High frequency noise when turning power supply on

[qpit3a]

I constructed a +/- 12 v power supply with 7812 and 7912 regulators.  I believe that the circuit is standard (the board is from Electronics Australia - it is old).  However I have recently purchased a DSO and thought to look at transients not really expecting to find them, but there are seemingly huge transients generated when the 240v dp switch is turned on.  I was testing with a load of 20 ohms but the attached images are for a load of 10k looks pretty much the same.  Is what I'm seeing normal and something I haven't seen before due to lack of the test equipment or is there something I should be doing to filter out this stuff?  I replaced the electrolytic capacitors on the power supply board and included 0.1 ceramics, but I suspect at this frequency the noise might just couple into all the wiring.  I wonder if I should be filtering the mains after the switch.

[mcinque]

You could be watching at the switch arcing inside and emitting radio interference, captured by the ground cable of the oscilloscope probe.
Did you try to measure it without the long ground cable and using the spring provided with the dso probe?

[Jwillis]

Switches "bounce".By adding a high voltage suppression capacitor in parallel that may filter that out . Look up switch bouncing and how to deal with it.

[qpit3a]

Thanks, I got a X grade capacitor and also ran the 240v wires one turn on a small toroid parts reclaimed from an old server power supply..  Seems to have mostly fixed the issue but then my repeated power cycling has caused one of the regulators to start oscillating never seen before.  However I can now hear the switch arciing when I turn it on, it wasn't a cheap switch but is it worth finding a better quality one?

[mcinque]

All switches arcs inside.
if it oscillates, you may have unwittingly created a BIG inductance+capacitance oscillator instead a suppressing filter.
You could observe what happens with the DSO but be careful: oscilloscopes and 220V mains are dangerous (in the worst scenario you could fry your DSO or even electrocute yourself). I suggest to use an isolation transformer (or at least an autotransformer 220V/12V) between the mains and the probes to float the measurement. "Use one hand to probe/measure and sit on the other hand".

[qpit3a]

I bought a proper mains filter to try, and I never use the DSO on 240v, although I did wonder if the spikes might be risky for the DSO even on the low voltage side of the circuit. Surely not.

[mcinque]

Which filter did you purchase? Do you have its specs and schematics (and values)?

AFAIK, theoretically at the hot side 220V mains and a big inductance with no clamping you could potentially have a spike while powering the transformer that can go over the limits of the dso (~300Vrms), so if the probe is inadvertently at 1X you could easily fry the input channel. Even with the probe at 10X if the voltage spike arcs inside the voltage divider of the probe it could reach the DSO input.

At the cold side you should be safe, but I learnt at my expenses that when dealing with 220V mains powered circuits, it's better to consider them always as live   

[qpit3a]

Ti datasheet says 79xx needs output capacitor > 25uf to prevent oscillations.. i took them of  as an experiment... No component values for the filter . Just going to do a rebuild and move low voltage away from high and put more toroids on wiring . And experiment. I thought power supplies were simple.