Sorry to post this beaten to death question again. I did search here and in other forums, but just get more confused about choices and details of circuitry.
MOSFET, Power FET, Hi Side Switch, SSR....I'm a little overwhelmed by the choices and specifics.
The attachment shows what I am trying to do. Very simple on/off SPST switching, but 18 switches needed, 9 each for two different types of load.
Main reasons for looking at options to relays are size (very limited circuit space is available) and reliability. One load-type is a resistance heating wire that will quickly overheat if the control device gets stuck "ON". I read that relay contacts can weld shut.
The circuit will operate only a few times a year, with maybe a few dozen total cycles each time.
Can a simple circuit diagram be shown for a recommended device please? From reading, I'm thinking an N-Channel enhancement mode MOSFET is the way to go, but really not sure.
Thank you.

 

How much room is "very little"?

Logic level N channel MOSFETs would do. Depletion mode MOSFETs are rarely used, so enhancement mode is assumed.

In theory AO3400 would work, but I've never tried to operate one with a 4A drain current. They're surface mount devices. Are you able to work with them?

For inductive loads, you need to use a snubber diode.

Can a simple circuit diagram be shown for a recommended device please?

Will post a schematic shortly.

 

For the solenoids a good-old ULN2003/2803 or the MOFSET version. They will switch 500mA and have the diode built in.
For the heaters - how many are switched on at the same time? If it's just one or two, then the thermal inertial of the circuit board will deal with any heat from the MOSFETs, so you could use a very small footprint MOSFET. Even if all 9 switch on together for 5 seconds the temperature rise will be manageable with no heatsink.
The snubber diode is never a bad idea even for the heating elements - they may be coils, and have significant inductance.

 

I see a lot of references to relays in these electronics forums so wanted to ask this.

When I was studying electronics back in Liverpool in the early 80s, relays were of course common but do come with problems, one of which is lots of electrical noise and limited frequency of operation and mechanical wear; but they do offer great isolation.

Back then we did not have the choice of some of the more recent and impressive semiconductor alternatives so I'd expect to see less of relays today, except maybe for serious very high current loads, so why do we still see so much talk of them among hobbyists? surely they're pretty dated nowadays?

So what about one of these? they'll handle a DC load up to 10 Amps and they're opto-isolated too!

 

I still think there is need or advantage sometimes to using a relay, for example the Ice cube version typically come in four NO/NC contacts.
This means that triggering from a solid state device, four functions can be switched or registered with one output.
If a BEMF diode is fitted, noise is minimum, also the latching kind can be had that has a 'memory' on removal or loss of power.
Max.

 

So what about one of these? they'll handle a DC load up to 10 Amps and they're opto-isolated too!

If it's for DC use, why does it claim to have "zero-crossing detection"? Seems dubious to me.

 

I see a lot of references to relays in these electronics forums so wanted to ask this.

When I was studying electronics back in Liverpool in the early 80s, relays were of course common but do come with problems, one of which is lots of electrical noise and limited frequency of operation and mechanical wear; but they do offer great isolation.

Back then we did not have the choice of some of the more recent and impressive semiconductor alternatives so I'd expect to see less of relays today, except maybe for serious very high current loads, so why do we still see so much talk of them among hobbyists? surely they're pretty dated nowadays?

So what about one of these? they'll handle a DC load up to 10 Amps and they're opto-isolated too!

 

How much room is "very little"?

Logic level N channel MOSFETs would do. Depletion mode MOSFETs are rarely used, so enhancement mode is assumed.

In theory AO3400 would work, but I've never tried to operate one with a 4A drain current. They're surface mount devices. Are you able to work with them?

For inductive loads, you need to use a snubber diode.

Will post a schematic shortly.
View attachment 231046

Dennis, thank you. The schematic is just what I needed. Surface mount is out....70yrs old...the eyesight is a bit lacking

 

For the solenoids a good-old ULN2003/2803 or the MOFSET version. They will switch 500mA and have the diode built in.
For the heaters - how many are switched on at the same time? If it's just one or two, then the thermal inertial of the circuit board will deal with any heat from the MOSFETs, so you could use a very small footprint MOSFET. Even if all 9 switch on together for 5 seconds the temperature rise will be manageable with no heatsink.
The snubber diode is never a bad idea even for the heating elements - they may be coils, and have significant inductance.

Ian, I need to do more looking, but the circuits I see with the ULN2003 have it feeding relays...oh, just understood. For low current load, just switch the load directly. Got it I think. Thank you.

 

An adequately sized decent quality mechanical relay has a life of thousands of cycles. And that relay can survive a load voltage spike that will arc across the open contacts. My central air conditioning has two relays, one for the half HP blower motor and one for the 240 volt 1HP compressor. They have lasted 15 years so far.
A poor quality relay will not last for many cycles at it's rated load current.
And that spike protection diode hung across the coil will slow the relay release which leads to a lot more arc on the current carrying contacts.
Are these for starting a heater system of some kind? Those are very short on times, especially for the solenoid

 

For "Cannot get stuck In ON mode", what I use in the industrial control devices I build, is have one port pin being toggled each time through the loop. Do not use PWM!!!!
You want this signal run via the software loop, not built in hardware.
This "AC" signal is coupled via a capacitor to a rectifier, and the resultant DC is fed to a Schmitt trigger. Then, the output of the Schmitt trigger drives a FET or relay that turns enables all the outputs.
So, if the control loop hangs for some reason, the "DC" drops and the outputs are turned off, independent of the actual output port pins.
When the system starts up, there will be a short delay until the "DC" signal gets to operation levels, so take that into account.

 

Edge-triggered retriggerable monostables such as 74HC123 can be used in a similar fashion. It looks like a lot of hardware, but the first time is saves the output device from burning out it looks much more like value for money!

 

You can avoid using that shunt diode by choosing a higher voltage rated mosfet device, or even by slowing the turn off time to a couple of milliseconds.

 

If it's for DC use, why does it claim to have "zero-crossing detection"? Seems dubious to me.

Yes that crossed (no pun intended) my mind too, so many of the products are poorly described, likely due to translating Chinese.

 

Thank you for the amazingly fast responses. Much appreciated. After considering the above, I attached what I came up with for the heaters and for the solenoids. There are nine of each, so this plan is much smaller than using even small relays.
This circuit is for a ship model animation project...firing cannons. Three carefully timed functions produce a very reasonable cannon fire effect: heater heats wick to vaporize glycol solution (it's like a very big e-cigarette), solenoid valve releases air blast to result in smoke cloud and super bright LED flashes to simulate muzzle flash. See prototype system at

Questions please....
In Heater circuit: Is R value and location correct? Is internal diode sufficient for and in proper orientation as snubber?
In Solenoid circuit: Is device selected OK? Are internal diodes sufficient and in proper orientation?
Thank you again.

 

ULN2003: Pin 9 connects to V+ to complete the flyback diode circuits.
MOSFETS: 10k will work - so will any other value from 1k to 100k.
You could easily use a dual MOSFET in a SO8 package to save space, considering the low duty cycle.

 

Ian0 - corrected ULN2003 pin 9 shown. Correct?

 

Ian0 - corrected ULN2003 pin 9 shown. Correct?View attachment 231140

Yes.

 

Yes.

THANK YOU. Can't wait to try this.
Will post results.

 

I think that the ULN2003 was designed by Sprague as a solenoid driver for dot-matrix printers (remember them?). That would explain why it has 7 outputs as one would normally expect 6 or 8, because it was originally a 7x5 dot matrix.

How do you produce the deafening noise of canon fire?