Getting appropriate voltage for relays

[abbey road d enfer]

Wouldn't a dedicated regulator help with reducing the "click" noise while switching? Within reason of course.

It may or may not, more likely not.
If a regulator improved this, it would be because of severe mistakes in the implementation.
Clicks due to relays have basically two origins:
One is the current variation propagating through the ground and polluting the audio ground. A good designer shouldn't let that happen.
The other is electromagnetic propagation of the voltage surge that happens when opening the circuit and the magnetic energy in the coil has to be released and transforms into radiowaves. Thorough design includes diodes and RC snubbers or soft-switching.

 

[JMan]

Thank you all for the ongoing discussion. As always, your thoughts are more interesting than the problem itself!

I’ve figured out what was going on, and I guess predictably, it was a couple of very human errors on my part. I still might toss that resistor in there just to drop the voltage a bit, but clearly I misidentified the issue.

 

[JohnRoberts]

Back in the day for bypass relays I generated a separate PS using a small reservoir capacitor from the unregulated rail so as soon as power was interrupted the cap would discharge and switch the relay to bypass mode before the regulated rails collapsed. A simple rc driving a transistor slowed the turn on relay drive a suitable time delay for the circuits to stabilize.

JR

 

[murrayatuptown]

While it intuitively seems like a DC relay coil would be inherently polarized (like wondering if the actuator would push rather than pull with opposite coil polarity), I have only ever physically encountered two kinds of DC relays that suggested or were marked with polarization: relays with integral diodes across the coil (less frequently, relays with both integral driver transistors AND transient suppression (like Teledyne TO-5 can relays), and relays with specs that called them 'polarized'. The ones I saw specifications for (I had to look them up for someone, and didn't literally have one) that said 'polarized' didn't literally say 'latching', but I think they had high G force ratings. Maybe they latched, also, but the ability to stay closed without contact bounce seemed to be the focus.

Panasonic describes (their) polarized vs. unpolarized relays in this literature, and their focus seems to be on latching types. Single-coil and dual-coil latching types exist (not referencing Panasonic, just in general). Difference between Polarized and Non-polarized Relays | Automation Controls | Industrial Devices | Panasonic

Some relays are shielded (like on multiplexer cards with several or many). Most of those I've seen (repairs for welded contacts or contact resistance exceeding end-of-life rating) didn't have coil polarity markings, and neither did the circuit hoards, but it might have been desirable to maintain the original behavior. The board layout prevented accidental reversal, and that's a different scenario altogether than incorporating a single relay that wasn't present previously.

'Not liking certain voltage': I often test a relay I haven't found specs for (yet) with a variable bench supply. If I'm lazy and just use clip leads and no transient protection diode, I have seen a regulated supply 'freak out' when the relay is cycled repeatedly. The voltage digital display shows transient voltage changes, and I immediately am embarrassed I was too lazy and didn't think about the power supply being offended by the transients. I usually think driver IC's or transistors and EMI are what the diodes protect. Then I respect the bench supply for a while until I get lazy months later. I should probably make a pair of test clips with diodes soldered so I have no excuse for thinking it's too much work to do a 30 second relay experiment.

If you do have a single diode across a relay coil, then there is no question at all that the application itself is now polarized.

A DC with too much coil voltage (within 'reason', which might be on a data sheet) applied will run hotter than it was intended but should operate in short term testing).

If you add series resistance to operate a lower-voltage DC relay from a higher supply voltage, it reduces the L/R time constant. That's probably not an issue unless you're driving a relay intermittently with pulses.

The pull-in minimum voltage spec is nice to be aware of but don't put your money on it...might depend on temperature. The actuating force from lower coil flux might be lower, but I always think I would like to run a relay between the minimum pull-in voltage and 'rated' if I'm going to have the coil powered long term and don't know if it has a duty-cycle rating...so it runs cooler.

Someone told me decades ago to not listen to people who say relays aren't reliable...like any other component, they have limits that need respect and it's usually a misapplication that results in a relay failure.

I've welded both reed relay and power switch contacts with my own ignorance.

 

[Newmarket]

That's why I wrote "full full switch-on voltage".

Got it. So I seem to have "scanned" what you wrote a bit too quickly (although there's one too many "full" in your self quote.
Probably due to having explained this numerous times to other people over the years who always seem to want to point to the relays as the cause of a problem (albeit in non-audio applications).
Thanks for clarifying the point.

 

[Newmarket]

While it intuitively seems like a DC relay coil would be inherently polarized (like wondering if the actuator would push rather than pull with opposite coil polarity), I have only ever physically encountered two kinds of DC relays that suggested or were marked with polarization: relays with integral diodes across the coil (less frequently, relays with both integral driver transistors AND transient suppression (like Teledyne TO-5 can relays), and relays with specs that called them 'polarized'. The ones I saw specifications for (I had to look them up for someone, and didn't literally have one) that said 'polarized' didn't literally say 'latching', but I think they had high G force ratings. Maybe they latched, also, but the ability to stay closed without contact bounce seemed to be the focus.

Panasonic describes (their) polarized vs. unpolarized relays in this literature, and their focus seems to be on latching types. Single-coil and dual-coil latching types exist (not referencing Panasonic, just in general). Difference between Polarized and Non-polarized Relays | Automation Controls | Industrial Devices | Panasonic

Some relays are shielded (like on multiplexer cards with several or many). Most of those I've seen (repairs for welded contacts or contact resistance exceeding end-of-life rating) didn't have coil polarity markings, and neither did the circuit hoards, but it might have been desirable to maintain the original behavior. The board layout prevented accidental reversal, and that's a different scenario altogether than incorporating a single relay that wasn't present previously.

'Not liking certain voltage': I often test a relay I haven't found specs for (yet) with a variable bench supply. If I'm lazy and just use clip leads and no transient protection diode, I have seen a regulated supply 'freak out' when the relay is cycled repeatedly. The voltage digital display shows transient voltage changes, and I immediately am embarrassed I was too lazy and didn't think about the power supply being offended by the transients. I usually think driver IC's or transistors and EMI are what the diodes protect. Then I respect the bench supply for a while until I get lazy months later. I should probably make a pair of test clips with diodes soldered so I have no excuse for thinking it's too much work to do a 30 second relay experiment.

If you do have a single diode across a relay coil, then there is no question at all that the application itself is now polarized.

A DC with too much coil voltage (within 'reason', which might be on a data sheet) applied will run hotter than it was intended but should operate in short term testing).

If you add series resistance to operate a lower-voltage DC relay from a higher supply voltage, it reduces the L/R time constant. That's probably not an issue unless you're driving a relay intermittently with pulses.

The pull-in minimum voltage spec is nice to be aware of but don't put your money on it...might depend on temperature. The actuating force from lower coil flux might be lower, but I always think I would like to run a relay between the minimum pull-in voltage and 'rated' if I'm going to have the coil powered long term and don't know if it has a duty-cycle rating...so it runs cooler.

Someone told me decades ago to not listen to people who say relays aren't reliable...like any other component, they have limits that need respect and it's usually a misapplication that results in a relay failure.

I've welded both reed relay and power switch contacts with my own ignorance.

Suffice to say that measures to counter the "Back EMF" on turn off are essentially mandatory.
If not implemented you may (literally) hear it acoustically. Some relays have an integral diode so are inherently 'polarised'.
Simple diode solution is very effective. It increases Turn-Off time but this is not usually an issue for audio applications. Other solutions (RC snubber etc) have other trade offs.

 

[Newmarket]

Wouldn't a dedicated regulator help with reducing the "click" noise while switching? Within reason of course.

Why would that be the case in general ?

 

[JohnRoberts]

I did a deep dive into relay clamp diodes within the last year to suss out my sump pump controller failure where the same drive transistor failed twice. There is a lot of info on the WWW from a search. I just replaced the blown npn with higher voltage part I has sitting in my back lab, and it is working still (so far).

Long story short using different clamp diodes (small signal vs power rectifier) alters the relay on/off switching time. Using relays for fast switching is not very sensible but relays have been around for a very long time... Back in the day when they talked about bugs in computer programs, they were literally talking about bugs (insects) in relay contacts.

JR

 

[ccaudle]

I just did this on my mix board I'm laying out. 16 volt rails = 32 volt swing. 3 9v relays in series = 27v 16.7mA per relay. For a total of 50.1 mA = Two 50 ohm resistors

A bit off-topic to the rest of the discussion going on, but that drawing is of three relays in parallel, so the entire 32V is across each relay. I don't think that is what you intended for 9V relays. Well, 32V - 5V dropped across the two resistors, so about 27V across your 9V relays.

 

[57sputnik]

The Panasonic DS2Y, now discontinued, is polarized and will not operate with reversed polarity. It does not have an internal clamp diode.

With a 28V supply powering a 24V relay I would suggest a 3.9V series Zener as others here have. 28V on a 24V coil will make it run hot and increase out-gassing. In addition to the series Zener there should of course be a clamp diode across the coil.

A series resistor to limit voltage assumes the relay will have fixed coil resistance. It won't because it will vary with model and will change due to the temperature coefficient of the coil. You certainly can use a resistor in a "one-off" situation. Run it at full coil voltage rather than the must-operate voltage.

I ran some experiments pulling in a DS2Y relay weakly and found that the contact distortion - some of it very high order - increased significantly at signal currents in the 1 mA RMS range. That's different scenario than pulling in the relay hard and relaxing current but its something to consider.

For boards having a lot of relays I usually put a dedicated 7824 regulator and provide an isolated relay common. For boards with only a couple of signal relays a 78L24 with the +input tied to +15V and the 7824 "ground" tied to -15V provides a +24V supply with the common being -15V. (It measures +9V relative to ground.) The "differential" connection to the bipolar rails returns the switching current to -15V instead of audio ground. I haven't found this approach to be "clicky" as long as the bipolar rails are well-bypassed and the circuit has reasonable PSRR.

I don't put a capacitor across the coil because that will create an in-rush current at turn-on if there is no series resistance. Without series resistance a capacitor may create clicks rather than suppress them. If the capacitor is large enough it will also increase wear on the controlling switch. Let the diode control the coil turn-off spike. When the controlling switch is open the current loop during turn-off is confined to the diode and coil.

The above has worked for me. As JR oftens says "YMMV."

FWIW here's a simple turn-on mute circuit using the TL431 and a relay with a series Zener. The supplies need to hold up for about 10 ms at turn-off.

https://www.proaudiodesignforum.com/forum/php/viewtopic.php?f=6&t=1271

 

[JMan]

Okay, so the practical message for my situation that I'm extracting from the higher level discussion is that my relay should work at 28v (which it now is), but that it is not ideal to run it at this higher voltage. An early suggestion was to add a series resistor to drop the voltage, and the subsequent suggestion was to use a zener diode instead.

In my case, the relay is not using the negative rail of my power supply, but is instead being fed the +28 directly from the psu with the other side of the coil connecting to psu ground. So if I'm reading the above correctly, a 3.9v zener between the relay and ground, anode toward ground (?), will accomplish what I want?

(Oh, and yes, I do have the flyback diode across the coil. That part at least I got correct!)

 

[JohnRoberts]

As has also been shared , relays require higher pull-in voltage than hold voltage, so a dropping resistor with an electrolytic capacitor in parallel could satisfy higher pull in voltage, and lower hold voltage.

JR

PS; As Wayne observed too little hold current is not good either.

 

[57sputnik]

"So if I'm reading the above correctly, a 3.9v zener between the relay and ground, anode toward ground (?), will accomplish what I want?"

Yes, anode of Zener to ground, cathode to coil. (Like my drawing only no TL431 and no -15V.)

 

[Bobby Baird]

---

A bit off-topic to the rest of the discussion going on, but that drawing is of three relays in parallel, so the entire 32V is across each relay. I don't think that is what you intended for 9V relays. Well, 32V - 5V dropped across the two resistors, so about 27V across your 9V relays.

Good point. Carry on.

 

[warpie]

This is probably a bit off topic but what flyback diode would be good for a 24V relay? Is a 1n4148 sufficient or is it better to use a more beefy one?

 

[elskardio]

This is probably a bit off topic but what flyback diode would be good for a 24V relay? Is a 1n4148 sufficient or is it better to use a more beefy one?

The diode has to handle a reverse voltage equal to the power supply voltage and it has to handle a peak current equal to the relay coil current.
The 1n4148 is more than adequate.

 

[Newmarket]

This is probably a bit off topic but what flyback diode would be good for a 24V relay? Is a 1n4148 sufficient or is it better to use a more beefy one?

1N4148 is fine. But I find the glass body package too susceptible to damage - I've had them sort of break and disintegrate during rework - so I tend to specify 1N400x types or SMT equivalents as appropriate.
fwiw the 'see thru' glass body makes them sensitive to light although that is not an issue in this application.

 

[warpie]

Very good points, thank you very much!

 

[warpie]

I have a few simple questions about relays and I thought I would ask here instead of starting a new thread. I hope that's OK...

1) Consider a 24V relay with minimum operating voltage @18Vdc. My understanding is that if you operate the relay on a lower voltage (i.e 20 or even 18Vdc) the power consumption would be lower, which is a good thing I guess. But what is the disadvantage comparing to operate the relay at 24Vdc?

2) If there are a few relays in parallel, do they all need a dedicated flyback diode, or one would diode be sufficient for all the relays? My guess is that the former is the correct approach but I'm not sure.

3) and finally (and most silly question). If the relay(s) is connected between a positive voltage and GND, does it make any difference whether you connect the switch across the positive rail or the across the ground? (i.e switching +Vdc or GND).

FWIW, in all the above cases/ scenarios a dedicated voltage regulator would be used for the relays, leds, etc...

Thanks!

 

[Bo Deadly]

I have a few simple questions about relays and I thought I would ask here instead of starting a new thread. I hope that's OK...

1) Consider a 24V relay with minimum operating voltage @18Vdc. My understanding is that if you operate the relay on a lower voltage (i.e 20 or even 18Vdc) the power consumption would be lower, which is a good thing I guess. But what is the disadvantage comparing to operate the relay at 24Vdc?

The disadvantage is that it might not work. There is a threshold voltage required to properly energize the coil. If that is 18V, then that should be fine. But if your supply dipped a little for some reason like because the supply is linear and there was a momentary "brown-out", you might get a bunch of relays that flip momentarily.

The advantage of running nearer to the threshold voltage would be that you would save power. But you can save a lot more power in another way. There are actually two different currents for a relay: the switching current and the "holding current". The holding current is usually a lot lower than the switching current. If you make the supply to the switch that feeds the coil into an RC, the switching current will come from the capacitor but, after the cap drains, the holding current comes through the resistor. That resistance is now in series with the coil which reduces current and thus power. You can actually save a lot of power that way but it's never done because it's just not really necessary and because the caps have to be pretty big. However, there might be a justification for this now that MLCCs are so small and cheap. You can get a 22uF MLCC the size of a grain of rice for less than $0.50 in quantity.

2) If there are a few relays in parallel, do they all need a dedicated flyback diode, or one would diode be sufficient for all the relays? My guess is that the former is the correct approach but I'm not sure.

If they are all switched together you can use one flyback diode. But you should probably use a beefier rectifier diode and not a little signal diode. Any relay coil that is switched separately needs it's own flyback diode.

3) and finally (and most silly question). If the relay(s) is connected between a positive voltage and GND, does it make any difference whether you connect the switch across the positive rail or the across the ground? (i.e switching +Vdc or GND).

Technically it doesn't matter but personally I would put the switch common on GND so that you don't have a high inductance coil and diode dangling off of GND. This is almost certainly of zero concern in practice but if it doesn't matter either way, I would switch GND.

FWIW, in all the above cases/ scenarios a dedicated voltage regulator would be used for the relays, leds, etc...

Using a separate regulator is not really going to do much for you. If you are trying to minimize the possibility of transient noises and such there are other things that are more important. The most important would be using a separate ground. I have always used a DGND or "dirty ground" to for all things non-audio like digital and relays. If you do that and follow the supply follows return rule, any transients should not be a problem at all.

Note that the best way to drive relays is with a special driver IC like L9823 that limits EMR and has diodes and other protections builtin. Similar for LEDs.