Dear Friends:
I built a points welding controller for automatic welding machine.
everything was good except lcd display,if any inductive load runs or enabled it shows garbage characters and unreadable Chinese characters.
i used 20 Cm ribbon cable to connect it in 4-Bit mode.
main controller is Atmega8, 8Mhz, good power supply and insulation Technic and good layout applied in design.
any idea to solve this problem?
Running LCD parallel interface through ribbon cable means asking for trouble. Connecting LCD ground or display frame to chassis metal must be absolutely avoided, it might help to place a small capacitor between E pins and ground at the display. You can also verify command and data register writes by read-back and reinitialize the display in case of discrepancies.
Running LCD parallel interface through ribbon cable means asking for trouble. Connecting LCD ground or display frame to chassis metal must be absolutely avoided, it might help to place a small capacitor between E pins and ground at the display. You can also verify command and data register writes by read-back and reinitialize the display in case of discrepancies.
i connected lcd body with metal sheet through screws to fix it, as you notified, it maybe a major reason of my problem,i will try to isolate it.
when i fixed lcd directly to controller board there was no problem at all, but when i fixed it into box problems start .
another point that i had three controllers and 3 lcds in one box, every controller is isolated and works standalone(power, lcd,...)
The metal frame holding the LCD glass is often connected to LCD ground, even if it's not it has a large coupling capacitance and very low creepage and clearance to LCD circuit. The latter is even worse because inductive overvoltage might cause a flashover between frame and other circuit nodes than ground, probably damaging something.
Even with isolated ground, the ribbon cable connected LCD is still prone to pick up interferences, caused e.g. by arcing contactors or welding arc itself. From a professional electronics viewpoint (particularly referring to ESD withstand standards), the LCD frame should be covered by a window so that electrostatic discharge to any point of the front panel doesn't cause circuit malfunction.
An GPIO interface to LCD is however slow, slowing down the "E" input by a filter capacitor can reduce the interference susceptibility. Using a LCD with serial interface could further improve the noise margin.
The metal frame holding the LCD glass is often connected to LCD ground, even if it's not it has a large coupling capacitance and very low creepage and clearance to LCD circuit. The latter is even worse because inductive overvoltage might cause a flashover between frame and other circuit nodes than ground, probably damaging something.
Dear FvM:
Thanks for your valuable support.
i did some tests on LCD and found this points:
1-Holding points are float, not connected with VSS.
2-There was two capacitors C1,C3 between VDD,VSS missed (i think those for bypass)
3-LCD metal frame also float, not connected with any pin.
what's your opinion about that?
do you suggest to put missed capacitors, what is a suitable value for?
Dear Friends:
I solved a problem as follows:
1- Grounding box(because a high voltage respect to human body was measured on box body)
2-initialize LCD before sending data(to clean it if any funny characters appear)
i know it's not a perfect solution, but in my next designs i will apply a full control on LCD like 8Bit data mode, hardware reset, R/W control and etc to avoid problems like this.