arduino making 7805 output noisy

Please see images for the setup that I use. The mega is powered from a 7812 and is generating a pwm wave at pin 7 (10us on time and 100us offtime) using delayMicroseconds and toggling the pin.

I took the measurements across the 7805 output with the arduino disconnected from 7812. It has a 15mv p-p noise (see image, its on 20mv scale). Then I took measurements with the mega connected to 7812 and generating the PWM. the 7805 output has become very noisy (even at 10 times the scale of 200mv)… How do i avoid/suppress this. The 7805 is supposed to power a dac and this noise is causing issues. Both the regulators have output caps at 0.1uf. The cap bank at the trafo output is 140kuf.

If you have noise you need more / better decoupling and a better layout for your circuit. Are the caps ceramic? Have you got any big ones? Have you got caps on all the devices?

How are you getting the ground to your scope probe, it might be an artifact.

Some 7805 data sheets show capacitors on the input and output of the chip to reduce voltage oscillation that can occur.

I have 0.1uf electrolytics at the output of both the regulator. There is a 10uf and 470uf (recently added to test) caps at the input of the regulators.

doors666: I have 0.1uf electrolytics at the output of both the regulator.

Rearly? They are useless for noise suppression they need to be ceramic. Also you need some large caps on the output as well. What about layout?

Hi, ALL spec sheets show capacitors on the input and output of the LM78 series regulators. You need 2 caps on each of the regulator outputs, a 0.1uF ceramic and 10uF or up to 47uF electrolytic capacitors. And they should be as close physically and electrically to the regulator IC as possible. Also with 18Vac measure how much your unregulated DC is, you may need heat sinks on the regulators due to the voltage drop across them. Especially the 7805.

Tom..... :)

ps, guys, what level of programming, electronic theory, electronic assembly do you have?

ps, guys, what level of programming, electronic theory, electronic assembly do you have?

Me? Only 45 years at it that's all.

doors666: Then I took measurements with the mega connected to 7812 and generating the PWM. the 7805 output has become very noisy (even at 10 times the scale of 200mv).. How do i avoid/suppress this.

How much current is being controlled by the PWM signal?

Do you have a groundplane and good high-frequency layout (ie minimum stray inductance)?

Are you aware that standard scope probes pick up induced voltages need swtiching/logic circuits that are spurious. A low-impedance probe properly connected is the only way to see anything like the true signal on a power rail.

You are seeing switching transients on the timescale of 10's to 100's of ns - stray inductance and capacitance dominate the picture in that regime.

Me? Only 45 years at it that's all.

No Grumpy_Mike.. Doors666 I was referring too.

Tom.... :)