TSC2003: Regarding TSC2003 sensor reading failure

sinngo kitahara

Part Number: TSC2003

Tool/software:

I have a question about how to start up the TSC2003.
When using the TSC2003, it is rare that the sensor value cannot be read correctly. Figure 1 shows the waveforms of SCL (yellow) and SDA (green) at this time. SDA is stuck low at the data area (red circle in the figure).

I assumed that the startup method was incorrect, so I took a software solution based on "Powering Up the TSC2003 Correctly" (toggle SCL 9 times or more after startup, then issue a start).

Figure 2 shows the waveforms at that time.
However, the symptoms of the TSC2003 do not improve. Are the above solutions correct? Are there any other possible causes?

over 1 year ago

0 Daveon Douglas over 1 year ago

TI__Mastermind 27710 points

Hi,

I can look into this issue further tomorrow, do you have a schematic view of I2C bus connection? Additionally, an oscilloscope view of the power up sequence would help as well.

0 sinngo kitahara over 1 year ago in reply to Daveon Douglas

Prodigy 10 points

Thank you for your reply. The CPU and TSC2003 are directly connected, and no other devices are connected to the I2C bus line. Does the startup sequence refer to the relationship between the TSC2003 power supply and the SCL and SDA pull-up waveforms?

0 Daveon Douglas over 1 year ago in reply to sinngo kitahara

TI__Mastermind 27710 points

Hi Sinngo,

Yes the power-on sequence and I2C bus are related for proper read/writes. If the ramp doesn't occur correctly the device may hold the SDA pin low as shown in your image.

Since you're able to replicate the issue; When i2c bus error occurs are you able to conduct any writes or is a hardware reset (power on/off) required? Also, can you share your schematic

0 sinngo kitahara 11 months ago in reply to Daveon Douglas

Prodigy 10 points

Hi,

I've attached an image of the power-on sequence. Blue is the TSC2003 power supply, yellow is SCL, and green is SDA.

Units where this phenomenon occurs will reoccur even after a reset. I've attached a circuit diagram.

0 Daveon Douglas 11 months ago in reply to sinngo kitahara

TI__Mastermind 27710 points

Hi Singo,

Have you tried making SDA and SCL lines high and stable before the TSC supply (+VDD) ramps up greater than 0.9V. Changing PENIRQ pull up value to lower resistor can make the touch detection more intentional.

Also in schematic, where is X/Y driver circuit?

0 sinngo kitahara 11 months ago in reply to Daveon Douglas

Prodigy 10 points

Hi,

This time, I am not using the touch detection function. Therefore, I have not implemented the XY driver circuit. Will the SDA line be unlocked by changing the resistance value of PENIRQ? With the current circuit, the SDA and SCL lines cannot be started up faster than the power supply. I would like to deal with this in software, but is there any other way?

0 Daveon Douglas 11 months ago in reply to sinngo kitahara

TI__Mastermind 27710 points

Hi,

I can't speak to a software solution and we no longer have a Linux support expert for TSC2xxx devices. I can provide troubleshooting from a hardware perspective.

But to answer your questions, no the SDA line and PENIRQ output are not related.

0 sinngo kitahara 11 months ago in reply to Daveon Douglas

Prodigy 10 points

Hi,

I understand. Are there any other solutions besides adjusting the voltage startup time of the SDA and SCL lines and changing the resistance value of PENIRQ?

0 Daveon Douglas 11 months ago in reply to sinngo kitahara

TI__Mastermind 27710 points

Sinngo,

In general, a few ways to improve sensor accuracy through hardware lie in reducing noise, improving REF voltage stability, and proper calibration.

I can update thread soon with more detail

0 sinngo kitahara 11 months ago in reply to Daveon Douglas

Prodigy 10 points

Hi,

I understand the contents. I cannot change the hardware, so I would like to change the IC.

Are there any TI temperature and voltage sensors that are pin-compatible with the TSC2003?

0 Daveon Douglas 11 months ago in reply to sinngo kitahara

TI__Mastermind 27710 points

Hi,

Looking into application notes,

0 Mir Jeffres 11 months ago in reply to Daveon Douglas

TI__Genius 14185 points

Hi Sinngo,

Have you tried changing your power supply sequencing? In this app note about the power up sequence (https://www.ti.com/lit/an/sbaa186/sbaa186.pdf), it specifies that "the TSC2003 SDA and SCL pins must be high before the device power-supply +VDD pin ramps up to 0.9 V", as shown in this diagram:

From your earlier post of your power up sequence, you powered on the I2C bus and your VDD at the same time. See if you can delay the power supply of VDD to ensure that the bus is high before power on.

Another note, I noticed that your SCL and SDA lines had a slope on the edges in your first post, I wonder can you reduce your I2C pullup resistors? Normally we suggest 4.7k for those, and I see you are using 10k, you could also go even lower than 4.7k if the clock edges are still sloped. I wonder if there are issues reading the I2C because the clock is not clean enough, there is a 1000ns slope requirement for standard I2C communication and if you have multiple devices on the same bus, the bus capacitance is higher and you need a lower pullup resistor to allow for the same steep slope on the clocks. You can read more about I2C resistor values here: https://www.ti.com/lit/an/slva689/slva689.pdf but see if you can start with 4.7k, and you could go down to around 1k if that does not help the slope. Let me know if this still doesn't fix your issue.

Best,
Mir Jeffres

0 sinngo kitahara 10 months ago in reply to Mir Jeffres

Prodigy 10 points

Sorry for the late reply. And thank you for your reply. Since it was difficult to delay the power supply of VDD from a hardware perspective, I only changed the I2C pull-up resistor.
As a result, I tried changing the pull-up resistor of the I2C line from 4.7 kΩ to 1 kΩ, but the symptoms did not improve.

0 Daveon Douglas 10 months ago in reply to sinngo kitahara

TI__Mastermind 27710 points

Hi,

Please adjust your power-on sequence to follow datasheet as shown above

Data converters

Data converters forum

TSC2003: Regarding TSC2003 sensor reading failure