2x16 LCD junk character issues

[panneerrajan]

Hi,

We are using 2x16 Character LCD as 4 data control (GPIO) option and its working fine.
LCD VCC is 5V and control 4 data line signal is 3.3V logic from MCU.
LCD placed at handheld device, through GPIO from MCU. we have used 0.5 meter ribbon cable from MCU unit and its working fine.

But while powering ON nearby load, My LCD display is showing junk character / garbage value. after re-initialization in Software only we are able to get proper value.
we already provided 0.1uF & 100uF capacitor at LCD VCC point.
please provide your suggestion.

Thanks,
Rajan

 

[nikhilmahasvar]

Check the routing of the cable whether it is routed near to any transformer, inductor, high voltage wire etc. Check whether your VCC is getting noisy with load.

 

[KlausST]

Hi,

0.5m ribbon cable is a good antenna....it will receive a lot of HF noise around.

This is why nowadays every device needs to comply with EMI (not to send out heavy HF signals) and EMC ( not to be too sensitive to received HF signals. This is your case).

In your case it's just showing different characters, but what may happen with an elevator control, or an airbag control?
*****

My - simple - solution is to use a shielded ribbon cable.
My - better - solution is use differential signalling. Usually one don't want to do this with a lot if signal, therefore one tends to use a serial interface. Maybe UART combined with RS485/RS422....and a microcontroller to (re) generate the LCD signals.

For sure you can do a lot of trial and error.
* common mode choke, HF filters at the signal ends, adding a resistor at the signal end (some kind of termination to make the signal line low impedance fir high frequency signal)

Klaus

 

[panneerrajan]

Dear Klaus,
For your information,
For testing purpose I connected LCD directly to the CPU FRC connector point (bypassed FRC cable). but same issue is occurring after some time.
As you said we will try with series resistor.

Thank you for your support.
Rajan

Hi,

0.5m ribbon cable is a good antenna....it will receive a lot of HF noise around.

This is why nowadays every device needs to comply with EMI (not to send out heavy HF signals) and EMC ( not to be too sensitive to received HF signals. This is your case).

In your case it's just showing different characters, but what may happen with an elevator control, or an airbag control?
*****

My - simple - solution is to use a shielded ribbon cable.
My - better - solution is use differential signalling. Usually one don't want to do this with a lot if signal, therefore one tends to use a serial interface. Maybe UART combined with RS485/RS422....and a microcontroller to (re) generate the LCD signals.

For sure you can do a lot of trial and error.
* common mode choke, HF filters at the signal ends, adding a resistor at the signal end (some kind of termination to make the signal line low impedance fir high frequency signal)

Klaus

 

[andre_luis]

LCD VCC is 5V and control 4 data line signal is 3.3V logic from MCU.

Did you even check on each datasheet if the electrical voltages are compatible within an acceptable range? It would be expected that you verified the maximum/minimum of both the input and the output of the MCU/LCD, which are working on different TTL ranges; 3,3v and 5v.

 

[panneerrajan]

Did you even check on each datasheet if the electrical voltages are compatible within an acceptable range? It would be expected that you verified the maximum/minimum of both the input and the output of the MCU/LCD, which are working on different TTL ranges; 3,3v and 5v.

Hi,
Data signal we can drive from 3 to 5V level and its working fine with 3.3V data signal.
while power on nearby load, that time only LCD display character goes to junk character.

Thanks
Rajan

 

[FvM]

LCD with parallel data bus is rather susceptible to ESD and electrical noise. Besides having a long cable there's often a problem with ground loops or capacitive coupling through module mounting or the metal display frame touching chassis ground. Isolated module mount and common mode chokes (ferrite core) for the cable may reduce the noise sensitivity.

Ultimately, you might think about periodically reading back the display content and reinitializing and reloading the display in case of inconsistency. Or use a less noise sensitive LCD module with SPI interface.

 

[panneerrajan]

LCD with parallel data bus is rather susceptible to ESD and electrical noise. Besides having a long cable there's often a problem with ground loops or capacitive coupling through module mounting or the metal display frame touching chassis ground. Isolated module mount and common mode chokes (ferrite core) for the cable may reduce the noise sensitivity.

Ultimately, you might think about periodically reading back the display content and reinitializing and reloading the display in case of inconsistency. Or use a less noise sensitive LCD module with SPI interface.

Hi FvM,

Thank you for your support,
Please refer above comment No:4. we will check with choke or series resistors and let you know status.
For differential signal, SPI interface need to change my design and PCB.
I try to resolve this issue with exist design (4 data GPIO interface logic)

Thanks,
Rajan

 

[KlausST]

Hi,

I said to use a resistor as some kind of termination..to make the node lower HF impedance.

But a series resistor causes higher impedance on a node. It may reduce ringing, but it won't help you to suppress induced noise.

--> according the shown characters you can find out if there is
* wrong_high on the data line --> add a pulldown resistor
* wrong_low on the data line --> add a pullup resistor
* added characters caused by noise at the control lines --> focus the bebuggin on the control lines (mainly "E")

Btw: What is the signal line state and the data line state when there is no communication with the display?
High? Low? Z? Activity?

Klaus