I need add Bluetooth to my hardware. I never worked with Bluetooth before, so I was wondering which approach I should use. I some did quick research on internet and found some solutions using a Bluetooth IC like CC2540 and CC2541 from TI. I also found solutions using a module such as the RN41 and RN42 from Microchip as well as solutions using a dongle.
I need to add bluetooth functionality to my hardware, however, I want it inside the system with the hardware, and not using USB if possible. That leaves the question between the module or chip.
With the CC2541, I will need to make the antenna on my PCB, but the RN41 comes with an integrated antenna. The price of a RN41 is 4.5 times the CC2541 price. Probably I will need some hundreds of units, so… I'm wondering, besides the antenna, how much more work would I have using the CC541 instead of the RN41?
I read this article on Digikey, and here is a quote that stood out to me:
If your team has sufficient RF expertise and the software skills to
integrate the controller’s application code with the code required to
support the protocol stack, investing 12 to 18 months of development
time will be handsomely repaid by a more efficient design and lower
production costs.
I would like to know what is necessary to develop a Bluetooth solution and integrate it to an embedded design? What do I need to add Bluetooth to my hardware, starting with an IC and not using a module?
Best Answer
Let's start with the bottom line: What is the total price difference between the IC solution and the module solution (Taking only the CC2541 vs. the RN42)? This turns the question into a balance between NRE and BOM spending.
Using optimistic numbers for your sales, let's look at prices per unit at 1k volume. I'm going to use the manufacturer's 1k prices. This gives us a better price per part, but really this is well suited for spreadsheet analysis:
This gives us a naive budget difference of over $100,000 for making 1,000 devices. The question you must ask yourself is "will developing for the CC2541 cost me less than $101,100?*"
(*) Insert your actual pricing numbers.
Note that your margin shrinks rapidly as you sell fewer and fewer units, or the closer the two are in price. Overestimate your sales volumes and you may have spent a bunch of time and money that you couldn't afford in order to develop a cheaper solution that not enough people want.
If I take the Digikey timeline figure at face value and pay one to two people to work on this at $50k/year, that puts labor costs between $50k and $150k, but quite possibly more. Don't forget to pay yourself! Don't forget that they qualified their statement with
How long do you think it would take to get the module working? Probably not nearly as long as the IC, but the product development may still be months.
Keep in mind that there are additional costs to the CC2541 besides just a longer development cycle. You may have more hardware revisions due to difficulties with the antenna system. You will need FCC certification, which will be more expensive for the IC. There are many variables. Assume that it will be harder than you thought, and will take longer than you thought, especially if you are unfamiliar with the technology.
Given these numbers, below 1,000 units I would probably stick with the module.