Touchscreen controller hangs during ESD testing

Hi Matthias,

Kernel 3.8 is not supported by TI. Have you tried with one of the TI EZSDK releases? On the hardware side - what TVS have you tried? Have you tried to shield the TS wires?

Hi Biser,

I don't this issue is software related. On the other side the TI 3.12 kernel does not support 5wire touch out-of-the-box. But that's another issue. But migration to 3.12 is work in progress.

Our TVS are WE-TVS, part# 824013, würth elektronik.

http://katalog.we-online.de/pbs/datasheet/824013.pdf

We tried to shield the TS wires using some copper foil - as far as this was possible. It does not seem to improve the issue.

Matthias

Hi Matthias,

How is the TVS connected? Do you have it referenced to the 1.8V ADC power supply?

Hi Biser,

originally we connected the part against 3.3V. But during diging behind this issue we also tried a connection to 1.8V. This does NOT change anything in the weird behavior. We also tried some caps (1nF) on the TSC lines. No improvement.

It sounds like the radiated emissions from the ESD test is coupling into the person touching the screen and this energy is coupled into the touch screen which injects unexpected current pulses into the AM335x device. The current pulses are apparently strong enough to disrupt normal operation of TSC in such a way that it locks-up the internal logic circuits. Unfortunately, the TSC does not have a soft reset that allows this module to be reset by software. The only way to reset this module is by applying a system reset.

I'm not aware of any other reports of AM335x TSC locking-up as you described so there may be some system level dependency that needs to be understood and resolved. For example, I saw something similar on another TI device once and we found a metal bracket on the board floating that was building up charge and coupling into the device inputs. We were able to solve the issue by grounding the metal bracket.

It sounds like you have explored several filter options without any success. I think this is the correct approach.  Does this happen only when touching the screen near the corners where the resistance to the AM335x AIN terminals is lower than it would be in the center of the screen. 

Is this ESD test a industry standard test or a custom test to validate your system? If possible, can you describe the ESD test or point me to an industry standard.

The BBB and capes were designed to be low cost evaluation products with minimum noise immunity precautions. So it is not be a big surprise to hear these boards are having issues with this type of test.

Are you having the same problem with the smaller display?

Regards,
Paul 

Hi Paul,

I am with your explanations.

The ESD test is acc. to DIN EN 61000-4-2. We are doing contact discharge on a vertical coupling plate that is placed near the device. So no direct discharge onto any point of the device itself. We did the test in our laboratory and also at an accredited external laboratory.

Of course the BBB is only an evaluation product. But there isn't any ESD protection on any TI evalboard with CPU at all and we didn't even find a hint that suggests any procedures towards ESD protection.  Our others similiar designs do not show such behavior.

We also did the test with a 7" 4wire touchscreen. Same results!

We are currently evaluating an external I2C touchscreen controller just to see if it solves our problems.
But the additional cost is not acceptable.

Regards,

Matthias

Hi,

we temporarily connected a ar1021 i2c touchcontroller (microchip) onto our board and did the ESD test again. We did not run into any issues with this configuration. So I am sure it's an AM335x issue with very very sensitive AIN pins and no proper fix for this issue so far.

Matthias

Hi,

I stumbled on this post researching an issue we are seeing with our custom am335x design where the ADC is failing and we were suspecting ESD as the possible culprit.  If this is in fact the case even performing a power down / power up of the ADC via REG_CTRL 0x40 will not reset the ADC?