LED bicycle light

[Ryan]

I'm designing a bicycle light as the darkness is becoming longer and
longer this time of year.

I have selected the Luxeon III LEDs as they have good power in a small
package. I plan to use 3 of these LEDs in a single housing with the
Fraen trilens. Each LED will have its own power supply. These
require 4.5V at 700mA.

So my question is regarding the power supply. The battery is a NiMH
10.8V pack. I would like a constant current switch mode supply and
the maximum efficiency possible. It would be desired, but not
required, to run at lower than 700mA.

My thoughts are to implement a hysteretic synchronous buck converter
in a Cypress PSoC. I would use one of those fancy IRF FETKY devices
with both N MOSFETs and diode in one package. I'd need a charge pump
for the high side MOSFET too.

Would a hysteretic synchronous converter provide the best efficiency?
Is there a better topology? Has anyone done such a thing in the PSoC?

Any input is appreciated.

Thanks,
Ryan

 

[Ban]

I'm designing a bicycle light as the darkness is becoming longer and
longer this time of year.

I have selected the Luxeon III LEDs as they have good power in a small
package. I plan to use 3 of these LEDs in a single housing with the
Fraen trilens. Each LED will have its own power supply. These
require 4.5V at 700mA.

So my question is regarding the power supply. The battery is a NiMH
10.8V pack. I would like a constant current switch mode supply and
the maximum efficiency possible. It would be desired, but not
required, to run at lower than 700mA.

With the LT1618 you can implement a tiny constant current switcher with
options of current adjustment and on/off switch. Only a few external
components needed. See design notes in EDN.

 

[Ryan]

With the LT1618 you can implement a tiny constant current switcher with
options of current adjustment and on/off switch. Only a few external
components needed. See design notes in EDN.

Hi Ban,

That design note (Linear-Tech #340) described an interesting
configuration for a boost (operating as buck) converter, although the
efficiency is only rated at 70%. I'm hoping for over 90% with
synchronous conversion. Thanks for the heads-up though!

All,

Does anyone else have any comments on an efficient buck converter
design for driving a 3W LED? Does the synchronous hysteretic
converter sound plausable? I would actually prefer to make my own
converter, just would like some confirmation on the topology.

Perhaps this is child's play for all of you, and I should go to s.e.b.

I'm sorry if I'm not interesting enough, or perhaps a "leech", or
maybe just not very political. I hope you all forgive me.

Ryan

 

[John Fields]

Does anyone else have any comments on an efficient buck converter
design for driving a 3W LED? Does the synchronous hysteretic
converter sound plausable? I would actually prefer to make my own
converter, just would like some confirmation on the topology.

Perhaps this is child's play for all of you, and I should go to s.e.b.

I'm sorry if I'm not interesting enough, or perhaps a "leech", or
maybe just not very political. I hope you all forgive me.

 

[Guy Macon]

Yeah, bite the hand that feeds you; that's the way to do it!

That comment certainly took away *my* motivation to help him.

 

[Spehro Pefhany]

That comment certainly took away *my* motivation to help him.

I'm sick of LED supply controllers anyway. They introduce new chips
for them every week, it seems, and the application notes are full of
them. Doesn't Luxeon suggest anything appropriate?

Best regards,
Spehro Pefhany

 

[Tim Mitchell]

I'm sick of LED supply controllers anyway. They introduce new chips
for them every week, it seems, and the application notes are full of
them. Doesn't Luxeon suggest anything appropriate?

I don't know of many (any?) of the LED controller chips which will cater
for the Luxeon's 350mA. They all seem to be for mobile phone
backlighting.

 

[Spehro Pefhany]

I don't know of many (any?) of the LED controller chips which will cater
for the Luxeon's 350mA. They all seem to be for mobile phone
backlighting.

Some 1/4-VGA color LCD displays are using LED backlighting these days.


Best regards,
Spehro Pefhany

 

[Ryan]

At least I provoked a response. I've been a lurker on-and-off on this
newsgroup for years. I don't ever remember it being as bad as of
late. It could be selective memory.

Many other engineering groups are much more civilized, why is the EE
group so nutty and off-topic?

In any case my response was directed at those who made the comments I
alluded to and not everyone on the group. I know there are still some
respectable people here.

Ryan

 

[Dirk Bruere at Neopax]

At least I provoked a response. I've been a lurker on-and-off on this
newsgroup for years. I don't ever remember it being as bad as of
late. It could be selective memory.

Many other engineering groups are much more civilized, why is the EE
group so nutty and off-topic?
ARSE!

In any case my response was directed at those who made the comments I
alluded to and not everyone on the group. I know there are still some
respectable people here.

Thank you.

--
Dirk

The Consensus:-
The political party for the new millenium
http://www.theconsensus.org

 

[Ryan]

I'm sick of LED supply controllers anyway. They introduce new chips
for them every week, it seems, and the application notes are full of
them. Doesn't Luxeon suggest anything appropriate?

Best regards,
Spehro Pefhany

I think you're missing the point. I'd like to make my own switching
controller. I would prefer not to use a COTS chip, but might if I
fail in my design. Also I could glean information from datasheets to
help with the design.

I think most people groan when they see the LED bicycle light subject.
In hindsight I could have eliminated mentioning the LED altogether.

Ryan