Recently I did some work on my car electric circuits. I see many relays are used in car circuits. These relays are used for simple switching, and I wonder why these circuits are based on relays and not on transistors or other electronic components usable for switching purposes. I thought that transistors were cheaper, smaller and more reliable than classic el-mech relays for switching.

Note: In car applications a 12 V car battery is used to power the coil of a relay and the same 12 V power is what's switched by the relay. Sometimes a relay switches just another signal line, i.e. without any high power load on it. And still, I can see no transistors in there. So I believe there is a solid reason why it is done this way, and I must be missing something here. :-)

  • 4
    \$\begingroup\$ How old is your car? Have you tried doing some research on the change over of relays to something solid state? \$\endgroup\$
    – Andy aka
    May 3 '16 at 13:07
  • 32
    \$\begingroup\$ I still remember people calling transistors "three-legged fuses"... \$\endgroup\$ May 3 '16 at 13:11
  • 5
    \$\begingroup\$ I'm just thinking about a guy in a garage trying to figuring out why something don't work anymore. He pushes a button, waits for the "click" of the relay, but he won't hear anything. Is it broken, or is it an transistor? This might not be the reason, but it's probably a pro for relais. It's way easier to tell if they're broken. \$\endgroup\$
    – jawo
    May 3 '16 at 13:16
  • 3
    \$\begingroup\$ Once the car will have 100% SSR instead of electromechanical, the people will be yelling: where have the good old relays gone? \$\endgroup\$ May 3 '16 at 13:36
  • 3
    \$\begingroup\$ @Sempie It's about time garage guys discover OBD systems. \$\endgroup\$ May 3 '16 at 14:24

Relays are much more stable temperature-wise: a sealed relay has essentially the same characteristics at -30°C and +70°C, both temperatures being common for cars. A transistor works quite differently at -30°C and +70°C, so the schematic has to be designed to account for those variations.

I once worked on a product with temperature range starting at -55°C, which used both relays and semiconductor devices. The funny part about the design was that below -20°C only the relay part was powered, which activated air heaters and would only switch on the semiconductor part once the temperature reached 0°C.

Relays also offer galvanic isolation, which effectively confines faults. Common failures like short circuits usually damage only one relay, whereas in transistor-based circuits several devices along the way would be affected. I bet people still want their car's motor running even when the air conditioner or a window lifter dies.

  • 12
    \$\begingroup\$ It's worth noting that isolation can be provided with solid-state devices, usually with opto-isolators (or solid state relays, which are basically an opto-isolator and a big transistor in one box). \$\endgroup\$ May 4 '16 at 8:06
  • 3
    \$\begingroup\$ @SomeoneSomewhere The original question is about transistors. SSRs are nice, but more expensive than conventional relays. \$\endgroup\$ May 4 '16 at 9:59
  • 10
    \$\begingroup\$ Also, the major advantages of transistors over relays are that they're small and quiet (no moving parts). These are very important when you have to fit millions of them into a smartphone, less so for the few dozen or so you need in a car. There's plenty of room for as many relays as a car needs under the hood, and the little noises they make are completely masked by the sound of the car itself. \$\endgroup\$ May 4 '16 at 20:08
  • 11
    \$\begingroup\$ The noise the relays in a car actually used to be an unintentional usability feature. The typical "blinker" sound originated from the sound the blinker relay made when switching the blink-lights on and off. This "feature" told the driver the blinker is on. Now that relays became quieter, the effect is often reproduced through the car stereo system. \$\endgroup\$
    – Philipp
    May 5 '16 at 11:12
  • 2
    \$\begingroup\$ @Philipp Good point on usability as when I switched from old Nissan to a new Toyota I was missing that click-click sound for a few months. \$\endgroup\$
    – zmechanic
    May 6 '16 at 12:22

A car is (still, in the modern era) an electrically harsh environment. The "12V supply" is typically 13.5 to 15 V, and may occasionally spike to 80V or more. There may also be some high-frequency junk on the wires from the spark plugs.

Relays put up with that abuse somewhat better than transistors, at least for "a similar price point."

  • 29
    \$\begingroup\$ @DmitryGrigoryev, If the relay costs $1.00 and there was a suitable, equally reliable, solution for $0.97, you can be sure the engineers at Ford/Chevy/Toyota/etc would choose it. \$\endgroup\$
    – The Photon
    May 3 '16 at 16:22
  • 13
    \$\begingroup\$ Without attempting to disparage any particular auto maker, @ThePhoton, in General, some auto makers will save the 3¢ even if it's less suitable or reliable. \$\endgroup\$
    – FreeMan
    May 3 '16 at 20:12
  • 3
    \$\begingroup\$ What could cause a spike to 80v (or higher)? \$\endgroup\$ May 3 '16 at 21:08
  • 12
    \$\begingroup\$ @DigitalTrauma shutting off your blower fan quickly. All the energy stored in the motor's coils dumps back into the power rails. \$\endgroup\$
    – rdtsc
    May 3 '16 at 22:12
  • 2
    \$\begingroup\$ Note also that incandescent lamps often draw up to ten times their running current when cold. Relays can easily be specified to handle this surge in current. A mosfet will have to have sufficiently low on resistance to handle this. A bipolar will need lots of base drive. \$\endgroup\$
    – Richard
    May 4 '16 at 13:36

The voltage in a car is only about 12V, which means that even moderately powered components can draw large currents. The dash in my car is illuminated by four 12V 2W bulbs. They draw a current of 666mA just to light up the dash! If you look at the fuses for all the circuits in your car, even the smallest will be 5A. Most will be 10-20A range and some even more than that. The reason relays are so popular is because they are durable, have low contact resistance, and are (sometimes) cheaper than solid state components that can handle the same current. Many modern cars actually do use solid state relays but they come in the same type of brick packaging so as not to confuse any mechanics.

  • 3
    \$\begingroup\$ Actually, solid-state relays are more durable, but that is not needed in cars. Most relays activate once per motor start, so in 10 years of driving twice a day they will be at 7200 cycles or so. \$\endgroup\$ May 3 '16 at 21:17
  • 7
    \$\begingroup\$ That's absolutely correct. I should have said "relays are durable enough". Mechanical relays hardly ever fail unless there are other electrical problems. Most outlive the car they are used in. \$\endgroup\$ May 3 '16 at 21:21
  • 2
    \$\begingroup\$ @DmitryGrigoryev depending on the car, that's not true. The second biggest relays are the Fan, Lamps, and Horn relays, all of which cycle multiple times, more than once per ride (the horn depends on road rage). \$\endgroup\$
    – Passerby
    May 5 '16 at 6:21
  • 1
    \$\begingroup\$ @Dmitry Grigoryev: From my experience, your calculations are about 4 times lower that "normal." Around 30,000 cycles is closer to the minimum number of cycles. \$\endgroup\$
    – Guill
    May 6 '16 at 6:08
  • 2
    \$\begingroup\$ @guill Most relays exceed 30,000 cycle life expectancy \$\endgroup\$ May 6 '16 at 6:11

Relays are just well-tried. I think the answer is simple as that.

Mechanics often want a simple solution, which works well. It's just in the mind of the people to switch a circuit with relays.

My brother mounts some special equipment on trucks for work. Always he needs to switch something he comes up with: "Oh, let's put a relay in there!". And he's right. Why not? It's often the simplest solution and it works fine.

In some situations the acoustic feedback can be used as well. Just think about your indicator.

  • \$\begingroup\$ Acoustic and temperature (warm = active) feedback does work to a certain point. I have also seen relays with built-in LEDs to indicate the state, though not in cars. \$\endgroup\$ May 3 '16 at 22:18
  • \$\begingroup\$ @robin I've heard that the life of relay is comparatively less.. Are the automotive standard relays are far more reliable ? Or is it ssr that you are talking about? \$\endgroup\$ May 4 '16 at 20:42

Relays have a few advantages over transistors.

  1. They have a very low on-resistance. In a high power low voltage system this is valuable as it improves efficiency and eliminates the need for heatsinking.
  2. They have very low leakage when off. So their outputs can be connected to unswitched battery power without worrying about draining the battery.
  3. They are robust against spikes, surges, temperature variation etc. The environment of a car is a harsh place both electrically and thermally and designing transistor circuits to survive it requires a fair bit of extra work.
  4. They provide isolation. While the grounds in a car do all eventually come together isolation is still useful to keep control of where exactly returnc currents flow.
  5. They can be easilly used for either high-side or low-side switching and their inputs can be either high or low side switched. Using N-channel fets (the better performing type) for high-side switching requires a gate drive voltage above the main power supply voltage.

Relays also have a few disadvantages over transistors.

  1. The coil requires a fair bit of power.
  2. The cycle-life is limited
  3. They are physically large

So for high power stuff that switches infrequently relays generally win. For complex and/or high speed control stuff solid state electronics generally win.


Can I keep this real simple? The car has a set of conditions under which it much operate. And as others above have indicated, they are pretty significant. The requirements include a solid performance pattern for component quality.

Bottom line... the reason for relays over solid state devices is summed up in a single word.


It's cheaper to provide relays than it is to provide solid state devices to perform that function. When solid state devices come down in price below that of relays, then the automotive manufacturers will switch over to solid state devices. Cost is a major driver when it comes to decisions like this.


A 10 amp relay costs 10 times as much as a 10 amp automotive grade power mosfet. – Passerby

So, I'm not quite sure that's a fair comparison. I'm thinking most automotive relays carry more current than that. (25 to 60 amps?) Additionally remember we're totally isolated between signal and output with the relay. Assume the Automotive OEM's would want to keep that isolation concept. What the cost of a opti-isolated 30 amp mosfet going to run? Oh, and better place that into a nice plug in container, best if its a perfect drop in replacement for an existing relay.

Additionally there is one more factor related to PRICE. Remember there are suppliers out there somewhere who have invested millions and millions of dollars to automate and tool up their relay manufacturing plants. Additionally they employ a whole bunch of folks. What do you think those guys charge for relays? Answer: As much as they think they can get away with. Guess what happens when the automotive Original Equipment Manufacturers (OEM's) tell those folks they are thinking of switching over to Mosfet technology... you guessed it, there is a immediate drop in price. It may not matter that they are losing money, they've got huge amortization and fixed costs and lots of employees.

It would not surprise me that a true price advantage of Mosfets over relays gets delayed in implementation for purely business reasons for a period of say five years. It would also not surprise me if some of those businesses had sweetheart deals with the OEM's (e.g. I want to invest a lot of money in my process to reduce my labor costs and therefore my price to you, but I'll only do that if I have a five year contractual commitment for product.)

  • 3
    \$\begingroup\$ A 10 amp relay costs 10 times as much as a 10 amp automotive grade power mosfet. \$\endgroup\$
    – Passerby
    May 5 '16 at 6:19
  • 1
    \$\begingroup\$ most automotive relays carry more current than that. 25 to 60 amps? Varies. Main switched power, sure. But Fans, Pumps, Headlights, Horn typically have 5~20. \$\endgroup\$
    – Passerby
    May 5 '16 at 7:22
  • 1
    \$\begingroup\$ Additionally remember we're totally isolated between signal and output with the relay. Not the way they are wired. Most are powered by the same source that the switched circuit connects to. Isolation doesn't seem important, based on the wiring schematics I've seen. Typical main engine, headlight, and horn i.stack.imgur.com/TDWEC.png i.stack.imgur.com/9mkUy.png From a 99 Camry. No isolation. When everything is mostly 12V and the same battery... \$\endgroup\$
    – Passerby
    May 5 '16 at 7:24

Relays will be used in place of semiconductor components due to following reasons

  1. More Durable - wide operating temperature ranges, spikes.
  2. no need of heat sink.
  3. very low on state resistance.
  4. Isolates load completely in off state.
  5. Low cost when compared to same current conducting semiconductor.
  6. Service point of view - (most automobile relays and fuses are plug and play...) - which is difficult when we use semiconductor (it has to be removed from PCB)

Another thing about relays as compared to semiconductors/transistors is that relays are more mechanic friendly to replace as to all transistor electrical systems for automotive repair.


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.