Spikes in AC-DC power supply

[mr.newbie]

I have a trivial circuit with a couple of ICs and NPNs working out of a 12V power supply and through a 7805 to get 5V, with bypass caps properly distributed across the circuit. I'm a bit surprised by the huge spikes being caused by outside interference in the AC line. Switching a nearby light on/off may cause waveforms of 2~3V to be seen at the 7805's output.

I understand spikes are to be expected and handled with caps and inductors, but I'm surprised at the magnitude of the interference. Is that something usual that people handle with additional circuitry to protect against, or would a bad supply be a more reasonable explanation?

Thanks,

 

[Tahmid]

2-3V spikes are much larger than normal.

"Isolate" the input of the regulator from the input voltage with a diode and use a large bulk capacitor. Use a small electrolytic capacitor (10uF to 68uF) from VIN to ground and another from VOUT to ground. Use 0.1uF capacitors from VIN to ground, VOUT to ground and also before the diode to ground.

Hope this helps.
Tahmid.

 

[mr.newbie]

Thanks for the hints. I've already tried the capacitors, even a relatively beefy one (220uF) between VIN/GND on the regulator, without much success. I haven't tried to isolate the supply with a diode yet, though. I'll try that and see how it goes.

 

[Tahmid]

Another thing I'd tried once:

After the diode and before the VIN of the regulator, connect a resistor. Try with that as well. And post your results.

 

[mr.newbie]

Will do, thanks again.

 

[mr.newbie]

Okay, so I've run a few tests, with 10uF and 2200uF caps between VIN and GND, a a diode to isolate the voltage regulator from the supply, or a resistor in the same place, and took screenshots that triggered on the light switching. Lastly, I've also checked results with a ferrite bead for the sake of experimentation. Results follow.

With just the usual 10uF capacitor:



With the 10uF and a 1N4007 diode:



Just the 2200uF cap:



The 2200uF cap plus the 1N4007:



The 2200uF cap plus a 100ohm resistor:



Lastly, 2200uF plus ferrite bead:



Looks like the real change comes from attaching the large capacitor, and even that leaves behind large spikes.

 

[mr.newbie]

Out of curiosity, here is another shot showing V before and after the 7805, with 2200uF on VIN/GND and multiple bypassing caps on the power rails at the inside:



Again the triggering was done by switching a nearby light on.

 

[Tahmid]

Can you try with another 7805? Maybe there's something wrong with the one you're using. It seems quite unusual. How close is the device to the light?

 

[tpetar]

What is distance between capacitors and regulator on input and output ? Keep close to regulator. Keep regulator close to bridge rectifier.

Capacitors are new, or they are adult ?

Where you buy 7805 ? (maybe fake, or bad copy)

What you put on output, what capacitor value ? Make sure to put one 100nF on end.

Make some sharp and focused (not blured) photo of all, trafo, bridge rectifier, capacitors, regulator as they looks like now.

 

[mr.newbie]

Can you try with another 7805? Maybe there's something wrong with the one you're using. It seems quite unusual. How close is the device to the light?

I did try to replace each one of the ICs to see if it was misbehavior, but not really.

The light switch is very close to the power supply (plug and switch are on the same wall outlet).

What is distance between capacitors and regulator on input and output ? Keep close to regulator. Keep regulator close to bridge rectifier.

Capacitors are new, or they are adult ?

Where you buy 7805 ? (maybe fake, or bad copy)

What you put on output, what capacitor value ? Make sure to put one 100nF on end.

Make some sharp and focused (not blured) photo of all, trafo, bridge rectifier, capacitors, regulator as they looks like now.

Capacitors are surrounding the 7805 very closely, in addition to the usual bypass capacitors along the power rails.

Got 7805 from the best local eletronics shop, so probably not fake. I've tested several capacitors and setups before (see attached images above).

I'm starting to wonder if perhaps the light switching is indeed making a big enough dent on the AC current that a supply just can't keep up with it, so perhaps the house wiring is to blame.

I don't have that exact setup in place anymore for a photo as I'm now introducing a darlington in the setup to conclude the appliance I'm preparing, but I'll send a photo along once I get in a good place again. Thanks for the suggestions.