high-side power MOSFET switch controlled by TTL

I have a motor connected to ground. The switch needs to be in between the source voltage and the load (motor). The switch needs to be controlled by a TTL signal. Is this the circuit I need?

http://www.electronics-tutorials.ws/transistor/tran59.gif

Or do I need to replace that on/off switch with a transistor and control the transistor with the TTL signal as a switch. Any circuit drawings would be nice.

Thank you.

I use IRL520 MOSFETS that take a TTL Trigger.

Don't I need a P-Channel mosfet for high side switching though? What circuit would I use? I'm new to all this.

Maybe this will help "http://forum.allaboutcircuits.com/showthread.php?t=39162(http://forum.allaboutcircuits.com/showthread.php?t=39162)":http://forum.allaboutcircuits.com/showthread.php?t=39162

and "http://arduinotronics.blogspot.com/2012/03/mosfet-control.html(http://arduinotronics.blogspot.com/2012/03/mosfet-control.html)":http://arduinotronics.blogspot.com/2012/03/mosfet-control.html

Didn't really help. The motor I am trying to control is tied to ground. Nothing I can do about that. I can only change the source side. So I would need to use a circuit like he is using there with the p-channel. I don't really understand that circuit though and was wanting something simpler but he is using a transistor like I imagined I might have to do.

There's really not enough information here. The motor voltage is critical. I assume it's DC but it could be anywhere from 5 to 100 volts, or more.

If it was AC, which doesn't seem likely, then an opto-isolated triac such as the MOC3043 (3063?) could be driven with a 5v signal and drive a triac like a BTA12-600BW that would switch 120VACor 240VAC, so long as the motor wasn't really honkin big. This requires only a three external resistors.

For DC you could possibly turn on an n-mosfet connected to ground with a ttl signal, so long as it was a logic level mosfet. It would also have to be rated for the motor voltage. Then that could pull down the input to a p-ch mosfet to turn on the motor. To avoid too high a voltage from the supply to the p-mosfet gate I'd put a high value resistor (R1) say 100k,between the n-mosfet output and the p-mosfet gate with a zener diode, larger than the p-mosfet on threshold, say 5.6v for a logic level p-mosfet, in parallel with a 1k resistor (R2)Then when the n-mosfet was on it would pull the p-mosfet gate 5.6v lower than the motor voltage and R2 would drop the rest of the voltage. When the n-mosfet was off R1 would pull the p-mosfet gate up and shut the p-mosfet off. The only issue here is that when the n-mosfet is off R1 and R2 will pull the n-mosfet up to the motor voltage so the n-ch mosfet has to be rated for the motor voltage.

I'm assuming that you're not trying to switch the motor quickly, such as with a high frequency pwm. In that case you'd have to consider the mosfet response times, a whole other subject.

If your motor voltage is very high you might be able to use an opto-isolator. I'd have to have more information so I could make a more intelligent suggestion.

Thank you for your reply. This application is for several different motors and even some that aren't motors (constant 12V) but most are at or around 12V or less. So let's just go with 12V as the voltage. No pwm. Just simple on and off. The current rating is going to be 20A up to 40A. I hope this helps you better see what I am doing. If you have any more questions about my setup I would be happy to answer them. I'm new to this so I might not have answered all of your questions so don't hesitate to ask.

20-40Amp continuous or 20-40A surge? If 20-40A continuous then this needs to be designed to handle a much higher surge/stall current.

Since the is a simple on/off (rather than a motor speed control) then the attached circuit should work -- but selection of the MOSFET device is contingent on the first question (above).

The diode across the motor needs to be able to handle the highest possible back EMF current from the motor windings.

Also, your description is a bit confusing. You say the "switch" needs to be controlled by a TTL signal -- are you talking about a _relay_? If so, then why not just use a relay (assuming one can be found that has a coil voltage of 5 volts and can handle the current involved)?

I am actually switching the control from relays to semiconductors just as a learning experience. Relays are big bulky and loud and I am trying the alternative. One of my circuits is running 20A and one is 40A. Those are maximum ratings. The fuses built in never blow so those are absolute maximum ratings.

Fuse size isn't a good indicator of maximum load. They respond very slowly and will pass instantaneous loads many times their ratings. If you look up their time response you may be amazed. Semiconductor devices aren't so forgiving so you could have problems.

As an example incandescent lights commonly have 10 to 20 times the inrush for about 20 ms. Two 60W 12V bulb (headlights) draw 10A after they're warm but 100A to 200A initially, limited a bit by the resistance in the wiring. This won't bother fuses but can kill solid state driver circuits rated for much more than 20A. Motors vary more but can have a surprising amount of inrush. This is basically their stall current.

Hi Nick, it would help if you can provide a circuit diagram of your existing and proposed system as far as you have currently developed it. Include motor voltage, power and current ratings, and a diagram of the control system you have been using.

I'm willing to do this by trial and error. It's okay if I burn up a few mosfets. The motors I'm working with are fairly small, old and unmarked. I know nothing about them. The relays I have been using are only rated for 10A max and they have been working fine for almost 6 months now so I think a 40A mosfet will be fine and if not I will find out when it gets too hot and starts to smoke and I'll go from there. Remember this is not for a job or school project where I will get fired if I screw it up. This is a learning experience for me so it is not crucial that everything is perfect right away.

Thank you so much for the circuit. I think it is exactly what I need. What transistor do you recommend?

The way I'd pick a mosfet is to go to your favorite supplier (digikey, mouser, etc.) and look for the specs you need, but be careful. There are lots of "40A" mosfets, meaning that they'll handle 40A for an instant in the best of all possible worlds. It's like rating a 1/2A use for 40A as it will pass 40A for an instant. Use the on resistance and ohms law to figure the real current.

If you want to use a thru hole part (easier to prototype) you probably want a TO-220 package though they are getting harder to find. I'd assume you will get 1 watt of heat dissipation without a heatsink if there's some airflow. 1W at 40A is 25 mV. R = E/I so the mosfet resistance should be 25 mv / 40A = 0.625 mohms, pretty small.

You can add a heatsink so you can use a higher resistance mosfet. You may also want to look at the motors and see if you really need 40A. 40A at 12V is over 1/2 HP. I don't know what inrush you expect but you probably don't need 40A continuously so you can increase the on resistance a bit. However, if your on resistance is too high you will get reduced voltage when the motor starts, increasing the time the mosfet will be drawing lots of current.

A quick look at Mouser shows:

STP310N10F7 MOSFET N-Ch 100V 2.3 mOhm 180A

which has 4 times the on resistance I recommended above. Without a heatsink it should handle about 10A continuously, more with a decent heatsink.

Note that relays are much better about inrush, it takes quite a bit to weld the contacts together, at least right away, so a 10A relay can't be replaced with a 10A mosfet.

If you pick a mosfet yourself be careful not to confuse mohms with ohms. Sometimes the selection guide is wrong (check the datasheet) and if the search engine orders the parts by on resistance it may think that 0.1 ohms is way less than 10 mohms.

And for Q2 (the bipolar transistor), a PN2222A or 2N3904 or similar will do the trick.

sorry - I didn't reread the discussion nor did I remember it, so I forgot that you need a p-mosfet, not an n-mosfet. P types have lower current ratings to you'll have to search harder and spend a bit more money for one, but most of my arm waving still applies.

An alternative to a more expensive p-mosfet is to use an n-mosfet but pull the base up to a voltage higher than 12V. That's just a matter of connecting the pull up resistor in the above schematic to a higher voltage (about 5 to 12v higher) if you have one.

If not, you can make a higher voltage (it needs very little current) if you AC couple a charge pump circuit to +12V. You can use a processor output pulled up to 5v or a 555 timer to make a square wave. Connect this through a capacitor to the charge pump circuit. A charge pump is made of a few small capacitors and diodes. You can find circuits on the web.

I have used this to drive a large 25A continuous 120VDC solenoid with poorly filtered rectified AC so I know it works. I could find my old schematic and make you a drawing, but I'm lazy and you probably don't want it anyway. If you do, let me know.

Here's a P-channel MOSFET that might do the trick:

IXTP140P05T

It's $5.88 at Digi-key

And, here's the selection guide I found this on:

http://www.ixyspower.com/store/Family.aspx?i=4

the IXTP140P05T is 9 mohms from what I read and at 40A it has to dissipate over 14W, a bunch, possibly beyond a passive heatsink. Even at 10A it drops 3.6W which would require a decent heatsink and add to the cost. That's why I'd consider the bother of a charge pump so an n-mosfet could be used. Of course, if the actual steady state current is way below that (if the initial stall current is 40A the motors could even be 4A) then this will work fine,

There is also the possibility of paralleling a number of MOSFETS -- they present a linear resistance, so they parallel very well (fairly even distribution of power).

The problem with mostets in parallel is that they increase the current linearly so four 9 mohm p-mosfets aren't about equal to a 2.3 mhom n-mosfet - it gets expensive fast. I also noticed a caveat about paralleling mosfets on the web. If may have only had to do with switching time, though that can affect the instantaneous heat generation and cause failures. I don't remember much (in general) but I think the fix was fairly simple. If you're going to parallel mosfets I'd look for something on the net.

I've done it for slow switching cases like this, and it works just fine -- but it was for currents in the neighborhood of 10A, so untried for 40A or above.

And, you're right about the cost escalation.

Wow, maybe a relay is the smart choice after all.

Relays have their problems as well. The coils draw a bit of current and have inrush problems as do motors. You still need a pnp transistor to drive the relay just like the circuit driving the motor. The relay contacts are noisy.

I recommend a diode across the coil, in addition to one across the motor. The coil has inductive kick back like the motor, just not as bad. When driving inductive loads I use a diode and a boucherot cell in parallel with the diode (2 to 5 ohm resistor in series with a 1.0 mf axial ceramic cap). This has solved reset problems for me a couple of times so now I always add one.