Re: Input protection

From: sleeper75se (sleeper75se_at_yahoo.se)
Date: 2002-01-11 01:53:41


--- In buildcheapeeg_at_yahoogroups.com, "Joerg Hansmann" <info_at_jhansmann.de> wrote:

Hello Joerg and Jim,

Regarding my optical receiver, let me explain how I was thinking (if
I can correctly retrace my steps). I did not do a complete analysis.
This was touch and go, pen and paper. It worked on the first try
though.

The datasheet for the photo diode said a typical light-current would
be around 18uA. The dark-current is around 10nA, so, being
conservative I chose a value around 9uA as my threshold.

I downloaded this datasheet from the reseller, and now it appears to
be an older revision. The latest datasheet from the manufacturer
states 35uA typical, but this is for infrared, not red light, so
maybe the low threshold was not such a bad idea.

BC847B (a SMT-part) has a minimum HFE of 200 =>
which means that 9uA * 200 = 1.8mA or more passes through the
collector-emitter of Q1 when the transmitter LED is on.

At this current, Q2 must be turned off, which means the voltage drop
in the pull-up resistor for Q1 should be at least 23.4V (A drop from
+12V to 0.6V above -12V). At 1.8mA current a 13K resistor would be
enough, but I chose a slightly larger one to be safe. Q2 is
essentially a switch that Q1 can turn on and off.

Saturating Q2 like this is not good in the face of the limitations in
current and relatively high capacitances in the circuit. But it
works. Looking at an oscilloscope, one can see that I'm fairly close
to the limit of what is possible, or perhaps even over it, because
the recovery time of Q2 not very good. The only thing that can charge
the base-emitter capacitance in Q2 is the 15K resistor sitting on Q1,
and that takes time. Fortunately, the receiver circuits in my PC are
pretty forgiving it seems. This circuit is excellent at, say 19200,
but it could be better at 115200 bits per second.

It's a hack that someone better versed in transistor design can
improve, but perhaps the simplest solution is to run the whole thing
on 5 volts. The reduced swing should improve the speed. A serial port-
powered DS275 (the transciever suggested by Doug a while back)
following this circuit could then do the last bit.

By the way, I added the diode so that current wouldn't pass in the
other direction if anyone programmed the power lines improperly, but
now I realize it probably is completely unnecessary, so reconsider
the design without it.

Latest datasheet for the photo-diode is here:
http://www.everlight.com/pdf/PD15-22C.PDF
And a datasheet for BC847 can be found here:
http://www.semiconductors.philips.com/acrobat/datasheets/BC846_847_3.p
df

I hope you guys can give me some ideas on how to improve it. :-)

Regards,

Andreas



This archive was generated by hypermail 2.1.4 : 2002-07-27 12:28:36 BST