How to beat this noise?

[onemilimeter]

I built a simple Microchip PIC18F4331 microcontroller based circuit board. The input voltage of the circuit board is +5VDC, which is supplied by a switching-mode AC/DC adapter.
The AC/DC adapter shares the same 4-way socket extension with the Weller WTCP 50 as shown in attached image.

[1] When the Weller Solder Station (WTCP 50) is switched (Switch-W50) from OFF-To-ON,  the microcontroller circuit board will "die", or in other words, it will be reset into UNKNOWN state.
[2] When the +5VDC input voltage is monitored using a scope, "disturbances" are observed when the Switch-W50 is switched from OFF-To-ON.
[3] I believe that these "disturbances" have caused the microcontroller circuit board to be reset into UNKNOWN state.
[4] In fact, I don't understand why it will be reset into UNKNOWN state. I thought the microcontroller, if it's reset, it will start executing the code from reset vector again.

Kindly suggest how to beat the "disturbance" appeared on the +5VDC input voltage.

[Alex]

It is weird that this is happening, the regulator inside the AC/DC adapter should be removing this conducted noise.

Seeing the waveform that is one huge damped oscillation! Max Pk-Pk about 6 Volts! What is channel 1 connected to?

Anyway, you can either increase your supply voltage an provide local regulation eg 78L05, or you can add a filter to the power supply of the MCU depending on what you have connected to the MCU. An RC filter might be suitable all that is needed.

In fact, just add a 1uF capacitor in parallel with a 100nF capacitor as close as possible to the MCU supply pins. If it still wont work, add a 10uF cap too. All in parallel ofc.

[Psi]

First thing i'd do is try a different switchmode 5V powerpack. Are you sure the one you currently have is switchmode? Maybe it's a fake. It's quite odd to get that kind of effect through a switchmode powersupply.

Another thing to look at it the AC mains filter capacitors in your gear.

Mains powered devices quite often have a mains capacitor inside (often marked as 250VAC at ~ 0.1uF - 0.4uF) these mains caps help to smooth the line and let the device draw really high current for a very short time without pulling the mains voltage down and creating ringing/spikes.

Quite often these caps die and, since they are only across the mains, the device quite often still works and you dont notice. With a dead filter capacitor the device can sometimes produce this type of noise effect on your line when it's switched on.

If it was me i'd open up the soldering iron and check the design has a filter capacitor and that it's not dead. Your 5V DC powerback might also be missing one of these capacitors and adding one could help (or it could have a dead one). Of course this requires working on mains circuitry so care must be taken.

You really want to design your circuit so it is able to run fine with other devices that may have bad AC filter caps. So fixing your soldering iron (if thats the problem) doesnt really fix your circuit even though it may stop the issue.

[onemilimeter]

It is weird that this is happening, the regulator inside the AC/DC adapter should be removing this conducted noise.

Seeing the waveform that is one huge damped oscillation! Max Pk-Pk about 6 Volts! What is channel 1 connected to?

Anyway, you can either increase your supply voltage an provide local regulation eg 78L05, or you can add a filter to the power supply of the MCU depending on what you have connected to the MCU. An RC filter might be suitable all that is needed.

In fact, just add a 1uF capacitor in parallel with a 100nF capacitor as close as possible to the MCU supply pins. If it still wont work, add a 10uF cap too. All in parallel ofc.

Channel-1 is connected to the one of the MCU's PWM output pin.

At the moment, a 0.1uF capacitor is connected parallel to the MCU's supply pins. I will try adding more capcitors and hope it will work.

[onemilimeter]

First thing i'd do is try a different switchmode 5V powerpack. Are you sure the one you currently have is switchmode? Maybe it's a fake. It's quite odd to get that kind of effect through a switchmode powersupply.

Another thing to look at it the AC mains filter capacitors in your gear.

Mains powered devices quite often have a mains capacitor inside (often marked as 250VAC at ~ 0.1uF - 0.4uF) these mains caps help to smooth the line and let the device draw really high current for a very short time without pulling the mains voltage down and creating ringing/spikes.

Quite often these caps die and, since they are only across the mains, the device quite often still works and you dont notice. With a dead filter capacitor the device can sometimes produce this type of noise effect on your line when it's switched on.

If it was me i'd open up the soldering iron and check the design has a filter capacitor and that it's not dead. Your 5V DC powerback might also be missing one of these capacitors and adding one could help (or it could have a dead one). Of course this requires working on mains circuitry so care must be taken.

You really want to design your circuit so it is able to run fine with other devices that may have bad AC filter caps. So fixing your soldering iron (if thats the problem) doesnt really fix your circuit even though it may stop the issue.

I believe the AC/DC 5VDC adapter is switching-mode type because it's quite light.
I will check if there is any dead AC mains filter capacitor in the soldering station.

[onemilimeter]

I'm thinking of buying an AC mains line filter.

Kindly advise how to select the right line filter (e.g. to beat the noise as captured by the scope shown in my first post).

[onemilimeter]

If two identical AC line filters are connected in series, is the effective attenuation doubled?

[Alex]

The switchmode supply will most likely have a small filter in it. You really dont need to invest more money. Try a small inductor in series with the PSU and higher capacitance in parallel.

[Zero999]

My guess is the noise is being coupled  from the AC to DC side via the Y2 capacitors inside the regulator. You could try connecting the 5V and 0V rails to protective earth via 100nF capacitors, although this is a pain as the SMPS will be double insulated.

Failing that, try ferrite beads and filter capacitors as mentioned above.

[onemilimeter]

If the common (or the return of +5V) of the MCU board is monitored, similar noise is observed too.

Which type of noise is it? Common-mode or differential-mode?

[scrat]

If the common (or the return of +5V) of the MCU board is monitored, similar noise is observed too.

Which type of noise is it? Common-mode or differential-mode?

I don't get how you did the measurements... In the figure, you put GND to the -, where is the reference (GND) connected is this last measure?

[onemilimeter]

If the common (or the return of +5V) of the MCU board is monitored, similar noise is observed too.

Which type of noise is it? Common-mode or differential-mode?

I don't get how you did the measurements... In the figure, you put GND to the -, where is the reference (GND) connected is this last measure?

My apology to make you confused. In my last measure, the GND probe is connected to the '-', and the signal probe is connected to the '+'.
To monitor the '-' signal, I place both the GND probe and signal probe to the '-'. Is the setup OK?

[scrat]

Measurement made with both probe ends at the same point should give no voltage, unless there is radiated noise. Try measuring with the probe shorted (a loop) an disconnected from any point of the circuit. If the noise is still present, you're measuring a radiated noise.

If there is no noise, there is still the possibility that you are modifying the circuit by introducing the probe. You can try using a so called "low inductance probe" that some scopes have (simply a spring-like GND contact) or at least twisting the GND cable on the probe, to reduce coupling (by reducing the magnetic loop area).

[onemilimeter]

The switchmode supply will most likely have a small filter in it. You really dont need to invest more money. Try a small inductor in series with the PSU and higher capacitance in parallel.

The inductor should be wound in differential or common mode?

Thanks.

[Alex]

None. These terms don't apply here.

No need to make your own anyway.

Something like this will be fine.

http://www.coilgun.eclipse.co.uk/images/prototypes_pages_images/cs_p01A/inductor_modification_1.jpg

http://www2.electronicproducts.com/images2/Z62PULSE0806.gif

Remember to add more capacitance to the MCU's pins too.

[onemilimeter]

None. These terms don't apply here.

No need to make your own anyway.

Something like this will be fine.

http://www.coilgun.eclipse.co.uk/images/prototypes_pages_images/cs_p01A/inductor_modification_1.jpg

http://www2.electronicproducts.com/images2/Z62PULSE0806.gif

Remember to add more capacitance to the MCU's pins too.

Should I opt for mH, uH, or nH inductor? Thanks.

[Alex]

a few hundrend uH, depends on how much capacitance you put. Just try until it works.