I gave up using PWM for a digital to analog converter and bought a real=20
10-bit DAC.

On 11/16/2012 3:00 AM, David Meiklejohn wrote:
> As some of you will know, I write (and sell) PIC tutorials:
> www.gooligum.com.au/tutorials.html.
>
>  =20
>
> I've just started working on a PWM lesson, using the 16F684, and am think=
ing
> about what examples to include - and I'd really appreciate any
> thoughts/feedback!
>
>  =20
>
> Simple single-output PWM is easy enough.  To show the effect of changing
> duty cycle, dimming an LED works well.  And to show the effect of changin=
g
> period, using the PWM output to drive a piezo speaker is effective - I've
> written a little example that changes the frequency as a pot is adjusted.
>
>  =20
>
> Simple (no dead band) half-bridge output is also easy to demonstrate -
> driving a piezo push-pull fashion makes it obviously louder.
>
>  =20
>
> However - people want to know how to drive motors.  I've had requests for
> tutorials about driving RC servos, but I think I'd do that as a separate
> topic.  Just good old-fashioned brushed DC should do to start with.
>
>  =20
>
> Motors aren't quite as simple to drive as an LED though, and this is wher=
e
> I'm wondering what the best approach is.
>
>  =20
>
> For single-output, I could specify a MOSFET and flyback diode, or a motor
> driver such as an L293D.  Any thoughts?
>
>  =20
>
> For half-bridge motor control, I'm thinking that the L293 is the way to g=
o.
> The two complementary PWM outputs on the PIC connect to two inputs on the
> L293, with the motor between the L293's outputs.  It would go full speed
> forward with 100% duty cycle, stop at 50%, and full reverse at 0%.  That'=
s
> the theory, anyway.  I've never used a half-bridge PWM for motor control
> before.  That arrangement seems quite inefficient, because it's pushing
> current through the motor 100% of the time (50% one way, then 50% reverse=
d),
> even when its stationary.  That also seems to be true of the "half-bridge
> output driving a full-bridge circuit" shown in the 16F684 data sheet.  Am=
 I
> right?  But does it work well enough anyway?
>
> Also I'm not sure, if I go with the L293 arrangement, whether dead-band
> delay is relevant.  It certainly would be best to mention it in any
> tutorial, and I understand about shoot-through current and can explain wh=
y
> that's to be avoided, but if I'm not using MOSFETs to build a half-bridge=
,
> is there any practical way to explain the potential need for dead-band
> delay, that relates back to the L293 circuit?
>
>  =20
>
> As for full-bridge, I'm leaning toward using discrete MOSFETs, although i=
t
> would end up being a little cumbersome if breadboarded.  I'm assuming
> P-channel on the high side and N-channel on the low, to avoid the need fo=
r
> drivers - that for a low-power (small hobby DC motor) app, the 20 mA that
> the PIC pins can supply is adequate to switch logic-level MOSFETs quickly
> enough.  I'm also wondering whether, in a low-power application like that=
,
> the MOSFET's body diodes will be able to handle the flyback current, or i=
f
> external diodes are necessary.  My aim, since this is a tutorial on using
> PWM, is to keep the electronics as simple as possible, but no simpler.  B=
y
> the way, I'm assuming a single 5V supply for PIC and motor, but obviously
> with some decoupling around the motor.  Again, not designing for field
> reliability, just a simple lash-up, but I don't want the PIC resetting ev=
ery
> few seconds due to power supply noise, either.
>
>  =20
>
> Does any of this make sense?  Anyone have any better suggestions on
> demonstrating PWM in a simple way, that doesn't use a lot of hardware tha=
t
> students may not have (or buy cheaply)?
>
>  =20
>
>  =20
>
> Thanks in advance,
>
>  =20
>
> David Meiklejohn
>
> www.gooligum.com.au
>
>  =20
>
>  =20
>

--=20
John Ferrell W8CCW

That which can be destroyed by the truth should be.
P.C. HODGELL


--=20
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.