Test/Review: LD-1 5A PWM-less driver - led4power (Tests in aYezl Y3)

74 posts / 0 new
Last post

Pages

Tom E
Tom E's picture
Online
Last seen: 51 sec ago
Joined: 08/19/2012 - 08:23
Posts: 14928
Location: LI NY
Test/Review: LD-1 5A PWM-less driver - led4power (Tests in aYezl Y3)

Look what I got in the mail today:

Oh wow - the fun will begin. The board has maybe a couple of solder splash's, not much, pretty clean. I'll clean it up with isopropyl alcohol a little first.  No docs with the board, but he has everything covered in the driver thread OP: http://budgetlightforum.com/node/29434.

Not sure yet host - Maybe a Yezl Y3 with the tail and e-switch, or a Small Sun ZY-T11 e-switch type. I feel it's best served by an e-switch light, but certainly is designed to go both ways.

If anyone has any suggestions for test methods, I'll listen, but may not have the equipment or knowledge. HKJ will do his usual thorough driver test on this, so I'll look at mostly the usability and basic performance and functions.

 

Wed 07/23/2014:

 This morning I was able to solder on some 20 AWG leads, add a temporay spring and test it on my simple bench setup. It works: all 4 std modes, etc. My bench setup is 1 XML/aluminum MCPCB with 22 AWG leads/clips, 1 XM-L2/Noctigon with 22 AWG leads/clips mounted on a large aluminum piece, and a simple one cell battery holder wired with 22 AWG wire leads/clips. Sorry no pics yet.

Results

Mode       XML         XM-L2

low-low   0.02A       0.02A

low         0.09A       0.09A

med        1.13A       1.13A

hi            3.48A       2.92A

Conclusions:

  • I can't get to 5A, most likely due to the test fixture setup with all these long 22 AWG wires, alligator clips, poor qualty battery holder (expected)
  • the lower amps on the XM-L2: expected because of all the voltage drops in the test fixture - XM-L2 on copper have a higher Vf demand
  • what's really nice is the lower modes stayed dead-on consistent, and this is unusual I believe. I think a typical PWM driver would scale all modes, not just hi mode. This indicates why this driver is superior to PWM based drivers.
  • In percent ratio of amps mapped to 5A as 100%, the modes are: 0.4%, 1.8%, and 22.6% which maps very closely to the LD-1 specs of: 0.3%, 1.5%, and 20%. Slightly higher, but not sure if his specs were in amps or output - perhaps led4power can explain

I did some debating on what to mount this in, and decided to do a HD2010 initially in a copper pill (simple power switch). I got a VOB copper pill all prep'ed already, so should go pretty fast. Plan is to move it on to a Yezl Y3 (power switch and e-switch), probably it's final resting place.

 The LD-1 measures 17.18mm in diam, 1.63 mm in height using my digital caliper. A Nanjg is 16.90 x 1.55, and a BLF17DD is 17.01 x 1.57 mm. I found the LD-1 pops right in to a P60 pill and a stock Convoy C8 pill, but requires sanding down to fit any of the copper pill, or brass C8 pills I have. So, got the file going on the edge and got it down to bout 16.90 - 16.95, and now it's fits in real nice to the HD201 VOB copper pill, and all the others.

Thu 07/24/2014:

Got the board mounted in a HD2010, in a VOB copper pill. Tried a bunch of cells to check max amp readings using just the bare pill. Good news is I can hit 5.07A on several cells, but other cells can't quite make it. I have a XM-L2 U2 1A on Noctigon screwed down to the smoothly sanded pill top, using GC Extreme grease, 20 AWG LED wires. Also used the infamous "copper" spring from FastTech -- the spring does still add resistance, so added 22 AWG copper wired down the middle, and now I get higher amp  readings: Powerizer went from 4.93A to 5.07A, so this proves those copper springs are not as good as copper wiring a spring yourself.

Cells Making 5.07A: Powerizer, LG HE2, Sam 25R, SONY VTC5

I'm only getting 4.84A on a fairly new EVVA 26650 unprotected (KK cell) at 4.20v from MtnElectronics, so that is sort of bad news not being able to get the full 5A. But, there may be some other variables going on because I'm getting some inconsistencies with these readings. I'm passing along my finding to led4power in pm's.

Sun 07/27/2014:

 I applied the resistor change led4power recommended as a result of HKJ's testing. I decided to enable turbo mode via the moon #2 pad and see how it works.

The way turbo mode works is it replaces hi mode with what should be a DD mode. I ran some tests on a few cells, here's results. "Hi" is moon #2 OFF, "Turbo" is moon #2 ON:

Cell                      Hi         Turbo

Powerizer 4.19v:   5.07A    5.29A

LG HE2 #2 4.21v   5.07A   5.35A

SONY VTC5 4.18v  5.07A   5.14A

These are the identical cells, so I'm pretty confident these #'s reflect the differences. However, there are lots of fluctuations between cells. I think what these results show is if you want a turbo mode for a single XM-L2/copper configuration, this driver appears to not be able to produce as much amps as the BLFDD17 driver for example, which can do higher with comparable wiring, LED, etc. I would suspect the FET comfy chose for the TINY13A based FET drivers has lower resistance. Also even being able to get to 5A may be a challenge for most cells in a single XM-L2 or XP-G2 on copper configuration.

 

Current Controlled medium mode vs. a PWM medium mode

To compare to a Nanjg PWM based driver, I measured the LD-1 in medium mode at 1.13A against a medium mode in a 12x350 7135 PWM driver in the same flashlight (HD2010).

LD-1 turbo mode configured, medium mode                      1.13A,  lumens: 483 @start, 479 @30secs

Nanjg 4.2A driver (measured 4.34A @tail), medium mode: 1.32A,  lumens: 381 @start, 371 @30secs

*** the LD-1 is 29% brighter, even though it is running on 14% less power. I'd say this proves how really poor PWM modes are on a higher powered light, and how efficient true current controlled regulation is.

Current Controlled medium mode vs. a PWM low mode

LD-1 turbo mode configured, low mode 0.09A, lumens: 34

Nanjg 4.2A driver (measured 4.34A @tail), low mode: 0.16A, lumens: 44

The LD-1 is using 44% less power, and produces 23% less brightness.

I'm not sure why these results are closer, compared to medium mode. Might be some non-linear scaling, plus the fact the medium modes were much closer in amps.

Tue 07/29/2014:

For some reason, I thought the LD-1 was set for non-turbo, but the amps reading is above 5A, so thinking clearing the solder on the moon#2 pad did not work, though I did test for continuity. Anyway, I wanted to do a test similar to HKJ's test for brightness % (http://budgetlightforum.com/node/33101), but for mine, I'll use lumens in the PVC lightbox. I decided to use a slightly depleted LG HE2 that records a clean 5A at the tail. I realize HKJ's amps are measured at the LED, so may not be the same.

Here's the results in the HD2010 (copper pill, XM-L2 U2 1A/Noctigon, 20 AWG):

Amps (HKJ)    Lumens OTF    % of max   HKJ % of max

5.0A (5.0A)        1,544                   100%         100%

1.13A (1.1A)       466                      30%          32.4%

0.09A (0.083A)    37                       2.4%          2.64%

0.02A (0.022A)    21                       0.45%        0.57%

These #'s are dang close! Might be our measurements for the "high" mode are not quite compatible - not sure how he measured output, but I'm pretty confident in the consistency of ranges of low to very high readings in my PVC lightbox (my high may be higher?). Or could be simply tolerances. May not matter much, but the scaling is very consistent.

Sun 08/02/2014:

I have the LD-1 installed in a fully assembled, upgraded Yezl Y3 (with e-switch). Further details can be found here: BLF - Yezl Y3 Picture Breakdown. Here's a pic of the driver:

It's still in turbo mode - can't seem to get it back to non-turbo. Main reason of using the Y3 is to have a decent host than can take the high amps, and so I can test the more advanced functionality of the firmware (multi-mode sets) that are only available with an e-switch. 

The piggyback mounting went pretty well. It's a nice low profile driver. As you can see above, I stripped off all components from the old driver and traced the switch to an MCU pad, and routed a wire (the blue one) from the MCU pad on the old driver to the e-switch pad on the LD-1.

Some Measurements:

LG HE2 @4.18v      5.46A

KK 26700 @4.21v   5.32A

On the KK 26700, lumens:  1,513 @start, 1,452 @30 secs (1,454 OTF), throw read @5 m: 111 kcd

Modes and Mode Switching

 The tail switch and e-switch work pretty good together. Using the first (default) mode set, you can easily change modes via either switch, and of course the e-switch can turn the light ON and OFF as well - nice combo. I did notice the e-switch delays mode changing, thinking this is on account of supporting a double-click. It feels a little unresponsive to a single click as a result compared to other drivers.

 In attempting to test the other mode sets, I'm kind of stuck. Thought I had it working in ramping, but could not figure it out - seems like there are delays, and in attempting to figure it out -- I'm stuck now - can't change mode sets, and not sure what exactly mode set I'm on now. It's acting as a two mode light now, with both modes pretty low, but close.

Update: Fixed the problem! I probably overrode the temp threshold setting, so the light was thinking it was constantly over-temperature. I got a much better feel for ramping mode now. It works great in the lower 50% range, but in the upper half it's very hard to notice the ramping at all - I'm sure that's because in output %, it doesn't change that much, as compared to the lower 50%.

Ramping is basically a single mode, but allows you to get the single mode to any level of brightness you want. I did notice that if you ramp it down to the lowest setting and use that, when you turn the light back on, it very briefly flashes a higher level of brightness. I reported this to led4power.

 

Edited by: Tom E on 08/04/2014 - 07:06
Tom E
Tom E's picture
Online
Last seen: 51 sec ago
Joined: 08/19/2012 - 08:23
Posts: 14928
Location: LI NY

reserved

djozz
djozz's picture
Online
Last seen: 2 min 34 sec ago
Joined: 09/07/2012 - 17:04
Posts: 18078
Location: Amsterdam

looking forward to the test! I got mine in the mail today as well, I hope to do some testing on thursday Smile

RaceR86
RaceR86's picture
Offline
Last seen: 5 years 9 months ago
Joined: 02/22/2012 - 08:32
Posts: 3777
Location: Norway

Mine is not here yet. I think you guys bribed someone. Silly

Ill probably be doing most of my testing in a light with a typical mechanical switch.

BLF LED database – collaboration spreadsheet and latest news about where to buy LEDs
http://budgetlightforum.com/node/19342

Werner
Werner's picture
Offline
Last seen: 2 years 2 months ago
Joined: 10/19/2012 - 15:00
Posts: 3679
Location: Germany

Why is there a border around the ground ring? Was the millinglayer set wrong or is it smaller in reality?

Halo...
Halo...'s picture
Offline
Last seen: 5 years 3 months ago
Joined: 12/15/2011 - 02:39
Posts: 3304
Location: Halo island

Werner wrote:
Why is there a border around the ground ring?

Oshpark does that sometimes
FlashPilot
FlashPilot's picture
Offline
Last seen: 21 hours 22 min ago
Joined: 05/10/2010 - 16:07
Posts: 4814
Location: USA

Subscribed… and happy testing! :bigsmile:

Ouchyfoot
Ouchyfoot's picture
Offline
Last seen: 9 months 2 hours ago
Joined: 09/01/2012 - 06:15
Posts: 5089
Location: Canada

I’ve been waiting for this. Get busy!

musicmagic
Offline
Last seen: 6 years 9 months ago
Joined: 10/28/2013 - 09:53
Posts: 984
Location: CT

Tom E wrote:

Look what I got in the mail today:




that package said authorized personnel only… can I see your ID please?

If you can’t blind them with your brilliance, baffle them with your bullcrap.

The real currency in the world is not money, it’s trust.

led4power
led4power's picture
Offline
Last seen: 1 day 5 hours ago
Joined: 12/29/2012 - 09:48
Posts: 1331
Location: Croatia,EU

OSHpark do that all time,at least in my case.They trim 20mils from edge on all copper layers.

WarHawk-AVG
WarHawk-AVG's picture
Offline
Last seen: 1 year 5 months ago
Joined: 01/04/2014 - 06:47
Posts: 5071
Location: H-Town

Ah, that explains why on some boards I see the top copper layer actually extending beyond the milling layer…to compensate for that cutoff by the manufacturers…thank led4power, can’t wait to see these hit production! Smile

Very interesting design, and I really like how you did the “stars” solder bridge locations…very cool!

Tom E
Tom E's picture
Online
Last seen: 51 sec ago
Joined: 08/19/2012 - 08:23
Posts: 14928
Location: LI NY

OP update added - 07/23 dated, round #1: amps measured across all std modes.

 

Werner
Werner's picture
Offline
Last seen: 2 years 2 months ago
Joined: 10/19/2012 - 15:00
Posts: 3679
Location: Germany

I always made the copper bigger than milling because of laziness so I never had this problems…I also did some where the milling was in the middle of through holes(thought that would increase GND contact) and they milled perfectly through it.

Is this just a newer phenomenon from Oshpark, did I only had luck or was it the point that I made the copper ring bigger than milling the point?

Rufusbduck
Rufusbduck's picture
Offline
Last seen: 7 months 1 week ago
Joined: 04/04/2012 - 15:34
Posts: 10389
Location: Golden state

Request for a run time comparison:
Same cell charged to same voltage
Match current output current between this driver and a Qlite 3A w/ one extra chip and pwm set to 1.1A.
Plot current vs voltage and time when each falls out of regulation.
Plot end of charge for both and calculate efficiency.
If you can get this driver up to 4A then add chips to a Qlite to match and then run the test.

Three Tanna leaves to give him life, nine to give him movement. But what if he eats the whole bag?

Scott

HKJ
HKJ's picture
Offline
Last seen: 1 week 5 days ago
Joined: 05/24/2011 - 12:23
Posts: 7447
Location: Copenhagen, Denmark

It is a bit unfair to compare them at same current, the LD-1 will be brighter than the pwm driver.

If you compared it to a 5A pwm driver, then the lower settings will be about 50% brighter.

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

ToyKeeper
ToyKeeper's picture
Online
Last seen: 4 min 47 sec ago
Joined: 01/12/2013 - 14:40
Posts: 10786
Location: (469219) 2016 HO3

That’s a good point… perhaps if someone does this test, it should be compared at the same lumen level instead of the same current?

(seems like the current would probably not be a very interesting test anyway, since any two devices at the same current will tend to have similar runtimes regardless of what they actually do with the power)

HKJ
HKJ's picture
Offline
Last seen: 1 week 5 days ago
Joined: 05/24/2011 - 12:23
Posts: 7447
Location: Copenhagen, Denmark

ToyKeeper wrote:
That's a good point... perhaps if someone does this test, it should be compared at the same lumen level instead of the same current? (seems like the current would probably not be a very interesting test anyway, since any two devices at the same current will tend to have similar runtimes regardless of what they actually do with the power)

Except that LD-1 can maintain full output to a lower battery voltage.

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

Rufusbduck
Rufusbduck's picture
Offline
Last seen: 7 months 1 week ago
Joined: 04/04/2012 - 15:34
Posts: 10389
Location: Golden state

Agreed, I’m mostly interested in independent confirmation of efficiency differential at equivalent lower mode output modes.

Three Tanna leaves to give him life, nine to give him movement. But what if he eats the whole bag?

Scott

HKJ
HKJ's picture
Offline
Last seen: 1 week 5 days ago
Joined: 05/24/2011 - 12:23
Posts: 7447
Location: Copenhagen, Denmark

Rufusbduck wrote:
Agreed, I'm mostly interested in independent confirmation of efficiency differential at equivalent lower mode output modes.

As I wrote above, you get about 50% extra brightness compared to a 5A pwm driver. This is because the led has higher efficiency at lower currents.

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

comfychair
comfychair's picture
Offline
Last seen: 5 years 6 months ago
Joined: 01/12/2013 - 05:39
Posts: 6198

You could also use a triple XPG2 as the test LED, to mostly avoid the high-Vf-limiting-max-current issue.

Rufusbduck
Rufusbduck's picture
Offline
Last seen: 7 months 1 week ago
Joined: 04/04/2012 - 15:34
Posts: 10389
Location: Golden state

HKJ, I was agreeing that the comparison should be based on lumen output rather than just current but that does make for a more complicated test.

Three Tanna leaves to give him life, nine to give him movement. But what if he eats the whole bag?

Scott

syracuse
syracuse's picture
Offline
Last seen: 5 months 1 week ago
Joined: 08/05/2013 - 00:43
Posts: 295
Location: Thailand

I also plan to do the comparison based on light output, but I’ve not receive my sample yet.

HKJ
HKJ's picture
Offline
Last seen: 1 week 5 days ago
Joined: 05/24/2011 - 12:23
Posts: 7447
Location: Copenhagen, Denmark

Rufusbduck wrote:
HKJ, I was agreeing that the comparison should be based on lumen output rather than just current but that does make for a more complicated test.

You can do a fairly good estimate without doing the test, but if you have a 5A pwm driver it is not that difficult to setup:

Mount the led and a luxmeter at some fixed position (A luxmeter with PC logging is preferred).

Run the test with the pwm driver.

Use the LD-1 in ramp mode and adjust it to same brightness on luxmeter.

Replace battery used for adjustment with fresh battery and run the test again.

 

It could be interesting to do, but I do not have a 5A pwm driver for it.

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

Tom E
Tom E's picture
Online
Last seen: 51 sec ago
Joined: 08/19/2012 - 08:23
Posts: 14928
Location: LI NY

Hhmm, gotta check my driver stock. I could build up a Nanjg to ~5A, but I do have this driver from Richard: http://www.mtnelectronics.com/opencart/index.php?route=product/product&path=67&product_id=214, but it's 2-3 cells. Didn't realize you gained 50% over a PWM mode - I knew it was better, but not by that much - cool!

HKJ
HKJ's picture
Offline
Last seen: 1 week 5 days ago
Joined: 05/24/2011 - 12:23
Posts: 7447
Location: Copenhagen, Denmark

Tom E wrote:

Hhmm, gotta check my driver stock. I could build up a Nanjg to ~5A, but I do have this driver from Richard: http://www.mtnelectronics.com/opencart/index.php?route=product/product&path=67&product_id=214, but it's 2-3 cells. Didn't realize you gained 50% over a PWM mode - I knew it was better, but not by that much - cool!

That is because the XM-L2 has low efficiency at 5A.

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

ToyKeeper
ToyKeeper's picture
Online
Last seen: 4 min 47 sec ago
Joined: 01/12/2013 - 14:40
Posts: 10786
Location: (469219) 2016 HO3

I wonder what the efficiency difference might be at 1.4A, or 1.5A, or 1.9A.

The 1.9A light I just put together with a Nichia 219B is really nice, and the PWM is fast enough it looks almost the same as full current control… but the runtime won’t be as long. I never really use it above 800mA or so (average), so I could reduce its overall power to increase efficiency, but it’s nice having a turbo mode once in a while. The higher power is also nice for strobes, because a 0.3ms pulse looks brighter at 1.9A than it would at 1.4A.

For a bike tail light, I went with 700mA because djozz’s measurements show that as more efficient than 350mA or 1050mA on a red XP-E2. I don’t need it that bright, but it should increase the overall runtime at the same brightness levels, even when using PWM.

comfychair
comfychair's picture
Offline
Last seen: 5 years 6 months ago
Joined: 01/12/2013 - 05:39
Posts: 6198

The efficiency difference is because a 50% (or whatever) constant current requires a much lower Vf, compared to a PWM'ed 50% which is full current for half the time and none for half the time (assuming both drivers do the same current at 100%). The average current might be the same for both drivers but the voltage isn't and the total power used comes from both volts & amps.

Tom E
Tom E's picture
Online
Last seen: 51 sec ago
Joined: 08/19/2012 - 08:23
Posts: 14928
Location: LI NY

Update added to OP. I'm probably not going to have time to do battery time tests, least not as thorough as I'd like.

MRsDNF
MRsDNF's picture
Offline
Last seen: 3 months 2 weeks ago
Joined: 12/22/2011 - 21:18
Posts: 13473
Location: A light beam away from the missus in the land of Aus.

Thanks for all the testing Tom E. You are trend setting here.Wink

 

djozz quotes, "it came with chinese lettering that is chinese to me".

                      "My man mousehole needs one too"

old4570 said "I'm not an expert , so don't suffer from any such technical restrictions".

Old-Lumens. Highly admired and cherished member of Budget Light Forum. 11.5.2011 - 20.12.16. RIP.

 

RaceR86
RaceR86's picture
Offline
Last seen: 5 years 9 months ago
Joined: 02/22/2012 - 08:32
Posts: 3777
Location: Norway

Im still waiting for mine. Nothing in the maibox today either.  :_(

Getting my fix here in the meantime. No need to rush the review though.

BLF LED database – collaboration spreadsheet and latest news about where to buy LEDs
http://budgetlightforum.com/node/19342

Tom E
Tom E's picture
Online
Last seen: 51 sec ago
Joined: 08/19/2012 - 08:23
Posts: 14928
Location: LI NY

Just fyi... I wasn't sure how turbo mode works. He is saying basically it's DD - very high amps, more than the 140% increase he listed. Basically turbo takes it out of current controlled/limiting mode -- this is a very nice feature!

Also, HKJ found an issue in high mode using higher powered cells that needs a cap to be replaced, and once done, fixes the issues that HKJ was seeing. So, I'm holding off more testing til I get the cap and can mod the driver. It's a tiny 0402 -- I never worked with a cap that small, so may have a "super tech" at work do the replacement. He would have no problem doing this, so less risk than me doing it Embarassed.

This HD2010, btw, is a total PIA for this experimenting because of the poor clearance for the bigger wires, big flat reflector bottom... Also the plastic LED alignment piece they ship with HD2010's now is too small for the opening, making it almost impossible to center the LED... Ugh. The older HD2010's I have didn't have this promleb- the LED alignment pieces fit much better.

Pages