Test/Review of 1-2* Lithium 2-Group 3/5-Mode 2.8A LED Flashlight Driver (LD-29)

23 posts / 0 new
Last post
HKJ
HKJ's picture
Offline
Last seen: 2 hours 54 min ago
Joined: 05/24/2011 - 12:23
Posts: 6678
Location: Copenhagen, Denmark
Test/Review of 1-2* Lithium 2-Group 3/5-Mode 2.8A LED Flashlight Driver (LD-29)
1-2* Lithium 2-Group 3/5-Mode 2.8A LED Flashlight Driver (LD-29)

DSC_4065 DSC_4067
DSC_4066 DSC_4068
DSC_4069 DSC_4070
DSC_4071 DSC_4072
DSC_4073

Driver is from Fasttech.com

Official specifications:
  • Product Type: Flashlight Driver
  • Constant Current: No
  • Driver Type: Buck driver
  • Input Voltage: 3-8.4 -volt
  • Output Current: 2800 mA
  • PCB Height (w/o Components): 1.8 mm
  • Reverse Polarity Protection: Yes
  • Flashlight Modes: 2-Group 8-Mode Configurable
  • Mode Count: 5
  • Mode Memory:No
  • Depth: 9 mm
  • Diameter: 20.5 mm
  • Product Weight: 3.2 g


Perfect for Cree XM-L series LEDs -- especially with T6 and U2
Dual PCB layout, 10mm total height
Designed for Cree XM-L LED emitters
Also works with four MC-E's and four SST-P7's connected in parallel
2800mA maximum current output
Low voltage warning battery protection mode triggered when voltage drops below 3V (in 1-lithium configuration) or 5.8V (in 2-lithium configuration): automatically switches to the lowest output setting, and blinks periodically
Modes:
Group 1: Lo (5%) > Mid (50%) > Hi (100%)
Group 2: Lo (5%) > Mid (50%) > Hi (100%) > Strobe > SOS

To toggle mode group:
Switch to Lo (5%) mode and count exactly 5 seconds. As soon as the flashlight blinks, half press (or quickly toggle power) to go into the other mode group.



Measurements

Tested with: Cree XM-L2
The driver uses buck mode
Driver diameter 18.6mm + 21mm
Driver height 8mm
The driver has memory, the actual mode is stored when the light is off for a short time.
A off/on will select next mode, the off has to be rather long, not to select next mode.
Mode set 1: High, medium, low, Strobe, SOS
Mode set 2: High, medium, low
The driver can be used with 1 or 2 LiIon batteries in series.
Strobe is 15Hz with 43% duty cycle
Threshold between 1 and 2 cell is between 4.5 and 5 volt



High

1-2%20Lithium%202-Group%203-5-Mode%202.8A%20LED%20(LD-29)%20HL%208.4V

The drivers has very good regulation and good efficiency. It turns power down at 6.7 volt and battery warning comes on at 6 volt.

1-2%20Lithium%202-Group%203-5-Mode%202.8A%20LED%20(LD-29)%20HL%208.4VDriver

There is a maximum of 1.5 watt lost in the driver, while delivering 9 watt to the led.

1-2%20Lithium%202-Group%203-5-Mode%202.8A%20LED%20(LD-29)%20HL%208.4VLed

High%207.2V

No noise in the light.

LowVoltage%205.9V

The low voltage warning is a slow flashing.


1-2%20Lithium%202-Group%203-5-Mode%202.8A%20LED%20(LD-29)%20HL%204.5V

With one cell the stabilization only works down to 3.7 volt, that means no stabilization for half of the runtime on a single LiIon cell.
The driver has a low voltage warning at 3 volt.

1-2%20Lithium%202-Group%203-5-Mode%202.8A%20LED%20(LD-29)%20HL%204.5VDriver
1-2%20Lithium%202-Group%203-5-Mode%202.8A%20LED%20(LD-29)%20HL%204.5VLed

High%203.6V
LowVoltage%202.8V



Medium

1-2%20Lithium%202-Group%203-5-Mode%202.8A%20LED%20(LD-29)%20HL%208.4V%20medium

Medium has the same good stabilization and does not step down, it goes directly to low battery warning at 6 volt.

1-2%20Lithium%202-Group%203-5-Mode%202.8A%20LED%20(LD-29)%20HL%208.4V%20mediumDriver
1-2%20Lithium%202-Group%203-5-Mode%202.8A%20LED%20(LD-29)%20HL%208.4V%20mediumLed

The led power is down to 2 watt.

Medium%207.2V

1-2%20Lithium%202-Group%203-5-Mode%202.8A%20LED%20(LD-29)%20HL%204.5V%20medium

With lower led power, the driver can sustain the medium level down to about 3.2 volt and it has very good efficiency.

1-2%20Lithium%202-Group%203-5-Mode%202.8A%20LED%20(LD-29)%20HL%204.5V%20mediumDriver
1-2%20Lithium%202-Group%203-5-Mode%202.8A%20LED%20(LD-29)%20HL%204.5V%20mediumLed

Medium%203.6V



Low

1-2%20Lithium%202-Group%203-5-Mode%202.8A%20LED%20(LD-29)%20HL%208.4V%20low

The regulation is also good on low, but the efficiency is slightly down.

1-2%20Lithium%202-Group%203-5-Mode%202.8A%20LED%20(LD-29)%20HL%208.4V%20lowDriver
1-2%20Lithium%202-Group%203-5-Mode%202.8A%20LED%20(LD-29)%20HL%208.4V%20lowLed

Low%207.2V

1-2%20Lithium%202-Group%203-5-Mode%202.8A%20LED%20(LD-29)%20HL%204.5V%20low

With a single cell the driver works down to about 3 volt with full stabilization, before it goes into low voltage warning.

1-2%20Lithium%202-Group%203-5-Mode%202.8A%20LED%20(LD-29)%20HL%204.5V%20lowDriver
1-2%20Lithium%202-Group%203-5-Mode%202.8A%20LED%20(LD-29)%20HL%204.5V%20lowLed

Low%203.6V



Strobe

Strobe%207.2V
Strobe%203.6V

Strobe is 15Hz with 43% duty cycle, the driver cannot reach full brightness with a single cell


SOS
SOS%207.2V
SOS%203.6V


Conclusion

The driver is a very good stabilization (Except on high with 1 cell) and good efficiency. The mode change could be better.


Notes

I have tested a slightly different LD29 driver before, with nearly the same result.

How do I test a led driver
List of all tested drivers

My website with reviews of many chargers and batteries (More than 1000): https://lygte-info.dk/

Barkuti
Barkuti's picture
Offline
Last seen: 2 hours 5 min ago
Joined: 02/19/2014 - 14:46
Posts: 3586
Location: Alhama de Murcia, Spain

Hello! Smile

Took a few close-up pictures of an LD-29 received from FastTech a week or two ago, combined them in a picture:

 

 

Some changes can be seen, among them the usage of an R020 current sense resistor (which means 60mV Thumbs Up sense voltage). The daughter board features component slots for… reverse polarity protection MOSFETs maybe?

I have a modding question: what are my chances of considerably increasing the current and power output of this driver? Let's say I aim for 7,8A current output LoL by stacking 4x additional R050 current sense resistors. The onboard schottky is rated for 10A so it should survive (?). The inductor looks insufficient to me, I can swap it with a much bigger 2R2 (10,2mm side lenght versus 6,8mm of the stock 4,7µH unit) or maybe go the extra mile and install a motherboard scavenged super high current R60 unit.

Main problem, as far as my understanding goes, is the AON7403 MOSFET switch. Do you think stacking another one above it with a tiny heatsink could be enough? I've checked its VGS versus resistance graphs and it doesn't seems appropriate for 1S input operation at all… Facepalm

 

Cheers Party 

P.S.: is the output capacitor 6V rated? 

teacher
teacher's picture
Online
Last seen: 5 sec ago
Joined: 02/23/2016 - 19:04
Posts: 8771
Location: NE & SW Alabama

Good pictures & info Barkuti!! Very sharp & clear, thanks for sharing!! Thumbs Up

You never know how a horse will pull until you hook him up to a heavy load./"Bear" Bryant 

 .................................. "Slow is Smooth, Smooth is Fast" ...................................

       Texas Lumens Flashlights / M4D M4X Deals : sign up - save $$$$  

         Rudeness Level _ mΩ _ {width:70%} _ LightWiki _ LED Tint Chart  

      Xlamp size chart _ BatteryU _ Flashaholic? Need Professional Help???            TheOriginal _ TAB _ LightSearch _ BatterySearch _ 14500's _ DiCal 

 

                                             

Barkuti
Barkuti's picture
Offline
Last seen: 2 hours 5 min ago
Joined: 02/19/2014 - 14:46
Posts: 3586
Location: Alhama de Murcia, Spain

Measured the output cap value with my ZT102 multimeter, 154,1µF. Is that a tantalum capacitor? 

 

Party

Barkuti
Barkuti's picture
Offline
Last seen: 2 hours 5 min ago
Joined: 02/19/2014 - 14:46
Posts: 3586
Location: Alhama de Murcia, Spain

Wellp, went to the Alpha & Omega Semiconductor site to look for AON7403 alternatives:

  • AON7405, higher rated than the 7403 but still no good performance at low VGS.
  • AON7421, yes now we're talking. Nicely low RDS(ON) (< 9mΩ) even at VGS = −2,5V.
  • AON7423, out of production unit even better than the 7421 (?); maybe there are units lying around. 

 

BlueSwordM
BlueSwordM's picture
Offline
Last seen: 1 hour 4 min ago
Joined: 11/29/2017 - 12:34
Posts: 4210
Location: Canada

I mean, you could get this too:
https://www.digikey.ca/product-detail/en/toshiba-semiconductor-and-stora...

It’s in the same package, and has even low resistance at 6mOhm even at VGS = -2,5V.

My very own high current Beryllium Copper springs: http://budgetlightforum.com/node/64047
Liitokala Aliexpress Stores Battery Fraud: http://budgetlightforum.com/node/60547

Barkuti
Barkuti's picture
Offline
Last seen: 2 hours 5 min ago
Joined: 02/19/2014 - 14:46
Posts: 3586
Location: Alhama de Murcia, Spain

Nice find BlueSwordM, also in stock at Arrow (free Wink shipping): https://www.arrow.com/en/products/tpn4r712mdl1q/toshiba

TPN4R712MD.L1Q datasheet

Also cheaper than the AON7421, but way slower switching speed though. tr + ton + tf + toff is 198ns for the AON7421, 617ns for the TPN4R712MD.L1Q, and just 42,2ns for the AON7403. The buck converter frequency is probably ≈300KHz, I wonder to which point do these parameters matter.

I'd like to hear some experts first. Big Smile

 

Party 

Barkuti
Barkuti's picture
Offline
Last seen: 2 hours 5 min ago
Joined: 02/19/2014 - 14:46
Posts: 3586
Location: Alhama de Murcia, Spain

Besides changing the current sense value my wild guess is that I should swap the switching MOSFET, the inductor and maybe improve input and/or output capacitance (?). This should allow the buck driver to operate reliably at much higher currents because of much improved output noise and reliability.

Anyone? 

 

kikkoman
kikkoman's picture
Online
Last seen: 6 min 27 sec ago
Joined: 05/03/2012 - 08:08
Posts: 185
Location: Austria

That buck controller is an XC9221, you’ll find guidelines and the relevant formulas in the datasheet here.

I think 4.7µH is already on the low side for 300kHz, any lower might cause stability issues, but if you can fit a bigger size inductor anyways you can up the current rating at that value.
I don’t think switching speed of the FET is that big of an issue at 300kHz, but what matters is the input capacitance (recommended <1500pF), otherwise the controller can’t drive the gate fast enough. (That’s where the Toshiba shits the bed)

btw, Vishay and TI have very good FETs (like the CSD25402Q3A )

PS: If you’re using this in a 2S setup anyway, does the Vgs@2.5V matter that much? I don’t think so

Barkuti
Barkuti's picture
Offline
Last seen: 2 hours 5 min ago
Joined: 02/19/2014 - 14:46
Posts: 3586
Location: Alhama de Murcia, Spain

Thanks for the input kikkoman, seems the XC9221 wasn't designed for higher than 3A output and I want to make it go a lot beyond Crazy that LoL. Thank God at least the schottky Innocent is generous.

Both the TPN4R712MD.L1Q and the AON7421 have about 3x the maximum recommended input capacitance listed in the XC9221 datasheet. The CSD25402Q3A is a nice find but it is not available in single pieces at Arrow, so I have to keep looking.

 

kikkoman
kikkoman's picture
Online
Last seen: 6 min 27 sec ago
Joined: 05/03/2012 - 08:08
Posts: 185
Location: Austria
Barkuti wrote:

Thanks for the input kikkoman, seems the XC9221 wasn’t designed for higher than 3A output and I want to make it go a lot beyond Crazy that LoL. Thank God at least the schottky Innocent is generous.


Both the TPN4R712MD.L1Q and the AON7421 have about 3x the maximum recommended input capacitance listed in the XC9221 datasheet. The CSD25402Q3A is a nice find but it is not available in single pieces at Arrow, so I have to keep looking.


 


There’s its brother, the CSD25404Q3 – slightly lower Rds, slightly higher gate capacitance (but still quite low), and it’s a bit pricier. 0.95€ for 1.
I really can’t tell you though how ‘set in stone’ that current limit is. There’s no internal current limit, so aside from external components (= available space), I guess the gate driver is the one limiting factor.

Barkuti
Barkuti's picture
Offline
Last seen: 2 hours 5 min ago
Joined: 02/19/2014 - 14:46
Posts: 3586
Location: Alhama de Murcia, Spain

Thanks again kikkoman, I was even then hopelessly looking at the PowerPAK 1212-8 section at Vishay.

Guess will have to try that CSD25404Q3. The input capacitance is specified at 10V, it should be a bit lower at a bit less voltage, doesn't it? Also, the QX9221 shouldn't complain if a little bit above 1500pF… 

 

Smile 

Barkuti
Barkuti's picture
Offline
Last seen: 2 hours 5 min ago
Joined: 02/19/2014 - 14:46
Posts: 3586
Location: Alhama de Murcia, Spain

To round this off, took a peek at Murata's inductor catalog. Their FDA1254-H-4R7M=P3 seems to fit the high current bill (LoL with the model naming). €2,01 a piece… Facepalm

 

Cheers Party 

Barkuti
Barkuti's picture
Offline
Last seen: 2 hours 5 min ago
Joined: 02/19/2014 - 14:46
Posts: 3586
Location: Alhama de Murcia, Spain

Just ordered 3x FDA1254-H-4R7M=P3 inductors and 3x CSD25404Q3 MOSFETs. Didn't asked this before, but I guess improving input and output capacitance is also advisable, doesn't it?

Stock input capacitance is 22µF and I have 90+ pieces of capacitors like the stock one inside a drawer. Also a handful of 47µF tantalum caps which may go well at the output… 

Will have to put a name to the badass, maybe LD-29.9 LoL.

 

Mon, 02/11/2019 - 06:23

kikkoman
kikkoman's picture
Online
Last seen: 6 min 27 sec ago
Joined: 05/03/2012 - 08:08
Posts: 185
Location: Austria

There’s no real downside to increasing the input capacitance I think, but output is different. If the ESR is too low, you’ll get load transient issues. I wouldn’t go overboard with that.
If you have a scope, you can check if output ripple is within acceptable limits (if it is, then there’s no reason to increase C_out).

Free overnight shipping @ Arrow is great but there are distributors that are cheaper in the end (even with shipping) where comparable inductors cost 1/4 of the price.

Barkuti
Barkuti's picture
Offline
Last seen: 2 hours 5 min ago
Joined: 02/19/2014 - 14:46
Posts: 3586
Location: Alhama de Murcia, Spain

Thanks again for contributing kikkoman.

Could add another input capacitor in parallel without much trouble.

What do you mean with output load transient issues?

No scope right now but since I aim for 3x the stock current output, I'll fit some extra capacitance.

With regards to getting a scope, will have to consider this. Something affordable Big Smile and good. I have not much idea of the features I should mainly care for with the exception of price LoL.

kikkoman wrote:

Free overnight shipping @ Arrow is great but there are distributors that are cheaper in the end (even with shipping) where comparable inductors cost 1/4 of the price.

Examples? 

 

Cheers Party 

Barkuti
Barkuti's picture
Offline
Last seen: 2 hours 5 min ago
Joined: 02/19/2014 - 14:46
Posts: 3586
Location: Alhama de Murcia, Spain

Just noticed I'll have to build a saddle for the MOSFET swap LoL:

 

Facepalm InnocentSilly

 

 

Barkuti
Barkuti's picture
Offline
Last seen: 2 hours 5 min ago
Joined: 02/19/2014 - 14:46
Posts: 3586
Location: Alhama de Murcia, Spain

https://gifyu.com/album/shA 

 

Connected my small 219B plus heatsink testing mule to it and supplied power (≈8.2V) with DC supply, sticking a normally closed push button in the middle for mode change.

Of course I set some current limit  in the supply just in case. The emitter started being lit very faintly, and its output started ramping up until the supply hit the current limit at more current than it should pull with the stock sense resistor value. I used the switch a few times and it seemed to enter some sort of weird strobe mode at some point, but continued hitting the switch. I raised the current limit a bit and then I left it doing what it was doing once or twice more for some seconds and all of a sudden it all stopped, with some very subtle burning smell I'd say. 

Could it be that the buck converter was unable to properly drive the MOSFET gate?

I have not dismantled the driver corpse yet LoL.

 

Innocent 

Barkuti
Barkuti's picture
Offline
Last seen: 2 hours 5 min ago
Joined: 02/19/2014 - 14:46
Posts: 3586
Location: Alhama de Murcia, Spain

I was toying with the idea of modifying another one of these, found a well suited inductor in the Bourns SRP6050CA-4R7M. However, I have a few qualms with the newer LD-29. The AON7403 MOSFET's low VGS (3 - 4V) performance seems much worse than that of the FDS4435 according to datasheet curves. I wonder how much of an impact will this have for 1S operation due to the much higher input current handling. Not being a scaremonger, just a little tired of burning stuff up. It would also help if manufacturers would stop spoiling their products (are you hearing LDCH?).

 

Cheers 

Barkuti
Barkuti's picture
Offline
Last seen: 2 hours 5 min ago
Joined: 02/19/2014 - 14:46
Posts: 3586
Location: Alhama de Murcia, Spain

Been peeking into some small package p-channel MOSFET datasheets with low input gate capacitance. There are some better offerings over the stock AON7403, the Infineon IRFHM9391TRPBF for example. However, all of them exhibit noticeably lower performance at less than –4.5V of VGS, and already a lot lower at –3.5V. I am surprised at how this driver did not have some problems on high mode with less than 4V of input, but of course this is due to the relatively low input current. Sucks, really.

 

Sat, 04/06/2019 - 20:53

Lexel
Lexel's picture
Offline
Last seen: 1 hour 21 min ago
Joined: 11/01/2016 - 08:00
Posts: 5038
Location: Germany

Hi,

you may have a look on 3×3mm footprints I found there some way better P-FETs
you can PM me if you found the right
a slight hint which makes you think its a N-Channel buts its indeed a P-FET

I have looked also at ultra low Vgs N-channel FETs I found one that has down to 1.5V almost no increased Rds

btw your infineon datasheet link is not working
it is a really bad choice looking on Vgs graph for 1S drivers

Barkuti
Barkuti's picture
Offline
Last seen: 2 hours 5 min ago
Joined: 02/19/2014 - 14:46
Posts: 3586
Location: Alhama de Murcia, Spain

I have fixed the link Lexel.

 

What's this?

The XC9220 which is onboard the LD-29 drives p-channel MOSFETs (I think or believe it uses that).

 

Smile 

 

Lexel
Lexel's picture
Offline
Last seen: 1 hour 21 min ago
Joined: 11/01/2016 - 08:00
Posts: 5038
Location: Germany

those are 2 very good fets
the one with red graph P-channel
the other N-Channel dont mix em up together

I dont write the type as its sort of a lot research I did to find them but you should be able to find them

if you look for an 1/2S replacement FET you need one that stays at a relative stable Rdson down to 3V better 2.5V