softening incandescent inrush current for relay contacts

Go To Last Post
18 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

For a 12 VDC incandescent lamp total load of about 4.8A at steady state, there's an initial cold filament turn-on inrush current draw of maybe 50A until the filaments heat. Lasting for maybe 40ms, with a sharp decay..

This is being switched with a relay on the high side of the lamp (or array of lamps wired in parallel). Inrush is hard on relay contacts. I'd like to keep the relay small and spec'd lower than one with contacts rated 50A.

I know I could limit inrush putting a series resistor in line with the bulb. But then I need to get the resistor out of the line or I'll just be heating that up, and robbing power to the light, and probably need a hefty power resistor to prevent it going up in smoke. Space is a premium.

Was wondering if using a series inductor with the lamp load instead of a resistor would be better? Of the three basic elements RCL I have a good grip on R and C and my maths and theory of L are less developed, shall we say.

The relay switching frequency is 0Hz (constant on) up to 3Hz (flashing), square duty cycle.

My thinking is putting an inductor on the supply side to the relay NO contact, relay wiper to lamp load. I'd put a flyback diode over the inductor so when relay opens the inductor dumps back to +12V supply... to prevent arcing over the relay contacts.

Would any kind of inductor work? What value would it have to be to limit inrush current to something less than 20A.. and the dimensions of such a beast?

Or is this scheme folly? Inductor would interfere with incandescence and lamp lighting, or not be able to meet the 3Hz frequency, or not be able to limit current in a meaningful way to get inrush down to a level that makes the relay contacts happier? ...

Thank you!

Regards,
Scott

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The inductor is an option and a good one. Hummm, the math for reduce the initial current I can't remember right now, but the inductor peak current should handle the 50A....

Other option is to use an NTC, but I don't think there is a NTC to handle 50A.

Other option is to put a second series relay, with an resistor connected in the NC contact and after 1s you turn on it and then it cutoff the resistor and direct connect the lamp. So the lamp is connected to the relay's NO contact. Got it?

Regards,

Bruno Muswieck

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You put a 'keep alive' resistor from the lamp to gnd that keeps it right about 6V... almost glowing red. Its always hot, turns on instantly, no inrush, no thermal shock. Win Win. You're Welcome!

Imagecraft compiler user

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

bobgardner wrote:
You put a 'keep alive' resistor from the lamp to gnd that keeps it right about 6V... almost glowing red. Its always hot, turns on instantly, no inrush, no thermal shock. Win Win. You're Welcome!

That'd kill battery... wish I could.

But can't put a constant low current across the lamp.

..back to that inductor question?

Regards,
Scott

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You need some kind of 5A current source.
Inductor is not a bad idea but you need to pick the one that charges to ~5A in ~40ms (assuming the light bulb is considered a short initially).
di/dt = U/L so you need an inductance of under:
L=12V*4e-2s/5A = 0.1H (not 0.1mH but 100mH !)
with some really small coil resistance to not overheat.
Mind that this beast holds E = L*i*i/2 = 1.25J of energy when charged and will try to get rid of that when the circuit gets opened.

Quote:
Space is a premium.

Then use a transistor instead of a relay. Especially when it is supposed to blink.
You can ramp up the drive (open loop) to increase the voltage slowly. With a little bit of attention you can make the algorithm prevent preheating too long or too frequently and I suppose a 5A peak, 30V MOS in SO-8 should do the job reliably (just do not ramp up too slow).

No RSTDISBL, no fun!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Thanks for that math and nice explanation!

Hmmm.. 100mH to handle 5A, looking that up is not a cheap or small thing.

This is an impractical idea I will put to rest.

Regards,
Scott

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I second Brutte's advice - why not just go solid-state? Inrush currents are then easily controlled...

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
Hmmm.. 100mH to handle 5A

You must understand this is based on assumption a light bulb is a complete short for these 40ms. In reality it is not, also the voltage on a coil drops from 12V to 0V at the end of preheat period. So a 100mH is a theoretical limit and your coil must have an inductance smaller than that. In other words if you use a 100mH inductor then the current won't hit 5A in 40ms even with short instead of the bulb. So perhaps you need 1mH, perhaps 90mH, but never 110mH. Everything depends on the construction of the bulb.

What I can propose is, instead of buying whole bunch of 5A inductors or frying relays, you should get a set of E12 capacitors (or one variable cap) and use a protoboard to make a gyrator:

  • one trimmed capacitor
  • one raw op-amp (LM324N is perfect)
  • one raw beefy MOS transistor (10A, TO220)
  • bunch of resistors
A gyrator can "pretend" any inductor(actually it can pretend anything, including negative resistance).
No need to wind anything or buying 5A, 100mH stuff, no cores needed etc, just change cap to change the "pretended" inductance.
Gyrator "pretends" an inductor in the sense of U=L*di/dt.

No RSTDISBL, no fun!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Go on. You have been offered the best solution. You could even use a transistor to switch the "startup" resistor. Then use the relay to switch the main current.

Give Uncle Bob a coconut.

David.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

But, all of the above taken into consideration, switching a mechanical relay at 3 Hz to flash the lights is still a poor design.

Look at using an IGFet to switch the load.

Also, look carefully at the load, (the lights), and see if you can split it into several, separate, loads. Each one with a smaller current, but switched simultaneously.

JC

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I would definitely agree with Doc. 3Hz is unwise for any mechanical relay.

Switching a pre-heating resistor is simple. You just start this prior to your main switching.

5A or 10A main load is fairly easy to switch with FET or even a junction transistor.

Mind you, Brutte's gyrator sounds interesting. I suspect that it needs a bit of experimentation first.

David.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Let me get this straight:

-- space is important
-- power draw is important
-- cost is important

12V, 50/60W -- some sort of automotive bulb?
https://www.divinelighting.com/e...

That says 940 lumens.

For a corresponding draw halogen, I can't find lumens.

Automotive bulbs are "competitively" priced because of the volume. Say $2 to $5?

=> How many amps of LED drive do you need to get the same lumens? Would that be more power efficient? Probably more costly for the LEDs? Does it help with the inrush situation?

You can put lipstick on a pig, but it is still a pig.

I've never met a pig I didn't like, as long as you have some salt and pepper.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Automotive turn signal flasher?
http://static.ddmcdn.com/flash/t...

You can put lipstick on a pig, but it is still a pig.

I've never met a pig I didn't like, as long as you have some salt and pepper.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

In another current thread:

Quote:
The lights are 30W LED spotlights (3500 Lumens each).

If correct and lumens are lumens then it would appear that LED lighting would be more efficient?

You can put lipstick on a pig, but it is still a pig.

I've never met a pig I didn't like, as long as you have some salt and pepper.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

scottm wrote:
For a 12 VDC incandescent lamp total load of about 4.8A at steady state, there's an initial cold filament turn-on inrush current draw of maybe 50A until the filaments heat. Lasting for maybe 40ms, with a sharp decay..

If using the relay is unavoidable, then try:
Digikey: Inrush Current Limiters

You'd want one with at least 2.5 Ohms @ 25C and 0.1 Ohms or less at steady state current.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I have to agree with the comments about using a FET -- although the right one may be too spendy for you.

The FET channel is basically a resistor with the value determined by the gate voltage.

So, if you ramp up the gate voltage fairly slowly, you'll reduce that initial inrush. Also, you won't need the relay (unless it's an external part).

Issues would be channel resistance when it's fully on, and power dissipation during the ramp up. Power dissipation might be a deal breaker.

But, you could probably just use an RC circuit to ramp the gate voltage up, if everything else works out.

Good Luck,

hj

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

If you ramp the gate then the device has to operate in the linear region. Hence the concern about heat dissipation as HJ mentioned.

I would think PWM'ing the output to warm up the filament would be a "better" approach.

JC

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
I would think PWM'ing the output to warm up the filament would be a "better" approach.

I agree with JC. As you said, you will have a 3 Hz switching frequency, so a transistor is a better option. But on this option you will need to spend a little time on it.

Regards,

Bruno Muswieck