I was testing a 2-channels RS232 adapter for my FPGA board, when I saw this "miracle":
If any signal from the RS232 is at a positive voltage and the board is not powered, there is a leakage current to VCC that turns on the led (power indicator).
More, if a VCC = 3.3V is applied the board doesn't work for speed greater than about 300 bps.
Applying a VCC = 5V the board works as expected.
I've used a couple of SP3232E (marked SIPEX that should work @ 3.3V and 5V according to the spec, ah ah...) that I bought from Aliexpress more that one year ago and not used until now:
I really thought that they were fake RS232 IC... so I removed both the ICs and cleaned the pads...
Than I've ordered a new lot of SP3232E from TME (
www.tme.eu), and resoldered them
...and now the board works at 3.3V and 5V @ 115200 baud and without any "strange" current leakage when not powered...
Here the good SP3232E from TME (in the bottom) compared with the fake one from Aliexpress (upper in the photo and not soldered):
They look very similar but some differences are anyway visible...
Analyzing the behavior of the fake one, I thing that they are fake MAX232. They seem working at 5V, but the leakage current from the RS232 when not powered suggests not to use them at all...