Spike in power line relay

[Freshman]

I have this circuit below.







I have an issue where, when the input turns on the transistors, I see a spike in the nearby power supply rail. I have other power supply rails such as 3.3V and 5V. Whenever, the transistor turns ON, I see a spike of say 15V peak to peak on the 3.3V rail. This causes my ICs connected to 3.3V malfunction.

However, the spike of 15V peak to peak is not happening always. The amplitude of the peak to peak spike varies by say +/-5V

Attached in the circuit image.

Can someone let me know why this happens and what fix I can try?

[Phil1977]

It´s strange that the spike happens when the relay is switched on. Usually the back-EMF when switching off relays is causing trouble, but here it must be something else.

I´d do two things:
- Put some decoupling capacitors on the power rails near the relay driver
- Slow down the switching on of the transistor. E.g. by a 330pf-cap from base to emitter
- Put a clamping zener or TVS on your 3.3V rail to protect the digital components.

Are you sure that the spike is coming from the relay itself and not from it´s connected load? What kind of power supply do you have, is it possible it suffers from really bad cross-regulation?

[ramussons]

It is doubtful if the spike occurs from the shown schematic.
Is the relay operating some other equipment?
Where / How  is the drive to the transistor generated? That also needs to be investigated.

[tggzzz]

I have this circuit below.

I have an issue where, when the input turns on the transistors, I see a spike in the nearby power supply rail. I have other power supply rails such as 3.3V and 5V. Whenever, the transistor turns ON, I see a spike of say 15V peak to peak on the 3.3V rail. This causes my ICs connected to 3.3V malfunction.

However, the spike of 15V peak to peak is not happening always. The amplitude of the peak to peak spike varies by say +/-5V

Attached in the circuit image.

Can someone let me know why this happens and what fix I can try?

This could be a real effect, a consequence of how you are measuring the spike, or because of "stray/parasitic" components not included in your schematic.

What is the relay switching, and is that load connected to the components shown in the schematic?

Show photo of implementation, how you are measuring the spike, and the measurement.

[BeBuLamar]

Usually there is a spike in of negative voltage when you switch off but the diode should clamp it.

[Zero999]

Presumably the diodes are connected to the transistor's collector and the coil?

Here's the correct way of drawing it.

Note the bottom diode doesn't do anything. It never conducts.

[djococaud]

The 1N400x is way too slow to absorb the spike from the relay coil. Try a schotty diode instead (like 10MQ100N or similar)