Very Bad EMI

[Mr.Blank]

Somebody please help me. I have a problem with my circuit. My microcontroller resets or hang whenever an inductor load switch on or off. My microcontroller is just a few millimeters apart away from the inductive load. My suspect is that my microcontroller is susceptible from emi. Somebody help me eliminate this EMI.

 

[kellys_eye]

What arrangements have you got on the supply and data lines to/from the microprocessor (i.e. decoupling etc)?

 

[Harald Kapp]

Is there a flybck diode across the inductor to limit overvolatge at turn-off?
Show us a schematic diagram.

 

[Mr.Blank]

Here is my schematic, Ive already put some decoupling capacitors on the vcc of the microcontroller.. also the flyback diode parallel to the relay..

 

[Mr.Blank]

Ive use batteries as a supply to the microcontroller to isolate its supply voltage but the seven segments display was a garbage... i think its the radiated emi that affect my microcontroller but i dont know what to do to suppress it..

 

[Harald Kapp]

If it's radiated EMI, it should go away when you place the relays farther away from the µC. This is a first check to isolate the source of the problem.

Ive already put some decoupling capacitors on the vcc of the microcontroller

Are these really close, close as in right next to, the µCs power supply pins?
You may also want to add an inductor in series with R21 to improve filtering of any noise on the high voltage supply V+. Additionally it is a good (and proven) idea to filter also the connection from V+ to the relays to reduce noise on V+ when the relays are operated. In that case a buffer capacitor (electrolytic) of e.g. 220 µF or 470 µF should be placed next to the relays to supply the power surge when turning on the relays.

also the flyback diode parallel to the relay..

Unfortunately the flyback diode D2 is not directly parallel to the relays but includes resistor R10. Therefore the flyback current develops a voltage drop across R10 which may contribute to EMI. Place the diode directly across the relay coils, preferably one per coil to minimize the inductive loop area.

 

[BobK]

How are the ground and power signals layed out? If the relay and the micro share a common track, that could be the problem. They should have completely separate tracks to to the incoming power on the PCB. If this is not the case, try cutting the tracks to the relay and replacing them with separate wires.

Bob

 

[dorke]

Let me add:
1. Your μC U1 is missing de-coupling caps with shortest legs possible,this may be critical.
2. Try a smaller pull-up on U1 Pin-4(functions as MCLR) say 1k with shortest legs possible.
3. show a photo of your actual build of the circuit.
4. A DSO may help you find the source of the problem(if you have one).
You should monitor pins 1,8 and 4 on U1.


General Notes:

1. You have an arrangement of 2 relays in series,this isn't optimal.
Why do you do that?,what voltage and current are your relays ?What are does relays switching?

2. The VCC=5v isn't generated in an effective way-5W resistor used.

3. Why are you using 2 Trs? Q1,Q2,one should be enough.

 

[Mr.Blank]

If it's radiated EMI, it should go away when you place the relays farther away from the µC. This is a first check to isolate the source of the problem.


Are these really close, close as in right next to, the µCs power supply pins?
You may also want to add an inductor in series with R21 to improve filtering of any noise on the high voltage supply V+. Additionally it is a good (and proven) idea to filter also the connection from V+ to the relays to reduce noise on V+ when the relays are operated. In that case a buffer capacitor (electrolytic) of e.g. 220 µF or 470 µF should be placed next to the relays to supply the power surge when turning on the relays.

Unfortunately the flyback diode D2 is not directly parallel to the relays but includes resistor R10. Therefore the flyback current develops a voltage drop across R10 which may contribute to EMI. Place the diode directly across the relay coils, preferably one per coil to minimize the inductive loop area.

as you may see Sir there is a capacitor across V+ parallel to the relay(maybe the orientation is not correct) and also i put zener for break down voltage, .now i am putting flyback diodes across each relay...

please be patient on me sir

 

[Mr.Blank]

How are the ground and power signals layed out? If the relay and the micro share a common track, that could be the problem. They should have completely separate tracks to to the incoming power on the PCB. If this is not the case, try cutting the tracks to the relay and replacing them with separate wires.

Bob

Sir thats one of my problem but my device is fix already so as long as possible i dont want to redesign again the circuit but im open on additional components to be placed.

 

[Mr.Blank]

Let me add:
1. Your μC U1 is missing de-coupling caps with shortest legs possible,this may be critical.
2. Try a smaller pull-up on U1 Pin-4(functions as MCLR) say 1k with shortest legs possible.
3. show a photo of your actual build of the circuit.
4. A DSO may help you find the source of the problem(if you have one).
You should monitor pins 1,8 and 4 on U1.


General Notes:

1. You have an arrangement of 2 relays in series,this isn't optimal.
Why do you do that?,what voltage and current are your relays ?What are does relays switching?

2. The VCC=5v isn't generated in an effective way-5W resistor used.

3. Why are you using 2 Trs? Q1,Q2,one should be enough.

Sir dorke ..ive try youre suggestions but the problem is still the same ..
.. ive measure the voltage across the microcontroller vcc and i have found that there is a noise that can be read clearly when the oscilloscope is set @25ns per time division every time the load across is switch on/off..

.. 2 relays are used to handle high power.. ive also measure 5.11 volts across microcontroller vcc, 2 transistors are used to slightly isolate microcontroller pin from relay pin..

..

 

[Harald Kapp]

as you may see Sir there is a capacitor across V+ parallel to the relay

You mean C4? but thsi capacitor is also directly connected to th Vcc generation. Any noise on C4 can easily couple into Vcc.
AS I mentioned: use another capacitor right next to the relays, use an inductor in series to decouple V+ from the relay voltage.

 

[Mr.Blank]

You mean C4? but thsi capacitor is also directly connected to th Vcc generation. Any noise on C4 can easily couple into Vcc.
AS I mentioned: use another capacitor right next to the relays, use an inductor in series to decouple V+ from the relay voltage.

.. thank you so much sir Harald..i'll gonna place a cap...

 

[Harald Kapp]

... and an inductor

 

[Mr.Blank]

Finally Sir i've solve the problem, Ive focus that much on filtering than preventing noise to get in to the supply. Ive suppress the noise coming from the inductive load by putting ferrite bead across the load line. Ferrite bead prevents very high frequencies noise or EMI to get in into the circuit..

thank you so much Sir for helping me and giving me ideas.. Godbless..
Maraming salamat.

 

[Mr.Blank]

... and an inductor

thank you so much sir..

 

[Harald Kapp]

Glad it works now.