@ crutschow - If you think a 150/200W MOSFET will handle that, then I think we're on the right track. There is a possibility that it will have to be pretty much in limit mode for a second or two during a finale punch of a show though... There may be several output switching actions inside of a 1-3 second window. But I can use a heatsink and afterwards the system would be turned off till the next show...

Should I consider a heat protection circuit as well? How complicated would that be?


Also - I looked up the fairchild fet you used in that diagram and its not available.. at least not through digikey. and with a SMD component, how would i Heatsink it?

 

@ crutschow - If you think a 150/200W MOSFET will handle that, then I think we're on the right track. There is a possibility that it will have to be pretty much in limit mode for a second or two during a finale punch of a show though... There may be several output switching actions inside of a 1-3 second window. But I can use a heatsink and afterwards the system would be turned off till the next show...

Should I consider a heat protection circuit as well? How complicated would that be?


Also - I looked up the fairchild fet you used in that diagram and its not available.. at least not through digikey. and with a SMD component, how would i Heatsink it?

I don't think you need a heat protection circuit as long as you use a heatsink and the pulses are low duty-cycle.

I just used that MOSFET because it was available in the LTspice simulation program I used. Look in a supplier catalog, such as Digikey, for p-channel MOSFETs with the power rating you need and a heat-sink tab (such as a TO-220 case).

 

Ok - I did up a quick n dirty block diagram for how I envision this working.
As you can see there are parallel output drawing from this limited output.

There are times when only one output FET will be on, but there are times when 2 or 3 might be on simultaneously. and this is to be used in a live, real time, show environment. So using a simple fuse/breaker isn't practical if there is a wiring error.

The system needs to clamp the output to 5A and automatically recover from a fault rapidly. hope this helps clear up what i'm aiming for.

If you want to use high side switches, logic level devices are unnecessary, because you will need level translators to drive them. You might as well use devices that require 10V gate drive, because you have more than that available.
I have been envisioning low side switches, where logic level N-channel devices would be useful.

 

Ron H - Thanks for catching that, I'm a goof. YES I am planning on low side switching. Obviously I'm not an electronics genius. I'm trying to learn as I go.

Thanks for the patience and the help guys. This is looking very promising to me!

and Ron H - I did a bit more searching and I DID actually find a resistor that's small enough and rated at 100W that might work. So I'm back to the toss-up... The Resistor would be operating at about 65-75 Watts (from what I calc'd) depending on the loading, but is always in the mix... Where the limiter circuit's MOSFET wouldn't be really introducing heat to the circuit unless the load is trying to draw more than the 5amps right? If I'm triggering multiple effects off of one output FET (which is quite common) then they'd be wired series so I'd theoretically have anywhere from 1-6 amps...

I've got some more research to do I think.

 

So, this resurrects my question: Could an accidental short go from the current limiter straight to ground, bypassing the low side MOSFET switch?

 

The project electronics are going to be completely encased inside an assembled enclosure, so the only chance for a user to generate a short circuit, is to accidentally miswire the "pyro terminal" leads which will be user accessible. Otherwise a physical fault (broken wire, burned up component, etc) inside the case is the only other way to generate a short circuit. At least that's the idea.

In theory though, if a wire were to break free inside the case and find its way to a negative board trace or component lead, then yes i suppose there could be a short circuit across the supply through the limiter circuit.


I am getting myself a bit confused when looking at MOSFET ratings and such in the datasheets as well... the Pd rating is calculated based on the current i plan to pull through it, and the internal resistance right? (Pd=I²R right?) The charts are all showing pretty low (like 5-12milliOhm) ratings when ON - so that means the thing is pretty much negligible wattage when ON, but as it starts to try to limit current, the internal resistance climbs and the wattage will increase?
So at Full Limit of 5A - it'd be simulating ~4.8Ohms and that is 120Watts... Am I understanding that correctly?
Will any MOSFET that is within the voltage, current, and wattage ratings be good? (Provided I get the correct P or N type dictated by the circuit of course, and consider duration/heat sinks/etc)

 

You need to understand thermal specs.
When you see a device that is rated a 250W, you naively think (no offense intended) that it should work fine at 120W.
If you read the conditions of that spec, you will see that this is with the case held at 25°C. This is not possible without an infinite heat sink, or one cooled by a powered heat sink (Peltier, or circulating refrigerant).
The first 4 pages of Altera's AN-185 has all the equations required to determine the required heat sink for a given application.
I have to go to an appointment. Come back with questions after you have looked over the app note.
Also, think about what the maximum ambient temperature of your current limiter (closed box, summer, etc.)

 

If you want to protect against a short for a 200ms pulse then you will likely need little on no heat sinking.

If you want to protect against a continuous short-circuit then you will need a very large heat sink unless you go with a foldback or hiccup (momentary) type limiter.

 

Ron H - no offense taken - I know I don't know a lot. Worst thing someone can do is think they know everything.

I tried following a video yesterday on how to calculate thermal dissipation and heatsink capabilities to help size a heat sink for the device, as I realize that they reduce capacity as the temperature increases. (lower Imax with temp increase) I'm was considering a 40C ambient inside the box as a likely MAX since my environment rarely sees over 30C and that's during the day, when in fact i'd be using the circuit at night.
Oversizing the device to much higher than the expected 120W seemed logical to help extend the time the device could function, with increasing temperature, before it reached a critical failure point and self destructed. The MOSFET I was looking at had an 11A max limit at 75C if i remember the graph correctly.

That pdf looks much more informative than the video/audio i found via digikey. Thanks - I'll look it over and try to get more educated questions going.

crutschow - I like to overbuild things. just a good idea, so I'd have a heat sink on it no matter what. even if its just some stuck to some aluminum flatbar i put together with a thermal paste.

 

Below is the simulation of a simple two transistor, two resistor current limit circuit which has a drop of less then 1V before it limits. The limit current is approximately 0.7V/R2. The P-MOSFET can be any device that has an ON resistance of less than 50mΩ, a current rating of >5A, and a 40V or greater Vds rating.

Edit: The MOSFET shouldn't need any heat sink if the pulses are no more than a couple hundred milliseconds. For longer than that, you will.

View attachment 45938

Hi Crutschow,
I have a few mins break to take another look at all this and I'm happy with the circuit you came up with. Since i'm doing all this myself, i'm not quite ready to tackle SMD components, so I'm looking at through-hole.
Will this work for the FET? It appears to meet your suggested requirements of <50mOhm Rds, and >40Vds. Its rated for 74A Continuous, which is way overkill but i've still got the big Watt rating thing stuck in my head... am I totally wrong to be over killing the wattage rating to help compensate for possible higher temps?

http://www.digikey.ca/product-detail/en/IRF4905PBF/IRF4905PBF-ND/812139

 

Hi Crutschow,
I have a few mins break to take another look at all this and I'm happy with the circuit you came up with. Since i'm doing all this myself, i'm not quite ready to tackle SMD components, so I'm looking at through-hole.
Will this work for the FET? It appears to meet your suggested requirements of <50mOhm Rds, and >40Vds. Its rated for 74A Continuous, which is way overkill but i've still got the big Watt rating thing stuck in my head... am I totally wrong to be over killing the wattage rating to help compensate for possible higher temps?

http://www.digikey.ca/product-detail/en/IRF4905PBF/IRF4905PBF-ND/812139

The power rating of a power MOSFET is largely determined by the package. Without a heatsink the power rating is limited by the free-air rating of the package which, for a TO-220 package, is perhaps a couple watts. You can add a small clip-on heatsink, which will raise the dissipation rating to several watts.

 

Sorry - should have mentioned that. I'm was looking at ordering a bolt on, bolt to board heat sink with a 20W@60C dissipation rating. (http://www.digikey.ca/scripts/DKSearch/dksus.dll?Detail&itemSeq=119652427&uq=634832264399407350) which my limited brain convinces me should keep this thing safe in the utmost of brutal conditions...

I read through that info sheet Ron-H posted up, but i didn't understand it completely and have to look through it again.

 

Sorry - should have mentioned that. I'm was looking at ordering a bolt on, bolt to board heat sink with a 20W@60C dissipation rating. (http://www.digikey.ca/scripts/DKSearch/dksus.dll?Detail&itemSeq=119652427&uq=634832264399407350) which my limited brain convinces me should keep this thing safe in the utmost of brutal conditions...

I read through that info sheet Ron-H posted up, but i didn't understand it completely and have to look through it again.

OK, that should be fine. Be sure you use thermal grease to mount the transistor on the heat sink and minimize the package-to-heatsink thermal resistance.

 

No prob - I have a small container of thermal paste already in my BOM ready to go.

I also adjusted the .13 Resistor to .15 partially cause of availability, and partially because i was originally aiming for ~4-5amps limit which I think should keep me closer to that (4.666A if I read you're comment correctly about .7/R)

Much thanks for the help everyone- I'll post back up once i get the parts and test it out!

 

My simulations show a limit of 4.9A with a 0.15Ω resistor. If you increase the resistor value from the source to ground (R1) to 10kΩ from 1KΩ, the simulated limit drops to 4.5A. Thus you can tweak the value of the limit current slightly by varying the value of R1.

 

There is one thing I forgot and that is the maximum Vgs, which is 20V on most MOSFETs. This would be exceeded with a 24V supply in my circuit as shown. To keep Vgs from exceeding 20V just add a resistor from the 24V supply to the MOSFET gate with a value about 3 to 4 times R1.

 

120W resistors are not cheap.

In a recent thread, somebody found a car headlight rated for 24 volts, 100 watts.