Inductive Voltage Spike in Relay

[andrew.grimm3]

I have this schematic that I'm using to control a 120vAC fan using this relay. I understand the concept of inductive voltage spike that can damage solid state components when controlling the relay with DC. Is there a certain type of diode I should use, or can I just get by with your basic Silicon or Germanium diode in my application?

The schematic is attached...

Thanks,

 

[(*steve*)]

Just a basic silicon diode. A 1N4001 is almost always sufficient.

 

[MrEE]

why do you have 2 transistors? Can you invert the TTL logic? that will save you a transistor.

 

[Zork]

The best choose is to not use a diode, it reduces the lifetime of the relay, if you need a faster switching time, this isn't the right circuit either.
relays.te.com/appnotes/app_pdfs/13c3264.pdf

If the cost has to be low and the coil current isn't so high, the best you can do is use a faster diode, like an 1N4148, the problem of using a slower diode is that on the transition on to off the spike voltage, as you said, will be very high on this time and the diode will not do the switching fast (trr).

The problem is that on the transistor, there are capacitances like Ccb, that can charge on this 'spike' moment and make an base current to flow, this can possibly cause damage on the load and on the relay contacts.

 

[Harald Kapp]

We've been using relays with 1N4148 diodes by the millions for years without problem.
Where switching times are an issue we've used the zener diode circuit as shown in the article linked by Zork.

 

[CDRIVE]

The problem is that on the transistor, there are capacitances like Ccb, that can charge on this 'spike' moment and make an base current to flow, this can possibly cause damage on the load and on the relay contacts.

Which is probably why the ULN200X series drivers incorporate a reversed biased diode across the Ce junction as well as a common terminal that places another reverse biased diode across the coil load.

Chris

 

[CDRIVE]

We've been using relays with 1N4148 diodes by the millions for years without problem.
Where switching times are an issue we've used the zener diode circuit as shown in the article linked by Zork.

This would be a close contest against the 1N4001. I don't think I'd hesitate to use either.

Chris