| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
| |
0x65 is for rev3.2 touch panels, and 0x45 is for rev3.1
To resolve multi-touch issue,
1. Added debounce algorithm on switching mode at noisy environment.
2. Added hysteresis on rejecting events at noisy state, to prevent
frequent drop of events and to improve usability.
Change-Id: I62791274bb5e59ef2cf3300977da205c9914aaee
|
|
|
|
|
|
|
|
| |
- Modified input reference voltage from 0.8V to 0.6V of signal for
stable condition, and prevent malfunction at certain devices.
Change-Id: I7a550dde53c017927f96e5f2f1fbcbfc5ad3a588
Signed-off-by: Dima Zavin <dima@android.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
0x70 for rev3.2, 0x71 for rev3.1
0x61 exposes a hw bug where one of the bits in acc register jitters
and shows up as ghost touches.
Previous fw was 0x5e, but it was for 48Hz refresh rate. So, go back to 0x70
Change-Id: Ied7aaaebfb18aeb4f072e3f31898b8335a1760b6
Signed-off-by: Dima Zavin <dima@android.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
- changed signal frequency to suit 60fps LCD
- have implemented features which fixed issues related to queue.
- Changing frequency is made to gain more margin at noisy environment.
Since LCD at 60fps has harmonics noise of 77.7kHz. It is safer to use lower signal
frequency to avoid the noise. i.e if we use 416kHz, 5% of variation to either
direction cause 41.6kHz shift, while 34.5kHz is shifted at signal frequency 345kHz.
Change-Id: I2dbde4f3d5b88466db1e010a506030cc698c184b
Signed-off-by: Dima Zavin <dima@android.com>
|
|
|
|
|
|
|
|
| |
This should be equivalent to 0x70 but for 48Hz screen refresh
rate.
Change-Id: Icc552544bba9aefa72d63a081499dd5995087d18
Signed-off-by: Dima Zavin <dima@android.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
0x70 for rev3.2, 0x71 for rev3.1
Changes
- increased frame rate on idle state to improve response on state
transition.
- Added queue on event report to prevent touch off report without
touch on report.
Change-Id: I9857a93c64d7c904f72df8e6783ce880d769ec1e
Signed-off-by: Dima Zavin <dima@android.com>
|
|
|
|
|
|
| |
Go back to version 0x57/0x37 to work with 60Hz refresh
This reverts commit ce92f4bd9aff37d619dabafb8144856ab345fee1.
|
|
|
|
|
|
|
| |
Updates due to change of LCD frequency was made (48Hz vs 60Hz)
Change-Id: Ifdd322e20eba7062c43c19effbd7d3b6e7b24b25
Signed-off-by: Dima Zavin <dima@android.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Updates are described below.
- Suspend/Resume handling has been modified due to i2c failure on wake-up
state.
We have de-glitch function on SCL/SDA implemented which should be turned
off on sleep mode. (de-glitch is activated by internal oscillator and can't
be produced while we turn off oscillator at sleep mode)
Problem might has been caused by toggling de-glitch function on/off when
i2c communicate was on process.
And I think sleep/wake-up command could not be received at instance at
normal condition, but I scoped a signal transmitted to IC that is composed
of
sleep command and wake up command within 150us, which could cause IC
trouble for handling toggling de-glitch function.
Added an hardwired i2c reset command, right after toggling
de-glitch(sleep/wake-up) to prevent i2c failure.
Change-Id: I8512ba7b33485817dc4386c765b6f4228233da8d
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Changelog from Melfas:
- During initialization after changing mode (recovery from ESD state, or
sleep/wake-up, etc.) there was also hardware initialization process also.
Which could have caused an i2c failure if i2c was on process.
- Also modified tuning value for touches on bottom of screen to be more
sensitive
Change-Id: I7b3424f21b85bffb20d4e5a2dd88026c6654919b
Signed-off-by: Dima Zavin <dima@android.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Updates are made to fix ghost touches on noisy circumstance, and is
described below.
- Configured seed value for DSP
- Modified point merging algorithm which previous version was vulnerable to
ghost touches on noisy state
- State change between idle and active mode on noisy circumstance could cause
false reference track and has been improved.
- Added self-recovery logic on ESD state - not related to ghost touches.
Change-Id: I488ddb4219f827ea8cee88528871381b017d9cc5
Signed-off-by: Dima Zavin <dima@android.com>
|
|
|
|
|
|
|
|
|
|
| |
Fixed a bug that can cause 'missing reference value track'. i.e.
firmware loses tracking and gives a touch on still. The bug could
have produced a touch fail on thermal variation or long term usage
without putting TSP into sleep mode / wake up.
Change-Id: Ieee26f7e2aae12ab826048f55ed8232c1dd590c5
Signed-off-by: Dima Zavin <dima@android.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Updated firmware 0x53 correspond to rev.32 and 0x33 correspond to rev.31
are attached hereto.
Updates are described below.
- Added an verification code at the end of binary, so the firmware can detect
flash failure.
It will return value 0xFF when host tries to access version information in
case of flash failure.
- Adjusted sampling parameters to fix ghost touch issues.
Ghost touches can appear due to oscillator deviation of touch IC or thermal
effect.
Algorithm for to detect and avoid noisy frequency band is added.
Change-Id: Iee6b270aa84a148fcb755e07beae0b989ffbd726
|
|
|
|
|
|
|
|
|
|
| |
- Fixed bug on clearing queue during sleep/wake-up.
Writing sleep command while touch IC is pushing data into queue,
could have caused non-empty queue after wake up sequence.
Change-Id: I958836b96f63ff1170c7bde349c019eb596de22e
Cc: Sewoong Moon <swmoon@melfas.com>
Signed-off-by: Dima Zavin <dima@android.com>
|
|
Updates for 0x50 are described below.
- Analog parameters for sample circuit has been optimized to new
hardware.
- Embodied new feature that firmware automatically recognizes noisy
environment and changes it's operating mode.
- Precision on bottom of touch screen has been improved.
Change-Id: I0f01893ff77d902f84eae1d409afa86cde1f505f
Signed-off-by: Dima Zavin <dima@android.com>
|