LED drivers and Accessories you want, but don’t exist

488 posts / 0 new
Last post
Lexel
Lexel's picture
Offline
Last seen: 1 day 11 hours ago
Joined: 11/01/2016 - 08:00
Posts: 5533
Location: Germany

I found a nice calculation for MOSFET losses

Agro
Agro's picture
Online
Last seen: 2 min 43 sec ago
Joined: 05/14/2017 - 11:16
Posts: 5036
Location: Ślōnsk

We’re seeing more and more low-Vf 3V LEDs out there. Luxeon MZ, quad E21, Q8WP to name a few. These lose a lot of efficiency when driven linearly and could benefit from some good buck driver for 1s out and either 1-2s in. Or simply 1s in, that would be enough.

Enderman
Enderman's picture
Offline
Last seen: 1 hour 5 min ago
Joined: 11/03/2016 - 22:42
Posts: 3966
Location: Vancouver, Canada

Agro wrote:
We’re seeing more and more low-Vf 3V LEDs out there. Luxeon MZ, quad E21, Q8WP to name a few. These lose a lot of efficiency when driven linearly and could benefit from some good buck driver for 1s out and either 1-2s in. Or simply 1s in, that would be enough.

So literally just a 7135 driver?
The_Driver
The_Driver's picture
Offline
Last seen: 16 hours 52 min ago
Joined: 10/20/2016 - 05:51
Posts: 1428
Location: Germany
Enderman wrote:
Agro wrote:
We’re seeing more and more low-Vf 3V LEDs out there. Luxeon MZ, quad E21, Q8WP to name a few. These lose a lot of efficiency when driven linearly and could benefit from some good buck driver for 1s out and either 1-2s in. Or simply 1s in, that would be enough.
So literally just a 7135 driver?

The AMC7135 is a linear regulator.
Agro wants a real 1S buck regulator (with an inductor) like many production flashlights have. This improves efficiency (runtime) when the voltage difference between the LED and the battery gets larger.

Agro
Agro's picture
Online
Last seen: 2 min 43 sec ago
Joined: 05/14/2017 - 11:16
Posts: 5036
Location: Ślōnsk

Yes, that’s precisely what I want.
Q8WP driven to 1.25 A does ~124 lm/W with Vf of 3V.
Drive it linearly and you get about 96 lm/W throughout battery discharge.
Buck it with 95% efficiency and end-up with 118 lm/W.

Lexel
Lexel's picture
Offline
Last seen: 1 day 11 hours ago
Joined: 11/01/2016 - 08:00
Posts: 5533
Location: Germany

next monday placing an order in my board fab

Lexel
Lexel's picture
Offline
Last seen: 1 day 11 hours ago
Joined: 11/01/2016 - 08:00
Posts: 5533
Location: Germany

next monday placing an order in my board fab

led4power
led4power's picture
Offline
Last seen: 17 hours 53 min ago
Joined: 12/29/2012 - 09:48
Posts: 1227
Location: Croatia,EU

Agro wrote:
Yes, that's precisely what I want. Q8WP driven to 1.25 A does ~124 lm/W with Vf of 3V. Drive it linearly and you get about 96 lm/W throughout battery discharge. Buck it with 95% efficiency and end-up with 118 lm/W.

That calculation is not correct, correct calculation is:

Linear drive efficiency with fresh battery: n1=Vf/Umax=3/4.2=71.43%

Linear drive efficiency with empty battery: n2=Vf/Umin=3/3=~100%

 

Because discharge curve for typ. li-ions is fairly linear. no integration over time is needed to get average efficiency, it's simply average of n1 and n2 calculated above.

So, average efficiency of linear driver when driving LED with Vf=3V from 4.2V down to 3V is (71.43+100)/2=85,71%, and that means overall flashlight efficiency is 124lm/W*85,71%/100= 106,3lm/W.

Not as good as 95% efficient buck driver, but not that bad as some think. 1.25A is relatively low current these days,at higher currents efficiency of buck driver becomes similar to linear driver,but due to much higher internal resistances, buck driver will be in "out of regulation mode" sooner than linear driver, so there is almost no advantage at high current 1S setups.

 

 

 

led4power
led4power's picture
Offline
Last seen: 17 hours 53 min ago
Joined: 12/29/2012 - 09:48
Posts: 1227
Location: Croatia,EU

Agro wrote:
Yes, that's precisely what I want. Q8WP driven to 1.25 A does ~124 lm/W with Vf of 3V. Drive it linearly and you get about 96 lm/W throughout battery discharge. Buck it with 95% efficiency and end-up with 118 lm/W.

That calculation is not correct, correct calculation is:

Linear drive efficiency with fresh battery: n1=Vf/Umax=3/4.2=71.43%

Linear drive efficiency with empty battery: n2=Vf/Umin=3/3=~100%

 

Because discharge curve for typ. li-ions is fairly linear. no integration over time is needed to get average efficiency, it's simply average of n1 and n2 calculated above.

So, average efficiency of linear driver when driving LED with Vf=3V from 4.2V down to 3V is (71.43+100)/2=85,71%, and that means overall flashlight efficiency is 124lm/W*85,71%/100= 106,3lm/W.

Not as good as 95% efficient buck driver, but not that bad as some think. 1.25A is relatively low current these days,at higher currents efficiency of buck driver becomes similar to linear driver,but due to much higher internal resistances, buck driver will be in "out of regulation mode" sooner than linear driver, so there is almost no advantage at high current 1S setups.

 

Barkuti
Barkuti's picture
Offline
Last seen: 5 hours 26 min ago
Joined: 02/19/2014 - 14:46
Posts: 4413
Location: Alhama de Murcia, Spain

Let's @#$%ing face it: bearing in mind implied contact losses (and even if don't) the useful voltage window of li-ion cells overlaps operational emitter voltage windows. A multi-purpose 1S/1S driver should be able to accept input voltages from 2.5V, offering regulated output up to at least 3.8V. This means buck and boost.

And another thing: PWM is bad. Since little to no one complains, there's no @#$%ing care about. The usual high frequency (4+KHz) switching of 7135's is the last straw, it causes lower efficiency and much earlier loss of regulation. A proper LC filter at the output of linear regulators can fix these issues, providing proper stable DC to the emitter and doing away with the need to switch on/off faster than our poor 7135s can tolerate. This means an inductor in series with the emitter is required. Can someone bear with me and do some inductor calculations for 1 to 2KHz 7135 PWM frequency? Is this too low frequency to smooth out?

 

Cheers Smile

Agro
Agro's picture
Online
Last seen: 2 min 43 sec ago
Joined: 05/14/2017 - 11:16
Posts: 5036
Location: Ślōnsk

led4power wrote:

Agro wrote:
Yes, that’s precisely what I want. Q8WP driven to 1.25 A does ~124 lm/W with Vf of 3V. Drive it linearly and you get about 96 lm/W throughout battery discharge. Buck it with 95% efficiency and end-up with 118 lm/W.

That calculation is not correct, correct calculation is:


Linear drive efficiency with fresh battery: n1=Vf/Umax=3/4.2=71.43%


Linear drive efficiency with empty battery: n2=Vf/Umin=3/3=~100%


 


Because discharge curve for typ. li-ions is fairly linear. no integration over time is needed to get average efficiency, it’s simply average of n1 and n2 calculated above.


So, average efficiency of linear driver when driving LED with Vf=3V from 4.2V down to 3V is (71.43+100)/2=85,71%, and that means overall flashlight efficiency is 124lm/W*85,71%/100= 106,3lm/W.


Not as good as 95% efficient buck driver, but not that bad as some think. 1.25A is relatively low current these days,at higher currents efficiency of buck driver becomes similar to linear driver,but due to much higher internal resistances, buck driver will be in “out of regulation mode” sooner than linear driver, so there is almost no advantage at high current 1S setups.


 


I did something wrong, your calculations look correct. Yes, at higher current the difference would be lower.

Either way, when I see that a different driver enables 5% higher sustained performance and 5-10% better runtime (depending on mode), I stop seeing the less efficient one as a top choice.

Barkuti wrote:

Let’s @#$%ing face it: bearing in mind implied contact losses (and even if don’t) the useful voltage window of li-ion cells overlaps operational emitter voltage windows. A multi-purpose 1S/1S driver should be able to accept input voltages from 2.5V, offering regulated output up to at least 3.8V. This means buck and boost.


Agree, buck-boost would be even more awesome.
Agro
Agro's picture
Online
Last seen: 2 min 43 sec ago
Joined: 05/14/2017 - 11:16
Posts: 5036
Location: Ślōnsk

Shocki, Lexel, have you considered porting support for the external thermal sensor from GXB172?

Schoki
Schoki's picture
Online
Last seen: 29 sec ago
Joined: 03/27/2017 - 15:34
Posts: 215
Location: Germany

It is possible to add external thermal support, but someone needs to write the firmware for it.

For those still waiting for the 17mm boost driver, I’m still working on it. It will look pretty similar to the one made by loneoceans.
But I still have problems with it, and need to solve them.

Niko
Offline
Last seen: 2 weeks 2 days ago
Joined: 02/08/2018 - 09:55
Posts: 153
Location: inside flashlight

Looking for one side driver with fet only and original nanjg 105 firmware (stars option)

CarryGun
Offline
Last seen: 1 year 6 months ago
Joined: 09/24/2017 - 05:04
Posts: 4
Location: Belarus

hello
where i can buy 12v booster driver to install xhp35 in my convoy s2+ ?

BlueSwordM
BlueSwordM's picture
Offline
Last seen: 11 hours 2 min ago
Joined: 11/29/2017 - 12:34
Posts: 5463
Location: Canada

@Carrygun, you can’t for now. You will have to wait until 17mm boost converters are available.

My very own high current Beryllium Copper springs Gen 3:
http://budgetlightforum.com/node/67401
Liitokala Aliexpress Stores Battery Fraud: http://budgetlightforum.com/node/60547

CarryGun
Offline
Last seen: 1 year 6 months ago
Joined: 09/24/2017 - 05:04
Posts: 4
Location: Belarus

it is possible to install xhp35 in s2+ in any way ? may be 20mm it is possible to “fit” ?

BlueSwordM
BlueSwordM's picture
Offline
Last seen: 11 hours 2 min ago
Joined: 11/29/2017 - 12:34
Posts: 5463
Location: Canada

Why do you want to install an XHP35 inside of an S2 anyway?

My very own high current Beryllium Copper springs Gen 3:
http://budgetlightforum.com/node/67401
Liitokala Aliexpress Stores Battery Fraud: http://budgetlightforum.com/node/60547

CarryGun
Offline
Last seen: 1 year 6 months ago
Joined: 09/24/2017 - 05:04
Posts: 4
Location: Belarus

just for fun
i love flashlights and convoy s2+

Schoki
Schoki's picture
Online
Last seen: 29 sec ago
Joined: 03/27/2017 - 15:34
Posts: 215
Location: Germany

With the current MLCC shortage, I can’t get the capacitors that I want for the boost driver, they’re all sold out. I will order from mouser soon though, and get the best capacitors they currently have.

eas
eas's picture
Offline
Last seen: 6 hours 14 min ago
Joined: 07/14/2014 - 18:53
Posts: 1361
Location: PNW
CarryGun wrote:
it is possible to install xhp35 in s2+ in any way ? may be 20mm it is possible to “fit” ?

I think a user (Jensen????) here managed to get either the H1-A or H2-C driver from into an s2+ by somehow mounting it behind/under the pill and, perhaps, grinding down the PCB to get it to fit.

H1-A is, I think, 20mm and boosts to 6v. H2-C is 22mm and boosts to 12v. I think, but I’m not sure, that the H1-A can be modded to higher compliance voltages.

gchart
gchart's picture
Offline
Last seen: 5 hours 13 min ago
Joined: 03/19/2016 - 11:57
Posts: 1808
Location: Central IL

I think I remember something like that, but all I could come up with a quick search was him building an S2+ into a MT-G2 with a H1-A. But yeah, he mounted the driver behind the pill instead of in it:

http://budgetlightforum.com/comment/1143398#comment-1143398

contactcr
Offline
Last seen: 8 hours 16 min ago
Joined: 05/19/2017 - 18:52
Posts: 2854
Location: US

BOOSTING” this thread to the top again in hopes of some life.

Remember that time we had 3-4 different people concurrently working on boost drivers then nothing panned out from any of them?

Good times Cool

Texas_Ace
Texas_Ace's picture
Offline
Last seen: 12 hours 38 min ago
Joined: 03/24/2016 - 07:44
Posts: 8569
Location: Everything is brighter in Texas

LOL, both of the boost drivers are still in the works, just on hold at the moment. Schoki is waiting for some components to get in stock at Muser last I heard and LoneOceans just got busy with life but still plans to do more work on it when he gets more time.

I still have a box for of components for prototyping whenever another prototype design is ready.

contactcr
Offline
Last seen: 8 hours 16 min ago
Joined: 05/19/2017 - 18:52
Posts: 2854
Location: US

I think Lexel and Richard (MtnE) also had something in the works. Ah well whats another 6 months at this point

Texas_Ace
Texas_Ace's picture
Offline
Last seen: 12 hours 38 min ago
Joined: 03/24/2016 - 07:44
Posts: 8569
Location: Everything is brighter in Texas
contactcr wrote:
I think Lexel and Richard (MtnE) also had something in the works. Ah well whats another 6 months at this point

Lexel was using Schoki’s design.

Interesting, I was not aware Richard was working on another boost driver.

contactcr
Offline
Last seen: 8 hours 16 min ago
Joined: 05/19/2017 - 18:52
Posts: 2854
Location: US
Texas_Ace
Texas_Ace's picture
Offline
Last seen: 12 hours 38 min ago
Joined: 03/24/2016 - 07:44
Posts: 8569
Location: Everything is brighter in Texas

contactcr wrote:
See below:

https://www.instagram.com/p/BcggNd2Dbm0/?hl=en

Very interesting, looks to be based on the same basic setup as the other drivers but impossible to say without seeing the other side.

I am guessing he ran into similar issues to the rest of us or it would be released already. It is proving very hard to get a stable driver in this form factor that can provide enough power for us.

BlueSwordM
BlueSwordM's picture
Offline
Last seen: 11 hours 2 min ago
Joined: 11/29/2017 - 12:34
Posts: 5463
Location: Canada

Even 20W 17mm boost driver would be great for us, so we can finally get our Convoy Nichia 144AM S2 ready.

My very own high current Beryllium Copper springs Gen 3:
http://budgetlightforum.com/node/67401
Liitokala Aliexpress Stores Battery Fraud: http://budgetlightforum.com/node/60547

contactcr
Offline
Last seen: 8 hours 16 min ago
Joined: 05/19/2017 - 18:52
Posts: 2854
Location: US

Honestly im just as interested in 20-22mm as 17mm and everyone seems to be going to larger drivers in lights that can take the heat anyways

Pages