Offline
Last seen: 26 min 55 sec ago
Joined: 05/19/2017 - 18:52
Posts: 2921
Location: US

I guess vinh/skylumen found some so we will know soon enough I’m sure

Agro
Agro's picture
Offline
Last seen: 3 hours 15 min ago
Joined: 05/14/2017 - 11:16
Posts: 5579
Location: Ślōnsk

Yes, Vinh and L4P have some already. .5 mm² variant is exciting news Smile

SKV89
Offline
Last seen: 33 min 52 sec ago
Joined: 12/10/2017 - 12:46
Posts: 3575
Location: US

I could not find the Boost HX on L4P’s website. Vinh definitely has it though. He just posted a pic of it.

JaredM
JaredM's picture
Online
Last seen: 3 min 2 sec ago
Joined: 10/31/2011 - 13:33
Posts: 768
Location: Pittsburgh, Pennsylvania
Agro wrote:
Yes, Vinh and L4P have some already. .5 mm² variant is exciting news Smile

Here’s the data sheet for the Boost HM (0.5mm²)

KW_CELMM1.TG

https://www.mouser.co.uk/pdfDocs/KWCELMM1TG_EN.pdf

JaredM
JaredM's picture
Online
Last seen: 3 min 2 sec ago
Joined: 10/31/2011 - 13:33
Posts: 768
Location: Pittsburgh, Pennsylvania
SKV89 wrote:
I could not find the Boost HX on L4P’s website. Vinh definitely has it though. He just posted a pic of it. !{width:50%}https://i.imgur.com/sh4EEqx.jpg!

Well hopefully he provides some test data or can lend a fellow member with the equipment one soon. Very curious if the 4040 Boost footprint allows the 2mm² to achieve the ~5.5A/mm² like its smaller brethren. Or at least close the gap.

Speaking of which, I really wonder how much the rectangular geometry is a functional design aspect allowing the larger die to reach such high current density. Please excuse my lack of knowledge here but is there a flow of current along the surface plane of the chip? Does anyone here have expertise in this area?

Jerommel
Jerommel's picture
Offline
Last seen: 5 days 7 hours ago
Joined: 01/04/2014 - 13:18
Posts: 6376
Location: the Hague, Netherlands

No idea really, but if it was all the same, i think we would have many round die power LEDs on the market.

Jerommel
Jerommel's picture
Offline
Last seen: 5 days 7 hours ago
Joined: 01/04/2014 - 13:18
Posts: 6376
Location: the Hague, Netherlands

JaredM wrote:
@Jerommel
Yes. Boost line seems to take a given chip and move it up one package size/class. This probably didn’t mean much for the 1mm² chip but should for the 2mm² . Where 1mm² saw a ~10% “boost” in current handling, I would expect 15-20% for the 2mm² . This might close the gap entirely between these two. The choice between would then be based on your host/driver/runtime restrictions.
Oh, i thought it was more than 10 % more current for the 1mm², but it makes sense that it will do more for the 2mm²
JaredM
JaredM's picture
Online
Last seen: 3 min 2 sec ago
Joined: 10/31/2011 - 13:33
Posts: 768
Location: Pittsburgh, Pennsylvania
Jerommel wrote:
No idea really, but if it was all the same, i think we would have many round die power LEDs on the market.

Well circles, in general, create processing challenges in all kinds of industries. Radii are just harder to deal with than perpendicular lines (again, generally).

I’m coming at this question/idea from a uneven current distribution caused by series resistances. This is the only reason I can intuit – other than tolerance stack up on defects – that there is a decrease in peak current density as dies get larger.

I know older generation CREE leds used to use grid style current spreaders, then they turned into arrays of dots (vias?). I’m assuming those aren’t used anymore with flip chip designs, either that or they just aren’t visible anymore? I guess it’s time I learn me up on some ol’ semiconductor anatomy.

Agro
Agro's picture
Offline
Last seen: 3 hours 15 min ago
Joined: 05/14/2017 - 11:16
Posts: 5579
Location: Ślōnsk

SKV89 wrote:
I could not find the Boost HX on L4P’s website.

He doesn’t sell them yet, but he said he had samples. And offered some as a prize in the current OL Contest.
Enderman
Enderman's picture
Offline
Last seen: 7 hours 4 min ago
Joined: 11/03/2016 - 22:42
Posts: 4021
Location: Vancouver, Canada

JaredM wrote:
I haven’t heard anything about it, but apparently they have another boost family emitter 0.5mm² named HM I think.

Agro wrote:
Yes, Vinh and L4P have some already. .5 mm² variant is exciting news Smile

The thermal pad is far smaller than any 3030 or 4040 LED.
Its overcurrent performance will be absolutely crap.
There is a reason we are using CULNM1 / CSLNM1 in flashlights and not the CELNM1.
U = 4040
S = 3030
E = 1915 (1.9mm x 1.5mm)
And that’s just the package size, the thermal pad is only a fraction of that.

Jerommel wrote:
Okay, so the ‘boost HX’ is basically the 2mm² white flat with a 4040 footprint.
So the 1mm² white flat with 4040 footprint will probably beat it in throw (like with the 3030 sizes)?

Yeah the 3030 1mm^2 white flat already beats it in throw, and the 4040 1mm^2 white flat beats the 3030 one.
My CSLNM1 reached 335cd/mm2 at about ~5.5A while the CULNM1 peaked at 349cd/mm2 at ~6A
This was using an LX1330b which overestimates the cd for cool white LEDs, but the point is the relative difference, the 4040 one is best.

Pages