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

487 posts / 0 new
Last post
Agro
Agro's picture
Offline
Last seen: 13 hours 22 min ago
Joined: 05/14/2017 - 11:16
Posts: 4423
Location: Ślōnsk

I’m contemplating tweaking my Q8 with XHP35 HI.
I’d like something that can sustain as many candelas and as many lumens as possible, so I need efficiency.

I don’t have the skills for such mod yet, but it doesn’t stop me from seeking options. Wink
2s->12V boost with XHP35 seems like a natural choice. Or even 4s->24V. Buck 4s->12V could work as well, just like boost 1s->12V.
Is there anything that would work well in such setting?
I guess BLF GT, but it won’t give good current to 4 LEDs, will it?
Something else?

Texas_Ace
Texas_Ace's picture
Offline
Last seen: 12 hours 15 min ago
Joined: 03/24/2016 - 07:44
Posts: 8468
Location: Everything is brighter in Texas

The BLF GT driver in stock form would be able to drive them at around 1.25A each, maybe a bit higher. If you upgraded some components it could do around 2A each.

Agro
Agro's picture
Offline
Last seen: 13 hours 22 min ago
Joined: 05/14/2017 - 11:16
Posts: 4423
Location: Ślōnsk

Not great, but not bad either. Turbo won’t be as high as possible, but 8A is significantly above what Q8 can sustain anyway. Will fall out of regulation before batteries are half emptied. And efficiency should be good, right?

Texas_Ace
Texas_Ace's picture
Offline
Last seen: 12 hours 15 min ago
Joined: 03/24/2016 - 07:44
Posts: 8468
Location: Everything is brighter in Texas
Agro wrote:
Not great, but not bad either. Turbo won’t be as high as possible, but 8A is significantly above what Q8 can sustain anyway. Will fall out of regulation before batteries are half emptied. And efficiency should be good, right?

2A to each xhp35 is where the trade off in lumens vs efficiency starts falling real bad.

You will have a big issue with falling out of regulation though with 4 LED’s and only 4 cells.

Lexel
Lexel's picture
Offline
Last seen: 2 hours 16 min ago
Joined: 11/01/2016 - 08:00
Posts: 5198
Location: Germany
Texas_Ace wrote:
The BLF GT driver in stock form would be able to drive them at around 1.25A each, maybe a bit higher. If you upgraded some components it could do around 2A each.

1.25A on stock GT driver with 500kHz frequency possible

I am working on an upgraded GT driver
I am discussing with del the needed changes
So far Buck frequency changed from 800 to 500kHz
input and output caps doubled, the design already has pads to fit 2 caps instead one
Diode 3A to 5A
inductor from 6A to 15.5A

6-8A should be no problem to get to
10A might be possible but that’s about the current the XHP35 get not much gain over 2A anyways

Texas_Ace
Texas_Ace's picture
Offline
Last seen: 12 hours 15 min ago
Joined: 03/24/2016 - 07:44
Posts: 8468
Location: Everything is brighter in Texas

Yep, with those changes it should handle 8A. The issue then becomes it falling out of regulation very quickly due to the voltage drop with only 4 cells.

Jensen567
Offline
Last seen: 2 months 2 weeks ago
Joined: 04/10/2017 - 16:59
Posts: 491
Location: MD/DC Area USA

You could get fully regulated 50W output using LTC3119 buck boost with 4S input and 12V output. Around 4.2A output is my guess. Bonus is that lower output power can also be used with 1S or 2S input.

Using 2S input the MP3431 chip could probably give you 8A output fully regulated. I don’t know of any existing boards for either IC though.

If you want to design a PCB there will be some design considerations for heat, especially for the MP3431. Realistically we can assume around 90% efficiency, maybe a bit better, but at 100W+ of output power that is still going to be almost 10W of heat into the driver. Most of this will be between the IC and the inductor. So you will need to design a PCB using the full available driver size, and ensure that the IC has a good heat path to the host through the PCB. Remote mounting the inductor directly to the host with thermal adhesive would also help, and it frees up PCB space as a bonus.

contactcr
Offline
Last seen: 3 hours 28 min ago
Joined: 05/19/2017 - 18:52
Posts: 2380
Location: US

Any update on the 17mm boost driver? Holding out on building a M1 XHP50.2 so I can go single cell route!

Lexel
Lexel's picture
Offline
Last seen: 2 hours 16 min ago
Joined: 11/01/2016 - 08:00
Posts: 5198
Location: Germany

I have contacted Schoki to get a 1S Boost driver with Narsil/Bistro done

My current goal is 12V/2A for 17mm and 20-22mm 12V 3-4A

We may go to a smaller MCU to get a bigger inductor and diode for higher currents,
but this complicate things and make building more expensive

Schoki
Schoki's picture
Offline
Last seen: 14 hours 39 min ago
Joined: 03/27/2017 - 15:34
Posts: 215
Location: Germany

Ordered new boards, they will arrive next year. But it should work then, maybe needs some firmware tweaking as well.

Lexel
Lexel's picture
Offline
Last seen: 2 hours 16 min ago
Joined: 11/01/2016 - 08:00
Posts: 5198
Location: Germany

17mm BLF 1S 6-12V Boost driver is in progress

This is a very powerful one, not cheap at all, but we target 8A for 6V LEDs, XHP35 can be as hard driven as desired

I got the board made in Diptrace I made improvements

power cap pad sizes increases to a 0805/1206 dual footprint, depending if 1206 fit to the driver cavity
other parts 0402/0603 Footprint
MCU/LDO caps 0805 size to use OTSM cap
and especially the thermal connection of the boost Chip got improved massive using the inner layers with direct thermal path through viases

No place for programming visases so a solder pads or pogo pin connector plate

ignore the 2. ring around it its just for production on a bigger board for easier separation

-X3-
-X3-'s picture
Offline
Last seen: 8 hours 37 min ago
Joined: 07/16/2014 - 03:28
Posts: 2701
Location: France, Angers

Really neat work with the implantation !
That big square is the coil I presume ?
I know it’s early stage development but do you have a guess at how how much it could cost once assembled/programmed for a lambda Customer like me ?

"-X3-, is there any place in your house without a flashlight ?" 

My Flashlight public album (mods, emitter swaps, eye candy)

My reviews channel (French language, Olight, Thorfire, Sofirn, Lumintop : 60+ lights tested)

My personal channel (including Olight SR mini, S1, S2, S1A and S-mini disassembly)

M4DM4X blog, saves you $$$ 

Mike C
Mike C's picture
Offline
Last seen: 1 day 16 hours ago
Joined: 01/22/2014 - 08:03
Posts: 2324
Location: Sweden

Using the ATtiny85 for this board?

Schoki
Schoki's picture
Offline
Last seen: 14 hours 39 min ago
Joined: 03/27/2017 - 15:34
Posts: 215
Location: Germany

It’s an ATTINY25V

Lexel
Lexel's picture
Offline
Last seen: 2 hours 16 min ago
Joined: 11/01/2016 - 08:00
Posts: 5198
Location: Germany

Also a Narsil variant will be coming as well, just needs further testing of a low standby current precision OP Amp

The bare material costs on this driver alone will be about 14$ From Mouser
Lets say about 6$ for assembly, customer support and testing
For first 50 drivers likely about 3$ to cover development

Shipping will be more expensive than TA drivers as its more than 10mm thick packed 4.5$

Its a small powerhouse, but the regulator can pull up to 19A from the battery, which gives an impressive output on 6 and 12V

For Narsil drivers above about 35mm the GT layout is likely the cheaper variant with same output and better efficiency on this the components are almost as cheap as regular multi cell FET+N+1 AMC based drivers

About correct in size comparism

MF02

TN42

17mm

1C3
Offline
Last seen: 1 day 17 hours ago
Joined: 03/30/2017 - 06:42
Posts: 77
Location: Italy
Lexel wrote:
Its a small powerhouse, but the regulator can pull up to 19A from the battery, which gives an impressive output on 6 and 12V

That’s really impressive, it means 50W output would be relatively easy right down to 3V, which in turns means 4000+ OTF lumens whatever the battery level on a single 18650 with the right led configuration.

That’s better than any production light I know of Big Smile if it ends up being produced I’ll definitely have to get one.

Texas_Ace
Texas_Ace's picture
Offline
Last seen: 12 hours 15 min ago
Joined: 03/24/2016 - 07:44
Posts: 8468
Location: Everything is brighter in Texas

I am so excited about Schoki’s Boost driver I can hardly contain myself.

I have actually been collecting parts and am now starting to look for hosts for this driver. These will be the first real lights I have built in almost a year, there was just nothing to significantly upgrade from what I had before.

Now I have plans to rebuild most of my collection around these drivers. The XHP35 is better then the XP-L or L2 in every way, more lumens, more throw, no tint shift. Then you have the xhp50.2 and xhp70.2 for some big flooder fun.

I have almost talked Schoki into letting me beta test some of these drivers…

Jensen567
Offline
Last seen: 2 months 2 weeks ago
Joined: 04/10/2017 - 16:59
Posts: 491
Location: MD/DC Area USA

I’d be happy to beta test as well if more testers are wanted. No problem paying for any prototype hardware.

The_Driver
The_Driver's picture
Offline
Last seen: 20 hours 19 min ago
Joined: 10/20/2016 - 05:51
Posts: 1384
Location: Germany

This will indeed be awesome!

Agro
Agro's picture
Offline
Last seen: 13 hours 22 min ago
Joined: 05/14/2017 - 11:16
Posts: 4423
Location: Ślōnsk

I have some question about further development of the Schoki’s driver….

1. We’ve seen GXB20 being optimized for 6V out. Lone oceans said that the driver would do 12V, but for optimal efficiency some components would have to be tweaked. Are there tweaks left to make the driver more efficient at certain output voltages?
2. How about low-profile inductors? Could they work sensibly with this driver?
3. Is there any component on the driver that supports temperature readout or are we back to the world of turbo timers?

Schoki
Schoki's picture
Offline
Last seen: 14 hours 39 min ago
Joined: 03/27/2017 - 15:34
Posts: 215
Location: Germany

1. If you have the driver configured to 6V, and you use the right parts from the beginning (with the output caps being able to handle >=20V), then you need to change two small resistors and the inductor to convert it from 6V to 12V.
The other way around, change the two small resistors and the inductor again.

To be honest, both inductor values work with both voltages, but the higher value inductor has a higher resistance, so it runs hotter. So just change the inductor for better efficiency.

2. Didn’t search specifically for low profile alternatives, but the XAL7030 is a lower profile version of the current inductor I’m using (XAL7070). 6V operation only then! Didn’t calculate ripple, but 1.5uH should do it (a bit lower than XAL707 with 1.8uH)

3. The MCU has a temperature readout. It is on the other side of the board, so we will have to set the maximum temperature lower. If that fails, the boost IC shuts off at 150°C.

Agro
Agro's picture
Offline
Last seen: 13 hours 22 min ago
Joined: 05/14/2017 - 11:16
Posts: 4423
Location: Ślōnsk

That’s better than I expected, thank you. Smile
Question…Lexel improved heat path from inductor and controller. Maybe it would be useful to review thermal path of the MCU as well?

Schoki
Schoki's picture
Offline
Last seen: 14 hours 39 min ago
Joined: 03/27/2017 - 15:34
Posts: 215
Location: Germany

No worries, the thermal connection of the MCU is pretty good. I made some new revisions of my board to make the heat path better. The board you can see in the thread was made to get a working board as fast as possible.
And remember, if the inductor and boost IC heat up, pretty much the whole board heats up, since those two parts cover a lot of surface area. And a 17mm diameter PCB is pretty small. So when those two parts heat up, every other part gets hot as well. And half of the MCU is directly under the inductor (and it’s a 4 layer board).

Lexel
Lexel's picture
Offline
Last seen: 2 hours 16 min ago
Joined: 11/01/2016 - 08:00
Posts: 5198
Location: Germany

Yeah one internal layer connects the BUCK controller to the whole board size

So far no BLF firmware supports external NTC to get more accurate temperature reading

TBone
Offline
Last seen: 2 days 11 hours ago
Joined: 01/27/2016 - 16:42
Posts: 341
Location: Europe

Great job guys. I will buy several of these.

Just a suggestion: If you do not need all pins of the MCU (and you find the space), could you please prepare external pads for those? Maybe with traces to GND that can be cut when needed.
Yes, no firmware today supports NTCs but this may change quickly once they become more common, as on led4power’s boards.

Lexel
Lexel's picture
Offline
Last seen: 2 hours 16 min ago
Joined: 11/01/2016 - 08:00
Posts: 5198
Location: Germany

There is literally no space to add such pads you suggest, but the pads of the MCU are wide enough you can solder very thin leads to them

The parts next to the MCU seem wider than the other, but in fact they are normal spaced and the others extremly close together

Pin 5 and 6 usually in 8 pin package are available as solder pads next to the ground ring

TBone
Offline
Last seen: 2 days 11 hours ago
Joined: 01/27/2016 - 16:42
Posts: 341
Location: Europe
Thumbs Up
Schoki
Schoki's picture
Offline
Last seen: 14 hours 39 min ago
Joined: 03/27/2017 - 15:34
Posts: 215
Location: Germany

You can solder small leads on the pads, you just need patience, and a fine solder iron tip. I mean really fine. Best you hold the leads down with some scotch tape while you solder them on. And watch out to put no force on the leads, or they could lift the whole pad.

TBone
Offline
Last seen: 2 days 11 hours ago
Joined: 01/27/2016 - 16:42
Posts: 341
Location: Europe

Yeah, that is why I hoped you find some space on the larger boards. It would make my life easier.

Texas_Ace
Texas_Ace's picture
Offline
Last seen: 12 hours 15 min ago
Joined: 03/24/2016 - 07:44
Posts: 8468
Location: Everything is brighter in Texas

I doubt we will see widspread use of NTC’s until we move to a more modern MCU with more pins. Like the 1616 or 1617. When we move to those, then lots of doors open.

Also, we are out of space on the tiny25 for anything else anyways, so it sure won’t be added to that.

Pages