Controlling High Current Loads without Relay

This is for experience. The goal is to control high current loads with a micro controller. Motors, large DC light bulbs, etc. Basically all with semiconductor devices (aka no relay). A relay is the obvious choice but I've already completed operating these loads using a micro controller and a relay. My goal is to do this without using a relay. It seems to be very tricky to do but I like challenges. I'm not getting much support on this project and I don't know why but I will continue on! :D

Here is what I have as a basic idea:
"Your text to link here...(Schematic)":http://i45.tinypic.com/3501lk1.png


I don't know how to size the cap or r2 but r1 will be replaced by whatever load. the SCR I am experimenting with now is a S6020L and the transistor I am experimenting with is a MJH6287. It's a PNP. I meant to order an NPN version but I got em mixed up. I want high true but pnp gives me low true. Not really a problem just more of a slight inconvenience to the way I normally think of things. Anyways. Is my though process on the right track here or am I trying to do something impossible?

Basically a positive pulse from the microcontroller will turn the circuit on and a different pulse from a different pin of the microcontroller will cause the transistor to close and force current the opposite direction of the circuit for a second causing the scr holding current to go low enough therefore turning it off.

You can't control DC current with an SCR. Let me explain. The SCR is a latch with 2 paths. The pulse can initiate the latch and the pulse can cause the current to go to zero through the A/K nodes. Although it can cause the current to go to zero, once the pulse is off, i.e., not driving current into the gate LOW. the latch will turn right back on and current will continue to flow.

An SCR was designed for AC circuits principally whereby the A/K voltage reverse and causes the total latch current to go to zero. In a DC application the only way to get the latch to turn off, is not with a gate pulse but it is by limiting the holding current to less than specified value.

Therfore in you case if you want the SCR not to conduct you have to hold the gate low actively. Aside from that the SCR will work.

It would be better if you drew the application circuit youintend to use. It doesn't have to be in finished detail with values text. What are you trying to do use AC and and control a motor or light bulb or are use trying to use DC and control an AC motor or light bulb.

That Transistor/SCR/CAP circuit is confusing in terms of your final solution. For DC application you don't need to use SCR and MOSFETs are better choices as switches also.

Thank you for your replies. I did intend on holding the gate actively low. That's not a problem. But a MOSFET seems to be a better solution in this case.

I am controlling DC with DC. A microcontroller (0 and 5V at 40mA max) used to control high current DC loads (12V at 1A to 30A+).

I will begin researching MOSFETS after chemistry class.

Thanks everyone so much for the replies.

BTW what is the difference between Id Tc and ID Ta?

Yes and he's doing that with the transistor and capacitor that momentarily shorts across the SCR to "limit the holding current". So, his circuit should work, but it's not practical.

And "holding the gate low actively" won't turn the SCR off (but I suspect that was a typo, and you meant to say, hold the Anode low).

No it isn't a typo. One error in the circuit, the cap which is charged + to - when the NPN is on, has no path to force the cathode of the SCR negative and cummutate it off. You need a diode across the NPN, anode to the emitter and cathode to collector. Now the current can flow through the SCR latched state back through the diode and to the negatively charged capacitor, that should turn if off since the latch has a relatively small charged compared to your capacitor.

"ID @ Ta=25°C" means the max Drain current at Ambient temperature 25 degrees C.

"ID @ TC=25°C" means the max Drain current with the Case (of the part) at 25 degrees C

Taken from the data sheet of one of the HexFETs that specifies Tc: "TC measured with thermal couple mounted to top (Drain) of part."

I beg to differ--the cap will discharge through the two 1K resistors while the transistor is off, priming it for when the transistor does go on, and thus will have [essentially] 0 volts across it after a few time-constants. When the transistor turns on, its Vce will go to around .1 volt and because the capacitor has zero volts across it, the voltage across the SCR's Anode-Chathode will momentarily go to .1 volt. This will cause most of the current to flow through the capacitor and transistor collector, thus unlatching the SCR.

To unlatch an SCR you have only to reduce the SCR's Anode-Cathode current below it's _holding current_. The cathode of the SCR doesn't need to go "negative" for this to happen.

Oddly, when I click on the link from the email I don't see your comment (at least I'm assuming it's your [Nick's] comment, because the email doesn't display the name of the commenter ) BUT, I will reply anyway, and risk looking like I'm talking to myself ;)

bq). Apparently, these things need to be either on or off because if they are in between, the magic smoke escapes quickly.

It's not that they (MOSFETs) need to be either on or off, it's that when they are "in between" (or otherwise known as being in their "active region"), they present a interim resistance to the current path and thus will tend to dissipate [far] more power.

You are attempting to use a MOSFET in a switching application (which is the most common use for these devices), so, yes, you want to drive them either all the way off, or all the way on.

bq(). I tested one by removing the wire from ground to 5V to turn the
motor on and off. Eventually it started smoking.

The gate of a FET has a very high impedance (it's basically one plate on a very small capacitor). As such, if left open (i.e. with no low impedance path to either a voltage high enough to turn it all the way on, or a voltage low enough to turn it all the way off), then it will "float" and be subject to electrostatic and electromagnetic fields that can easily drive it into it's active region -- thus the occurrence of smoke! And, thus, it's unwise to leave a FET, that is configured to drive large loads, with it's gate open. It is likely to self destruct, as you have witnessed.

Okay, so I’m using a NTD5867NL N-CHANNEL MOSFET. I’ve bough 25 and already burned up 3. I have no idea what’s going on. The motor I pulled used a 20A fuse so I assume it never pulled 20A or else the fuse would have broke. Is my part just not good enough? It got so hot it desoldered my connection to the source of the mosfet. This is after it started smoking. I even tried using a 15k pull-down resistor for the gate but it made no difference. No matter what, after a couple minutes of running the motor, the mosfet started to smoke and even if i grounded the gate pin directly it still would not turn off the motor after this point. What part should I get? I tested the part using an led and it works fine. It also works fine until that magic smoke comes out when using the motor. Maybe I need a better mosfet.

It looks like this part should handle the load OK (based on the Data Sheet), so, some questions:

1. Are you using proper Static Sensitive Device handling practices?
* http://en.wikipedia.org/wiki/Electrostatic-sensitive_device
* http://pic.dhe.ibm.com/infocenter/powersys/v3r1m5/index.jsp?topic=/iphak/static.htm
2. Do you have a rectifier across the motor (anode to the MOSFET drain and cathode to 12V)? A motor contains inductors that can generate high voltage spikes that will, likely, fry the MOSFET.
* ww1.microchip.com/downloads/en/AppNotes/00898a.pdf
3. From the data sheet I get a rough estimate of the rDS at Vgs of 5 Volt at around 45mOhm. At 18Amps that's I*I*R = P = 18*18*0.045 = 14.6 watts, which suggests the need for a heat sink (or a MOSFET that can take more current or a number of MOSFETs in parallel [such as the ones you have])
* www.altera.com/literature/an/archives/an185.pdf

bq). ...and even if i grounded the gate pin directly it still would not turn off the motor after this point

Then the MOSFET is probably fried.