1
\$\begingroup\$

Relays are expensive and have limited lifetime. Choosing a proper relay for a particular application is also important. In other words relays require a bit more of everything when compared to a simple electric switch. So why not instead of using relays, manufacturers somehow couple a small servo to a simple electric switch. Then use it for switching, by controlling the servo. We too can do it (though not perfectly) . What's the reason that relays are so popular and so common for power switching? What's the flaw in using a switch as suggested?

Edit: Using switch can be a one time investment instead of using relays which requires replacement. Also, the cheapest servo motor(premade) can do the work for us. Main idea is lifetime, which I think, may be a reason enough! Also, if this setup is fabricated there may be less chances of individual part failures.

\$\endgroup\$
9
  • 11
    \$\begingroup\$ when compared to a simple electric switch OK, oh but then you need to servo to operate the switch. I fail to understand how that is simpler, cheaper and more reliable than a relay. I think there are many more parts involved in making a servo than making a relay. \$\endgroup\$ Sep 24 '19 at 13:24
  • 18
    \$\begingroup\$ "Relays are expensive and have limited lifetime." You are suggesting that using a servo to control a switch is less expensive and has longer lifetime.I very much doubt that. \$\endgroup\$
    – Oldfart
    Sep 24 '19 at 13:25
  • 16
    \$\begingroup\$ It's like building a better mousetrap. What you described is a relay. With many more points of failure than the actual relay, you are trying to replace. \$\endgroup\$
    – vini_i
    Sep 24 '19 at 13:29
  • 11
    \$\begingroup\$ You do realize that the the switch in your device still wears out just like a relay contact, right? Thereby requiring the entire expensive, complex, less reliable, servomotor-switch assembly to be replaced at least as often as a relay, if not more. \$\endgroup\$
    – DKNguyen
    Sep 24 '19 at 13:46
  • 3
    \$\begingroup\$ The cheapest servo motor is certainly not going to be the most reliable. \$\endgroup\$
    – Finbarr
    Sep 24 '19 at 14:45
11
\$\begingroup\$

That's what a relay is.

In both cases you have a set of contacts carrying current. In both cases there is a mechanism for moving the switch: in a relay, this is done by a magnetic coil. All techniques available for switch construction can also be applied to relay contacts.

The lifetime issue is only visible on relays because they generally get switched a lot more often than humans can manage on a switch. If you look at lifetime specs in terms of number of operations they're usually similar.

\$\endgroup\$
6
\$\begingroup\$

Everything you say about relays is also true about switches. Switches have a limited lifetime, they have to be chosen for the task, and the "somehow" in "manufacturers somehow couple" is going to require individual engineering work and validation each time -- meaning, great expense.

Moreover, cheap hobby servos aren't nearly reliable enough, or, in a lot of cases, fast enough to replace solenoids as drivers for contactors.

The flaws in your thinking are:

  • You want to replace a relay with a switch plus an actuator, without realizing that a relay is a switch plus an actuator, and that we've had over a hundred years to optimize the relay.
  • You do not realize, or are glossing over the fact, that switches wear out just as relays do. And cheap wear out faster than good (generally expensive) ones do -- just as cheap relays wear out faster than good ones.
  • You do not realize, or are glossing over the fact, that the key word in the phrase "hobby servo" is "hobby". Hobby servos are not designed for high reliability. They are not designed to work over a wide temperature range, or for tens of thousands of cycles. Design a servo as reliable as a solenoid coil, and your "relay not a relay" idea gets more expensive than a traditional relay.
  • A servo + switch is going to be slow. Even relays are too slow for some applications; your servo+switch idea is going to be tens or hundreds of times slower than a relay with similar power handling capability.
\$\endgroup\$
1
\$\begingroup\$

There are a few solid reasons to use relays. Your system could do them, but never without extra cost and component count (which adds even more failure modes).

  • Very few applications these days actually wear out relays on a regular basis. Rapid switching can often be done using solid-state components (SSRs or simply FETs depending on the application) (but not with a typical servo) and relays can be over-specified. These two approaches take care of the major failure modes.
  • Relays are inherently insensitive to dirty power circuits. They're reliable even during cranking in automotive applications, when the rail voltage drops by up to around 20%. Try running a PWM control circuit (as used to drive cheap servos) and you'll realise the need for voltage regulation (these aren't the cheapest components to add). That becomes more complicated when you want to do
  • relay logic. This is pretty simple these days, as anything more complicated would be done using microcontrollers (etc.), but in something like a car you can put a relay at the far end of one circuit and use it to provide input to another. In the unlikely event of a failure this relay is user-replaceable. In a home heating system mains power is used to actuate relay coils, meaning you don't need to run a separate low voltage line*. In both of these cases you either need local power supplies/voltage regulators or extra cables.

* A 3-port valve as used in wet central heating systems may be close to what you describe but simpler: a low-speed mains motor moves the valve and a lever; the lever operates a switch that cuts off the motor. Others are spring-return.

\$\endgroup\$
6
  • \$\begingroup\$ I think this answer completely misses the point of the OP's question. \$\endgroup\$
    – DKNguyen
    Sep 24 '19 at 15:28
  • \$\begingroup\$ @DKNguyen in what way? Obviously I disagree but I can't see why an answer that lists advantages of relays over the OP's proposed solution "misses the point" \$\endgroup\$
    – Chris H
    Sep 24 '19 at 15:31
  • \$\begingroup\$ Your first and third bullet also hold true for the OP's proposed device. They are not reasons why you would go with a relay over the OP's proposed device. The only bullet point that attempts to differentiate the OP's device and relay is your second bullet point and it's still not a particularly compelling since it is still possible to design the OP's device to accommodate for noisy power. \$\endgroup\$
    – DKNguyen
    Sep 24 '19 at 17:01
  • \$\begingroup\$ @DKNguyen all at considerable cost and complexity. Clearly that needs to be blindingly obvious rather than implied \$\endgroup\$
    – Chris H
    Sep 24 '19 at 17:50
  • \$\begingroup\$ You wouldn't think so, but you're not taking into account how the OP's the primary mistake (among other things) in his line of reasoning is directly related to complexity. The critical point which is so blindingly obvious to you is clearly one that is lost on him and you think that implying it will be enough? \$\endgroup\$
    – DKNguyen
    Sep 24 '19 at 19:02
0
\$\begingroup\$

When using SPST switches, you need to manually switch them (as you most likely know) whereas, if you use a servo to switch on and off at high speeds, you may realize that the speed of the servo can only go so fast, when a human switches the switch multiple times, the delay between switches is almost never congruent, this is where a relay could come into use for making equal-delayed-switches. Relays would be useful in this case, because in your application running the circuit could tell it to switch for a given time, and be exponentially more accurate with its delays. Hope this helps!

\$\endgroup\$
0
\$\begingroup\$

You can use a switch to operate a relay with a "safe" voltage level while the relay controls a voltage of several thousand volts which is much safer than a switch directly controlling the several thousand volts...

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.