Arduino resets when solenoid shuts off

Question

I have an arduino hooked to a 2 relay board. Each relay is opto-isolated. The arduino and board each have their own 5V power feed. The relays are switching 12V to controll a motor and a solenoid valve.

Sometimes when the solenoid turns off, the arduino resets. From what I've read, it's likely an EMI problem caused by the solenoid. If I disconnect the solenoid, I don't see the problem.

I've seen various suggestions about using a varistor or a diode, but I don't know which is better or what values to try. Any recommendations?

I'm assuming that the varistor or diode should be as close to the solenoid as possible, shorting the power and ground lines, is this correct?

Thanks,

Chris

c4m4ch0 · Answer

Hello maranatha, you just need to connect the diode as it is in this image

Former Member · Answer

I put an IN4004 Rectifier Diode across the terminals of the solenoid and that seems to have solved the problem. I also found this page pretty educational: http://en.wikipedia.org/wiki/Flyback_diode

COMPACT · Answer

Relays can't be driven directly from RPi's GPIO pins. A GPIO pin needs be connected to a current amplifier to drive a relay. The current amplifier can be as simple as a NPN transistor such as a BC548 or equivalent. (It's a transistor acting as a switch for the relay). It's called a current amplifier because a tiny bit of current coming out of the GPIO pin causes a larger current to energise the relay coil). The transistor base is driven through a 1k resistor to limit the current out of the GPIO pin. The diode across the relay is to create a discharge path for the relay coil to minimise the back EMF. Note that the polarity of the diode is backwards to the energising current flow. The diode could be an 1n914. When a relay coil is energised a magnetic field is created around it. When the power is disconnected the magnetic field collapses to create a current. This collapse can create a large voltage between the ends of the coil (seen in old radios as a spark gap) which we don't want in this application so a diode is used to complete the circuit to keep the voltage low.

D_Hersey · Answer

Make sure your supplies have big, fast decoupling capacitors

mcb1 · Answer

The schematic for the relay board is here http://shop.emscdn.com/catalog/emskits/relayshield/kitv1/G5LA_E.pdf The 100uF is not adequate for powering via the USB ... BUT you made the statement that you have tried powering the Arduino from an external 7-12v supply. The Arduino will switch between USB and external power to whatever has the greater value. The external socket goes through a regulator so it needs 7v minimum and the regulator is 500mA on proper Arduino's. Some copies may not be supplying 500mA, and if the input voltage is not greater than the regulator dropout (2v) then weird stuff can happen. The relay specs are 72mA for the 5v version BUT this is after it has pulled in and is operating. It is very likely to draw 2-4 times than amount as it pulls in. ie 288mA An Arduino draws 46mA when running. So you could experiment by putting a large (more than 1000uF) cap across the Vcc and GND pins of the shield. This should help eliminate any voltage sag causing a reset. Also just to be sure can you attach the Arduino code (paste it, highlight/select it and use the >> Syntax highlight and C++) Eventually we'll sort out the issue. Mark

mcb1 · Answer

I asked because the ethernet could result in the RPi flaoting up. The input pins on an Arduino are very sensitive and would likely float high when the RPi was powered up. While this shouldn't matter as far as the program is concerned, it's most likely that it was causing an issue on the supply side, resulting in it crashing the controller and the program. The fact it didn't do it while the HDMI lead was plugged in supports this. Various supplies can be used for the different items, however connecting the grounds ensures they all work relative to one another. Your 1F cap is a bit over the top, and maybe a 10,000 uF aor even somethign in between 1000 and 10,000uF would ehlp ensure any relay switching is not going to cause issues. Personally I would have used a 2,200 or 4,700uF on that board. Mark

COMPACT · Answer

Room 101 - Tie unused input pins to a known stable state.

mcb1 · Answer

it's likely an EMI problem caused by the solenoid I think you might be confused with back EMF which is where the energy in an inductive load (relay, solenoid or coil) causes a spike on the power supply, whenever it is turned on or off. I suspect it is your 5v supply that is having the issue. You state there are seperate 'feeds' but this isn't the same as seperate supplies. Does it happen when you switch off the 12v supply, and run the sketch.? Sometimes it can be because the Arduino ground wire is also passing the 5v load ground. You should be connecting the Arduino ground to the power supply (wire #1) The relay board (if it has a ground) should have a ground to the power supply (wire #2) The ground on the 12v supply should also connect to the power supply (wire #3), although it isn't strictly necesary to connect it at all. While all three end up at the same point, the current flows are each contained in their own ground wire. It could be that your 5v supply doesn't have sufficent filtering for the surges created when inductive loads are switched. A large capacitor (4700 - 10,000 uF) across the 5v usually fixes the problem. Mark

michaelkellett · Answer

If you are switching a solenoid with a relay you can easily have EMI problems that affect things some distance away (many metres or more). Put a diode across the solenoid coil (it must be arranged so that no current flows through the diode when the solenoid is energisesd ie annode to positive connection , cathode (usually marked with a ring or line on the diode body) to negative. You should always put dioded across relay or solenoid coils unless there is a good reason not to. When you turn the solenoid on the current takes time to build up - there are no surges, when you turn it off the voltage across the solenoid changes to try to keep the current going - (this is what inductors do) - because the relay is open ciruit the voltage spike across the inductor can be 100s or even 1000s of volts and the rise time can be very fast. It is this very rapidly changing high voltage which can couple into things and upset processors. If you have a diode across the solenoid it gives the current a path and the voltage won't get very high at all (but the solenoid will take longer to drop out). Relay contacts often make and break several times (bounce) and it may help to put a small capacitor across the coil. (0.1uf disc ceramic is often a good start). MK

COMPACT · Answer

You also need enough power budget on the driving voltage as not cause it to sag and stop correct operation.

maranatha · Answer

Hi I just replaced the PIR sensor with a push button. But it didn't solve the problem. It works when the HDMI display is connected but the relay unlatches when the HDMI display is unplagged from the RPi. Replacing the RPi with a LED will not help me in debugging, since the problem is in my opinion with the HDMI port.

COMPACT · Answer

What happens with to Arduino behaviour if you power up the RPi directly from the battery pack without using the relay operating with and without the HDMI connection?

D_Hersey · Answer

You may also want to use some ferrite beads on your power feed

maranatha · Answer

Hi That is good point. I just tried that now few times. The arduino behavior is quite normal. It doesn't reset.

COMPACT · Answer

Trying to find the root cause first.

maranatha · Answer

Just quickly tested the following (in all conditions raspberry pi doesn't have HDMI display) I connected one of the power leads to the relay at a time (of course, the circuit is incomplete and doesn't power the RPi). The Arduino behavior doesn't change. When I push the button (the PIR sensor replacement), the relay latches and stay in that position until I manual reset the arduino. In the same setup, after pushing the button and the relay latches, I waited few seconds and connect the other power lead to the relay to power up the RPi. And it does it without interrupting. This one is still a mystery. After all these trials,(without having to do anything else), the original problem seems to be solved. I tried to power the raspberry pi few time using the button on the arduino and it did work. The relay didn't unlatch in the boot up process. I am not sure how long this will sustain. Will it work on a different environment, different desk, different devices around?

Arduino resets when solenoid shuts off

Top Replies