-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi folks -
While trying to make DM2 GPS work good, I became annoyed at the range of behaviours the GPS unit and finally spent the whole latter part of the day meddling. I don't think I found where the problems came from, but I found something that is less good than it should be. GPS stopped working altogether on my boards at the moment (unrelated to ripping the cans off the GPS section I believe) so I don't know where I got to. U7605 seems to be planned as an "automatic external antenna detect" thing, when the external antenna starts drawing current I guess the idea is to switch VCONT to select "OUTPUT" 1, the external antenna RF. However nobody cleaned up the output of the current sensor, eg, with a schmitt trigger to stop oscillatory behaviours for example. The output driving net VCONT shows 810mV with no external antenna connected. Also the output is not so low impedence itself yet drives PNP base Q7602 by 10K, affecting the source. RF mux U7606 (UPG2012TB) that actually selects the GPS antenna: http://www.necel.com/microwave/en/pdf/PG10218EJ01V0DS.pdf specifies low level input selecting "OUTPUT" 2 is min -200mV, typ 0 and max 200mV. The other active level is specified min 2.7V, typ 2.8V, max 3.0V. So you have to drive that chip hard to 0V or the power rail to select either input. But at our normal situation with internal antenna and no load on external antenna, we drive VCONT to 810mV (26% Vdd). - -Andy -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org iEYEARECAAYFAkgWHS4ACgkQOjLpvpq7dMr6GACdFmwHdZFVoERoKtqwFx9UVBOa b8MAnRsy+mLC6UKbt2oNCgk/Ikr4brhN =BYFg -----END PGP SIGNATURE----- _______________________________________________ hardware mailing list [email protected] http://lists.openmoko.org/mailman/listinfo/hardware
