Driving an Optocoupler from 74LS Series Logic

I am working on a project for a friend of mine. This device will switch
between N different electronic instruments and send the output to a single
output. In thinking about the design for this, I was going to use relays,
but then found optocouplers to fit our needs better (no noise, spikes, pops,
and isolates from logic circuit). So from Silonex, I ordered the NSL-32SR3
(http://www.silonex.com/datasheets/specs/104058.htm), which offers low
closed resistance (60 ohms) and high open resistance (25M ohms). However,
to driver the device, I need 20mA at 2.5V.

I am using a J-K flip-flop (74ALS109A) to hold each channel's toggle state
(on/off). I have a 5 volt power supply for project.

The question is, how do I interface the Q output of the flip-flop to the
optocoupler? Can I drive it directly from the Q output pin?

Thanks in advance!

 

Better is

\Q-------k|a----r------+5v

John

 

Most TTL compatable output pins have better pull-down capability than
pull-up. Connect the opto from the 5 volt supply to the output pin and it
should work. A low on the FF output will select the opto.

Jim

 

I should add some specifics...

This optocoupler consists of an LED input optically coupled to a photocell.
The photocell resistance is high when the LED current is "off" and low
resistance when the LED current is "on". See
http://www.silonex.com/datasheets/specs/104058.htm

So the circuit I want looks like this:

need 20mA @ 2.5V here
+-------------+
|
| J-K |
V +-------+
| Q | --------- circuit
here --------------- ----------- | NSL |----------
| | what to do?
LED side | 32SR3| photocell side
| |
+-----------| |----------
+-------------+
| +-------+

GND
[other inputs/outputs on J-K left out]


The photocell side is where I'll hook in the input from the instruments into
the common output.

 

Hi,

I made such a device for a stereo store back in Urbanna years ago. I
used CD4051's. Put +5 for VDD (as usual) then put -10 volts on VEE.
Control the selection (1 or 8) with the 3 address lines. These chips
have only about 200ohms in the on channel.

Note: do not use +/- 5 volts for the supplies. These devices have a
small 'crossover' notch at the mid voltage point. Using the -10 for VEE
moves the notch out of the way for low signal levels.

 

ULN2003 or cousins. Only catch is it inverts so you use not Q.
gg

 

What is the worst case OC output terminal resistance for the NSL-32SR3
across all LS09's, 500 ppm/oC power supply variation, device
characteristics, 0-70oC operating temps, and OC LED aging degradation
over 100,000 hours operation? Just sketch an approach if you can't come
up with hard numbers- take your time, you have five minutes:

Please view in a fixed-width font such as
Courier.


5V
|
+---+--+
| |
| |
/ /
2.2K 120
/ /
\ \
| |
| c
| |/
LS109 Q >-----+----| 2N3904
|\
e
|
OC ---
LED \ / ~

 

Your circuit dumps all available TTL pullup current *plus* the current
through the 120 ohms, *plus* the 2.2k pullup current, all funneled
into the LED... 45 ma maybe, but not very well defined. That much
current will make it degrade in a lot less than 100K hours.

That's a lot of parts just to damage one optoisolator.

John

 

Nah- the LS/ALS fades to less than 0.1mA at Voh>3V, the OC LED is Vf
2.5V @ If=20mA, and the transistor saturates to Vce~0.1-0.2V. The
circuit has far better defined current than your direct drive. Better
luck at your next interview-

 

Hmm... they must be using a visible LED, perhaps a green one. That
would make sense, IIRC, CdS peaks at wavelengths close to where the
human eye does. Note that the 2.5V is a MAXIMUM figure, not a typical
one.

Best regards,
Spehro Pefhany

 

Yes- the CdS wavelength response is very similar to human vision-and
this accounts for its popularity in light detection applications. The
2.5V should be Typ Vf, the datasheet is obviously in error-using W for
Ohms-and spec'ing 4mA reverse current at 4V- where they mean uA @100oC.

 

LS parts aren't spec'd to pull up hard, but they are actually pretty
good, and get better at max Vcc and temperature. There's nothing
between Vcc and an output but a darlington fed by a pullup. And the
LED here has a *max* drop of 2.5 volts, with no min specified. If
things stack up, this circuit could indeed push a *lot* of base
current. The OP wants 20 mA but opto spec sheet says it can stand only
25.

If you really want to use a bipolar transistor, this is a lot
easier...

+5
|
r
|
|
led
|
|
e
ttl-----------b pnp
c
|
gnd


or, for one more part,



+5
|
r
|
|
led
|
|
c
ttl-----r-----b npn
e
|
gnd

which is less sensitive to diode Vf variations.


As far as the interview goes, I already own the company.

John

 

I read in sci.electronics.design that John Larkin <[email protected]
techTHISnologyPLEASE.com> wrote (in <[email protected]
4ax.com>) about 'Driving an Optocoupler from 74LS Series Logic', on Sat,
3 Jan 2004:

I just bought your bank. Now, about this interview.... (;-)

 

Don't give me any lip, young man, or we'll withdraw all our accounts.

John

 

I read in sci.electronics.design that John Larkin <[email protected]
techTHISnologyPLEASE.com> wrote (in <[email protected]
4ax.com>) about 'Driving an Optocoupler from 74LS Series Logic', on Sat,
3 Jan 2004:

That should increase our liquidity by a significant factor, although it
will impact our cash flow. Have we really conned^H^H^H^H^H^H negotiated
you into paying THAT much interest? (;-)

 

Sorry, I'm old fashioned. We never borrow money and pay for everything
in cash. We save money, and pray for high interest rates.

The problem with borrowing money is that you have to pay it back.

John

 

They must have re-designed the family because there used to be something
like 100 ohms in the collector of the pull-up. I checked the TI logic
family handbook for the LS/ALS family and the output characteristic
curve is a total fictional joke- LS couldn't pull a unit load above 3V-
and by their graph the stuff can drive 20 ohm lines to full logic

K- that's better- you can count on LS staying on the low side.

 

I read in sci.electronics.design that John Larkin <[email protected]
techTHISnologyPLEASE.com> wrote (in <[email protected]
4ax.com>) about 'Driving an Optocoupler from 74LS Series Logic', on Sat,
3 Jan 2004:

Quite agree. I run my firm on the same basis.