Notices
Electronics, Lighting, & Gadgets HRM, GPS, MP3, HID. Whether it's got an acronym or not, here's where you'll find discussions on all sorts of tools, toys and gadgets.

LED Taillight Circuit

Old 01-06-20, 08:58 PM
  #1  
UniChris
Senior Member
Thread Starter
 
UniChris's Avatar
 
Join Date: May 2017
Location: Northampton, MA
Posts: 1,778

Bikes: 36" Unicycle, winter knock-around hybrid bike

Mentioned: 15 Post(s)
Tagged: 0 Thread(s)
Quoted: 866 Post(s)
Liked 360 Times in 260 Posts
LED Taillight Circuit

Anyone have an idea of a good current-mode regulator circuit for a taillight? I'm thinking something where I could add an MCU so I could make it steady or blink or better yet to a combo steady + blink, but the main issue is figuring a good efficient current mode regulator to give a few LEDs the right current regardless of battery state of charge.

And yes, it's easy to light an LED with a current limit resistor, but if I'm going to go to the bother of making something, I feel like doing it right, not in a crude way that's either wasteful or varies in brightness with state-of-charge.

Otherwise I'll probably just case mod a few cheap ones that have weak cases as their major design flaw. Or dig out the scope and spend some time figuring out what they're actually doing and how much of the battery they are wasting...
UniChris is offline  
Old 01-06-20, 11:31 PM
  #2  
znomit
Zoom zoom zoom zoom bonk
 
znomit's Avatar
 
Join Date: Sep 2006
Location: New Zealand
Posts: 4,408

Bikes: Giant Defy, Trek 1.7c, BMC GF02, Fuji Tahoe, Scott Sub 35

Mentioned: 6 Post(s)
Tagged: 0 Thread(s)
Quoted: 457 Post(s)
Liked 560 Times in 284 Posts
LM317 makes for a cheap and easy constant current supply...

LM317 constant-current power supply | LEDnique
znomit is offline  
Old 01-07-20, 09:20 AM
  #3  
VegasTriker
Senior Member
 
Join Date: Feb 2006
Location: Sin City, Nevada
Posts: 2,633

Bikes: Catrike 700, Greenspeed GTO trike, , Linear LWB recumbent, Haluzak Horizon SWB recumbent, Balance 450 MTB, Cannondale SM800 Beast of the East

Mentioned: 18 Post(s)
Tagged: 0 Thread(s)
Quoted: 459 Post(s)
Likes: 0
Liked 160 Times in 124 Posts
There is a flaw in your thinking and that is someone seeing your "brake light" is going to realize it means you are stopping or slowing. We have come to recognize brake lights on a vehicle because they are standard (located in the same place) so when you are behind a car or truck and see the brake light illuminated you know that the driver is slowing or stopping. No so for a bicycle. Virtually no driver or bike rider is going to correctly interpret the meaning of your light. IMHO brake lights or directional lights on a bike are useless. They are not big or bright enough in daytime and for signals, not far enough apart to be effective.
VegasTriker is offline  
Likes For VegasTriker:
Old 01-08-20, 03:39 PM
  #4  
noglider 
aka Tom Reingold
 
noglider's Avatar
 
Join Date: Jan 2009
Location: New York, NY, USA
Posts: 40,135

Bikes: 1962 Rudge Sports, 1971 Raleigh Super Course, 1971 Raleigh Pro Track, 1974 Raleigh International, 1975 Viscount Fixie, 1982 McLean, 1996 Lemond (Ti), 2002 Burley Zydeco tandem

Mentioned: 493 Post(s)
Tagged: 0 Thread(s)
Quoted: 6992 Post(s)
Liked 1,713 Times in 1,067 Posts
I agree with VegasTriker. I've thought about turn signals and brake signals. I would get one if I thought it would be useful.
__________________
Tom Reingold, tom@noglider.com
New York City and High Falls, NY
Blogs: The Experienced Cyclist; noglider's ride blog

“When man invented the bicycle he reached the peak of his attainments.” — Elizabeth West, US author

Please email me rather than PM'ing me. Thanks.
noglider is offline  
Old 01-08-20, 08:51 PM
  #5  
steelbikeguy
Senior Member
 
steelbikeguy's Avatar
 
Join Date: Feb 2008
Location: Peoria, IL
Posts: 3,272
Mentioned: 72 Post(s)
Tagged: 0 Thread(s)
Quoted: 1254 Post(s)
Liked 1,572 Times in 791 Posts
There is a basic two transistor current regulator that gets used in a ton of places. I've used it in tail lights before... although I can't find a quick schematic of it.

Instead, I do have a schematic where I used it in a proposed modification to a B&M taillight. The change replaces a series resistor and LED with the regulator and two series LEDs.



Steve in Peoria
steelbikeguy is offline  
Old 01-09-20, 12:36 AM
  #6  
gilesa
Senior Member
 
Join Date: Feb 2018
Posts: 88
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 41 Post(s)
Likes: 0
Liked 12 Times in 9 Posts
Some rather random thoughts on this:

I did not see the OP mention brake lights. Perhaps those replies were intended for another thread.

As a linear regulator is really a controlled, variable series resistor, can any meet the requirement for
"not in a crude way that's either wasteful or varies in brightness with state-of-charge"? Switch-mode?

Is constant-current with batteries really a good idea? It implies rapid dimming on charge exhaustion.
I have a decently-bright headlamp that runs from a single AA and goes very dim in less than a minute on NiMh.
I learned to carry a spare cell and fitted an extra "be seen" lamp.

A series resistor is not necessarily inefficient. Two NiMh cells (2.4V) and a 1.9V red LED is a popular choice
with almost 80% efficiency. That is probably about the best that low-voltage switch-mode without an active
(MOSFET) freewheel diode can do.

Any circuit choices should probably start by fixing the battery choice and series LED count. Or vice-versa.
In other words, the OP's question seems too general to answer.

A microcontroller seems over-elaborate unless you want to send messages in Morse to people behind you.
If chosen, it could also be a controller for a custom switch-mode regulator.

A pulsing rear light sounds like a nice combination of the properties of steady and blinking ones, but
I think I have never seen one. That seems like a reason for a custom build. Perhaps unfortunately illegal in many jurisdictions?

Last edited by gilesa; 01-09-20 at 12:53 AM. Reason: Restore intended spacing
gilesa is offline  
Old 01-09-20, 11:18 AM
  #7  
steelbikeguy
Senior Member
 
steelbikeguy's Avatar
 
Join Date: Feb 2008
Location: Peoria, IL
Posts: 3,272
Mentioned: 72 Post(s)
Tagged: 0 Thread(s)
Quoted: 1254 Post(s)
Liked 1,572 Times in 791 Posts
Originally Posted by gilesa View Post
Some rather random thoughts on this:
As a linear regulator is really a controlled, variable series resistor, can any meet the requirement for
"not in a crude way that's either wasteful or varies in brightness with state-of-charge"? Switch-mode?

....
as with any design, one must look at the requirements in order to determine the optimum solution.
A big consideration for this sort of application is the voltage range that the circuit will be subjected to.
If the voltage is tightly regulated, or if the voltage source doesn't vary much, then there isn't much current regulation required. A resistor might perform just as well as a regulator.

If the voltage varies quite a bit, then the designer has to worry about things like power dissipation and how well the regulator operates over the entire voltage range. I've designed voltage regulators that had to operate from about 5V to 80V. A linear regulator dissipated a lot of power with the 80V input, so a switcher was an attractive solution.

There is also the consideration of how much voltage the circuit needs to operate. As a current limiter, the resistor will operate even with a tiny voltage drop across it.
By comparison, the two transistor regulator that I showed will need about 0.65V across the sense resistor and about a minimum of 0.2V across the upper transistor's collector-emitter.
The option of using a LM317 as a current regulator requires even more voltage. As a voltage regulator, I think it requires 2 or 2.5V of headroom. I'm guessing that it will need a bit more when used as a current regulator.
The necessary voltage for a switcher to operate will depend on the specific device. Some can operate at very low voltages... I've used a Zetex boost converter, the ZXSC310, to drive a white LED. It can operate down to 0.8V, which makes it an excellent choice for operating from a single NiMH or alkaline cell.

I should mention that there are other ways to make linear regulators.. especially one that uses a voltage reference, an op-amp, and a pass transistor. The drop-out voltage is only the voltage drop across the pass transistor, which could be tenths of a volt. Not a bad option, but you do have more complexity, cost, parts cost, and concerns about oscillation, than you have with the two transistor design.

Steve in Peoria
steelbikeguy is offline  
Old 01-10-20, 05:29 AM
  #8  
gilesa
Senior Member
 
Join Date: Feb 2018
Posts: 88
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 41 Post(s)
Likes: 0
Liked 12 Times in 9 Posts
Steve,

Do you have any advice on hand-soldering those tiny switch-mode parts? I bought a similar part (ZXLD383)
and dual MOSFETs in that tiny TSOT package, but found the soldering very tricky. Or do you have a way to
use conventional surface-mount technique?

Thanks, Giles
gilesa is offline  
Old 01-10-20, 10:55 AM
  #9  
steelbikeguy
Senior Member
 
steelbikeguy's Avatar
 
Join Date: Feb 2008
Location: Peoria, IL
Posts: 3,272
Mentioned: 72 Post(s)
Tagged: 0 Thread(s)
Quoted: 1254 Post(s)
Liked 1,572 Times in 791 Posts
Originally Posted by gilesa View Post
...Do you have any advice on hand-soldering those tiny switch-mode parts? I bought a similar part (ZXLD383)
and dual MOSFETs in that tiny TSOT package, but found the soldering very tricky. Or do you have a way to
use conventional surface-mount technique?
My method for hacking together a board with surface mount parts is not for the faint of heart or those with a weak stomach!

I've had some success by taking a copper clad board and using a dremel tool (and cut-off wheel) to cut slots in the copper. This creates pads for mounting the parts and can sometimes serve to connect components. For simple SOT-23 parts, such as transistors, this is reasonable.
For the 5 terminal SOT-23 package of the Zetex boost converter, I have to carefully lift one pin and solder a wire to it (using 30 ga. wire wrap wire).

As an example, here are some photos of a headlight that I built that uses the Zetex boost converter for the standlight.
This is the board where I've built up the full wave rectifier, but have only sketched out the rest of the circuitry....



I should have cut all of the slots in the copper first.... it's hard to cut slots right next to parts that are already soldered.
Here's a photo of the completed board.....



Once I've got the board checked out and working, I cover it with a layer of Plasti-Dip. This serves two functions... it makes the board resistant to moisture, and it holds the wires in place, which keeps them from fatiguing from vibration and breaking.

Making a proper printed circuit board is a much cleaner method and allows you to use parts with fine lead pitches, but a hacked up copper clad board is a quick way to experiment and throw something together.
Just don't let anyone see it!

Steve in Peoria
steelbikeguy is offline  
Old 01-11-20, 04:41 AM
  #10  
gilesa
Senior Member
 
Join Date: Feb 2018
Posts: 88
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 41 Post(s)
Likes: 0
Liked 12 Times in 9 Posts
Thanks. That is an interesting technique, but looks to need a very steady hand on the Dremel!
I will try it for myself next time. It seems to work well enough to use SMT parts for everything,
while so far I use them only when forced to.

My current attempt has the SOT parts glued to the board, legs up,
and wiring carried on pads. Described as dead-bug/Manhattan style somewhere on the web.
My big problem was persuading the solder to wet the SOT package pins when attaching wires.
I think your technique would help there by immersing them in a tiny pool of molten solder.
gilesa is offline  
Old 01-11-20, 09:52 AM
  #11  
steelbikeguy
Senior Member
 
steelbikeguy's Avatar
 
Join Date: Feb 2008
Location: Peoria, IL
Posts: 3,272
Mentioned: 72 Post(s)
Tagged: 0 Thread(s)
Quoted: 1254 Post(s)
Liked 1,572 Times in 791 Posts
Originally Posted by gilesa View Post
Thanks. That is an interesting technique, but looks to need a very steady hand on the Dremel!
I will try it for myself next time. It seems to work well enough to use SMT parts for everything,
while so far I use them only when forced to.
there are so many parts that are only available in surface mount packages that it's good to have a way use them.
Of course, the inherent small size comes in handy for bike lights!

Originally Posted by gilesa View Post
My current attempt has the SOT parts glued to the board, legs up,
and wiring carried on pads. Described as dead-bug/Manhattan style somewhere on the web.
My big problem was persuading the solder to wet the SOT package pins when attaching wires.
I think your technique would help there by immersing them in a tiny pool of molten solder.
Dead-bug is a well established technique, but it is kinda messy and hard to keep track of what the parts are.
Plus, it's harder to change parts when you damage them.
I should mention that there are little SMD prototyping boards available. I've used the "Surf-boards" from Capital Advanced. These are small boards with pads for SOIC parts or other packages, along with some areas for mounting other basic parts. Very handy stuff! I often glue or solder them to my primary copper clad board.
These Surf-boards can be purchased from the usual suppliers. Info at www.capitaladvanced.com

here's a headlight circuit where I used a SOIC part mounted to a cut-down Surf-board....



Steve in Peoria
steelbikeguy is offline  
Old 01-31-20, 01:36 AM
  #12  
ChrisAlbertson
Senior Member
 
Join Date: Dec 2019
Location: Southern California
Posts: 158

Bikes: 70's frame, newer parts

Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 91 Post(s)
Likes: 0
Liked 36 Times in 26 Posts
Originally Posted by UniChris View Post
Anyone have an idea of a good current-mode regulator circuit for a taillight? I'm thinking something where I could add an MCU so I could make it steady or blink or better yet to a combo steady + blink, but the main issue is figuring a good efficient current mode regulator to give a few LEDs the right current regardless of battery state of charge.

And yes, it's easy to light an LED with a current limit resistor, but if I'm going to go to the bother of making something, I feel like doing it right, not in a crude way that's either wasteful or varies in brightness with state-of-charge.

Otherwise I'll probably just case mod a few cheap ones that have weak cases as their major design flaw. Or dig out the scope and spend some time figuring out what they're actually doing and how much of the battery they are wasting...

The MCU controlled PWM is the current regulator. This method is the best in terms of battery life. The MCU controls a MOSFET gate and the MOSFET is either on or off. You control current by controller the duty cycle.

The suggestion to use the LM317 was a good idea in 1990 but in 2020 linear regulators are considers power wasters. If it needs a heat sink you don't want it. Use an MCU analog pin to monitor battery voltage or use a DD/DC supply to make the volts from the battery constant.

Look at places like Digikey there str specialized parts that are made just for this exact purpose, controlling a LED with PWM. These are much easier to use than a generic MOSFET. Lookuop "led driver" as they can be directly controlled by a 3.3 volt uP pin.

For high power LEDs you place a temperature sensor on the LED's heat sink and software does not let the heat sink go over some limit.
ChrisAlbertson is offline  
Old 02-02-20, 05:03 AM
  #13  
gilesa
Senior Member
 
Join Date: Feb 2018
Posts: 88
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 41 Post(s)
Likes: 0
Liked 12 Times in 9 Posts
>Look at places like Digikey there are specialized parts that are made just for this exact purpose, controlling a LED with PWM. These are much easier to use than a generic MOSFET. Lookuop "led driver" as they can be directly controlled by a 3.3 volt uP pin.

I am not sure that is really true. When I looked, I found almost all switch-mode controllers are designed for constant-voltage output, and most others are for higher input voltages, apparently aimed at mains-powered lighting. I think I found one or two suitable controllers from each of Maxim and Linear Technology, plus the Zetex chips that are more suitable for a dynamo standlight. I also bookmarked this https://www.maximintegrated.com/en/d...es/4/4328.html which might be a perfect solution to OPs question, but two-chip.

And I still see little case for a microcontroller.
gilesa is offline  
Old 02-02-20, 11:31 AM
  #14  
ChrisAlbertson
Senior Member
 
Join Date: Dec 2019
Location: Southern California
Posts: 158

Bikes: 70's frame, newer parts

Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 91 Post(s)
Likes: 0
Liked 36 Times in 26 Posts
Originally Posted by gilesa View Post
>Look at places like Digikey there are specialized parts that are made just for this exact purpose, controlling a LED with PWM. These are much easier to use than a generic MOSFET. Lookuop "led driver" as they can be directly controlled by a 3.3 volt uP pin.

I am not sure that is really true. When I looked, I found almost all switch-mode controllers are designed for constant-voltage output, and most others are for higher input voltages, apparently aimed at mains-powered lighting. I think I found one or two suitable controllers from each of Maxim and Linear Technology, plus the Zetex chips that are more suitable for a dynamo standlight. I also bookmarked this https://www.maximintegrated.com/en/d...es/4/4328.html which might be a perfect solution to OPs question, but two-chip.

And I still see little case for a microcontroller.
The use for the microcontroller is to allow different modes, like "daylight flash" or "solid on with pulses" and you want it all controlled with one button. And while you are at it the controller can act as a battery meter and can refuse to do the high power modes on a low battery. I doubt you'd find a commercial LED bike light today that does not use a microcontroller to send PWM to a MOSFET. No one uses analog control as it wastes battery power.

I Googled for an example of this and found "Step #8 " on this page that does all this for about $3.
https://www.instructables.com/id/Cir...h-Power-LED-s/

The biggest cost item with a DIY bike light is going to be the battery and the charger. For any given power and reasonable (few hurs) run time the battery will cost more than the LED and the reflector. THose cheap eBay 18650 LiPo "6000 ma" batteries are all fake and have only 1000 ma or so. You have to buy name brand for $10+ each to get no more than 3000ma.

The trouble is the cost and Hassel of designing just one light might not be worth it. Maybe a better way to get a custom light is to disassemble a new light that you bought on sale at REI.
Then use those parts in yu custom housing.
ChrisAlbertson is offline  

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off


Thread Tools
Search this Thread

Contact Us - Archive - Advertising - Cookie Policy - Privacy Statement - Terms of Service - Do Not Sell My Personal Information -

Copyright © 2021 MH Sub I, LLC dba Internet Brands. All rights reserved. Use of this site indicates your consent to the Terms of Use.