Hi Guys

Im using an analog devices DG412 analog switch to turn on a 24VDC relay. The 5v logic signal from the micro will cause the DG412 to switch the 0V side of the relay thus energising the coil.

This works perfectly, but i want to add some protection. Because there is a connector to the outside world where an electrician will wire the relay, i want to protect against reverse connection. ie at the moment if +24V is applied to D1 of the DG412, it will blow, the current output is only 30mA max of the DG412.

How can i protect against this condition? I cant use blocking diodes because then my sinking signal wont switch the relay.. I could use a current limiting circuit? but i dont want to get too complex.

any thoughts much appreciated, hope the diagram makes sense. it is not quite complete, but IN1, IN2 etc is the logic signal from microcontroller. D1, D2 etc are connected to (-ve) side of relay coil, and S1,S2 etc are connected to the 24/0V level. VCC is +5V in this case, the supply for the microcontroller logic levels.

 

I am a bit amazed. These devices are made to do signal switching, not power. The only current rating I can find is for about 10ma through the switches. The switching terminals are also supposed to be interchangeable, so the polarity protection seems to be unnecessary.

 

yea my thoughts too, im trying to modify an existing design. im almost thinking mosfets would be better?

 

Oh yes. Just get a FET good for 60 volts and the relay current. You can find logic level FET's for under $1.50.

 

In your case using a BJT or a MOS won't make a big difference because the relay coil doesn't draw large current.

A BJT will switch faster in this case because the output current of the chip is just 10mA max.

 

Oh yes. Just get a FET good for 60 volts and the relay current. You can find logic level FET's for under $1.50.

An addition to Beenthere's suggestion is that ebay has lots of cheap MOSFETs, have a look.

 

excellent. thats why its been working so far i guess, the relay coil current is very low. the only issue i have is that my outputs need to be isolated from my 5V microcontroller supply. ie 24/0V cannot be connected to 5/0V. I would need to do this if switching the fet directly...

Im going to try using an opto isolator and simple voltage divider to bring the 24V down to 5v required by gate of FET... i think this should work and give me the isolation i need, ill post a pic of the circuit once i try it

 

what do i connect the COM signal to on the darlington array? +24V? I would like the isolation because this device will be working in a potentially noisy environment, with the 24V lines being an external source from the main PCB with potential noise here. i want to ensure this isnt going to make it onto my PCB..

 

Hello,

Did you think of an optocoupler as isolation ?
In this "telephone ring detector" circuit, the optocoupler is used to isolate the telephone line from the power circuit for the relays.
http://forum.allaboutcircuits.com/showthread.php?t=13430

Greetings,
Bertus

 

Ive also tried the attached circuit, which does not work. It will work if i apply a 5V signal to pin 15 of opto, but not if i try to use the voltage divider? Is it because of the current the FET needs? It worked with lower resistance values in the divider, but the resistors got too hot...

Excuse my ignorance, im not great on theory!

 

Hello,

What does this 24 / 0 mean ?

Greetings,
Bertus

 

sorry, thats the GND signal for the 24V supply, isolated from the 5V supply

 

Hello,

I have tried to draw my idea.

The gate of the fet needs about 10 Volts to switch.
The diode accross the relays protects the circuit agianst inductive kickback of the coil.

Greetings,
Bertus

 

ahh thank you mate, now i see where i went wrong! works perfect! thanks so much!!

now i think i can use a PTC and crowbar type setup to protect against reverse polarity, seeing as the FET can handle larger current than my DG412 could