TriLux Bike Light

Sawtooth

Newly Enlightened
Joined
Apr 16, 2006
Messages
104
Location
New Mexico
I finished this mod awhile ago. Guess I never got around to posting final pictures.
Since I finished this one, my wife wanted one too so I built a second one with some
improvements...

1) In the second mod, the pot is a 5K linear taper with switch.
2) Added an internal 15-turn trimpot to convert the linear taper to more of a log taper.
3) I added the recommended 220 microF cap across the input to the buckpuck.
Produces far less noise when listening to FM stations...

Otherwise the second light is the same. I'll probably add these features to the first light too.

One other thing I added to the first version is some reflective material. I used SOLAS grade silver reflective material to the exterior of the water bottle battery holder. And I added diamond grade white reflective material to the part of the PTS-D heatsink that can be seen from the exterior of the light. These felective materials are amazingly efficient and really enhance visibility at night. I'll post some pictures soon...
 

Sawtooth

Newly Enlightened
Joined
Apr 16, 2006
Messages
104
Location
New Mexico
Here is a beamshot comparison I did awhile back. The stock mag probably didn't
have fresh baterries, but after some runtime that is always going to be the case.

Beam1.png
 

andrewwynn

Flashlight Enthusiast
Joined
Apr 28, 2004
Messages
3,763
Location
Racine, WI USA
that is just amazing.. even though i make the same kind of lights.. it never fails to amaze me how little light the likes of a FIVE D mag has!

you will be very happy if you tweak that buck puck with an audio-taper pot vs the linear one.. the.. 15-turn trimmer won't fix the linear/log problem.. but it will make it much more useful.. if you have the trimmer in parallel with your main pot you can trim it to the point where it JUST reaches max output.. and in addition.. either figure out what resistance (i.e. 470ohm) is required in series with the main pot to set the minimum output.. i'm working on a light almost identical do this for my brother.. where i have the bare min set to 150mA.. it seems kinda high.. but that also includes the current going to the FAN!

(i have a fan+series resistor in parallel with the 3-lux... it's a 12V fan for the likes of a computer CPU... so it runs not at full speed but 'just enough'.. it also runs slower when the light is turned down..

(of course this particular light is not environmentally sealed.. it's a work-light, not meant for outside use.. at least not if it's raining!)

I have another similar puck-buck solution that has 4 emitters and IS environmentally sealed.. it uses a much bigger heat-sink that is vented to the outside even though the heat-sink is internal.. the parts inside are sealed so it can be used in the weather!

-awr
 

Sawtooth

Newly Enlightened
Joined
Apr 16, 2006
Messages
104
Location
New Mexico
Here are some additional photos:
dscf5165.jpg

dscf5168.jpg

dscf5171.jpg

dscf5173.jpg

dscf5175.jpg

dscf5178.jpg


Although I wouldnt rate the light IP69 (68?), I am confident it
can handle the weather just fine :naughty::
dscf5179.jpg

dscf1579.jpg



dscf1579.jpg
 
Last edited:

Sawtooth

Newly Enlightened
Joined
Apr 16, 2006
Messages
104
Location
New Mexico
Andrew: The 15-turn pot is not in parallel with the ext. pot. I have arranged
them in in such a way that I am getting a log-like taper out of the pair.

I have tried a number of different pots including: multi-turn precision wire-wound, conductive plastic, and carbon (film?). Some log, some linear.

My initial thinking was that I would get a sealed conductive plastic log taper pot ganged with a >1A switch. Easier said than done! Although you can get almost anything if you buy enough of them, I could not find what I wanted in single quantities... Very annoying! :sigh:

Also, many of the so-called log-taper pots are actually just two linear segments combined. And you really notice it when turning the pot up.

My solution was to stick with linear for the external, but convert it to an approximate log-like taper with another trimpot. It actually works alot better than any of the log-tapers I tried. I'll post a diagram of the circuit when I get a chance... (Its not a true log taper, but it is smooth and gives alot of dynamic range to the intensity.)
 

Lumbee1

Newly Enlightened
Joined
Jun 21, 2006
Messages
105
Sawtooth, so what's the deal with two handlebar lights? Are both T-bins? Do you have any beamshots with both lit?
 

andrewwynn

Flashlight Enthusiast
Joined
Apr 28, 2004
Messages
3,763
Location
Racine, WI USA
Unless you came up with some circuit magic.. you are still just using a linear solution... but what i think you did.. was fine-tune the set-points so that the 15-turn pot eliminates the wasted portion of the dial so it 'gets right to business'.. i would like to see the circuit though to see what you came up with!

®S sells a 10k audio taper pot with a switch.

You are 100% correct that most non-linar pots are actually two linear segments bent in the middle.. the do an ok job and that's the exact type i have.. usually called 'audio taper'.. far easier to fabricate than an actual log taper.. especially if using a stereo model!

The key to it working at least mostly smoothly is that the 'low end' of the scale will get you some precision.. i.e. you need to have control where the resistance is the highest.. the other end.. you need to change the resistance quickly to see any difference .. so on a linear scale what you will see depending on how you wire it.. is this..

A) very little change on the bright side.. and when it starts to dim.. it goes to off almost immediately.. it's very easy to dial in a precise level for the high.. impossible for the low.

B) very slow changes on the dim side.. and when it starts to get bright.. gets to high very quickly.. very hard to adjust say between 800 and 900 mA.. very easy for 100-200.. obviously this is the way to have it set up.

if you have a linear pot.. that is 'trimmed' properly.. you will be able to smoothly dial anywhere in between.. no 'tricky side'... however.. what you will see.. is that you have to dial most of the dialing to get from high to medium.. and much less dialing from medium to off.. because of the way apparent brightness is a log relation to mA.. you want to have a pot so that dialing from 0 to 10 on a dial.. if you are at 5.. the output will be still only like 15% not 50%.. it makes the linear dial change to an exponential change in output so that the 'brightness' follows a linear path.

Image-33EC322A454511DB.jpg

On a related topic.. here's my latest creation.. a low-volt cutout for the buckpuck.. 'take two'..

Of interest in this picture:

on the bottom right of the pot.. two white 30ga wires.. they go to a 6.65K resistor in parallel with the 10k pot.. the value was determined by first using a 100k pot, and dialing the pot to max.. and dialing down the 100k pot 'til i can start to see dimming and backing off slightly.. this trims the 'high end' of the pot to match the particular circuit.

follow the blue wire from the wiper on the pot.. it goes to a tiny white resistor that is soldered to the control pin on the buck puck.. this resistor is a 470ohm resistor.. also determined initially by using a trimpot (ala your 15 turn model).. 2k.. and adjusted for in my case.. 130-150mA.. (i need an extra 30-50mA for my fan).. this would normally be a little lower resistance.. the buck puck can stay stable as low as 50mA maybe less.. i think i use 40-50mA for low on megasonic.. but the influence of the fan.. it wont run stable lower than about 150 (could also be from my accident that killed the 5V reference pin on the buck puck).

On the far left.. soldered by the Vin pin to the switch 'out'.. you can clearly see the hole in the heat-sink of the replacement 5V regulator.. that i use to run the low-volt cutout.. which is that tiny tiny circuitboard on top of the buckpuck.

The other blue wire that goes to the control pin.. (behind the flagg-soldered resistor).. is the output from the LVC circuit.. it goes 'high' and applies 5V to the control pin when Vbat goes low.. if it's attached on the other side of the resistor.. rather than go 'off' it goes to 'low' when bat is dead.. nicer for bike-riding!

I would actually have a second resistor.. maybe 200-300ohm so i could have a manual low of 100mA but when vBat goes low.. it will go to 40-50mA.. this will be just enough light to come to a stop should the bat die while riding.

The beauty of this type of setup.. is that when vLow is hit.. the light shuts off and the bat. will recover.. if you had the light on high-beam.. you can go to 'low' and get an hour more light if you have a big bat pack (mine on megasonic is 60WH).. i get 5 days or something on low beam!

did you mention what if anything you are using to protect your batteries from over discharge in your circuit?

-awr
 

Sawtooth

Newly Enlightened
Joined
Apr 16, 2006
Messages
104
Location
New Mexico
lumbee1: 2 heads (er lights) are better than 1 for light output. Both are T-bins. I have tried
a UX0J Lux III in a different project and it actually doesnt even seem as bright
as some of the T-bins I have used. Don't know whats up with that, but I think
the T-bins I have are not much dimmer than much more expensive U-bins..
I will try to get some beam comparisons with both lights...

Long John: Thanks!

andrewwynn: You can indeed obtain non-linear log-like tapers with just 2 linear pots.
Here is a quick blurb I hacked together...
http://cybermesa.com/~nhenderson/Resist.pdf
 

Sawtooth

Newly Enlightened
Joined
Apr 16, 2006
Messages
104
Location
New Mexico
Another comment on log-tapers....

It seems that there really isn't a standard "log-taper" in the following senses:

(1) Many log tapers are really just piece-wise linear.
(2) The actual curves can vary quite a bit between different pots.

Log-tapers also seem to be much more limited in terms of availability and features like
ganged switches, etc...

Finally, although human sensitivity to sound and light are logarithmic,
this well-known fact doesnt necessarily make a purely logarithmic taper the best
solution. If you want your eyes to perceive a faster than linear increase in intensity, you
need something a bit different from a pure log-taper. I kind of like being able to fine-tune the taper with the trimpot.
 

andrewwynn

Flashlight Enthusiast
Joined
Apr 28, 2004
Messages
3,763
Location
Racine, WI USA
well hot dawg that is just really damn cool. adjustable bend to the taper.... i think i'll have to play with that for my next buck-puck solution... here's another tip though.. the buck puck really only adjust voltage between 1.6V and 4.0V.. (i.e. below 1.6 it's full power.. above 4.0 it's off).. that's just aproximate but that's pretty close.. so you'd want to tweak the top and bottom of the scale..

The input resistance is really low on that control pin.. maybe only 2k.. it really threw off my attempts to run it from an FET switch was driving me crazy! It could just be on my blown buckpuck.. however.. considering that the stock solution works with only a reisistor from Ref to control it has to be pretty low resistance for 5Kohm to be effective.

That's a really creative solution how'd you come up with that?

I didn't go through the details of reproducing this with forumulas but i can only imagine you made the chart by plotting the formulas.. now.. not quite sure what the value for the trimmer pot is.. you have between 1 and .05.. is that Kohm?

Now i want to make another buck-puck solution thanks for introducing us to this neat 'better than audio' taper.. that's way better because the 'audio' taper are not remotely 'log' (or exponential).. these curves are way better.. especially when you can tweak how fast the rate changes to suit yourself.

boy do i love being proved wrong.. because usually it means something reaally cool comes out of it!

-awr
 
Last edited:

Sawtooth

Newly Enlightened
Joined
Apr 16, 2006
Messages
104
Location
New Mexico
The "b" value is the fraction of the R value. I.e., I use R=5K Ohms, so b=0.1 would be
500 Ohms. I am actually not sure what I am using though - never measured it diectly.
The trimpot I use is a 10K 15-turn. I just tweaked it until I got something I liked.

Very useful info on the input ranges. I suspected as much. I'll probably add some additional tweaking on the ends in future...
 

andrewwynn

Flashlight Enthusiast
Joined
Apr 28, 2004
Messages
3,763
Location
Racine, WI USA
fortunately.. with being able to bend that curve enough.. you can get something very useful.. it does seem that it's 'upside down' though now that i'm looking at it more closely.. the 'fine tuning' or more shallow ramp would be more useful on the 4V side where you are trying to make small adjustments like from 50mA to 100mA... and on the high-output-side.. it would be ok if it swings past 1.6V quickly. Let us know if you end up tweaking the ends and what setup you use.. less work for us :-D

also.. i didn't get an answer from the low-voltage-cutout question.. are you using any protection against over discharge, since with buck or boost topology you risk battery damage if you try to use 'dimming of the light' to detect a dead battery..

check out this thread

(maybe i already linked it earlier.. oops if so)..

-awr
 

Sawtooth

Newly Enlightened
Joined
Apr 16, 2006
Messages
104
Location
New Mexico
andrewwynn said:
fortunately.. with being able to bend that curve enough.. you can get something very useful.. it does seem that it's 'upside down' though now that i'm looking at it more closely.. the 'fine tuning' or more shallow ramp would be more useful on the 4V side where you are trying to make small adjustments like from 50mA to 100mA... and on the high-output-side.. it would be ok if it swings past 1.6V quickly. Let us know if you end up tweaking the ends and what setup you use.. less work for us :-D

also.. i didn't get an answer from the low-voltage-cutout question.. are you using any protection against over discharge, since with buck or boost topology you risk battery damage if you try to use 'dimming of the light' to detect a dead battery..

check out this thread

(maybe i already linked it earlier.. oops if so)..

-awr


I think you are right about it being upside down. Potential at C should be V*R2/RT. I'll update the PDF.

Also, about the low voltage cutout. The battery pack has a PCB that is hooked into each cell. It is supposed to protect from under/over voltage and over charge/discharge. I also have a 3A inline fuse in the water bottle close to the battery. The pack is from batteryspace and is made from 4 2400mAh LG 18650's.
 

andrewwynn

Flashlight Enthusiast
Joined
Apr 28, 2004
Messages
3,763
Location
Racine, WI USA
most excellent on the LVC protection.. it's very important.. and the fuse is good but actually probably not even needed.. the prot. ckt ends up going infinitely open if a short happens.. nice pack!

-awr
 
Top