Standlight circuits (Will this circuit work?)

[Steve K]

I'm not fond of the idea of completely discharging the supercap, if only because it takes a while to charge it back up and the light will be dim while it is charging. When the supercap discharges through the LED, it only discharges a few tenths of a volt below the normal operating voltage of the LED. As a side note, this is also why the simple standlight circuit is not a very effective way to use the energy stored in a supercap.

One option would be to add a switch in series with the LED to interrupt the current from the supercap. The drawback would be that bad things could happen if you forget to turn the switch back on. If the switch is at the LED, then there wouldn't be anything to limit the voltage that gets to the supercap and you could damage it. If the switch is at the supercap, then you wouldn't be charging it when you start riding again. Nothing is damaged, but you won't have a standlight when you stop.

Another option would be to throw a bag over the headlight. This would have the effect of protecting the light from the effects of the weather, sun, bird poop, etc., and perhaps also reducing the attractiveness of the bike to thieves.

 

[atari2k6]

[...] The drawback would be that bad things could happen if you forget to turn the switch back on. If the switch is at the LED, then there wouldn't be anything to limit the voltage that gets to the supercap and you could damage it. If the switch is at the supercap, then you wouldn't be charging it when you start riding again. [...]

That's why I was thinking about a momentary switch to discharge the cap, so you won't have to remember to turn the standlight back on.

Thank you for your quick and clear answers!

 

[jdp298]

The Busch & Muller rear toplight turns itself off after 4 mins, but mine wasn't holding enough charge to last beyond 30s, so it had to go.

Letting the lights stay on isn't the biggest problem I've found, and even though my caps will always totally discharge, the inefficiency of them means that as soon as I move, I get decent light ouput. This is because the caps don't act as a 'black hole' until at a certain voltage. If you really need to stop them sooner, a push-to-make switch and a low-ish value resistor (~100 ohms) ought to do it. It also means that you couldn't forget to switch it back on, hence avoiding Steve's problem.

 

[johnnya156]

Hi there,

First of all, apologies for the being a total newb. Also, in the interests of full disclosure, I should probably add that I have a VERY limited understanding of electronics.

I am trying the upgrade an old dynamo headlight that is currently running a 6V 3W halogen bulb. What I'd like to do ideally, is retrofit an LED bulb similar to this:

http://www.ebay.co.uk/itm/252138590419?var=550976278922

... but also install some sort of Standlight Circuitry. I was hoping to build the circuit myself although having read through this thread, I now realise that it is way beyond my abilitly! So I was wondering if someone could point me in the direction of a cicuit design that would fit my needs? And also recommend someone who might build it for me? (I am based in the UK).

Many thanks,
Johnny

 

[Steve K]

I've seen various people use these bulbs and they appear to be happy enough with them. These are 1 watt LEDs, and that always seemed to be a rather modest amount of power, although compared to putting 3 watts into a very inefficient incandescent bulb.. well, maybe it's good enough?

Without knowing what's inside of that Nicelite bulb, I have to assume that it needs roughly 6v to operate. The only way to make it operate as a standlight is to wire a full wave bridge rectifier and supercap between it and the dynamo.

An alternative would be to build a low power standlight that is wired in parallel with the headlight and is powered continuously. That should be more efficient and more predictable, I suspect. It would still need to be packaged and mounted somewhere on the bike. I think that Pawel's circuit...
https://people.nscl.msu.edu/~daniel/back/nntec/standlicht.gif
should be suitable.
(this is from Pawel Danielewitz (sp?), who used to be active on the BikeCurrent list, and still haunts the I-Bob list now and then)

From a practical perspective, the LED headlights nowadays are really pretty good, and there is little advantage to converting an old incandescent light to LED. To do it well requires throwing away essentially all of the original headlight... the optics aren't suitable for the LED, the LED requires heatsinking, and you have to find space to install the electronics. I tried it once and figured out this lesson. I've converted battery powered lights, but only because I liked the mounting hardware and battery enclosure.

 

[johnnya156]

I've seen various people use these bulbs and they appear to be happy enough with them. These are 1 watt LEDs, and that always seemed to be a rather modest amount of power, although compared to putting 3 watts into a very inefficient incandescent bulb.. well, maybe it's good enough?

Without knowing what's inside of that Nicelite bulb, I have to assume that it needs roughly 6v to operate. The only way to make it operate as a standlight is to wire a full wave bridge rectifier and supercap between it and the dynamo.

An alternative would be to build a low power standlight that is wired in parallel with the headlight and is powered continuously. That should be more efficient and more predictable, I suspect. It would still need to be packaged and mounted somewhere on the bike. I think that Pawel's circuit...
https://people.nscl.msu.edu/~daniel/back/nntec/standlicht.gif
should be suitable.
(this is from Pawel Danielewitz (sp?), who used to be active on the BikeCurrent list, and still haunts the I-Bob list now and then)

From a practical perspective, the LED headlights nowadays are really pretty good, and there is little advantage to converting an old incandescent light to LED. To do it well requires throwing away essentially all of the original headlight... the optics aren't suitable for the LED, the LED requires heatsinking, and you have to find space to install the electronics. I tried it once and figured out this lesson. I've converted battery powered lights, but only because I liked the mounting hardware and battery enclosure.

Thanks for coming back so quickly Steve - I had sort of come to the conclusion that from a practical perspective, an off the shelf headlight makes a lot more sense. The only reason I haven't gone down that route (as yet!) is that I quite like the aesthetic of the current fitting - so was hoping I might be able to cobble together a homebrew alternative.

From what I know of the Nicelite bulb, it is indeed 6v, so if I wanted to go down the full wave bridge rectifier and supercap route, would you be able to point me in the direction of a suitable circuit?

 

[Steve K]

Thanks for coming back so quickly Steve - I had sort of come to the conclusion that from a practical perspective, an off the shelf headlight makes a lot more sense. The only reason I haven't gone down that route (as yet!) is that I quite like the aesthetic of the current fitting - so was hoping I might be able to cobble together a homebrew alternative.

From what I know of the Nicelite bulb, it is indeed 6v, so if I wanted to go down the full wave bridge rectifier and supercap route, would you be able to point me in the direction of a suitable circuit?

regarding aesthetics.... I just saw some reference to a nice retro looking LED light that B&M makes. That might be the best of both worlds.

regarding sources of suitable circuits... other than making it myself (I'd have to charge as much as the cost of a B&M light), I don't know of a source. The demand for this sort of thing is rather limited.

 

[kg6gfq]

I don't know this stuff as well as Steve, but a few thoughts:

1) It looks like the Nicelite bulbs are meant to be used with a voltage regulator that they provide. For a really simple standlight, I believe you could rig up a rectifier, and across its output put a supercapacitor rated for 6V in series with a smallish resistor (100 ohms?) in parallel with the light. It's not an efficient use of the capacitor, but it would give you a brief, somewhat dim standlight. A 1-2F capacitor costs just a few dollars, and resistors far less. If you want a brighter standlight, reduce the value of the resistor. For a longer one, increase the resistor value or get a bigger capacitor. The circuit would look something like this:

2) To use the capacitor more efficiently (for a longer, brighter standlight), things get more complicated. You could:
- Use an active switching circuit to charge the capacitor higher, or add a boost converter to run it down further than the LED would on its own. I think this would potentially work with the Nicelite. (Designing this is beyond my current level of comprehension.)
- Ditch the Nicelite and design a simpler circuit based on varying diode voltage drops, as discussed earlier in this thread. That's what I'm working on over here, with much help from Steve and Savvas. I'm planning to write up a guide to this sort of light design, but haven't gotten around to it yet.

As for finding someone to build it for you... I think there's another section of this forum for folks who make custom flashlights, but I get the impression it's more about machining than circuitry. Still, CPF is probably a good place to ask around.

Have you already seen the lighting system built by minisystem (who posted in this thread several years ago) to fit in a vintage S-A housing? Minisystem went on to develop a very elegant retro-styled dynamo headlight prototype, but doesn't appear to have gone into production.

All that said, B&M makes excellent lights. If they have something that suits your needs, that's likely a good option. Designing your own dynamo light is probably only worthwhile if you enjoy this sort of thing. (-:

Whatever you do, good luck!

-Darin

 

[johnnya156]

I've seen various people use these bulbs and they appear to be happy enough with them. These are 1 watt LEDs, and that always seemed to be a rather modest amount of power, although compared to putting 3 watts into a very inefficient incandescent bulb.. well, maybe it's good enough?

Without knowing what's inside of that Nicelite bulb, I have to assume that it needs roughly 6v to operate. The only way to make it operate as a standlight is to wire a full wave bridge rectifier and supercap between it and the dynamo.

An alternative would be to build a low power standlight that is wired in parallel with the headlight and is powered continuously. That should be more efficient and more predictable, I suspect. It would still need to be packaged and mounted somewhere on the bike. I think that Pawel's circuit...
https://people.nscl.msu.edu/~daniel/back/nntec/standlicht.gif
should be suitable.
(this is from Pawel Danielewitz (sp?), who used to be active on the BikeCurrent list, and still haunts the I-Bob list now and then)

From a practical perspective, the LED headlights nowadays are really pretty good, and there is little advantage to converting an old incandescent light to LED. To do it well requires throwing away essentially all of the original headlight... the optics aren't suitable for the LED, the LED requires heatsinking, and you have to find space to install the electronics. I tried it once and figured out this lesson. I've converted battery powered lights, but only because I liked the mounting hardware and battery enclosure.

 

[johnnya156]

I don't know this stuff as well as Steve, but a few thoughts:

1) It looks like the Nicelite bulbs are meant to be used with a voltage regulator that they provide. For a really simple standlight, I believe you could rig up a rectifier, and across its output put a supercapacitor rated for 6V in series with a smallish resistor (100 ohms?) in parallel with the light. It's not an efficient use of the capacitor, but it would give you a brief, somewhat dim standlight. A 1-2F capacitor costs just a few dollars, and resistors far less. If you want a brighter standlight, reduce the value of the resistor. For a longer one, increase the resistor value or get a bigger capacitor. The circuit would look something like this:

2) To use the capacitor more efficiently (for a longer, brighter standlight), things get more complicated. You could:
- Use an active switching circuit to charge the capacitor higher, or add a boost converter to run it down further than the LED would on its own. I think this would potentially work with the Nicelite. (Designing this is beyond my current level of comprehension.)
- Ditch the Nicelite and design a simpler circuit based on varying diode voltage drops, as discussed earlier in this thread. That's what I'm working on over here, with much help from Steve and Savvas. I'm planning to write up a guide to this sort of light design, but haven't gotten around to it yet.

As for finding someone to build it for you... I think there's another section of this forum for folks who make custom flashlights, but I get the impression it's more about machining than circuitry. Still, CPF is probably a good place to ask around.

Have you already seen the lighting system built by minisystem (who posted in this thread several years ago) to fit in a vintage S-A housing? Minisystem went on to develop a very elegant retro-styled dynamo headlight prototype, but doesn't appear to have gone into production.

All that said, B&M makes excellent lights. If they have something that suits your needs, that's likely a good option. Designing your own dynamo light is probably only worthwhile if you enjoy this sort of thing. (-:

Whatever you do, good luck!

-Darin

Thanks Darin, the ciruit you sent is exactly the sort of thing I was after (I even vaguely understand how it works!). If I was to attempt building it myself, what's the simplist construction method? (I am actually reasonably competant with a soldering iron).

Thanks also for sending the link to the S-A upgrade. It's a pity it doesn't seem to have made it past the prototype stage. Very cool nonetheless.

 

[Steve K]

I don't know this stuff as well as Steve, but a few thoughts:

1) It looks like the Nicelite bulbs are meant to be used with a voltage regulator that they provide. For a really simple standlight, I believe you could rig up a rectifier, and across its output put a supercapacitor rated for 6V in series with a smallish resistor (100 ohms?) in parallel with the light. It's not an efficient use of the capacitor, but it would give you a brief, somewhat dim standlight. A 1-2F capacitor costs just a few dollars, and resistors far less. If you want a brighter standlight, reduce the value of the resistor. For a longer one, increase the resistor value or get a bigger capacitor. The circuit would look something like this:

<..snip...>
-Darin

that seems plausible.... the biggest unknown is the performance of the regulator that the Nicelite folks sell. The critical thing will be to keep the voltage down to the level that won't damage the supercap.
The regulator is likely just two large zener diodes connected in series, anode to anode, or cathode to cathode. It's not a regulator per se, but a reasonable voltage clamp.

There aren't a lot of 6V super caps. I did find one on Digi-key that is rated for 6.3VDC and is 1F.
http://www.digikey.com/product-detail/en/DK-6R3D105T/604-1018-ND/970179
It is a bit over $4.

It still seems like a lot of work to make a 1 watt LED headlight with compromised optics. Almost anything that you buy should perform better.

 

[Savvas]

Darin,

I think earlier in the thread you may have mentioned difficulty finding suitable supercaps. There are plenty available on the 'bay and elsewhere. Search for '1F 5.5 super cap'. Put in Elna, NEC or Panasonic if you like. A 1F super cap works fine. Most are 5.5v which just need a series diode to reduce voltage a bit.

Personally I have found the Nicelite bulbs a bit feeble. I tried them in a relatively well designed Union HS-3 headlight (the state-of-the-art dynamo light before the SON E6 came along) and wasn't particularly impressed. I think a couple of XR-Es in series with some decent Ledil optics would do better. If you are set on using your old light enclosure, Minisystem has posted some excellent work in converting his light. I believe he used his own circuit and installed a separate optic inside the original reflector. Search for his thread from last year (I think).

If all you want is a stand light to use with your original light (putting aesthetics aside), maybe install the Nicelite bulbs front and rear and then combine with a couple of the excellent Smart lights using AAA Eneloops. This is the approach I have used on my somewhat neglectful son's bike (which he rides to university) and it's proved both reliable and cheap to run.

Savvas.

 

[johnnya156]

Hi guys,

Thanks for all the feedback. I think I'm going to have an experiment with the Nicelight buib/regulator combo.... And then think about adding the standlight circuit if I'm happy wth the results.

Assuming, I do go ahead and construct the ciruit, would someone mind explaining to me what my options are in terms of circuit construction methods?

Thanks,
Johnny

 

[Steve K]

my personal preference of construction technique is to cut up some copper-clad circuit board. I use a dremel tool with the cut-off wheel to make slots in the copper layer, which produces pads for mounting the surface-mount components. If I do a good job of the layout, it also provides the connections between most of the components. I use 30 ga wire-wrap wire to make the rest of the connections.

I've got pictures of my latest bike headlight that gives a hint at the process....
https://www.flickr.com/photos/kurtsj00/albums/72157649424813991

I sketch out the layout on the copper using a pencil, and mark the cuts in Sharpie prior to getting the dremel out.
I don't show it explicitly, but I usually coat the board with Plasti-Dip. This provides a moisture barrier, but more importantly, supports the thin wires so they won't vibrate, fatigue, and break.

A few thin coats of Plasti-Dip can do a tolerable job of protecting the board against the elements, but a housing made of aluminum works better. I cut mine out of L shaped aluminum extrusions that I buy at the hardware store... or you could buy a nice box from Digi-key or Mouser.
An intermediate level of housing can result from slipping some heatshrink tubing over the small Plasti-Dip'ed board. Some extra Plasti-dip at the ends, or hot-melt glue, will seal up the ends. For a small board, you could zip tie it into onto a wire bundle.
edit: I just remembered.. I've also put some RTV over the plasti-dip'ed board and then put the heatshrink over it. The excess RTV gets squeezed out of the ends, sealing things off rather well.

 

[johnnya156]

my personal preference of construction technique is to cut up some copper-clad circuit board. I use a dremel tool with the cut-off wheel to make slots in the copper layer, which produces pads for mounting the surface-mount components. If I do a good job of the layout, it also provides the connections between most of the components. I use 30 ga wire-wrap wire to make the rest of the connections.

I've got pictures of my latest bike headlight that gives a hint at the process....
https://www.flickr.com/photos/kurtsj00/albums/72157649424813991

I sketch out the layout on the copper using a pencil, and mark the cuts in Sharpie prior to getting the dremel out.
I don't show it explicitly, but I usually coat the board with Plasti-Dip. This provides a moisture barrier, but more importantly, supports the thin wires so they won't vibrate, fatigue, and break.

A few thin coats of Plasti-Dip can do a tolerable job of protecting the board against the elements, but a housing made of aluminum works better. I cut mine out of L shaped aluminum extrusions that I buy at the hardware store... or you could buy a nice box from Digi-key or Mouser.
An intermediate level of housing can result from slipping some heatshrink tubing over the small Plasti-Dip'ed board. Some extra Plasti-dip at the ends, or hot-melt glue, will seal up the ends. For a small board, you could zip tie it into onto a wire bundle.
edit: I just remembered.. I've also put some RTV over the plasti-dip'ed board and then put the heatshrink over it. The excess RTV gets squeezed out of the ends, sealing things off rather well.

Thanks Steve. That actually looks quite do-able! Maybe I'm in with a shot at doing this myself after all. I've not worked with copper clad circuit board before. Is there an alternative to using a Dremel to carve the board? I don't have access to one at the mo.

 

[Savvas]

With the copper-clad board you can just use something with a sharp point. I have used an old dental probe or scraper (donated by my dentist!) However it's rather difficult to do curves or 'T-juntions' - something the Dremel would do better. I have a very cheap tool just like a Dremel from the local car shop. Works OK though I haven't tried it on a circuit board yet.

Steve - have you outlined the functional design of this 2-LED-with-standlight circuit anywhere in a previous thread? It actually looks sufficiently 'simple' for me to get a handle on. I'm interested as to why you went with such a large capacitor (in both size and capacity)...?

ta,

Sam.

 

[Steve K]

Thanks Steve. That actually looks quite do-able! Maybe I'm in with a shot at doing this myself after all. I've not worked with copper clad circuit board before. Is there an alternative to using a Dremel to carve the board? I don't have access to one at the mo.

A dremel tool is pretty handy for a variety of things, so this might a good excuse to buy one.

An alternative would be to use the "toner transfer method". It relies on laying out the board on the computer, printing the layout onto a waxy sort of paper using a laser printer, and then ironing that onto the copper clad circuit board. The board is placed into an etching liquid (ferric oxide, I think).
I've bought the waxy paper from Digi-key, and had mixed success when trying to iron the toner onto the circuit board. For SOIC parts, it's probably better than a dremel tool, since I haven't been able to make pads for pins spaced 0.05" apart.

There are places that will build a small batch of boards. You have to learn how to use the appropriate software to draw the schematic and lay out the board. If I only need one board, then my preference is to just make my own.

 

[Steve K]

Steve - have you outlined the functional design of this 2-LED-with-standlight circuit anywhere in a previous thread? It actually looks sufficiently 'simple' for me to get a handle on. I'm interested as to why you went with such a large capacitor (in both size and capacity)...?

ta,

Sam.

Are you asking how the circuit works? It's basically the same as the standlight circuit using the nicad cell, but with the required tweaks to deal with a supercap that gets charged at 2.7V. It's interesting that the standlight can be charged faster than a nicad, but this can mean that the lower LED doesn't get much current. This might be a good option if you spend a lot of time at stoplights. If not, then reducing the charge current might be worthwhile.

I'm using a 100F cap because that's what a friend had available at a low price.
It works out well... providing roughly a 10 minute run time. It does dictate a larger light housing than would be required otherwise. It is lighter than the nicad I use in other lights, though.

I do like how I implemented a reed switch to disable the boost converter used in the standlight. Getting the "toggle" part with the magnet figured out was a bit more work than I expected. The most frustrating part was the fact that I started with some 3mm Cree LEDs (I forget which...), but I wasn't able to get the Ledil optics lined up properly and get the desired beam. No idea why. I ended up going to the old XR-E that I used in previous lights.

 

[jdp298]

Having had the latest and greatest B&M headlight identified by deNaranja in this thread, I've found it at RS and dug into the data-sheet. It appears to have the LEDs in series, each with a Zener to protect it. So using the principles of this thread, and my own standlight circuit (which no one else seems to like), I have drafted this: https://drive.google.com/open?id=0BwTUpbJp7NzNdzl6RTVZTE1TTEU

The thoughts or uncertainties I have are:
- Will SD2 allow me to fully charge the supercaps, or am I clamping the whole assembly to 5.5V?
- Can the LDO circulate round on its own output through the first LED? I'm pretty sure it won't but a second opinion won't hurt.

And yes the whole enclosure/reflector is a separate question I have to answer somewhere else.

 

[Steve K]

there are a bunch of aspects of that circuit that have me scratching my head.
Have you considered using the standlight circuit that was developed in the big standlight thread years ago?....
http://www.candlepowerforums.com/vb...rcuit-work-)&p=2706860&viewfull=1#post2706860
.. although that "diode voltage clamp" should be a zener with the cathode on top.

but.. back to the circuit that you've proposed....

- The attempt to regulate the charging voltage of the supercaps should be using a resistor in place of SD2. Of course, this circuit just dumps excess current to ground through the zener, so it's not at all efficient.

- I'm not sure why there are zeners in parallel with the LEDs. LEDs are static sensitive, but not in the direction that they conduct current. You could put a simple diode across the LED if you are worried about that sort of thing. A small cap works surprising well too, as I got to learn at work. With a 0.1uF cap across a white LED, I was able to hit it with a 15kV ESD discharge and not kill the LED.

- The LDO isn't doing anything that R2 can't do. I'd get rid of the LDO.

- I'm not sure why the current is being limited to 370mA, but a 18 ohm resistor doesn't seem like the right value. At 0.37A, there will be 6.7V dropped across the 18 ohm resistor. With 6.6V dropped across the LEDs and roughly 1V across the bridge rectifier, you appear to be expecting over 14V out of the dynamo.

- If you get rid of the zeners across the LEDs, you won't need SD1.

All in all, I think this circuit could be salvaged by turning it into the circuit from the standlights thread.