Where can I find an Embedded Developer?

You came to the right place, but in all honesty, a canned FM chip is probably the lowest cost option for you, that would also do a decent job. You need a high amount of gain and a discriminator (usu coil based for low cost, but can be PLL based for tight space requirements, albeit for a higher price). Filtering in general is hard to do in tight spaces, but it's needed if you want to receive a weaker station - having almost no filtering would possibly work for very strong local stations). Do you want a speaker in that device, or do you play it through the soundcard? I'm guessing the soundcard, to keep it cheap. You also compete with radio stations that have web feeds in a streaming media format. This kind of thing is commonly used among your target audience (young adults, techno-savy). But your device has value in workplaces that might filter out streaming media. Your idea is very cool! But I think you want it to be cheap enough that a station can buy thousands of them and give them out for free. This is going to be difficult. You also have to factor in the cost of a USB interface chip. You might also need to have software loaded on the PC, or else you might make a live self-running program on the USB device - this is very new technology. Some companies don't allow "regular" employees to access their USB ports due to security concerns. This is another gotcha.

Thanks Eric,

You pretty much hit everything on the head. Yes, I want to use the sound card. That is why I was leaning toward the Audio Class. Although, a Mass Storage device with the player in the flash has alot of appeal.

Yes, I want to keep it cheap enough to where it can be bought by the thousands. If the device costs too much (i.e. $20) I might be able to offset the the costs with other promotional items.

Thanks for the advice on the FM chip. Do you have any recommended chips? I have seen some priced at $3.50 in bulk Thanks again!

If you have some candidate chips in mind for the USB microcontroller, check the manufacturer's websites. Most manufacturers have some kind of referral program to help you find consultants in your area. There are also ads in Inventor's Digest and on their website for designers who take small jobs.

SHAMELESS PLUG: We are Cypress CyPros certified USB & PSoC consultants. We offer a discount to independent inventors (as advertised in Inventor's Digest).

Hello,

$20 sounds way too high for a USB-FM receiver.

This stuff requires tough negotiations. Just went trough that with a uC. You usually need to talk with the manufacturer directly and make it crystal clear that the design-win for them won't happen unless it's under $x.xx at so many thousand parts a month. It does help to point out a few competing chips. Sometimes it can feel like buying a car ;-)

Just FWIW: You will most likely need two engineers. One who can design the radio part and another one who takes the data and does the software/firmware part.

Another issue to contend with is noise. Where's the antenna going to be? What kinds of in-band "birdies" will the attached computer gear generate?

Regards, Joerg

Provisional patent is a bad idea. Can trip you up later on when you file the final patent. Skip the provisional and just file the patent.

What is "bulk"? IC-based FM receiver designs sell for under $5.00 _RETAIL_. The receiver IC would be about $0.75.

Hello Lewin,

That's because they buy in super duper bulk. A few hundred thousand per shipment or so. You really have to negotiate this stuff hard.

Specialty chips aren't always the answer. For many of my designs I ended up going discrete. Even on some switch mode supplies where the PWM controller chip just wouldn't come below 50c. The 20 or so parts that replaced it were under 25c including placement and all.

Regards, Joerg

Depends what you call super duper... which is why I asked. I am responsible for designs that use about 15,000 per month of receiver chips.

Did anyone mention the Silicon Lab USB FM Raido Reference Design?

It uses the Si4701 FM radio chip. Still over $3 for 10,000 last time I looked. A tad high for my taste.

Regards, Joerg

Hello Lewin,

That qualifies as super duper. 75c would be ok, anything above a Dollar is hard to justify. Unless you really, really don't have the space. At those qties you could almost have an ASIC made.

Regards, Joerg

I've contacted Silicon Labs. multiple times for the technical datasheet of the chip and for a place where to buy them (in small quantities). They haven't replied at all, so I imagine they aren't very keen on having small-time customers.

But if you're planning on having a huge quantity made, then I'm sure they'll be much more forthcoming towards you.

I, however, wasn't pleased with them. Going to use a Phillips TEA5767 in my USB FM project (thesis work)

Greetings, Antti Keskinen

Hello Antti,

That does not sound like a good business philosophy. Not supporting the next generation of engineers can quickly lead to falling design-in rates. Relying on a few large customers alone usually leads to problems down the road as can be seen with some European semiconductor manufacturers. Excellent products but disappointing sales.

Regards, Joerg

In America the electronics manufacturers are slowly beginning to understand the need to support the small design shops better. They are slowly understanding that their bulk sales will be off shore but a great deal of design work is done in the US and elsewhere. And if they do not support us someone else will.

Hello Nappy,

They began to understand that early on and some like AD were always supportive. But most EU companies do not have a clue in that respect. Consequently my design-in rate for their stuff went from around 30% to really close to zilch.

Regards, Joerg