[TriEmbed] Transistor as switch

[Pete Soper]

Shouldn't a saw be able to cut down a tree? Is there a difference 
between a saw that will cut a limb the size of a finger vs the base of a 
tree 12 feet in diameter? But also, what if it's a sweet gum that wants 
to grab the blade and kick the saw and its operator back into the next 
county vs a sawing operation involving balsa wood?

Simple DC motors are basically dead shorts while they're starting up. 
So, momentarily, the transistor has to be able to put enough current 
into the motor to get it started. Divide that by the gain ("Hfe" from 
the transistor data sheet) and you get the current that has to go into 
the front end of the transistor (to/from the base to emitter circuit, 
depending on NPN or PNP) to get the back end (to/from collector to 
emitter circuit) to conduct enough current. In the real world a "five 
volt" motor will often run at a lower voltage, and that corresponds to a 
lower starting current. The real requirements are all over the place, 
but a 2N3904 is a very, very wimpy transistor for many motors. But the 
other, important factor, is that as the motor's commutator (for a simple 
motor) is switching current from one set of windings to the next, the 
circuit delivering current is seeing high voltages spikes caused by the 
collapsing of the motor winding's magnetic fields (if I didn't word that 
right, please, somebody, jump in on this). The peak voltages that can be 
seen by the transistor in this situation can be fierce: way WAY WAY 
higher than what the transistor is rated to tolerate. This is why you 
see backwards rectifier circuits across the loads, as well as 
resistor/capacitor networks, varistors, and many other things designed 
to "snub" the back EMF (electromotive force) currents that come with 
inductive loads (imagine the coils of a motor being the prototypical 
inductance).

And relays are motors, as Carl said. And Robert G is right: MOSFETS 
(with internal or external protection diodes) were made for this 
application.

-Pete


On 06/15/2015 05:37 PM, Grawburg via TriEmbed wrote:
> Shouldn't a 2N3904 transistor be able to switch a 5VDC motor?
> My circuit is correct since I tested it first with an LED, then replaced the LED with a cheap little DC motor. (I did use a diode to protect the circuit).  I'm using the Pi for the necessary voltage on the transistor base.  Yes, I did check to see if the motor worked with 5V.
>
>
> Brian Grawburg
>
>
>
>
>
> _______________________________________________
> Triangle, NC Embedded Computing mailing list
> TriEmbed at triembed.org
> http://mail.triembed.org/mailman/listinfo/triembed_triembed.org
> TriEmbed web site: http://TriEmbed.org