I suspect that you may have bigger capacitors in the 5V side than in the
22V side, thus when you turn off the latter, your voltage regulator sees
0V in its input and a discharging 5V in the output, destroying it.
I'm not sure if this is your problem, but it indeed was in one of my PSU
designs. I fried a lot of voltage regulators until I simply added a 1N4001
diode from the voltage regulators' output to its input. This will prevent
any "back voltage" passing through the chip.

I hope it helps.. but if it doesn't, believe me it's always better to add
this diode when one builds a PSU.. expecially when one likes to put big
capacitors at the output of the voltage regulators.

However, in your specific case a 10K resistor and a 5.1V zener is all you
need.. remember you're talking about logic signals here (i.e. "data"), not
power.


At 16.00 27/02/2004 -0800, you wrote:
>Hi all,
>
>I have a project where a vacuum pump outputs a 22V signal indicating
>whether it is working or not.  If it malfunctions (0V) I need to stop the
>process.  The signal goes into a datalogger which is controlling everything
>(I wasn't allowed to use a PIC to set this up).  The datalogger needs a 5V
>signal, but I would like the pump to be electrically isolated because it's
>very expensive and a huge pain to fix/replace.  The way I decided to do
>this (one-time project) was use a 24V-5V DC/DC converter (DCP022405P).  I
>have a 2.2k resistor that goes from the converter's output to the digital
>input (it only tolerates very low current) and also a 500 ohm resistor that
>goes from the output to a LED for monitoring the status.  The inputs of the
>DC/DC converter come straight from the vacuum pump (the pump signal is
>current limited to about 50 mA).
>
>This worked great for about a month until someone unplugged the connections
>to move the setup around.  Then it wouldn't work -- somehow the chip just
>died.  I popped in another chip and it worked fine until the power went out
>last night.  When the power came back on the chip was fried again, no
>unplugging of anything.  I figured maybe some ESD the first time, but now
>I'm not sure, maybe the pump output goes screwy at times (it really
>shouldn't though).  Is there something I should be putting in between the
>pump output and the DC/DC converter input to protect it?  Some kind of
>resistor or RC combo?
>
>Thanks,
>Mark
>
>
>PS I thought about using an optoisolator and a 5V VCC, but the main power
>for everything is 12V.  So I would need a voltage regulator or zener to
>goto 5V for this IC.  Plus I had never used a DC/DC converter so I figured
>this was a good opportunity to play with one.
>
>--
>http://www.piclist.com hint: The list server can filter out subtopics
>(like ads or off topics) for you. See http://www.piclist.com/#topics

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.