Question about diode across relay coil

[J M Fahey]

Dave H said it all.
Most important: he measured.
350V peaks !! well, that doesn't surprise me.
Yes, they are probably microseconds wide, maybe a millisecond.
Enough to kill a few components.
On a mechanical switch, they don't have energy enough to melt its contacts, of course, but I'm sure each and every little spark causes a minute crater.
Just give them time.

EDIT:
As a side note, Marconi used spark transmitters to communicate across the Atlantic.
The Titanic used one of them too, and ships 350 nautical miles away heard that SOS.

[guitarmike2107]

yeah ok, thats what I thought..

Thanks

[kg]

I should have been more specific about the circuit I'm using. There are no transistors or IC's to damage. I'm using the heater supply rectified by 2 diodes. There is also a 470 uF capacitor between the supply end of the relay coil and ground. I imagine this would absorb any voltage spikes. The switch used to ground the other end of the coil is a heavy duty toggle switch suitable for use as a standby. It would be rated for hundreds of volts at least so I don't think arcing will ever be a problem with a 5V relay coil. However, if there's something I've missed here and I'm getting away with something I'd like to know.

Also, it is still not clear from the discussion whether or not the flyback diode is needed if there are no sensitive parts like transistors in the circuit. Some members have said it is not necessary other than suppressing pops, some have said it is because of possible arcing in the switch to ground the coil. Again, if your coil voltage is only 5V or 12V and you're using a heavy duty switch rated for 100's of volts is this a concern? Also would that capacitor I have absorb these spikes anyway? Could someone please clarify?

Thanks.

it's almost like you didn't read the thread.

either that or you're looking for reassurance that what you did is proper. you won't get that from me.

here's the bottom line: if it bothers you to put a diode in, don't use one.

if you want to do things right instead of doing them twice, solder in a 2 cent diode.

[R.G.]

here's the bottom line: if it bothers you to put a diode in, don't use one.

if you want to do things right instead of doing them twice, solder in a 2 cent diode.

+100. Otherwise, you'll be setting up a little repair mystery for yourself for someday after you've forgotten this discussion.

[GregS]

Originally Posted by kg
here's the bottom line: if it bothers you to put a diode in, don't use one.

if you want to do things right instead of doing them twice, solder in a 2 cent diode.
+100. Otherwise, you'll be setting up a little repair mystery for yourself for someday after you've forgotten this discussion.

Thanks.

[Enzo]

And my 2 cents worth:

All this discussion and angst over whether or not to include the good practice part - a 2 cent diode? Geez.

It is like a seat belt in your car. 99% of the time, you could leave the belt unused and drive home without a problem. If the need for it ever comes up, you should be wearing it. If the amp seems OK without a diode, fine, but for 2 cents, why not wear one anyway?


And why use it to protect transistors when you are using none? It is that whole good practice thing. You build one without and use a large toggle switch and it all works fine. LAter in life you build something that needs to click several relays to switch channels, so you use transistors. Never occurs that you need to change how you do it now, and transistors fail. Or late in life you include the relays and toggle in some much higher gain amp, and NOW you get the pops.

[GregS]

And my 2 cents worth:

All this discussion and angst over whether or not to include the good practice part - a 2 cent diode? Geez.

It is like a seat belt in your car. 99% of the time, you could leave the belt unused and drive home without a problem. If the need for it ever comes up, you should be wearing it. If the amp seems OK without a diode, fine, but for 2 cents, why not wear one anyway?


And why use it to protect transistors when you are using none? It is that whole good practice thing. You build one without and use a large toggle switch and it all works fine. LAter in life you build something that needs to click several relays to switch channels, so you use transistors. Never occurs that you need to change how you do it now, and transistors fail. Or late in life you include the relays and toggle in some much higher gain amp, and NOW you get the pops.

My reluctance to put one in has to do with the fact that it's a bit of a pain in the neck to do it at this point in time. To put it on the board how I'd want it requires desoldering a bunch of wires to turn the board over - so if it wasn't really necessary it would be a lot simpler to leave it. For future builds I'm not going to to worry about saving 2 cents. For this particular amp I'll put one in in an awkward way to save the trouble I already mentioned since it seems physical proximity to the coil isn't that important from other inforamtion in this thread.

[kg]

does lead dress inside the amp not matter? would you route your power supply and ground lines all the way around the outside of the chassis? do you use a 40' instrument cable?

if you answered no to any of those, then position and proximity are important.

however, if choosing between a less-than immediately proximal catch diode vs. none at all, the former is definitely preferable!

[GregS]

It would amount to a few extra inches of wire in a non-sensitve part of the amp so I don't think it will be a problem.

[mooreamps]

It would amount to a few extra inches of wire in a non-sensitve part of the amp so I don't think it will be a problem.

It won't...

-g