Fenix E01 unpotted & driver efficiency data

[chimo]

I received my two HA Nat Fenix E01s today. I have not dissambled a head (yet), but I measured the power going in to them and the voltage to the LED when in regulation.

Here is the data:

Head #1			Head #2		
Vin(V)	Iin(mA)	Pin(mW)	Vin(V)	Iin(mA)	Pin(mW)
1.7	65.0	110.5	1.7	81.2	138.0
1.6	67.8	108.5	1.6	87.4	139.8
1.5	69.0	103.5	1.5	88.4	132.6
1.4	73.5	102.9	1.4	93.0	130.2
1.3	77.2	100.4	1.3	96.8	125.8
1.2	79.8	95.8	1.2	102.9	123.5
1.1	84.4	92.8	1.1	108.5	119.4
1.0	92.2	92.2	1.0	123.0	123.0
0.9	59.5	53.6	0.9	72.3	65.1
0.8	26.3	21.0	0.8	25.3	20.2
0.7	19.6	13.7	0.7	19.1	13.4
0.6	17.5	10.5	0.6	17.0	10.2
0.5	14.9	7.5	0.5	14.7	7.4
0.4	11.8	4.7	0.4	11.7	4.7
0.3	8.4	2.5	0.3	8.5	2.6
0.2	5.2	1.0	0.2	5.2	1.0
0.1	2.3	0.2	0.1	2.3	0.2

It appears the head drops out of regulation at around 1V.

The voltage measured at both LEDs when in regulation was 3.2V (constant voltage regulation perhaps).

The minimum start up voltage was approx 0.3V ---> WOW!

A constant voltage regulation scheme would explain the different Power In figures above and the different run-times being measured.

 

[paulr]

Re: Partial Fenix E01 driver data

Good measurements, thanks for making them. LED current is very sensitive to the junction voltage. Do you think it changes with temperature? It might be interesting to stick the E01 in the fridge or freezer for a while, then make another measurement.

 

[chimo]

Re: Partial Fenix E01 driver data

Good measurements, thanks for making them. LED current is very sensitive to the junction voltage. Do you think it changes with temperature? It might be interesting to stick the E01 in the fridge or freezer for a while, then make another measurement.

Absolutely. If the driver is CV, the current should go up a little as the die heats up. Since these LEDs don't seem to be badly over-driven, the temp rise should not bee too great. If the driver is 75-85% efficient, the LEDs in my E01s are driven in the range 70-118mW. The max spec sheet power dissipation for these LEDs is 105mW, however, we know that some other similar lights blow that figure out of the water. If I had to make a swag at emitter currents, I would say between 25-30mA.

The one concern I have for this light is the long-term reliability of the board ground contact. It relies on a ridge in the body tube to contact the PCB. My concern is the PCB traces wearing through due to frequent use. For most people and usage, this situation may never occur. For others, it could happen within several months or a year. Unfortunately, problems of this nature usually get greatly blown out of proportion. Fortunately, a fix is fairly easy (just solder a new contact layer to bridge the trace cuts). I will watch this one.

 

[chimo]

Re: Partial Fenix E01 driver data

Here are some "what if" calculations.

I measured the current through 5 bare GS and 6 bare DS LEDs when driven at 3.2V. I plugged these numbers into Excel to derive some possible efficiencies. If I were to use my small GS sample set to form conclusions, I would ballpark converter efficiency to be around 50-75% efficient (I will have to crack one open to be more exact). I will recant my original guess at emitter current to around a 19-24mA drive level depending on Vf of LED)

 

[streetmaster]

Re: Partial Fenix E01 driver data

The one concern I have for this light is the long-term reliability of the board ground contact. It relies on a ridge in the body tube to contact the PCB. My concern is the PCB traces wearing through due to frequent use. For most people and usage, this situation may never occur. For others, it could happen within several months or a year. Unfortunately, problems of this nature usually get greatly blown out of proportion. Fortunately, a fix is fairly easy (just solder a new contact layer to bridge the trace cuts). I will watch this one.

I am worried about this too. I noticed wear on the contacts after a day of playing with the light. I think Fenix should have made the contacts a bit thicker. The ledge inside the body where the head makes contact is too abrasive in my opinion. I think this issue will haunt them in the near future. Not everyone who bought these lights has a soldering iron to fix their light when it wears out prematurely. This is one of the very few flaws I can find in the E01, at least with the two I own. I think I got lucky as far as the LED color and brightness. Both of mine came from the first batch. I got a third one for my Dad, which came from the second batch from Fenix, and it was totally blue-ish(not just the hot spot) and not as bright as mine. I don't know if this was due to being from the second batch, or just the (bad)luck of the LED lottery.

 

[Moat]

Re: Partial Fenix E01 driver data

Nice info, Chimo! :thumbsup:

I wonder if it would be possible to polish smooth the battery tube's negative contact ridge, in order to minimize the wear on the board's contact. I've done that on some E0's (much easier than the E01, it appears), and they seem to be holding up fine with regular use, so far (about a year).

Am I reading correctly that the DS LED's you tested @ 3.2V, are passing between 41-48 mA?? That's pretty amazing... wouldn't that be a (@ spec 20 mA) Vf of around 3.0V? I thought the data sheet lists the DS as 3.6 Vf (and the GS as 3.2V - which looks very close to what you got) ...?? Kinda the opposite of what I'd expect, IOW - thinking the DS @ 3.2V would only be passing maybe 8-10ish mA.

:thinking: Very efficient, if so!

 

[Curious_character]

Re: Partial Fenix E01 driver data

I received my two HA Nat Fenix E01s today. I have not dissambled a head (yet), but I measured the power going in to them and the voltage to the LED when in regulation.

Here is the data:

Head #1            Head #2        
Vin(V)    Iin(mA)    Pin(mW)    Vin(V)    Iin(mA)    Pin(mW)
1.7    65.0    110.5    1.7    81.2    138.0
1.6    67.8    108.5    1.6    87.4    139.8
1.5    69.0    103.5    1.5    88.4    132.6
1.4    73.5    102.9    1.4    93.0    130.2
1.3    77.2    100.4    1.3    96.8    125.8
1.2    79.8    95.8    1.2    102.9    123.5
1.1    84.4    92.8    1.1    108.5    119.4
1.0    92.2    92.2    1.0    123.0    123.0
0.9    59.5    53.6    0.9    72.3    65.1
0.8    26.3    21.0    0.8    25.3    20.2
0.7    19.6    13.7    0.7    19.1    13.4
0.6    17.5    10.5    0.6    17.0    10.2
0.5    14.9    7.5    0.5    14.7    7.4
0.4    11.8    4.7    0.4    11.7    4.7
0.3    8.4    2.5    0.3    8.5    2.6
0.2    5.2    1.0    0.2    5.2    1.0
0.1    2.3    0.2    0.1    2.3    0.2

It appears the head drops out of regulation at around 1V.

The voltage measured at both LEDs when in regulation was 3.2V (constant voltage regulation perhaps).

The minimum start up voltage was approx 0.3V ---> WOW!

A constant voltage regulation scheme would explain the different Power In figures above and the different run-times being measured.

If you try regulating the voltage of an LED, you'll have large differences of current from one LED to the next or from one slightly different regulator voltage to the next. You can also get quite a large increase in current as the LED heats up. These are because the current is an exponential function of the voltage across an LED -- tiny differences in voltage result in large variations of current. So any good regulator for an LED (which I'm sure the Fenix is) is a constant current regulator. Even moderate changes in current for one reason or another will cause very little change in the LED voltage. And because of the relatively constant LED voltage, a constant current regulator delivers very nearly constant power to the LED. Your measurements are consistent with a constant current regulator which has approximately the same efficiency over the operating range of 1.0 - 1.7 volts.

c_c

 

[chimo]

Re: Partial Fenix E01 driver data

Nice info, Chimo! :thumbsup:

I wonder if it would be possible to polish smooth the battery tube's negative contact ridge, in order to minimize the wear on the board's contact. I've done that on some E0's (much easier than the E01, it appears), and they seem to be holding up fine with regular use, so far (about a year).

Am I reading correctly that the DS LED's you tested @ 3.2V, are passing between 41-48 mA?? That's pretty amazing... wouldn't that be a (@ spec 20 mA) Vf of around 3.0V? I thought the data sheet lists the DS as 3.6 Vf (and the GS as 3.2V - which looks very close to what you got) ...?? Kinda the opposite of what I'd expect, IOW - thinking the DS @ 3.2V would only be passing maybe 8-10ish mA.

:thinking: Very efficient, if so!

Yes, the DS currents are for a 3.2V drive voltage. I was a little surprised at the differences.

 

[chimo]

Re: Partial Fenix E01 driver data

If you try regulating the voltage of an LED, you'll have large differences of current from one LED to the next or from one slightly different regulator voltage to the next. You can also get quite a large increase in current as the LED heats up. These are because the current is an exponential function of the voltage across an LED -- tiny differences in voltage result in large variations of current. So any good regulator for an LED (which I'm sure the Fenix is) is a constant current regulator. Even moderate changes in current for one reason or another will cause very little change in the LED voltage. And because of the relatively constant LED voltage, a constant current regulator delivers very nearly constant power to the LED. Your measurements are consistent with a constant current regulator which has approximately the same efficiency over the operating range of 1.0 - 1.7 volts.

c_c

Thanks cc, I'm aware of the V-I properties of LEDs and the subsequent drop in Vf as the die temperature rises and have preached the same thing many times over. It is good, however, to remind people regularly of this.

However, I am not as sure as you that the E01 uses a current regulation scheme. Do you have any additional details to share with us?

There have been several runtime plots done now for the E01 (thanks TIN). They show fairly flat regulation with very little initial output drop.

In my experience, with CC regulation, there is ususlly an initial drop in output (usually within the first few minutes until the die temp reaches a steady state). With CV regulation, as the die temp increases but the current will also raise slightly which should help increase output to account for the drop due to the increase in die temp. Now this all relies on reaching a steady state and the cct not going into thermal runaway. These are driven at fairly mild levels.

Did I mention that both my E01 heads have a LED drive voltage of 3.2V? The LED is soldered thru-hole so it can be measured from the bottom of the driver board.

I am not conclusively convinced either way that the E01 uses CV or CC. It will take cracking open and depotting an E01 head, doing some additional measurements and perhaps some reverse engineering to be certain. Cheers,

Paul

 

[streetmaster]

Re: Partial Fenix E01 driver data

So, when are you gonna crack one open?(I don't mean a beer, although that sounds good) I would give it a try, but there would be no sense in me doing it because I wouldn't know what to do with it once it was open.

 

[chimo]

Re: Partial Fenix E01 driver data

So, when are you gonna crack one open?(I don't mean a beer, although that sounds good) I would give it a try, but there would be no sense in me doing it because I wouldn't know what to do with it once it was open.

As soon as that "new light smell" wears off.

I have to go out of town on business next week and I doubt I'll get time this weekend, so it will be a while.

 

[chimo]

Re: Partial Fenix E01 driver data

Well, that new light smell didn't last too long. I heated one of the heads up, pushed out the pill and unpotted it. I took some pics and made some measurements. The LED is getting 25.22 mA. The efficiency was approx 78% with a fresh alkaline - pretty darn good. My wife is using the computer I normally post from so I will post the pics later.

 

[chimo]

Re: Fenix E01 opened up & Driver data

Here are the pics.

The pill

The driver

Inside the head. Note the LED o-ring groove.

Other data:
Resistor in series w/ LED is: 1% 2ohm
Drive current: 25.22mA
Boost freq: 79kHz
No mark on the inductor
Switcher IC top mark is: 3033
Schottky top mark is: SJ
Ceramic cap is under my thumb

 

[warlord]

Re: Fenix E01 opened up & Driver data

Here are the pics.

Other data:
Resistor in series w/ LED is: 1% 2ohm
Drive current: 25.22mA
Boost freq: 79kHz
No mark on the inductor
Switcher IC top mark is: 3033
Schottky top mark is: SJ
Ceramic cap is under my thumb

Interesting. The spec sheet for the led shows the luminous intensity as 31,000-44,000mcd @ 20mA so at 25mA it should be brighter than spec.

Also I'm noticing that the driver seems to have the same components as the E0 but I can't confirm this since I don't have one in front of me and the search function is disabled. Can you point out any differences?

I'm tempted to disassemble mine when it arrives and extend the LED out a bit further probably with a new GS led. Will the o-ring allow the LED to protrude much further?

Thanks for sharing the info and the pics :goodjob:

 

[chimo]

Re: Fenix E01 opened up & Driver data

Interesting. The spec sheet for the led shows the luminous intensity as 31,000-44,000mcd @ 20mA so at 25mA it should be brighter than spec.

Also I'm noticing that the driver seems to have the same components as the E0 but I can't confirm this since I don't have one in front of me and the search function is disabled. Can you point out any differences?

I'm tempted to disassemble mine when it arrives and extend the LED out a bit further probably with a new GS led. Will the o-ring allow the LED to protrude much further?

Thanks for sharing the info and the pics :goodjob:

Unfortunately I don't have an E0 so I don't know about any differences.

I was concerned at first about the LED being too reset, however, I now believe that Fenix may have the LED ideally positioned to redirect the lateral tip radiation to the front off of the reflector. I find the beam of the E01 more resembles a high-power reflectored LED. There is a hotspot and a gradual drop-off of output. With my Arcs, I get output ring bands. This makes items in the darker bands much harder to see.

I'm not sure if the head will allow the LED to be sunken much deeper without a little modification. There is a fair distance from the land on the inside of the head to the beginning of the taper of the reflector.

 

[Blue72]

Re: Fenix E01 opened up & driver data

What did you use to heat up the head?

Also, How did you unpotted the head?

 

[chimo]

Re: Fenix E01 unpotted & driver data

What did you use to heat up the head?

Also, How did you unpotted the head?

I used a heatgun to warm the head and slid the pill out by pushing on the LED.

I used the same heatgun (carefully) and a dental pick to unpot the pill.

Check my thread on unpotting a new ArcAAA for more details on the method.

 

[jirik_cz]

Re: Fenix E01 unpotted & driver data

Great work Chimo :thumbsup: Please can you measure efficiency from zero to 1.7V input voltage? Thanks.

 

[chimo]

Re: Fenix E01 unpotted & driver data

Great work Chimo :thumbsup: Please can you measure efficiency from zero to 1.7V input voltage? Thanks.

I'll be out of town on a business trip next week. I should be able to get to it after I get back. The light is re-assembled but I left it unpotted.

 

[Burgess]

Re: Fenix E01 unpotted & driver data

to Chimo --


Thank you for your time and effort
in bringing us this information.

:twothumbs
_