[TriEmbed] Question concerning a mechanical release

[Scott Hall]

I'm going to disagree with sing opto-isolators.  They are slow (non-squarewave
forming, but very rounded waves) and very noisy (interference with PWM signals,
and mis-triggers for other I/O).

I really like the solution provided in the pair of articles in Circuit Cellar
magazine, August & September.  Basically he uses a separate higher voltage,
higher current power supply suitable for driving motors, solenoids and any other
loads you might want.  Then use a 5V regulator which connected to the R-Pi GPIO
pins 2 & 4 through a 0.74A fuse to supply power to the R-Pi -- avoiding the
proven delicate miniUSB-B power connector.

Noting that the Gertboard uses level shifters, and you have to know which
direction the signal is going.  Instead the author uses a $1 dual-channel Si8421
isolator to buffer each pair of I/O lines.  Each side will operate with 2.7 to
5.5 V logic levels.  He also uses a bidirectional Si8602 isolator for pairs of
bidirectional I/O (like I2C).  And these chips are far faster than anything the
R-Pi can produce.  This will keep any back EMF or other spikes generated by the
I/O from working its way back through to the R-Pi I/O pins, and offer the
ability to drive regular signals while not hurting the R-Pi.  Note that the
outputs should have similar specs to those in an AVR chip (Arduino), depending
on which digital isolator you use.  That should be enough to drive a MOSFET or
Darlington pair directly.

I just noticed that Analog Devices makes 6-channel digital isolators for around
$2.  (ADuM_764x)

On 09/26/2013 10:41 AM, John Vaughters wrote:
>
> Also, on the power control discussion, one of the best circuit protection
> plans is to use opto-isolators. I did not see that mentioned. The basic
> concept is you only power an led witch turns on a photo transistor, which
> completely isolates your microprocessor from back EMF issues. This does not
> protect the other circuit from back EMF, so you still need to deal with it,
> but usually those components are not as sensitive. Here is the best Motor
> H-Bridge design I have found and it is entirely controlled through
> opto-isolators. 

-- 
Scott G. Hall
Raleigh, NC, USA
ScottGHall1 at GMail.Com

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.triembed.org/pipermail/triembed_triembed.org/attachments/20130927/caeef46a/attachment.htm>