12 ms·

Can anyone explain why there is no open hardware phone at this point? During a recent trip to Shenzhen, it was clear that all the components are readily availab

Can anyone explain why there is no open hardware phone at this point? During a recent trip to Shenzhen, it was clear that all the components are readily available. I have had one person suggest that it was the cost of FCC approval that was the holdup and not the technology. Any company that could afford the approval process would not want to open their design. I am not technically versed enough to know if this is correct, however. Does anyone else have a clear perspective on the issue?

yafujifide 13y ago
Can you give more information on the FCC approval process? What requires approval exactly? If people designed open-source hardware plans, then made the hardware, then shipped it, at what point is FCC approval necessary?
- pjc50 13y ago
  I don't believe FCC approval alone is the big bottleneck. You need to test its radio emissions in defined conditions, and that requires some expensive test gear (base station simulator ~= $100k), but realistically you need one of those anyway to debug the device you're developing. A more serious problem is trying to get all the relevant firmware and radio software stack without tripping over IP that is (a) patented and (b) will only be licensed to you under NDA. Qualcomm are particularly bad at this. So you need to choose a suitable SoC with favourable IP licensing. Writing your own "baseband" is a substantial piece of work; some people have embarked on it at http://bb.osmocom.org/trac/ http://bb.osmocom.org/trac/ Canonical's approach is sound here and their costs are realistic. Getting it done for free by people working in their spare time is very unlikely to happen.
ajb 13y ago
The components may be available, but that doesn't mean they are actually open. Low end phone makers just get their software from the SOC OEM, and they just want to ship a phone, not provide a platform for developers, so they don't need liberal licensing conditions.
skeletonjelly 13y ago
I bought an OpenMoko a number of years ago. It was a brave effort but a hard mountain to climb.

yafujifide 13y ago

Can you give more information on the FCC approval process? What requires approval exactly? If people designed open-source hardware plans, then made the hardware, then shipped it, at what point is FCC approval necessary?

pjc50 13y ago
I don't believe FCC approval alone is the big bottleneck. You need to test its radio emissions in defined conditions, and that requires some expensive test gear (base station simulator ~= $100k), but realistically you need one of those anyway to debug the device you're developing. A more serious problem is trying to get all the relevant firmware and radio software stack without tripping over IP that is (a) patented and (b) will only be licensed to you under NDA. Qualcomm are particularly bad at this. So you need to choose a suitable SoC with favourable IP licensing. Writing your own "baseband" is a substantial piece of work; some people have embarked on it at http://bb.osmocom.org/trac/ http://bb.osmocom.org/trac/ Canonical's approach is sound here and their costs are realistic. Getting it done for free by people working in their spare time is very unlikely to happen.

ajb 13y ago

The components may be available, but that doesn't mean they are actually open. Low end phone makers just get their software from the SOC OEM, and they just want to ship a phone, not provide a platform for developers, so they don't need liberal licensing conditions.

skeletonjelly 13y ago

I bought an OpenMoko a number of years ago. It was a brave effort but a hard mountain to climb.