Help designing circuit for dynamo-charged-battery-driven LED lights

seanspotatobusiness

Newly Enlightened
Joined
Jul 12, 2010
Messages
81
I would like to make a set of lights based on LEDs and powered by a bottle dynamo (I don't believe a hub dynamo is sufficiently superior to warrant the expense and hassle) with a battery between which can be charged during the day to help drive the LEDs at night.

There are circuits at this website which get pretty complicated but they're all about driving a number of LEDs directly whereas I've been advised to use an LED driver and in addition, I want to use a battery.

I don't want a manual switch for choosing between the bridge rectifier and Greinacher voltage doubler and I don't want the complication of the automatic switch either. Since I usually ride at 13-23 km/h, can I just use the voltage doubler? Then I could use a 6v dynamo to charge a 12v battery (either a NiCd battery pack or sealed lead acid) and in turn drive a few power LEDs via driver off eBay. My technology teacher at school always used to say "keep it simple, stupid!".

My proposed circuits:
bikelights.png

bikelights

(I'm sorry for the quality; I've tried about 10 different programs this morning and none have been easy to work with - this one was easiest but that really, really doesn't say much).
 

Steve K

Flashlight Enthusiast
Joined
Jun 10, 2002
Messages
2,786
Location
Peoria, IL
I think that your basic idea is reasonably sound. The details may need to be addressed, though.

Regarding the use of the voltage doubler: I haven't used these circuits, so I don't have detailed knowledge of their characteristics. I assume that they produce less current than the usual 500mA from a dynamo. You might try making a measurement of the current with a regular bulb as a load, although the waveform will cause some error for the average multimeter. A meter that could make RMS measurements would be ideal.

Anyway.... I only bring up the current because it's a factor for charging the battery. Some sort of charge regulation is desireable. This could be charge current regulation for a nicad or NiMH, or voltage regulation for a SLA. Without charge regulation, you'll need to use a larger battery. For instance, if the charge current is 300mA, you'll want to use a 3 amp-hour nicad. Nicad's can typically tolerate charging at C/10 indefinitely, where C is the battery's capacity as measured in amp-hours.

NiMH batteries are much less tolerant of trickle charging... you might want to go to a 10 amp-hour NiMH battery. For SLA's... I don't know. Something big is probably better.

Using the LED driver, which can operate down to 7v, with a 12v nicad or SLA presents problems in terms of battery discharge. Discharging a 12v nicad below 10v or so will damage the nicad. Discharging the SLA that far may permanently damage it. A LED driver such as the one you mention will maintain LED brightness up until the point where the 7v battery voltage is reached, so you'll never know when you hit the threshold of battery damage until you are already there.

Battery charging is also sensitive to temperature. You shouldn't charge if the battery is already pretty warm (not sure how often this will happen, but might if you live in Arizona or such). In cold temperatures, the battery chemistry slows down, and it takes much longer to charge.

About 10 years ago, I designed a light similar to yours, and the continuous charging killed the 800mA-hr nicad in a year.

About 8 years ago, I designed a light that took power from the dynamo, regulated the charge of a nicad battery, and then regulated the battery output to drive the headlight. It worked great in spring, summer, and fall, when I didn't need the light very often. In the winter, when the light was on most of the time I was riding, the battery didn't charge much at all and the light would go out shortly after I stopped. I did have a low-voltage shutdown feature, so at least I didn't hurt the battery.

I now use a simpler system where a low-power light comes on when I'm stopped. That makes it easier to charge the battery (a single AA nicad). The AA nicad will run the low-power light for over 30 minutes, so I've never discharged it. The LED driver for the low-power light can't operate below 0.8v, so it can't completely discharge the nicad (and discharging a single nicad cell really isn't bad for it).

Depending on your circumstances, your design could work just fine. The problems with charging and discharging batteries make it problematic enough to discourage manufacturers from building something designed for widespread use, though.

regards,
Steve K.
 

seanspotatobusiness

Newly Enlightened
Joined
Jul 12, 2010
Messages
81
Thanks very much for relating your experiences, Steve. I really appreciate it.

My old bike in the Netherlands used to have a light in the rear that would come on when stationary (when it felt like it) and that indeed sounds like a simpler solution than my initial plan.

I actually want to fit my lights in an old carbide bicycle lamp since my bike is kind of retro anyway and I like that (it's only 30 years but some people act as though I was riding a penny farthing!). There would be space in there for an XP-G when moving and a few cheap generic LEDs while stationary.

Where did you get your circuit that switches to using a battery when stationary?

Edit: I'm reading this one now. Should be enough to keep me quiet for a while. I'm interested in something MrAI said about charging supercaps inline with a diode to prevent their discharge into the power LEDs and then using a PNP transistor to enable them to discharge when the power stops. I think this would allow discharge of the supercap(s) through 5mm LEDs which would be preferable to trying to power the power LEDs. I could have power both front and back. Anyway, back to reading... :)
 
Last edited:

seanspotatobusiness

Newly Enlightened
Joined
Jul 12, 2010
Messages
81
I finally found this post with none appearing to supersede it in the realm of NiCd batteries. It's taken a long time and I don't remember, much less understand most of what I've read but I'm attempting to hash together a circuit design based on yours.

I anticipate using two XP-G LEDs (one front; one rear) while moving and eight (or so) cheap LEDs (four front; four rear) while stationary.

My circuit schematic (click to enlarge) is incomplete because it was not clear to me where the upper most line connects. I left it hanging in the air. I'm not certain how far the original values apply to my circuit, given that it will use modern power LEDs.

Is it still possible to buy the zxsc310?



I used EAGLE to make the circuit diagrams and there's no function for inserting a zxsc310 so I had to substitute it with a badly-arranged circle :/
 

Steve K

Flashlight Enthusiast
Joined
Jun 10, 2002
Messages
2,786
Location
Peoria, IL
I finally found this post with none appearing to supersede it in the realm of NiCd batteries. It's taken a long time and I don't remember, much less understand most of what I've read but I'm attempting to hash together a circuit design based on yours.

I anticipate using two XP-G LEDs (one front; one rear) while moving and eight (or so) cheap LEDs (four front; four rear) while stationary.

My circuit schematic (click to enlarge) is incomplete because it was not clear to me where the upper most line connects. I left it hanging in the air. I'm not certain how far the original values apply to my circuit, given that it will use modern power LEDs.

Is it still possible to buy the zxsc310?



I used EAGLE to make the circuit diagrams and there's no function for inserting a zxsc310 so I had to substitute it with a badly-arranged circle :/


Your circuit is close to correct, but did miss an important detail that I may not have emphasized. The charging of the nicad cell relies on the voltage generated across the yellow leds and the silicon diode in series with the nicad cell. The voltage across the yellow leds is about 1.8 (I think??), and the silicon diode has a voltage drop of 0.6v or so. This leaves about 1.2v to charge the nicad, which will get it at least half charged (depending on temperature).

Your drawing shows the standlight circuit in parallel with the headlight, which is going to apply to much voltage to the nicad cell. The proper arrangement would be to insert the standlight and red and yellow led arrays in series with the headlight. Or, use the standlight circuit shown in my headlight schematic:
http://www.flickr.com/photos/kurtsj00/3984952705/in/set-72157621965148305/
It's not as clever, and simply uses a voltage reference and transistor to set the charge voltage for the nicad. I haven't built it yet, but it should work <crossing fingers>.

I've been very happy with how my bike lights have turned out. Two Cree white leds work pretty well as a headlight, and the array of 10 red 5mm leds makes an outstanding taillight. For my commuting bikes, I've got one of these standlight circuits built into the taillight too. This video of my commuting recumbent shows how visible these lights are in the evening hours:
http://www.flickr.com/photos/kurtsj00/3851153642/in/set-72157621965148305/

For this reason, I'd encourage you to use two Cree's in front, and a 10 led array w/standlight in front and back. I don't know about using the voltage doubler, though... without the need to charge a big battery, there's no reason to use the doubler.
Hmmm... what did I do to get the right nicad charge voltage in the rear light where I used red leds? I think they were a newer chemistry with a forward voltage around 1.8v.

regarding the zetex boost converter, the zxsc310, they are available. I usually buy parts from Digi-key, www.digikey.com. These parts are about $1 in small quantities.

I should find time to post pics of the rear light I built for the recumbent. It has the standlight and array of 10 leds built into a bit of aluminum channel. Much more professional in appearance than most of the cobbled up stuff I make. :)

regards,
Steve K.
 

BrianMc

Enlightened
Joined
Nov 4, 2009
Messages
940
....I've been very happy with how my bike lights have turned out. ...video of my commuting recumbent shows how visible these lights are in the evening hours. ... regards, Steve K.

Thanks SeanPB, I learn something new each time someone designs a dynamo lighting system. There is one maybe two in my future. The output of the LEDs at 3 Watt is approaching what I want, and my batteries are getting older.

Steve:. Thanks for the video. It takes more time than most people would think unless they have done it. Very interesting.

I find that as long as you don't have the sun shining low directly at your tailights washing them out especially if you don't have polarized lenses, :cool: (blocked by those woods, for example), the evening visibility of tailights is very good, because the contrast of them versus the background light level is up, and drivers eyes are more dilated. Under those conditions larger may beat out brighter. Night is gets easier to see the lights, again. The hard test is how much they show up in full daylight if you need that, and I do. BTW artistically I LOVE the video angle, fall colors, and the ambient light.

Extrapolating it to the driver's view is a problem, though. It is sort of a squirrel trying to run into your front wheel view. :eek: If you are ever tempted a repeat video, I would love it if you could consider placing the camera at driver height or close, such as on the hood or dash of a taller vehicle that is placed in the road in the close to the correct lane relationship. Unfortunately, that likely requires recruiting another person for vehicle duty.

This comparing different lights with different videos under different lighting from different angles isn't straight forward. I can't match you camera's resolution right now. :broke: And so it goes... we're DIY and doing the best with the tools we have.

I might have my biases, but there are few daylight videos of bike lights in action on the road, so my hat is off to you for doing it at ALL! :twothumbs

Ride safe. Ride on!

Brian.
 

seanspotatobusiness

Newly Enlightened
Joined
Jul 12, 2010
Messages
81
Your drawing shows the standlight circuit in parallel with the headlight, which is going to apply to much voltage to the nicad cell. The proper arrangement would be to insert the standlight and red and yellow led arrays in series with the headlight.
I've adjusted my circuit to put the battery in series with the headlights - thanks!

I've been very happy with how my bike lights have turned out. Two Cree white leds work pretty well as a headlight, and the array of 10 red 5mm leds makes an outstanding taillight. For my commuting bikes, I've got one of these standlight circuits built into the taillight too. This video of my commuting recumbent shows how visible these lights are in the evening hours:
http://www.flickr.com/photos/kurtsj00/3851153642/in/set-72157621965148305/
Indeed, your headlight is incredibly bright! Nonetheless, I'm not sure whether I want both Crees at the front. I did read that the XP-G is brighter and more efficient than the XR-E so I think that one might should be enough? Your using two XR-E in that video, right? Any idea how close to maximum output?

I don't know about using the voltage doubler, though... without the need to charge a big battery, there's no reason to use the doubler.
Yeah, I wont be using the voltage doubler.

regarding the zetex boost converter, the zxsc310, they are available. I usually buy parts from Digi-key, www.digikey.com. These parts are about $1 in small quantities.
Thanks - since your example proved that "Goolge Shopping" was failing in some way, I searched an alternative way and found a UK electronics supplier :)


In your circuit, do the stand-lights constantly light up while moving? I'm unfamiliar with inductors (among many other things) but would I not be right in saying that current flows through the inductor to the stand-lights constantly while the dynamo is spinning?

Edit: I can see that I must be wrong but I don't see why - the inductor offers only half an ohm of resistance?

Here's what I hope is the corrected circuit:
 
Last edited:

BrianMc

Enlightened
Joined
Nov 4, 2009
Messages
940
Simplistic (wrong) overview:

Resistor: adds resistance to either AC or DC circuits.

Capacitor is resistive to DC, lets AC pass. Also used as a current smoothing device that stores electrical potential and fills in dips in current like in AC power supplies, or a short circuit for unwanted EMI AC noise (filter).

Inductor resists AC lets DC pass.

http://en.wikipedia.org/wiki/Inductor

You can look up the rest.

My understanding: the inductor prevents the standlight from operating until the source AC stops, then the capacitor drains through the inductor to the standlight until it is empty or the bike resumes and power is generated and the Capacitor is recharged.
 

Steve K

Flashlight Enthusiast
Joined
Jun 10, 2002
Messages
2,786
Location
Peoria, IL
Indeed, your headlight is incredibly bright! Nonetheless, I'm not sure whether I want both Crees at the front. I did read that the XP-G is brighter and more efficient than the XR-E so I think that one might should be enough? Your using two XR-E in that video, right? Any idea how close to maximum output?

Two LEDs at the front is a popular choice and the light output is not so much that I'd choose to cut it in half. I am indeed using two XR-E's, and they are being driven by the rectified dynamo output, which is approximately 500mA. They are rated for 700mA, I believe.

In your circuit, do the stand-lights constantly light up while moving? I'm unfamiliar with inductors (among many other things) but would I not be right in saying that current flows through the inductor to the stand-lights constantly while the dynamo is spinning?

pin 3 of the zetex IC is the shutdown pin. When the voltage at pin 3 is low (under 0.5v, maybe?? check the datasheet), the IC is turned off, and the standlight is off. I simply use the "AC detector" circuit to take the voltage across the group of yellow LEDs, rectify it and let it charge a 0.1uF cap, and then use that voltage to drive a small transistor. When the dynamo is running, there is enough voltage on the 0.1uF cap to turn on the transistor, which pulls the voltage at pin 3 of the Zetex IC down to about 0.1v or so. This turns off the standlight.

Alternately, I have a switch to ground at pin 3 too. I use this to disable the standlight when the bike is parked or when riding in the daytime.

I'll have to take a look at the corrected circuit and get back to you....

regards,
Steve K.
 

Steve K

Flashlight Enthusiast
Joined
Jun 10, 2002
Messages
2,786
Location
Peoria, IL
I've adjusted my circuit to put the battery in series with the headlights - thanks!
....<snip>.....

Here's what I hope is the corrected circuit:

Well, you've moved part of the standlight in series with the headlight, but most is still in parallel. Plus, you aren't using a bank of parallel leds to generate the voltage that is used for charging the nicad cell. Also, the Zetex can't drive two parallel strings of four leds (at least I haven't tried it).

If you go back to my original schematic:
https://www.candlepowerforums.com/posts/2714168&postcount=79
you'll see that there are two groups of 10 leds wired in parallel. The anodes of one group are wired to the cathodes of the other group, and visa versa. This just means that each group will carry half of the dynamo's AC current. These two groups of 10 LEDs are wired in series with the dynamo and headlight (the Lumotec).

On my Powers bike, I actually have a standlight circuit built for each group of 10 LEDs. In each standlight, I use the voltage across the group of 10 leds to charge the nicad cell, and control whether the standlight is on or off. The nicad voltage is then used to provide power to the Zetex IC circuit.

Does this clarify matters?

regards,
Steve K.
 

seanspotatobusiness

Newly Enlightened
Joined
Jul 12, 2010
Messages
81
Hi. I've dropped the super cap from the design since I don't think they're worth their cost and size!

Thanks for that explanation of capacitors inductors, Brian - makes sense!

The current isn't flowing through the inductor to the LEDs like I thought, 'cause the voltage is insufficient!

According to the datasheet (p12),

"The ZXSC310 offers a shutdown mode that produces a standby current of less than 5uA when in operation.
When the voltage at the STDN pin is 0.7V or higher the ZXSC310 is enabled, hence the driver is in normal operation. When the voltage at the STDN pin is 0.1V or lower the ZXSC310 is disabled, hence the driver is in shutdown mode. If the STDN pin is open circuit the ZXSC310 is also enabled."
What I can figure from that is that when your switch is closed, your stand lights are disabled. However, I don't understand how your AC detector functions; when AC is present, the transistor is open and the stand lights are disabled - fine - but when AC stops, the transistor closes, isolating pin 3, and leaving it at 0V - why would that activate the chip?

I also don't know how it applies increased voltage over the LEDs - where is its output? VCC is used to accept input voltage and Idrive simply switches on a short-circuiting transistor to ground? Sheer madness!


Edit: I noticed that you sneaked another post in there while I was composing this! I'll have to take some time to figure out my next move!
 
Last edited:

seanspotatobusiness

Newly Enlightened
Joined
Jul 12, 2010
Messages
81
On my Powers bike, I actually have a standlight circuit built for each group of 10 LEDs. In each standlight, I use the voltage across the group of 10 leds to charge the nicad cell, and control whether the standlight is on or off. The nicad voltage is then used to provide power to the Zetex IC circuit.

Are you saying that I need 10 regular LEDs to create a circuit to charge my battery or that's just how it works in your circuit so mine can't copy that method?

I don't think that the circuit would have worked with the battery in series with the LED's 'cause it needs to pass 500 mA constantly. I moved the battery to be parallel with one of the lights and added a cell to it. The power LEDs will have 3.0-3.1V across them, and 0.7 of this will be eliminated by the diode, leaving 2.4V to charge the battery.

 

Steve K

Flashlight Enthusiast
Joined
Jun 10, 2002
Messages
2,786
Location
Peoria, IL
Are you saying that I need 10 regular LEDs to create a circuit to charge my battery or that's just how it works in your circuit so mine can't copy that method?

I don't think that the circuit would have worked with the battery in series with the LED's 'cause it needs to pass 500 mA constantly. I moved the battery to be parallel with one of the lights and added a cell to it. The power LEDs will have 3.0-3.1V across them, and 0.7 of this will be eliminated by the diode, leaving 2.4V to charge the battery.


I'm saying that you need approximately 1.3v to charge the nicad cell. How you get that voltage is up to you. For me, the drop across the yellow LEDs did the trick very nicely. In my later design, I rigged up a reference voltage with a current amplifier. Not as energy efficient, but functional. The fact that the 10 LEDs make an outstanding taillight (or a decent "parking light" in the front) is just an added benefit.

If I haven't been clear about charging the nicad cell.... if you put 2.4v across the nicad cell, you'll kill it pretty quickly.

regards,
Steve K.
 

seanspotatobusiness

Newly Enlightened
Joined
Jul 12, 2010
Messages
81
I understand re: the NiCad charging voltage, but it's okay 'cause I've thrown in another cell. There are two now, so the voltage across them will be 2.4. With that in mind, do you recon this circuit would work?

I'm gonna order some breadboard and check it all out (will power a bottle dynamo with an electric drill on low speed).

I'm also hoping to get a basic old oscillopscope on eBay, so I'll be able to probe it and see what's going on in there (I can't fathom how that chip is driving current through the LEDs 'cause I don't see what pin it could be coming from!) If you don't know yourself, maybe in a couple of weeks, I'll be able to tell you! :)
 

HOLONYAK

Newly Enlightened
Joined
Jun 19, 2010
Messages
47
In the process of performing this work, please provide the portion that highlights the battery portion. Specifically, I would like to run a Schmidt Edelux or Busch & Muller Ixon IQ Speed (or two) with cells that can be charged by my Pila charger. Assuming it involves something other than direct-wire...
 

pe2er

Enlightened
Joined
Dec 22, 2007
Messages
380
Location
Europe - Holland - Almere

There is an error in how you connect the generator and C1 in your last published circuit - you did it correctly in the first circuit...
I'm also hoping to get a basic old oscillopscope on eBay, so I'll be able to probe it and see what's going on in there (I can't fathom how that chip is driving current through the LEDs 'cause I don't see what pin it could be coming from!)
The current comes from the inductor. The IC briefly connects the coil to ground through the transistor, causing high current to flow through the coil. After the transistor opens, current in the inductor will still flow, causing a currrent through your LEDs. The voltage across the coil will add to that of the battery because they are connected in series.

In other words, the coil stores energy when the transistor closes and the coil it is connected across the battery, to release it into the LEDs when the transistor opens.

Did not look up the specs for the integrated circuit you use, but normally there is i diode between the coil and the load (i.e. the LEDs) in an LED boost circuit. That seems to be missing from the circuit you drew...

EDIT: OK, so you go me curious. Id did look up the datasheet of the ZXSC310
Here is the typycal application circuit:
naamlooskz.jpg

It seems you have omitted C1 and D1. Might work, these components serve to smoothen the current through the LEDs (and make them flicker less).

One last thought; You use two 4 LED strings in parallel. The front will be white LEDs, the rear red. They have different Vf, so this means that only the LEDs with the lower Vf (the 4 red LEDs) will light up. the White LEDs will have less current and thus reduced intensity.

Another last thought:
I'm gonna order some breadboard and check it all out (will power a bottle dynamo with an electric drill on low speed).
Here is what I did on my Dynamo powered amoeba clone
 
Last edited:

seanspotatobusiness

Newly Enlightened
Joined
Jul 12, 2010
Messages
81
In the process of performing this work, please provide the portion that highlights the battery portion. Specifically, I would like to run a Schmidt Edelux or Busch & Muller Ixon IQ Speed (or two) with cells that can be charged by my Pila charger. Assuming it involves something other than direct-wire...

Sorry, but could you rephrase your request? I don't understand. What would you like me to provide?



There is an error in how you connect the generator and C1 in your last published circuit - you did it correctly in the first circuit...

Nice catch! Thanks!

The current comes from the inductor. The IC briefly connects the coil to ground through the transistor, causing high current to flow through the coil. After the transistor opens, current in the inductor will still flow, causing a currrent through your LEDs. The voltage across the coil will add to that of the battery because they are connected in series.

That's very crafty! Thanks again :D

EDIT: OK, so you go me curious. Id did look up the datasheet of the ZXSC310

It seems you have omitted C1 and D1. Might work, these components serve to smoothen the current through the LEDs (and make them flicker less).

Yeah, I think it will still work. Further down in the data sheet, it suggests leaving out the diode and capacitor to maximise battery life. I'll add them to my order anyway, since they're pretty cheap.

Not sure that I have the patience to solder on one of those tiny little capacitors though.

One last thought; You use two 4 LED strings in parallel. The front will be white LEDs, the rear red. They have different Vf, so this means that only the LEDs with the lower Vf (the 4 red LEDs) will light up. the White LEDs will have less current and thus reduced intensity.

Another last thought:

Here is what I did on my Dynamo powered amoeba clone

I will get around it by using the same LEDs front and back and relying on a red filter!

I'll check out your work now :) Thanks again!


Edit: I now have three designs which I'm going to explore -







The option to abort launch expires in 09:52... 51...!
 
Last edited:

HOLONYAK

Newly Enlightened
Joined
Jun 19, 2010
Messages
47
Sorry, but could you rephrase your request? I don't understand. What would you like me to provide?

I would like to do what you are doing, however I don't have a dynamo. I would like to run (lights designed for a dynamo) with rechargeable cells. I wanted to know if anything fancy was being done to maintain the appropriate input voltage...and if so, I want to copy it. :)

From the instruction manual:

Connection to rechargeable battery
If you connect the headlamp to a rechargeable battery (entirely at your own risk)
make sure to observe correct polarity, i.e. "plus" to the earth inlet and "minus" to
the current inlet. Maximum brightness is obtained at a voltage range between 7.2 and
7.5 V (7.5 V must not be exceeded).
 

seanspotatobusiness

Newly Enlightened
Joined
Jul 12, 2010
Messages
81
I would like to do what you are doing, however I don't have a dynamo. I would like to run (lights designed for a dynamo) with rechargeable cells. I wanted to know if anything fancy was being done to maintain the appropriate input voltage...and if so, I want to copy it. :)

From the instruction manual:

Connection to rechargeable battery
If you connect the headlamp to a rechargeable battery (entirely at your own risk)
make sure to observe correct polarity, i.e. "plus" to the earth inlet and "minus" to
the current inlet. Maximum brightness is obtained at a voltage range between 7.2 and
7.5 V (7.5 V must not be exceeded).

Do you already have a battery? Circuits 1 and 2 in my last post avoids the issue of regulating voltage by using a number of NiCd cells which have low internal resistance that remains practically constant throughout discharge. If they're 1.2V fully charged, they will be (very, very close to) 1.2V near complete discharge. If you use 6 cells in series, you get your 7.2V. The problem is that if the voltage across your lights and battery don't reach 7.2V (because your bulb lights up and conducts at 6V, for example), then I don't think that the battery would ever get charged (it's gonna need at least 1.2V across each battery to charge (and preferably no more than that after they've reached capacity 'cause then they'd overcharge but I don't think that's an issue here)). Ultimately, I will test my circuits and report back, but I guess the characteristics of your own light (what sort is it?) will affect whether or not your results would differ.
 
Top