Do relays need decoupling?

Thread Starter

jpanhalt

Joined Jan 18, 2008
11,087
The attached circuit shows a 0.1 uF decoupling cap across the relay. Is that common practice? The intended use is with a relatively low current miniature relay with a 33 mA coil at 12V.

I have also seen smaller (e.g., 0.001 uF) caps across the diode, which I assume is for snubbing. Again, how common and necessary is that?

These might be unimportant details. I am just seeking advice from other's experience. Thanks.

John
 

beenthere

Joined Apr 20, 2004
15,819
The diode is certainly a good idea. The cap will make a path to ground for noise from the field collapsing, so it can't hurt. I don't normally do it on a small relay.
 

Thread Starter

jpanhalt

Joined Jan 18, 2008
11,087
Thanks for the quick response. That is the same conclusion I came to. The 12V circuit contains a 555 monostable, no logic chips, and a v. regulator for a 5V supply. Even with a blip in voltage, the 555 is pretty well isolated, so I was going to leave it out, unless someone saw the need for it.

John
 

Ron H

Joined Apr 14, 2005
7,063
Thanks for the quick response. That is the same conclusion I came to. The 12V circuit contains a 555 monostable, no logic chips, and a v. regulator for a 5V supply. Even with a blip in voltage, the 555 is pretty well isolated, so I was going to leave it out, unless someone saw the need for it.

John
Due to the inductance of the relay, there should not be a significant current spike from the power supply, so it seems reasonable to omit the decoupling capacitor.
 

SgtWookie

Joined Jul 17, 2007
22,230
Well, there's another reason for a cap (but connected across the diode.) Unless they're really fast diodes, it takes them a while to turn on and conduct. During that time, you can get a really high voltage spike across the coil.

Usually, a smaller cap in the range of 200pF-500pF across the diode is sufficient to supress the transient until the diode can begin conducting.
 

Audioguru

Joined Dec 20, 2007
11,248
The capacitor decouples the power supply to the 555, not to the relay.

The datasheet for the ICL7555 explains that the ordinary 555 produces a very high supply current spike of 400mA (shoot through) when it switches. The datasheet for the LM555 recommends a minimum of a 1uF and a 0.1uF supply bypass capacitors for it.
 

Thread Starter

jpanhalt

Joined Jan 18, 2008
11,087
SgtWookie: That's what I meant by the 1 nF across the diode. I called it snubbing, which maybe was an imprecise use of that term. I am using a 1N4148 diode. The relay is a Schrack 33 mA. Do you think the snubbing capacitor's benefit is worth it? From the standpoint of designing the board, it is quite a bit easier to route the parallel ("snubber") than the decoupler, and it's only one more SMD to solder. So, just a gut feeling or intuition will move me to include it. I am sure the circuit will work without it, but I like things to be durable.

AG: The 555 is decoupled locally near its power pins on the other end of the board. I was worried about a dip in the 12V line from when the relay turned on.

John
 

Thread Starter

jpanhalt

Joined Jan 18, 2008
11,087
Its max current is only 33mA. Can't your power supply handle it?
I don't know, AG. I thought about using the photocell scavenged from a handheld calculator. ;)

In reality, I'm just learning this stuff. I know a little theory and am light on practice. This is the second and hopefully last version of the board. The first, proof of concept, worked with a few, off-board modifications. So, at this point, it is time to go past just getting it to work and start to incorporate good practices.

The members on this forum have much more practice experience and have seen many more schematics designed by pros than I have. My question was not a "right/wrong" or "works/won't work" type of question. It was meant to find out whether it was a good practice or not. As for the real supply, it produces 4 A easily. There is, of course, inductance in the traces to the relay, which would have the effect of delaying response of the voltage in the trace to the relay and could result in a dip in the 12V line. That's why I used the term decoupling, which was by analogy to the capacitors put on the power pins of the 555. I realized the relay turns on much more slowly than the 555. That's why I asked the question.

John
 

SgtWookie

Joined Jul 17, 2007
22,230
OK John,
The 1N4148's are fast computer diodes. I couldn't even find a spec for turn-on delay time. Recovery time is 4ns max. This leads me to believe that a snubber cap may not be necessary.

But if you have the room on your PCB, you might go ahead and put in the pads for one, just in case you start blowing those 2N7002's. It's easier to plan ahead for possible retrofits, than to try to cobble something on there after the fact - especially if you've had a large run of boards made.
 

Thread Starter

jpanhalt

Joined Jan 18, 2008
11,087
It is done. All have have to do now is print, photo-expose, etch...and my taxes.

I included a 1 nF snubber cap across the diode.

Thanks. John
 

tronics

Joined Apr 16, 2008
14
yes a diode across it prevents a high emf thats produced smoking any transistors around, also a cap and act as an r.f supressor to remove noise.
 
Top