mouldy.org

I should point out that these early posts are going to be out of order; my actual progress is way ahead of what I've written up, and some posts (like this one) are a convenient dumping ground for my own information.

The idea I've decided on is:

Design, manufacture and sell a driver module for high-power LEDs. It should take any input voltage and be able to run a string of up to three Luxeon K2 LEDs at full power. It should be physically small and lightweight so that it can be integrated into battery-powered products.

There are some implicit specs there:

• it'll be a buck-boost converter
• there will be some limit to the voltage range it can accept, to be determined by the components being used
• it'll need to be able to supply 1.5A at about 13V

First, time to determine - using facts and evidence - whether there is a need for this product. First, the information I've been carrying in my head.

• I want it for my own projects. So I'm a customer.
• I've seen devices such as the Buckpuck, Boostpuck, and a few DIY options. Other people think this is a good idea.
• I haven't seen anything that can run the Luxeon K2 at full power. It comes out pretty soon - within a few weeks.
• A company that I bought LEDs from once will be getting the K2, but admit that they don't have a driver for it. They expect that it will be a new product soon.

So some people want it. I have no idea how many. There's also a good chance that a similar product will come into being really soon - I have the opportunity to be a market leader, but the threat that someone will beat me to it.

Applications?

What is my market? I'm starting with "people who want to run high-power LEDs", but there are plenty of solutions for that already. "People who want to run the Luxeon K2 at full power" is good, because there's nothing there yet (to my knowledge).

Narrow it down at all? The DIY crowd will likely be a good beta customer; there are lots of people building funny custom lights using LEDs, and I bet a lot of them are champing at the bit to get the extra power that a Luxeon K2 offers.

Who's manufacturing LED products? I know of bike lights, and that's where my own interests lie - so they're definitely a possibility. Car headlights? Most high-end lights are HID derivatives nowadays - LEDs aren't so compelling. They might find some use for indicators and tail-lights, though. Torches, definitely, but they're a niche product right now. They also tend to build their own converters to fit their own space restrictions. Google also provides the ideas of

• Headlamps - a definite possibility, since they need power efficiency, reliability and durability
• Halogen lamp replacement in domestic lighting. The Luxeon K2 would bring a single lamp closer to (but still a long way off) the brightness of a halogen. I don't personally understand why you would replace a halogen lamp with an LED - halogens are brighter, prettier, and you don't care about power efficiency in the home - but there seems to be a lot of interest.
• Lamp replacement for power-sensitive areas - caravans, space shuttles, solar homes. It's a possibility; the K2 is certainly compelling there.
• Garden and mood lighting. These are always pretty. Some are solar powered, so power efficiency is an issue. This allows them to be deployed without digging up and running cable through your garden. There's also the safety issue with running mains cable underground - I suspect that a lot of them use something like a 12V power bus and each light has its own LED driver. Maybe a battery charger circuit would be useful here?
• Flex lights or fiber optic lights. I don't see any compelling reason for LEDs here - except for battery-powered devices - but there's some interest.
• Dental lamps. I bet this was unexpected for the Lumileds folk. Dentists need high-powered light of a certain wavelength to cure new types of fillings. It's probably difficult to filter a traditional incandescent properly, and specifically coloured lasers are insanely expensive, so Lumileds has jumped on this one and produced an LED with the correct wavelength (and triple the price). Awesome idea. Good for me? Probably not. I have no desire to get into markets where that level of quality control is needed. This is a very niche application that has been around for a while - the LED was produced for a specific application - so I have no doubt that the driver role has already been filled. I still love the idea.
• Safety lights. Usually low power, so not much need for the Luxeon K2 there. Having a flexible boost converter might be useful when your batteries run low, though.
• Long distance optical communcation systems. Probably more of a laser application.
• Small video projecters. Wikipedia suggests that some small projectors use LEDs for light. I find this implausible - they're just not that bright - but the K2 will make this more plausible.
• TV and display backlighting. Lumileds is pushing this one. Again, I'm unconvinced. I can't think of a way for my converter to be a compelling choice for them, particularly since they're in a cost-sensitive market. There's the HDR TV thing, but they'll be looking at a large array of small LEDs, not a small array of large LEDs.
• Lighting for case mods. There's definitely cash being thrown around there, and a need to make basic electronics easier. None of the existing high-power LED manufacturers have approached it at all - so there's an opening.
• Display backlighting for mobile phones and PDAs. I'm certainly not going after this market - they use tiny LEDs, have tiny amounts of space, and massive engineering budgets. There are plenty of chips to handle this already from the major semiconductor companies.
• Pseudo-flashes for phone cameras and tiny digicams. They might like the K2 there due to the brightness, but they're getting along alright using conventional LEDs. In the space you could fit a K2 and a driver, you could put a normal flash, and save power in the process. It's also a very cost-sensitive market. No go.

So the top markets look like battery-powered devices. This makes sense, because if power supply was a non-issue, you'd use a halogen, krypton or HID lamp. This research al so suggests two products - a basic "single LED at 1.5A driver" with no configuration options, and a more complicated "multiple LEDs at 1.5A with dimming and configuration" option. This is, unsurprisingly, what has happened with the LuxDrive Buckpuck and Bucktoot converters.

Competitors?

LuxDrive is right up there. They make a range of buck and/or boost converters for high-powered LEDs. So far, they don't have a 1.5A buck converter, and they don't have any boost converter rated for more than 350mA. Is there a technical problem with making high-powered boost converters? Could it be that they just couldn't fit it into the space available?

A number of established semiconductor companies - Allegro, Maxim, Linear, Sipex - make drive ICs. Few have a rating greater than 350mA. There are a few high-power control IC's, but they don't do the whole job - you need to build a circuit to use them. So I might be able to use those ICs and sell complete converter boards. They're difficult to find, and there's a lot of information to dig through - so there's a definite marketing gap there.

ledsales.com.au sell a boost converter board for 1W LEDs. It looks like they manufacture it themselves. They're more likely to be useful as a sales channel down the track.

Mightex sell 'LED controllers' - one of which is capable of 1A continuous, 3.5A pulse - but they're targeting very different markets to what I am. There's not much information on the website, but from what I can ascertain, their converters aren't very portable.

The Sandwich Shoppe sells buck and boost converters in similar markets to what I'm targeting. Marketing is weak - I only found it because I knew about it already. Their stuff goes up to 1A. It looks like it's just one guy designing and selling the products. They already have some traction.

Solus Corporation has some 350mA and 700mA buck converters.

TaskLED makes drivers - including a 1000mA boost converter called the FatMan. They most closely match what I'm trying to do. Again, there's not a lot of marketing, but they have some traction with customers.

Customers?

There are two broad classes of customers that I can think of: individuals building their own custom light (DIYers) and OEMs that want to integrate the device into their own product.

DIYers will probably only buy one item at a time. They'll be bothered more by shipping costs, but less by the device cost, so long as it solves their problem. They'll be a bit more tolerant of bugs that come up in the early stages.

I haven't explicitly considered the problem that I'm solving: to make it easy to use Luxeon K2 LEDs. I save the development cost of making your own converter. I don't anticipate that this will be massive, which sounds like a problem in theory. Risk reduction? For most individuals, designing your own converter isn't an option; it's too time consuming, difficult, and most individuals don't have the ability or inclination to do so. Most components have to be ordered in large quantities or with high shipping cost, which is undesirable if you just want one part.

OEMs will want larger quantities, but at a lower price. They'll probably be fussier about physical appearance and size. There'll probably be custom design work there. It's a difficult customer to approach right at the start with an unproven product - unless I find someone who must use the Luxeon K2. There's simply nothing out there that can drive them at full power.

Individuals will be a better initial customer, but OEMs will be more lucrative in the long term. 

I've had a lot of trouble finding numbers for this stuff on the free Internet - primarily, how many customers there are for this stuff. I expect that it's not a lot - hundreds of units per year. This doesn't concern me too much, because I'm aiming small at the moment. If I can sell one hundred units, awesome. It's not a lot of cash in absolute terms, but it should cover my development costs and teach me a lot in the process. 

Customer location

Most customers will be in the United States. This is something I can't get around easily. Overseas shipping is a price penalty - but not terrible. The exchange rate favours US customers, so it's probably not as bad as I think it will be.

I think I'll completely own the Australian market. Buckpucks and Fatman drivers are horrendously expensive once you factor in the exchange rate and international shipping costs. I know there's a small DIY LED light industry over here.

The device itself will be quite small, and so can probably be posted as a letter rather than a parcel. This helps shipping costs. 

If I can find a US distributor, that can reduce the shipping penalty and provide a sales channel to boot. This is definitely something that I should pursue. 

I think geography will be a disadvantage, but not crippling.