Standlight circuits (Will this circuit work?)

[syc]

Re: Will this circuit work?

In essence, Alex's circuit is ending up being very similar to mine. Here's the schematic for my first front standlight...

[ snip ]

You'll note that I use a switch to disable the standlight when the bike is parked. It's not desireable to fully discharge the nicad.

I see that this stuff is old hat for you guys
Looking back, bandgap also mentioned doing dynamo charged batteries for years in a thread that Martin started on Charging Li-Ion with ripple current.

What I was wondering is, if we took advantage of the LM317L and protected li-ion batteries instead of supercaps, could we dispense with the zetex and use a circuit like this?

If I understand the LM317L resistors, the 500 and 240 ohm resistors should set a nominal max charging voltage of about 3.9V (with a worse case of 4.3V if the parts are out of spec enough) which is a voltage that provides a long lifetime on the li ion batteries, and the protection circuit on the li ion cell would prevent over discharge (or is that only for too much current draw and not deep discharge?). Add in a current limiting resistor on the output so that we keep the discharge within a reasonable range, and then a transistor switch and manual switch to only discharge when the bike is stopped and the switch set to "on".

Is this too simplistic again? I didn't see the LM317L mentioned in any battery charging threads, so I'm just guessing what works and what doesn't. Is 100ma too low of a current for charging a li ion battery? Or is this a viable solution for a "standlight" that can actually last for a while?

Or do you basically need a version of Alex latest circuit but with the components to activate the shutdown pin on the zxsc310 and maybe the switching transistor across the feed to the bottom cree?

 

[Bandgap]

You have to turn off the resistive divider load from the battery with a transistor.

I agree with using protected Li ion cells.

But even with those - here is a SAFETY NOTE TO EVERYONE.

Do not make your own charging circuits for Li ion cells unless you really know what you are doing.
They can explode if charged incorrectly.
Thinking you know what you are doing is not enough.

And to keep the battery healthy, you really want to prevent the leds discharging the cell below 3V0 or 2V7 - what ever is specified.
A white led on its own almost does this automatically.
A boost circuit will need disharge protection

All that taken into account, and remebering that LM317 circuits can need some post-making tweaking - take a look at

http://www.electronicsweekly.com/bl...in-wonderland-liion-charging-with-an-l-1.html


Steve

 

[rageahol]

what about LiPO4 cells instead? less power, but at least you dont run the risk of rapid-venting-with-flame....

 

[Bandgap]

Re: Will this circuit work?

....(with a worse case of 4.3V if the parts are out of spec enough) .....

That transistor circuit is a bit crazy and, in my opinion, needs sorting.
For the conventional circuit, the collector and emitter conections need swapping, and the short from base to 0V needs to be a resistor selected to approriately satutate the transistor. And the transistor need either a base resistor, or a resistor in series with the diode.

And this circuit opens up a current path back through the 317 (through the new diode) which will turn the 317 reverse protection diodes on (and probably blow them up).
To solve this, put a diode in series with the input of the 317, or replace the existig diode with two diodes, both connected to the base of the transistor like yours is, but with the anodes connected one to each side of the dynamo.

AND, can I ask a favour of everyone here, there is a genie that needs keeping in its bottle.
PLEASE don't put Li ion charging circuits that use non-precisions components into the public domain without a health warning against each circuit.
Someone who does not know what they are doing may stumble across such a circuit, build it, and have a cell catch fire or explode.
Newbie/Jarhead - a fine contributor to CFP and someone who really knows leds and Li ions - put repeated dire warnings up abouth Li ion cell abuse on CPF.

The best thing to do with your circuit Syc is put two resistors in series, instead of the 500R (which I don't think is obtainable), then you have a good chance to find a combination that sets the output just where you want it. A variable resistor is another option, but these can self-adjust under vibration.

4.3V may eventually be enough to cook the internal protection circuit.
4.2V should be considered the absolute maximum that is deliberately applied to a Li-ion cell - or 4.1V depending on the cell anode material.

 

[syc]

Re: Will this circuit work?

Hi Steve,
Looking at the drawing again, I see that I did an incredibly very bad job of copying out the parts from Steve K's original circuit. I've fixed the transistor and added the resistor from his original transistor switch. I also put in the diode as you described and an appropriate warning label.

Sorry for the original mess!!

Does this circuit seem reasonable?

Steve

That transistor circuit is a bit crazy and, in my opinion, needs sorting.
For the conventional circuit, the collector and emitter conections need swapping, and the short from base to 0V needs to be a resistor selected to approriately satutate the transistor. And the transistor need either a base resistor, or a resistor in series with the diode.

And this circuit opens up a current path back through the 317 (through the new diode) which will turn the 317 reverse protection diodes on (and probably blow them up).
To solve this, put a diode in series with the input of the 317, or replace the existig diode with two diodes, both connected to the base of the transistor like yours is, but with the anodes connected one to each side of the dynamo.

AND, can I ask a favour of everyone here, there is a genie that needs keeping in its bottle.
PLEASE don't put Li ion charging circuits that use non-precisions components into the public domain without a health warning against each circuit.
Someone who does not know what they are doing may stumble across such a circuit, build it, and have a cell catch fire or explode.
Newbie/Jarhead - a fine contributor to CFP and someone who really knows leds and Li ions - put repeated dire warnings up abouth Li ion cell abuse on CPF.

The best thing to do with your circuit Syc is put two resistors in series, instead of the 500R (which I don't think is obtainable), then you have a good chance to find a combination that sets the output just where you want it. A variable resistor is another option, but these can self-adjust under vibration.

 

[ktronik]

I really should read the whole thread...but martin & I worked on this a while back, to get me a standlite & portable power source, all from super caps...

some problems I ran into while making my dynobatt standlite is that the cap has to charge first & only when charged will the power be given to the LEDs...

So Martin & I came up with this circuit...

with this circuit the supercap / capbattery will be trickle charged, thus giving power to the main LEDs straight away...this way you can have a bigger capbattery bank & run a USB power out or run a bflex to a single LED, as I did. I had the bflex set to 'flash' a single LED, on another light.

Using the bflex powered from the caps, made much better use of the capbattery voltage drop, from the Vf of 3 or 4 LEDs, right down to the minimum voltage of the bflex driving 1 LED...

we also put a 'blead' resistor on the output of the cap's... on the dynobatt I had a switch, to switch it in or out, thus giving full cap output or whatever setting I wanted...

worked really well as I could then use the standlite for changing a 'flat' or whatever...

food for thought...

K

 

[Bandgap]

Syc.

I am looking forward to hearing your results.

Looks like you will be finishing yours before I finish mine.

Couple of comments.
The resistor chain will still drain the cell - take a look at the link I put in the previous post.

And, - I suspect I was not clear.
The diode from the base to ground (in series with the resistor) should not be there.
It should be in series with the input of the 317. - although this does drop an extra 0.7V in the charging circuit.
To save the extra 0.7V, I prefer the idea of replacing the top diode with two - with pointy ends together, and blunt ends to either side of the dynamo - on the dynamo side of the bridge rectifier.

 

[Alex Wetmore]

Using the bflex powered from the caps, made much better use of the capbattery voltage drop, from the Vf of 3 or 4 LEDs, right down to the minimum voltage of the bflex driving 1 LED...

If you read the rest of the thread you'll see that Steve Kurt (and later me) are using the ZXSC310 as a LED driver for the standlight. This is similar to how you are using the bFlex.

The bFlex seems like an odd choice since the minimum input voltage is 4 volts. The supercaps max out around 5.5 volts. So you aren't using much of the supercap's capacity. A driver that runs on lower voltage will give you a longer runtime or allow you to use a smaller (physically and capacity) supercap. The MaxFlex would give you about double the runtime just because it can run off of 2.5 volts. The ZXSC310 is great because it runs off of just .8 volts.

alex

 

[syc]

I am looking forward to hearing your results.

Looks like you will be finishing yours before I finish mine.

Not really, I have only just got my ghetto-style PCB and cree's mounted on the heatsink for the basic circuit I prototyped in post #70. And the second one I build will probably be closer to Alex's current design. It will be a while before I get to a battery backed dynamo light - but it seemed like an interesting design exercise, and I wanted to get it all mapped out now, while it is the topic of conversation.

Couple of comments.
The resistor chain will still drain the cell - take a look at the link I put in the previous post.

And, - I suspect I was not clear.
The diode from the base to ground (in series with the resistor) should not be there.
It should be in series with the input of the 317. - although this does drop an extra 0.7V in the charging circuit.
To save the extra 0.7V, I prefer the idea of replacing the top diode with two - with pointy ends together, and blunt ends to either side of the dynamo - on the dynamo side of the bridge rectifier.

Okay, sorry for not reading things more closely. I went back and read the article you linked to - I ended up trying to integrate the transistor switch used in that article to cutoff the resistor chain when the dynamo is off - hopefully I didn't screw up the cut & paste this time! Can someone sanity check this one for me? Its possible to get a base current that can properly switch both transistors right?

I still don't "get it" on the 2 diodes pointed at each other that you described to avoid the voltage drop into the LM317L - what I'm imagining would simply block any current going through either way - I don't suppose you could sketch something and send it to me?

 

[syc]

I really should read the whole thread...but martin & I worked on this a while back, to get me a standlite & portable power source, all from super caps...

I read that thread a while back, and thought it was an interesting design. What got my attention about that setup was with all the caps, it seemed to be creeping into the size range of a battery based solution. Reading the discussions around this topic, it seemed that going to batteries would provide more power storage once the solutions got to similar sizes.

And on a tour, it might be handy to be able to be able to trickle charge your Li Ion batteries. I picked charging voltage based on battery lifetime and possible sloppy tolerances - if this were to be used as a general charging solution, it would be better to cherry pick the components and try to nail a 4.1V charging voltage. Maybe change the settings to just work with NiMH batteries in series instead of Li Ion.

 

[ktronik]

If you read the rest of the thread you'll see that Steve Kurt (and later me) are using the ZXSC310 as a LED driver for the standlight. This is similar to how you are using the bFlex.

The bFlex seems like an odd choice since the minimum input voltage is 4 volts. The supercaps max out around 5.5 volts. So you aren't using much of the supercap's capacity. A driver that runs on lower voltage will give you a longer runtime or allow you to use a smaller (physically and capacity) supercap. The MaxFlex would give you about double the runtime just because it can run off of 2.5 volts. The ZXSC310 is great because it runs off of just .8 volts.

alex

Hey Alex,

I am sorry I have not read the whole thread... & sorry that this post ended up off topic... sorry everyone

In answer to your question, my capbattery pack on my dynobatt runs 3 LEDs, thus is a ~11v capbattery pack (using 2.5v /55F caps)... but when the caps drop below Vf for 3 LED's, I still have power left in my caps... I also wanted to run a flashing LED light when I stopped, had a spare bflex, so for this app it worked to drain most of the power left in the caps...

the zetec is a great idea for this app... I will read the rest, so I can contribute better...


Syc:

I wanted to avoid a batt overcharge / undercharge circuit (not needed for the way caps work) & loved the fact that a cap has 1 mill cycles instead of a few hundred...

K

 

[Calina]

Re: Will this circuit work?

I see that this stuff is old hat for you guys
Looking back, bandgap also mentioned doing dynamo charged batteries for years in a thread that Martin started on Charging Li-Ion with ripple current.

What I was wondering is, if we took advantage of the LM317L and protected li-ion batteries instead of supercaps, could we dispense with the zetex and use a circuit like this?



quote]

I can't see the picture on flickr.
Is it just me?

 

[syc]

Re: Will this circuit work?

I can't see the picture on flickr.
Is it just me?

No it's not just you - I generally delete old circuits that are flawed and replace them with the corrected versions. But the links to flickr for the images don't track the replacement images. Bandgap requested that I post these schematics with big Surgeon General's Warnings on them, which is very reasonable considering that we're talking about Li Ion battery charging.

I could go back and replace the link, but then the comments correcting the circuit would no longer apply. So I left it blank this time around.

 

[Bandgap]

I just remebered something Steve K pointed out to me a while ago.

Li ion cells are a bit touchy (I don't know how touchy) about being charged at temperatures below 0C - which is one of the reasons he sticks to nickel cells as where he lives it gets a bit chilly in the winter.

EDIT: I feel the need for another HEALTH WARNING

"At cold temperatures the pores in the SEI are effectively closed. If charging a cell at severely low temperatures is attempted, lithium metal will plate on the surface of the SEI, resulting in two dangerous conditions, either of which can cause the battery to enter thermal runaway."
From www.powermanagementdesignline.com/1...DIRKQSNDLQSKHSCJUNN2JVN?printableArticle=true

Steve

 

[znomit]

Re: Will this circuit work?

To maximize the amount of charge stored on the cap, I scribbled out this little circuit:



I don't have the parts laying around to try this myself. This won't be as good as the boost regulator, but will nearly triple the usable stored charge if it works as planned (3.5-2.5) becomes (5.4-2.5). Hope it helps.

Tried this out today on an MC-E. Works very very well. Probably triple the usable runtime from the bare supercap. Definitely be in my next build.
:thumbsup:

 

[FrontRanger]

Re: Will this circuit work?

Tried this out today on an MC-E. Works very very well. Probably triple the usable runtime from the bare supercap. Definitely be in my next build.
:thumbsup:

Right on. I always enjoy reading about your builds, so please do post it.

 

[syc]

Re: Will this circuit work?

Tried this out today on an MC-E. Works very very well. Probably triple the usable runtime from the bare supercap. Definitely be in my next build.
:thumbsup:

I've been quite happy with how it charges up the supercap, especially considering how simple it is.

Here's the latest iteration of the LiIon trickle charger/standlight circuit. I think I've addressed the issues bandgap pointed out to me, but that doesn't mean I haven't screwed something else up :huh:

 

[Calina]

Re: Will this circuit work?

Tried this out today on an MC-E. Works very very well. Probably triple the usable runtime from the bare supercap. Definitely be in my next build.
:thumbsup:

Did you build it? Any news on stand light circuits?

Actually this post is just a bump: this thread is to good to let it go away in oblivion.

 

[znomit]

Re: Will this circuit work?

Did you build it?

Summer in NZ so light building not quite so urgent!
Heres what I'm thinking.... an MC-E with the upper dice powered by a full wave rectifier and the lower on doubler. Both charge up a supercap and three 5mm leds (a much wider beam than the MC-E). Gives me a real dipped beam and a small light head.

 

[Steve K]

Re: Will this circuit work?

This is wandering off topic, but has anyone actually built up a MC-E and tried out the idea of a high and low/dipped beam? I love the concept, but wonder how well it actually works out. I'm behind on my standlight project, and I'm already contemplating building up a new light using the MC-E! I feel sorta guilty.

Steve K.