May I be so bold as to propose an alternative (Or minor modification) to Kai's 220V detection circuit? I would propose to eliminate the use of the transorb and the half watt resistor as a means of making the circuit smaller and lower cost to fabricate. Here is a modified copy of Kai's schematic:



As you can see from Kai's original version this one has the front end resistor boosted up to 22K ohms which permits its power rating to be significantly lowered. I also replaced the transorb with a pair of 1N4148 diodes which are much smaller and much lower cost devices.

I modeled the circuit in a small free Spice package using the schematic shown below. Due to the limited selection of component models available in the Space package I had to make some educated guesses at suitable alternative parts for the two transistors.



I found from the modeling exercise that the wattage rating of the front end 22K resistor can now be a much more PCB friendly 1/10 Watt device. The model also showed that for nominal component values as shown that circuit will detect the AC output of the transformer down to voltage of about 1V peak. The following diagram shows the waveforms predicted by the modeling program for the input node, the filter node, and the output waveform. Note that the output is at a low level in the ABSENSE of the AC input....so if this was considered an AC present detection circuit the output would be active HIGH for AC detect.



If anyone is interested in goofing around with the model circuit some more and expirimenting with the circuit component values here is the link to the text file for the model schematic. (Save this text file to your disk as an .ASC file).

To be able to use the free Spice package you have to download it from the Linear Technology Web site. The package is known as LTSpice or SWCad-III. Here is a link to investigate the free LTSpice download.

Michael Karas