Switching inductive load

[martin_t]

Hi all.
I have a question. I have to switch a contactor/relay. It has to remain open for fifty secs every two minutes or so.
The coil resistance is 58 ohm and it takes 0.50 amps aprox (36Vdc) , and is connected to ground.
Some people suggested using relays, but I wanted to do it all solid state.
The signal comes from a flip-flop (15Vdc) and goes into the N-Mosfet (actually it's an IRF830).
Should the zener be 36V?
This is what I came up with. Any suggestion is welcome.



Thanks very much in advance.

Martin

 

[srizbf]

is there any difficulty in connecting the relay(contactor) to drain of n-channel mosfet ?

 

[martin_t]

Yes. It's a few meters away and it's hardwired.
Besides, I'm not allowed to change the wire configuration.
So no, I only have one terminal to control the relay. The other side is put to ground.

Martin.

 

[D.A.(Tony)Stewart]

you will need twisted pair drive between 15V CMOS gate and MOSFET and perhaps some filters for EMC ( both ingress and egress for current spikes). All switched current needs to be either shielded and/or in twisted pairs.

If the contactor coil current doesn't induce control currents in the gate, then the switched currents will likely interfere.
Control cables must be at 90 right angles to high currents being switched.

The clamp diode should be across the contactor coil terminals to reduce the area of the loop ( stray pulsed H fields noise) and not on the FET Source, although both ends is ok too. This loop will resonate (LC) as current declines from 0.7V clamp/coil loop.

As you Contactor is 36V/0.5Adc, it's Rs or ESR is 72 Ohms. RdsOn should be <1% of this while Rgate 1~ 10x depending on noise immunity.

Grounds may need to be improved with flat braid strap to reduce ground noise.

 

[KlausST]

Hi,

I'd change some things.
R18 should be 10k also. The 1st dissipates/wastes too much power.
The IRF540 with 33A rating is an overkill. FDN304 or BSS138 will do.
The zener should be about 10k type.

******
Instead of using a lot of discrete parts i'd use an intelligent high side switch. The benefit is that it is short circuit proof.
BTS432 is one that may fit your needs.
Manufacturers have interactive selection guides to find the best for your application.

Klaus

 

[chuckey]

I think C14 is in the wrong place, I would be happier if it was across D12.
Frank

 

[FvM]

I think C14 is in the wrong place, I would be happier if it was across D12.

C14 will greatly increase output rise- and fall times and may be wanted in some cases to reduce switch generated EMI. Smart power switches have typically similar smooth edges. Obviously also increase switching losses, which is meaningless for the intended switching frequency. Generally I would omit it.

A capacitor parallel to the load will cause additional inrush current and is usually unwanted.

 

[KlausST]

Some corrections to my post#5

The 1st dissipates/wastes too much power.

The 1k dissipates/wastes too much power.

The zener should be about 10k type.

The zener should be about 10V type.

Klaus

 

[martin_t]

Hi all. Thanks very much for your input.

SunnySkyguy: no twisted pair needed. The flip-flop, mosfets, everything except the contactor is on the PCB.

KlausST, I´d like to use a smart switch too, but I was asked to use generic parts. The fact of being more discrete is intentional.
Regarding the zener, do you think a BZX55C10 would be ok?
The IRF540 is in fact an IRF830. Still overkill, but...

FvM, as you said, since losses are not a concern here, C14 is there to slow down the turn on a little bit.

If you see anything else that could be improved, feel free...
Thanks very y much.

Martin