Orin Eman wrote: > http://datasheets.maxim-ic.com/en/ds/MAX6816-MAX6818.pdf > > Note the following statement: > > "Driving an input to +25V will cause > approximately 0.32mA of current (up to 2.6mA for eight > inputs) to flow back into the VCC supply. If the total system > VCC supply current is less than the current flowing > back into the VCC supply, VCC will rise above normal > levels. In some low-current systems, a zener diode on > VCC may be required." > You will find this on all 15KV protected devices. The current has to go somewhere. One of my first tasks at my new job was to determine what was wrong with several failed IDEC PLCs. The serial port had quit working. Although IDEC would not supply me with a schematic, I still had to try to troubleshoot it. It turns out that the PLC is used in an environment with very noisy universal motors with carbon brushes. If these brushes are defective, the EMI radiation becomes horrendous, and the ground path can't be easily made low enough to discharge the energy. From time to time, the brushes become worn out. IDEC makes a peripheral, a combo touch screen / LCD color monitor. The serial port in the PLC simply drives this touch screen. But a new PLC always worked, so the failure was not taking place in the monitor. I took apart the PLC and determined that the SIPEX MAX232A-like device with internal 16KV protection had failed. I replaced it with an ST version of the same thing and all was well until the NEXT time the brushes wore out. It seems that the internal power traces were destroyed, as the DC-DC charge pumps had collapsed, too; all pins out of the device read 0.0v against GND. IDEC does not take special steps to crush EMI spikes imposed on the serial cables. I am designing a small PCB with heavy TVS on every pin, and series resistors large enough to reduce the spike size. The spike will be crushed into heat, rather than be absorbed thru the chip and system's VCC system. --Bob -- http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist