Seperate Circuit to power siren?


I got a Arduino Uno for christmas and having fun playing around/order the odd part from eBay.

I want to power a siren switched on/off through the Arduino, now I’ve done some Googling before opening this thread and believe I have the right solution but would like someone to confirm before I fry anything.

I hope to power this siren using a 9V battery

Option 1: Use the 4N35 in the starter kit, I don’t think this is the right option because the draw of the attached siren is 90ma and I’ve found a datasheet for the 4N35 that seems to indicate I shouldn’t put more than 50ma through it.

Option 2: Use the 4N35 combined with a MOSFET, this seemed to be suggested when other people were talking about driving a 9V motor… but I can’t find any other details

Option 3: Use a relay again struggling to find details

I’ve attached the 4N35 Schematic, hopefully someone can point me in the right direction.

4n35.fzz (6.85 KB)

Looking at the circuit someone else provided, I think when I said MOSFET I meant a normal transistor… The below is what they suggested for a motor. I’m thinking I can rip the motor part out (with C1 and D1) and replace it with my siren.


Option 4. Use a transistor to switch the siren. Replace the motor it the diagram with your siren:-

Perfect many thanks Mike

The problem with that circuit using an opto-isolator is that - it is not isolated! If you connect the grounds on both sides of the isolator, it is effectively reduced to a transistor.

If you use an opto-isolator, the ground for the LED side should not be connected to the ground for the transistor side. At least directly. (Note the difference in Mike's diagram.) It may be that there is a remote connection between the ground of the power supply used to power the siren to that of the Arduino power supply in which case the opto-isolator does its job, but the connection should certainly not be made at the opto-isolator.

This seems to be an obvious fault in many of the relay boards offered for the Arduino which assert to be "opto-isolated" and indeed do use opto-isolators, but to little actual effect. If you are going to connect the grounds together, you might as well just use a transistor (or a "logic level" FET which would provide slightly more actual isolation and perform the switching function with a single component).

Thank you for the advice, out of intrest why would the ground not matter on the transistor circuits but does on the optos?

mossj88: Thank you for the advice, out of intrest why would the ground not matter on the transistor circuits but does on the optos?

One of the purposes behind using optical isolation is so that you can have a circuits communicate between systems with different ground references. For example, part of the system could be connected to your mains power and drawing energy from it, but could be floating up at 100V, and then you want to get the information from that circuit to something useful, like a computer. Your computer does not want to be floating at 100V, and connecting the two would probably kill the computer. However if you use an isolation circuit in between you can transfer the information without damaging the circuits.

I imagine the idea behind the relays using optical isolation is to also prevent risk of kickback and noise from transmitting back on the uC. I've seen opto-isolators used to significantly reduce noise between the low voltage control and high power sides of an H bridge.

mossj88: Thank you for the advice, out of interest why would the ground not matter on the transistor circuits but does on the optos?

It's the other way around.

If having the ground directly connected does not matter, then that implies that you do not need isolation; if you really need the isolation (to avoid impulses being conveyed to the ground connection), then you simply cannot have the grounds connected and therefore must use the opto-isolator.

It has a lot to do with the way the grounds are in fact, connected, which ground goes to which other ground and in what order. If that is correctly, organised, then a relay simply driven by a transistor with the "kickback" diode connected directly across the relay will work just fine. Without that kickback diode, an impulse can be conveyed from the collector of the transistor back to the base and perhaps affect the operation of the microcontroller.