3 ms·

Can someone explain to me why don't we have open source WiFi chips? Are they so complicated that we can't get a open hardware pcie adapter?

by stein1946 3y ago

miah_ 3y ago
Likely due to regulations, fcc, certification costs, etc.
zamadatix 3y ago
The hard part of open hardware isn't the technical complexity by itself. I'd even go so far as to say there is no such thing as something so complicated we can only make a closed implementation of it, just see Linux. The questions are more like "Do you want to fund the hardware investment, including legal certifications since this is consumer radio technology, on the hope it'll sell enough to make it back?", "Do you want to continuously design new versions to new standards every 5 years so it doesn't just become obsolete anyways?", "Do you want to maintain the firmware and driver, answering all of the issues that come about connecting to thousands of other devices across thousands and thousands of pages of standards behavior you or they implement poorly?", and, moreso, "Do you want to do this for the rest of your life all out of principle that client Wi-Fi chips, which the high end of the latest generation go up to a whopping $16.55 per in bulk, sometimes have bugs and you can't go in and just fix the particular bug you want to? Now sure, maybe you can convince your closest 10 dozen friends interested in open source hardware to not put everything on one person but at the end of the day there are a few people it's going to be a significant burden on and they have to see it as a worthwhile burden otherwise nobody is going to run the project. It'd probably be easier (though not easy) to stir up enough support to convince an existing manufacturer to make their firmware open source instead.
rcxdude 3y ago
Even much simpler chip functions tend not to have an open-source implementation. Open-source doesn't mesh very well with the economics of chip fabrication, is the main reason (in general it doesn't work with hardware as well, but chip fabrication is probably the most inaccessible of hardware types). And yes, wifi is complex. Probably the only function more complex is cell phone radios.
PeterisP 3y ago
It would be complicated and expensive (yes, WiFi is that complicated), and you likely couldn't sell that open hardware adapter so there's no motivation to make one. The law states that "an intentional or unintentional radiator must be constructed such that the adjustments of any control that is readily accessible by or intended to be accessible to the user will not cause operation of the device in violation of the regulations" (https://www.law.cornell.edu/cfr/text/47/15.15 https://www.law.cornell.edu/cfr/text/47/15.15 ) So if the manufacturer makes a device where changing the firmware is "readily accessible" to the user and there is an open source firmware available that can circumvent the FCC transmission restrictions (for example, change the power limits or channel limits for wifi physical layer), then that would be grounds for FCC refusing to certify that device, as it is not permitted to make, import or sell general unrestricted transmitters to the general public (there are certain exceptions for licensed operators, ham radio, experimental use by manufacturers etc). It's similar to other clauses that prohibit manufacturers from making it easy for the user to modify the equipment - e.g. 15.203 (https://www.law.cornell.edu/cfr/text/47/15.203 https://www.law.cornell.edu/cfr/text/47/15.203) "the use of a standard antenna jack or electrical connector is prohibited." so that the user can't easily replace the antenna with a different one from what was certified.
pabs3 3y ago
There is some open source firmware for very old WiFi chips: https://wiki.debian.org/Firmware/Open#Radio https://wiki.debian.org/Firmware/Open#Radio There is also some FPGA based open source WiFi SDR things: https://github.com/open-sdr/openwifi https://github.com/open-sdr/openwifi
jxj 3y ago
To my understanding, the main difficulties are: 1. The investment needed for a real "free" Wi-Fi chip is much higher than most people's (FPGA, DIY, enthusiast and garage guys) imagination. I won't go through all the complications (not only design, but also tool chain, cell libs, etc.) here. Just one hint: think how many thousands pages are there in the Wi-Fi standard pdf, and how many test cases/vectors would be extracted from that before the actual tape out. 2. The chip business is purely the game of volume – if we want to not just burn money, but actually make it successful from the business point of view. We discussed with some businessman, investors and big companies in the chip/Wi-Fi industry regarding the "free" Wi-Fi chip business opportunities and failed to convince any of them. In other words, so far, no serious man actually believes that the "free" Wi-Fi chip could be financially successful. (Keep in mind the game of volume). A key question to you: Will you buy a Wi-Fi dongle (with free Wi-Fi chip) with higher price and lower performance/functionalities (This is the result of low investment and low volume compared to those big players: Qualcomm, Broadcom, Mediatek, Realtek, etc.)? Even if you say you will, could you estimate how many people are there like you in this world? For chip: NO volume == NO business. This is Xianju who created the opensource Wi-Fi chip project openwifi: https://github.com/open-sdr https://github.com/open-sdr It is still in FPGA status, not a real chip yet. We have been working on this for 7 years (starting from the time project setup internally).

Can someone explain to me why don't we have open source WiFi chips? Are they so complicated that we can't get a open hardware pcie adapter?

miah_ 3y ago

Likely due to regulations, fcc, certification costs, etc.

zamadatix 3y ago

The hard part of open hardware isn't the technical complexity by itself. I'd even go so far as to say there is no such thing as something so complicated we can only make a closed implementation of it, just see Linux. The questions are more like "Do you want to fund the hardware investment, including legal certifications since this is consumer radio technology, on the hope it'll sell enough to make it back?", "Do you want to continuously design new versions to new standards every 5 years so it doesn't just become obsolete anyways?", "Do you want to maintain the firmware and driver, answering all of the issues that come about connecting to thousands of other devices across thousands and thousands of pages of standards behavior you or they implement poorly?", and, moreso, "Do you want to do this for the rest of your life all out of principle that client Wi-Fi chips, which the high end of the latest generation go up to a whopping $16.55 per in bulk, sometimes have bugs and you can't go in and just fix the particular bug you want to? Now sure, maybe you can convince your closest 10 dozen friends interested in open source hardware to not put everything on one person but at the end of the day there are a few people it's going to be a significant burden on and they have to see it as a worthwhile burden otherwise nobody is going to run the project. It'd probably be easier (though not easy) to stir up enough support to convince an existing manufacturer to make their firmware open source instead.

rcxdude 3y ago

Even much simpler chip functions tend not to have an open-source implementation. Open-source doesn't mesh very well with the economics of chip fabrication, is the main reason (in general it doesn't work with hardware as well, but chip fabrication is probably the most inaccessible of hardware types). And yes, wifi is complex. Probably the only function more complex is cell phone radios.

PeterisP 3y ago

It would be complicated and expensive (yes, WiFi is that complicated), and you likely couldn't sell that open hardware adapter so there's no motivation to make one. The law states that "an intentional or unintentional radiator must be constructed such that the adjustments of any control that is readily accessible by or intended to be accessible to the user will not cause operation of the device in violation of the regulations" (https://www.law.cornell.edu/cfr/text/47/15.15 https://www.law.cornell.edu/cfr/text/47/15.15 ) So if the manufacturer makes a device where changing the firmware is "readily accessible" to the user and there is an open source firmware available that can circumvent the FCC transmission restrictions (for example, change the power limits or channel limits for wifi physical layer), then that would be grounds for FCC refusing to certify that device, as it is not permitted to make, import or sell general unrestricted transmitters to the general public (there are certain exceptions for licensed operators, ham radio, experimental use by manufacturers etc). It's similar to other clauses that prohibit manufacturers from making it easy for the user to modify the equipment - e.g. 15.203 (https://www.law.cornell.edu/cfr/text/47/15.203 https://www.law.cornell.edu/cfr/text/47/15.203) "the use of a standard antenna jack or electrical connector is prohibited." so that the user can't easily replace the antenna with a different one from what was certified.

pabs3 3y ago

There is some open source firmware for very old WiFi chips: https://wiki.debian.org/Firmware/Open#Radio https://wiki.debian.org/Firmware/Open#Radio There is also some FPGA based open source WiFi SDR things: https://github.com/open-sdr/openwifi https://github.com/open-sdr/openwifi

jxj 3y ago

To my understanding, the main difficulties are: 1. The investment needed for a real "free" Wi-Fi chip is much higher than most people's (FPGA, DIY, enthusiast and garage guys) imagination. I won't go through all the complications (not only design, but also tool chain, cell libs, etc.) here. Just one hint: think how many thousands pages are there in the Wi-Fi standard pdf, and how many test cases/vectors would be extracted from that before the actual tape out. 2. The chip business is purely the game of volume – if we want to not just burn money, but actually make it successful from the business point of view. We discussed with some businessman, investors and big companies in the chip/Wi-Fi industry regarding the "free" Wi-Fi chip business opportunities and failed to convince any of them. In other words, so far, no serious man actually believes that the "free" Wi-Fi chip could be financially successful. (Keep in mind the game of volume). A key question to you: Will you buy a Wi-Fi dongle (with free Wi-Fi chip) with higher price and lower performance/functionalities (This is the result of low investment and low volume compared to those big players: Qualcomm, Broadcom, Mediatek, Realtek, etc.)? Even if you say you will, could you estimate how many people are there like you in this world? For chip: NO volume == NO business. This is Xianju who created the opensource Wi-Fi chip project openwifi: https://github.com/open-sdr https://github.com/open-sdr It is still in FPGA status, not a real chip yet. We have been working on this for 7 years (starting from the time project setup internally).