New converter design. What should it be?

[dat2zip]

I've got the TPS61030 stabilized in a 14mm footprint and compatible with the emitter board to make a sandwich. It requires more work, but, I'm satisfied with the design and stability in a constant current regulator design.

I'm thinking of switching to focus on the TPS61020 series for a single cell design.

I've setup a poll here to gather some info.

You can vote your opinion and post any additional thoughts and suggestions.

A few possiblies exist.

One would be to make it a voltage regulator since there are less parts. Offer two or three votlages to mate with the appropriate Luxeon Vf grades. Maybe three will be sufficient to give good results.

Another possibility is to run it wide open, but, that might be too much current for some applications.

Lastly, a full constant current design.

Please vote the poll and post any suggestions or comments in this thread.

Wayne

 

[kj]

Are you going to replace the current lineup (BB and various MMs) with this new converter?

 

[dat2zip]

No, these will be new products.

The TI chips are low voltage only and would be suited for the 1W and 3W only. The 5W and higher voltage options are out of the question with these.

A 2 cell would still use the current Badboy until I can find a true better replacement.

That's why I'm not too sure I can afford to make the 2cell 1A version. It would cut into my own market.

the 1 cell version offers some new possiblities that are not being offered. The possibility of constant current regulation of 150mA+ down to 1V is something that is not available on the present market.

 

[kj]

Thanks. I chose the AA option since that's for the low voltage and the circuit diameter is currently 14 mm /ubbthreads/images/graemlins/smile.gif

BTW, Madmax series can work with one AA battery and delivers roughly 200~350 mA to LED, though it's not the constant current, right? If yes, I think another area that is not covered right now is a converter that can deliver over 400 mA from one AA battery. Do you think it's feasible?

 

[shiftd]

a coupla comments

on the first part of the poll, is it single cell? meaning if it is AA then it would be 1.5 V max only? or can it accept higher V?

on the second part, personally i would not care. if this is the case, then what choice should i choose?

 

[dat2zip]

Shiftd,

If you don't care about the 2nd part, I would think you would vote no. No anode contact would mean that this board would not be compatible for usage in making a sandwich module or use in the ECan which I think would be a real shame not to have this feature.

Then again, we shall see what CPF'ers in the poll vote for.

Personally I would vote for the N cell. If it this small, it would be able to fit in anything larger as is or with adapators. The same can not be said going the other direction. If the board was 17mm in diameter it would not be useable in a 2AA conversion like a sandwich for example.

kj,

You are correct about MM+ delivering 200+. Problem with that is a high Vf will not fair as well as a low Vf. The new converter if it was constant current would be able to maintain regulation at the top current setting down to 1V. That's really a rather impressive specification and at 90%+ efficiency is awesome.

That's why I'm discussing it here since there are tradeoffs in size, efficiency and useability.

I'm pretty certain I will not be able to package it in the N cell diameter or AAA diameter unless I dump the anode contact and am able to put components on both sides.

 

[shiftd]

Ahh, so there is some trade off. Smaller design would not allow anode contact, while larger board cannot get into smaller light.
For now, i would say the biggest diameter is about the size of the AA. anything smaller would be great but would make soldering harder, wouldn't it?

 

[vcal]

Someone here mentioned a NimH N cell with 500mAh capacity. -Which I would think would be an absolute minimum (currentwise) for the operation of a converter.
-Anybody remember who that was?

IMO-the alkaline version of the N cell just doesn't have all that much overall capacity either, -being even less than the AAA alkaline.

We are talking about driving Luxeons?

 

[Rothrandir]

i put 123 and yes, but it really depends on the kind of output...

i assume a great range of outputs is possible from a single 123, or maybe even 2aa (alks having a higher resistance).

most often i would use a single 123, but i also would like to run it on a single aa cell (like mmm), but it needs to be at least 350ma output for me to consider that, otherwise i'd just stick with the madmax for that application...

 

[dat2zip]

This converter board ideally would be more targeted for a keychain light which when underdriven runs cool and doesn't really require a lot of heatsinking.

Considering the NLS is running 80-120mA and is plenty bright for that size, this should be considered a target for this.

If you want beef a different converter board is in order.

But, considering there are converters boards like Madmax which does pretty well even tho it's not flat regulation is pretty close over a limited voltage range.

The AA size is doable on one side of the board.

But, I think the sweet spot on this design would be usage on 1.5V or a true singe Alkaline solution.

Yes, true, it will run on 2AA or a single CR2, 123, but it will be still underdriving the LED.

I'm still mulling over all the ideas and I have yet to make an decision. This may or may not help to clear this up. It could easily get out of control with to many options and parameters to jell into anything concrete.

 

[kj]

Hmm, the N cell diameter is possible if you can put components on both sides. That's interesting. If you are going to that way, I want you to provide an optional thin anode board to make the anode contact. BTW, if you aim at the N cell size, can I assume the main difference between ARC mania's Micro Converter and your new converter is the efficiency? Also, you converter may drive a LED at a constant current.

If you are going to make the AA size board, I may want more power... I used to use NiMH and one single NiMH AA would drive Madmax Lite over 4 hours and Madmax Plus about 2 hours. Here, I would want a 500mA output with 1 hour runtime solution. I know it will not work well with an Alkaline AA, though...

 

[dat2zip]

Interesting...

Well, with the rating of the part, at the 1.5V single cell the internal switch would limit the max output to the LED.

For a 2 cell application there is more headrooom and it's possible the same design in a two cell could deliver .5A or more.

The graph shows at 1.7V it can do around 600mA with a 3.3V output. So, it's possible to have a 500mA version for a 2 cell and hold regulation below 1.7V depending on the Vf of the Luxeon.

At this point, there isn't much difference if we compare Madmax to the new design. Both are very efficient.

I could convert Madmax to a Constant current design, but, I won't. It's not cheap. The TI IC is much cheaper in small volumes making it very atractive to migrate in that direction which is why I want to persue this.

 

[cy]

In rundown batt tests, MM+ R2H will light up and happily run at 1.1volts in an almost dead single 123. So I see why this design would overlap MM.

I don't see much gain with single 123. What would be the max output at led w/single AA?

IMHO sammies and all things modular are the wave of the future.

 

[jtice]

Wayne,

If I recall correctly, you are also working on a higher effeciency BB board?
If so, that takes care of the 123 option.
And I would then focus on a 1AA option.

I think the voting has a flaw.
PPL are voting for the most used, and one they want the most.
Well, they might not be taking into account what is already out there, and what is in the works.
If you are working on a better BB already, then theres not much sence in doing a 123 version.

If you arent working on one, then I suppose the 123 option is where I would focus my efforts.

Please dont take offence to this, as I do not mean it to be rude.
But, I would work on something totally different.
IE: A more powerful, more effecient buck/boost. (something I am dieing to see happen)
Or, a multi level board the size of the BB.

These would break off into another direction for you, and give you a much wider variety of products.
While making your currents designs better, is always good,,, making something new and different might be even better.
Especially a good buck/boost. If you could get a really good buck/boost, you would only have to make those, instead of the two seperate BB and DB boards. I would think in the long run, making the one board would be more cost effective.

I do realize that I am asking alot. Its all about small steps, right? /ubbthreads/images/graemlins/wink.gif

Something you might want to do is...
Sit back and say, which board the least effective. (least effecient, least powerful, etc) lol, that would probably be the Wizards.
And then improve it, so that it can catch up to the others.

Actually, this has me curious.
I wonder wht ppl would want more.
A little more effecient, more powerful Wizard, or a kick butt BadBoy. /ubbthreads/images/graemlins/thinking.gif

Just some food for thought.

Thanks for all your hard work Wayne. /ubbthreads/images/graemlins/thumbsup.gif

-John

 

[Rothrandir]

a more efficient wizard would be a dream come true! but i'm not sure how possible it is at this time /ubbthreads/images/graemlins/frown.gif

kind of like asking for a 5ghz cpu? i dunno...

honestly i have no idea as to the plausability of a more efficent buck/boost design, but i think the current technology or parts might not be suitable...

 

[jtice]

That may very well be true Roth.
Keep in mind though, that the board can be made thicker.
As long as it fits in an E-can, I think it would be fine.

Keeping a anode contact on the bottom is pretty much out of the question though. /ubbthreads/images/graemlins/frown.gif

 

[Christoph]

A little more effecient, more powerful Wizard is what I would love to see.
chris /ubbthreads/images/graemlins/grin.gif /ubbthreads/images/graemlins/thumbsup.gif

 

[dat2zip]

Great comments.

Keep them coming. I'll take this into consideration.

Buck-Boost. Would it be possible to set the Vin min around 2V? If a new Wizard was made, there are more possible ICs if we can set the Vin low to be not so low. There are many that cut out around 2.4V.

I would imagine the Wizard would be used with battery combinations around Vf. Like a 3 AA design or a rechargeable Lithium battery.

Also, while we are on the subject, I would think then that having some option or feature for low battery detect status or LED would be manditory. This is possible too.

I see several possilities.

1) I finish Nexgen BB (and call it something else...). This replaces the Badboy in the 1W/3W arena. The current Badboy would still be used for all other higher voltage conbinations like a 5W, two 1W in series etc.

I have not decided on this decision yet. So, it is pending.

2) I build a ultra small low end tiny boost regulator for the single 1.5V cell arena. That's what this poll was for but anything open for discussion. No biggie.

I'm not clear about this since I'm not exactly sure what mods this would go into. If for example more CR2s and NLS were on the horizon and the keychain mod market needed this. Then I would persue this. Maybe, a good host will be found that we can upgrade for dirt cheap. This might open up to a bigger market.

3) sounds like a more efficient Wizard would be nice. I agree that would fill a nitch that the straight buck or straight boost can't solve.

 

[jtice]

Limiting the Vin to 2V.... hmmmm

Well, I suppose that is ok,
I just need a solution, so that I can use a head on a 1x123 or 2x123 body. 3x123 would be a nice bonus, but not needed.
Limiting it to 2Vin might be tricky to get good runtime out of a single 123.
But then, my BB500 dropped out of regulation before the 123 got down to 2 V so, I guess its fine.

I like the better Wizard idea, ALOT.
Second I would love to see a smaller board, that can be used with an AA cell.
If possible to get it to except the Vin form a cr2, the PERFECT.
That could replace the very low powered Arcmania boards.

The small lights like the Larry CR2, KeyLux, and NLS are very popular, and I think we will be seeing alot more of their kind.
A small Arcmania size board running upwards of 500mA+ and that could suck an AA cell DRY, would be excellent.

 

[Rothrandir]

smaller keychain converters are nice, but unless somone steps up to the plate to make the bodies, i'm not sure how much demand you'll get /ubbthreads/images/graemlins/frown.gif

even if you have 300 n-cell bodies, that's not going to be enough to justify a run of converters is it?

personally, i'm most interested in a more efficient buck/boost, and, well, more efficient everything /ubbthreads/images/graemlins/wink.gif

the badboy/madmax/wizard/downboy line is pretty dam great as it is! increasing efficiency would be icing on the cake.

of course options like indicator lights are nice, but i can't see myself using them in every project...