16 Apr 2011 |
Posted by Michael Krufky | 8 Comments.
I've applied a few patches to the tda18271 driver that may improve tuning in certain channel ranges.
tda18271: fix calculation bug in tda18271_rf_tracking_filters_init
There was a misplaced parenthesis in the tda18271_rf_tracking_filters_init function, causing improper calibration. This has potential to improve tuning on some channels, and possibly reduce the time it takes for the RF tracking filter calibration to take place.
tda18271: prog_cal and prog_tab variables should be s32, not u8
Changeset name speaks for itself
tda18271: fix bad calculation of main post divider byte
03 Oct 2009 |
Posted by Michael Krufky | 0 Comment.
There are some TDA18271 bug-fix patches awaiting merge that improve tuning performance significantly. Here's a run-down:
fix overflow in FM radio frequency calculation
Multiplication by 62500 causes an overflow in the 32 bit freq variable,
which is later divided by 1000 when using FM radio. Prevent the overflow
by scaling the frequency value correctly upfront.
fix signedness issue in tda18271_rf_tracking_filters_init
Fix a signedness issue in the function, "tda18271_rf_tracking_filters_init"
which caused the RF tracking filter calibration values to be entirely bogus
07 Sep 2009 |
Posted by Michael Krufky | 2 Comments.
Until now, we've been leaving the slave tuner output and clock outputs of the TDA18271 tuner enabled at all times, even while in standby mode. This was the safest choice, as some tuners may need those features enabled while others do not. Disabling one of these outputs can cause adverse effects when done on a board that needs such features enabled.
I've added a new configuration option to describe which output features are allowed to be disabled when the device enters low power standby mode. I've tested this on a number of boards, keeping the clock output enabled while disabling the slave tuner...
30 Aug 2009 |
Posted by Michael Krufky | 0 Comment.
Some performance issues with the HVR-22x0 boards have lead me to take another look at tweaking the TDA18271 driver.
Steve and I have found that tuning performance improves when the RF tracking filter calibration is performed in a specific order between the master and slave tuners.
I created a new attach-time configuration option that allows for the bridge driver to request for the RF tracking filter calibration to occur at startup of the tuner driver.
After some testing, we found that this alleviates the "second tuner fails to lock" issue.
Until now, this type of configuration was available...
Yesterday I commented (link):
"In an unrelated note Iâ€™ve also made more changes to the TDA10048 driver. Michael found a case where ..."
Background: Michael's TDA18271 tuner driver supports multiple output frequencies in an effort to give slightly better noise protection (and thus performance improvements) to it's attached demod. Theory being is that we can squeeze and extra SNR point or two just by adjusting the I/F between these two parts when moving between DVB-T 6, 7 and 8MHz. (Misnomer: While I've explain these as output frequencies, inside the industry we call these input Frequencies...
I've been testing the HVR-2250 model #88021 and trying to reproduce the problem where users are experiencing a full system hang. This was occurring shortly after the driver loads if you did not have any SAA7164 firmware files present in your system. Good news, I found and patched the bug. Thanks to everyone who'd reported this.
The saa7164-dev tree http://kernellabs.com/hg/~stoth/saa7164-dev/ contains this patch so feel free to test / comment.
The saa7164 -dev tree also contains a fix related to model 88021 support, another good reason to try this out.
I'm still experiencing a syslog 'Hauppauge...