Voltage spikes in power supply

Hi there,

Following is the schematic for a power supply what I have put together for my amplifier build.

24v will power a TDA7498 Class-D 100Wx2 Stereo Digital Power Amplifier @ 4Amp max.
12v will run a preamplifire. 500mA max.
5v will run an Arduino Pro mini, 20x4 LCD and a real time clock module. I would say max of 500mA. Currently 7805 runs a bit warm (not hot).

Issue I'm facing is,
When I switch-on with "SW", I see +12v and +5v rails power spikes way above there intended values.
E.g.: +12v rail max peak around 20v and 5v rail max peak around 8v.

I had issues with many Arduino pro mini boards where it stopped working and some, I had to bun boot loader again to function.

So any thoughts to why I get high voltage spikes, even with liner regulators and please, if my design is not correct, I'm open for suggestions to fix and improve.


That's not a standard topology for a power supply, in particular the heavy pulse currents at switch on could lead to brief reverse-biasing of the L7812 and L7805 (you can't assume the secondary voltages are what you expect during these pulses).

I'd add a diode from centre-tap to these regulators.

And are the regulators genuine parts?

Can you confirm you are seeing these spikes with an oscilloscope. How long are spikes and how quickly do they decay?

I hope so. But not sourced from renounced vendors though.

This is on the +5v rail.
Please do let me know if my scope settings are wrongly set to capture these.
I'm still learning.

Switch on

Switch off

Are you using a 10meg, shielded probe with the ground connection right at the grouns of the circuit you are monitoring?

Also, with your 28 VAC, your circuit showing 24 volts should actually be 39.59 volts, if you measure it.

Should I implement a DC soft starter?

could you kindly explain a bit further here?

probe is on 10x mode. This is model that came with the uint.


Could be correct, if you followed the compensation procedure and have set the probe to 10meg Ohms.

Yes yes. 10x mode and compensation procedure done.

Sorry the correct measured value is 25.23VAC not 28VAC.
But still how come the voltage increases above VAC value when converted to DC?

The peak voltage of a sine wave is about 1.4 times the RMS voltage and the capacitors charge-up to the peak. Then there is about a 1.4V drop across the diode bridge.

1 Like

But back to your original scope pictures.
The second, switch off, shows you are picking up stray environmental noise. Do you have any wires connected to the 5 volt output? Looks like it.
The first picture, switch on, shows inductance limiting the charging of the capacitors. Likely because of what is connected to the 5 volt line.

There's only one orientation of a diode that makes sense for the regulators - it will protect them from negative spike (which can easily destroy a 780x regulator).

1 Like

Because AC is measured as rms, not peak.

1 Like

Looking at your PCB, you have two electrolytics on one regulator and two ceramics on the other.

You should have the electrolytics on the input of each reg and the ceramic (0.1uF) I hope on the output of the regs.

Tom.. :smiley: :+1: :coffee: :australia:

1 Like

From other thread, your 5V regulator is oscillating.
Have you fitted the caps correctly as in my previous post?

Tom... :smiley: :+1: :coffee: :australia:

Not yet Tom, will do and give my feedback soon, thanks.

ON the diagram there is a 100nF before and after both regulators but on the pcb the 2 electrolytic caps (according to the size probably around 100uF) is not in the diagram.

With the regulators been able to supply up to 1A should the output Caps not be closer to 1000uF?

Sorry this was mentioned in post 16 .. did not saw the note in the photo.