Oops, I accidentally replied to Mychaela directly, here's our conversation so far, please others join in with thoughts.
On Sun, Mar 10, 2019 at 09:09:31AM -0800, Mychaela Falconia wrote:
Hi Craig,
Thanks again for your details Mychaela. I will read this very carefully and keep it in mind as I try to port osmocom-bb to fernvale/mediatek. I already expect a good dose of refactoring in layer1 so maybe I will touch on fixing the issue.
Have you got yourself a CMU200 instrument or not yet? A CMU200 is absolutely required if you wish to have any chance of success at your idea of OBB on MTK.
I do have a Racal Instruments 6103E GSM Digital Radio Test Set which seems to work well so far. I have tested some SIM800 modules (mtk6261) with the internal BTS. Not sure if that will fall short or not but I'll take things one step at a time. When I need a CMU200 I'll probably get one. :)
Another question: is there any way to run any official MTK-based firmware, however non-free and blob-laden it may be, on Fernvale hw?
Probably. I know from reading that Bunnie and Xobs who designed fernvale did run standard firmware/OS on some similar 6260 based devices. Maybe we could track down a good firmware and test on fernvale.
For what it's worth the fernvale isn't exactly my main target. I would much rather use a SIM800 module or eventually design of my own.
Has anyone actually done it and posted howto instructions? If not, then how do you know if the RF hw on your Fernvale kit is actually any good and not physically defective? Whoever physically manufactured those Fernvale boards sold by Sysmocom, have they actually tested their RF hw and how? And what about calibration - have they performed individual per-unit calibration of various RF parameters on those Fernvale kits like every standard GSM MS hardware manufacturer is required to?
I don't have any clue about these details. Will have to rely on Harald or maybe inquire of Bunnie and Xobs.
I calibrate my FCDEV3B boards at production time, use the calibration procedure to also serve as a test of the RF tract, and ship every unit with turnkey-working official firmware and a fully valid and legitimate IMEI, fully fit for operation as an MS on public GSM networks - do they do likewise or not? If not, then they are substandard.
M~
Thanks for your efforts, Craig
Craig wrote:
I do have a Racal Instruments 6103E GSM Digital Radio Test Set which seems to work well so far.
Sorry, I don't have any experience with that instrument, and therefore won't be able to tell you whether or not it would make an acceptable substitute for a CMU200. The CMU200 is the only instrument I have first-hand experience with.
Another question: is there any way to run any official MTK-based firmware, however non-free and blob-laden it may be, on Fernvale hw?
Probably. I know from reading that Bunnie and Xobs who designed fernvale did run standard firmware/OS on some similar 6260 based devices. Maybe we could track down a good firmware and test on fernvale.
Firmwares are made for boards, not chips, and there is no such thing as "a good firmware" which you could just "track down". You cannot take fw built for one board and expect it to work on a different board, so if the makers of the Fernvale have not made a customized version of MTK's firmware specifically for their board in the same way how I maintain my own version of TI-based fw specifically for my FCDEV3B, then you are basically screwed on their hw, and won't have any way of testing if their RF hw even works.
[RF testing and calibration]
I don't have any clue about these details. Will have to rely on Harald or maybe inquire of Bunnie and Xobs.
My guess is that they probably skipped all RF test and calibration steps which all standard GSM MS manufacturers are required to do, and just ship completely untested RF hardware.
This Fernvale is the most misguided hardware project I have ever seen, and I have seen a lot of dumb ideas in my professional career of about 20 y so far.
M~
baseband-devel@lists.osmocom.org