Maker Pro
Maker Pro

spdif and toslink switch

J

JESSD

Jan 1, 1970
0
I'm designing a spdif/toslink (digital audio) switch and would like
some advice. The switch can convert from optical to coax, coax to
coax, optical to optical, or coax to optical automatically. I've got
a circuit that works for the output portion that can drive both the
optical and coax. The input portion is a little trickier.

This is what I have so far:

+--10K--+
10 nF | |
|| | |\ | |\
COAX/SPDIF|-----------+---||---100R--+--| O--+---+---| O---- (output
stage)
| || |/ | |/
75R 10K
| 74HCU04 |
--- ---

I want to add in the Optical audio receiver to the front end of this
circuit without affected the 75 ohm termination resistance needed for
the coax connection. The user would only connect either the coax or
optical to the input, never both.

The output of the optical TOSLINK receiver is TTL.

I thought of just connecting the optical output through a diode to the
coax input. My thinking is that the termination resistance would
remain the same since the diode should look like an open circuit one
way. See below:

Optical Output (TTL)
|
\/ Diode +--10K--+
-- 10 nF | |
| || | |\ | |\
COAX/SPDIF|-------+---+---||---100R--+--| O--+---+---| O---- (output
stage)
| || |/ | |/
75R 10K
| 74HCU04 |
--- ---

When the user plugs in the optical cable the diode would let the
signal through. When not connected the coax shouldn't notice the
extra circuitry hanging on since the diode would always be reverse
biased.

Any suggestions/comments or other options to consider are appreciated.
Thanks in advance.
 
M

Michael Schwingen

Jan 1, 1970
0
I want to add in the Optical audio receiver to the front end of this
circuit without affected the 75 ohm termination resistance needed for
the coax connection. The user would only connect either the coax or
optical to the input, never both.
The output of the optical TOSLINK receiver is TTL.

What is the idle output level of the receiver, ie. with no signal at the
input? Low or High?
I thought of just connecting the optical output through a diode to the
coax input. My thinking is that the termination resistance would
remain the same since the diode should look like an open circuit one
way. See below:

Hm - how about connecting the TTL optical input signal to the lower end of
the 75R terminating resistor?

+--10K--+
10 nF | |
|| | |\ | |\
COAX/SPDIF|-----------+---||---100R--+--| O--+---+---| O---- (output
stage)
| || |/ | |/
75R 10K
| 74HCU04 |
| ---
Optical Input(TTL) ---+


Depending on output impedance of the Toslink receiver and idle polarity, you
might have to insert an additional inverter or buffer (better CMOS than
TTL), and maybe lower the 75R value so that the sum of output impedance and
resistor is again 75R.
Any suggestions/comments or other options to consider are appreciated.

Remember to put isolation transformers on the coax outputs. These are easily
made using a small core for a few parts, or readily available for higher
volumes. For my home wiring (about 20m of S/PDIF running through CAT5
cabling), I got good results using 100BaseT ethernet transformers (ST6118T
or similar) which I had around.

cu
Michael
 
S

Stefan Heinzmann

Jan 1, 1970
0
JESSD said:
I'm designing a spdif/toslink (digital audio) switch and would like
some advice. The switch can convert from optical to coax, coax to
coax, optical to optical, or coax to optical automatically. I've got
a circuit that works for the output portion that can drive both the
optical and coax. The input portion is a little trickier.

This is what I have so far:

+--10K--+
10 nF | |
|| | |\ | |\
COAX/SPDIF|-----------+---||---100R--+--| O--+---+---| O---- (output
stage)
| || |/ | |/
75R 10K
| 74HCU04 |
--- ---

I want to add in the Optical audio receiver to the front end of this
circuit without affected the 75 ohm termination resistance needed for
the coax connection. The user would only connect either the coax or
optical to the input, never both.

The output of the optical TOSLINK receiver is TTL.

I thought of just connecting the optical output through a diode to the
coax input. My thinking is that the termination resistance would
remain the same since the diode should look like an open circuit one
way. See below:

Optical Output (TTL)
|
\/ Diode +--10K--+
-- 10 nF | |
| || | |\ | |\
COAX/SPDIF|-------+---+---||---100R--+--| O--+---+---| O---- (output
stage)
| || |/ | |/
75R 10K
| 74HCU04 |
--- ---

When the user plugs in the optical cable the diode would let the
signal through. When not connected the coax shouldn't notice the
extra circuitry hanging on since the diode would always be reverse
biased.

Any suggestions/comments or other options to consider are appreciated.
Thanks in advance.

If you use and AND gate or OR gate you can control the impedance of both
inputs independently. Choice of gate type is dependent on which logic
level you get from an idle input.

Cheers
Stefan
 
J

JESSD

Jan 1, 1970
0
What is the idle output level of the receiver, ie. with no signal at the
input? Low or High?

The output of the toslink receiver in idle mode is high, but I planned
on using an inverter to change that to low.
Hm - how about connecting the TTL optical input signal to the lower end of
the 75R terminating resistor?


That's something I didn't think about. I wonder what the output
impedance is of an inverter? I also worry that it won't be consistant
from inverter to inverter. I'll try this out though.
+--10K--+
10 nF | |
|| | |\ | |\
COAX/SPDIF|-----------+---||---100R--+--| O--+---+---| O---- (output
stage)
| || |/ | |/
75R 10K
| 74HCU04 |
| ---
Optical Input(TTL) ---+


Depending on output impedance of the Toslink receiver and idle polarity, you
might have to insert an additional inverter or buffer (better CMOS than
TTL), and maybe lower the 75R value so that the sum of output impedance and
resistor is again 75R.

Yes, that is a good idea. Why is CMOS better than TTL?

Remember to put isolation transformers on the coax outputs. These are easily
made using a small core for a few parts, or readily available for higher
volumes. For my home wiring (about 20m of S/PDIF running through CAT5
cabling), I got good results using 100BaseT ethernet transformers (ST6118T
or similar) which I had around.

I would but the cost of the transformers I have seen is a little too
high. I'll check out the ST6118T you mentioned. What is the
disadvantage of not using the transformer?
 
M

Michael Schwingen

Jan 1, 1970
0
That's something I didn't think about. I wonder what the output
impedance is of an inverter? I also worry that it won't be consistant
from inverter to inverter. I'll try this out though.

I would guess some 10 Ohms - but you should better measure the result.
However, with the 74HCU04 used as shown, you will need to control the exact
type of inverter anyway.
Yes, that is a good idea. Why is CMOS better than TTL?

Without looking it up, I would expect a better (lower) "L" output level.
I would but the cost of the transformers I have seen is a little too
high. I'll check out the ST6118T you mentioned. What is the
disadvantage of not using the transformer?

The possibility for ground loops.

I know Pico has ready-made S/PDIF transformers, but nearly any ring-shaped
core with a few windings of insulated wire (with the ratio matched to your
output stage) will usually work fine.

cu
Michael
 
J

John Woodgate

Jan 1, 1970
0
I read in sci.electronics.design that Michael Schwingen <michaels@heisen
berg.ccac.rwth-aachen.de> wrote (in <[email protected]
Aachen.DE>) about 'spdif and toslink switch', on Fri, 2 Jan 2004:
I would guess some 10 Ohms - but you should better measure the result.

Indeed. If you put 10 ohms on the input, you must measure -10 ohms at
the output. It's an inverter!
 
J

JESSD

Jan 1, 1970
0
Stefan Heinzmann said:
If you use and AND gate or OR gate you can control the impedance of both
inputs independently. Choice of gate type is dependent on which logic
level you get from an idle input.

I originally tried or'ing the output of the optical receiver and the
output of the inverter feedback network. The problem was the output
of the inverter feedback network started oscillating when there was no
coax input present. When I managed to get rid of the noise and the
oscillation, the level would sit at 2.5V. I wanted the output to be
low when no input was present so that I could properly or the signals
together.

Jessie
 
J

JESSD

Jan 1, 1970
0
I would guess some 10 Ohms - but you should better measure the result.
However, with the 74HCU04 used as shown, you will need to control the exact
type of inverter anyway.

I tried it out and it worked. I measured 60 ohms using the 74HC04
using only 1 inverter. If I used all 6 inverters on the 74HC04 in
parallel I could get that number down quite a bit I suspect.
Without looking it up, I would expect a better (lower) "L" output level.


I'll have to get a cmos chip and try it out to see.

Jessie
 
S

Stefan Heinzmann

Jan 1, 1970
0
JESSD said:
I originally tried or'ing the output of the optical receiver and the
output of the inverter feedback network. The problem was the output
of the inverter feedback network started oscillating when there was no
coax input present. When I managed to get rid of the noise and the
oscillation, the level would sit at 2.5V. I wanted the output to be
low when no input was present so that I could properly or the signals
together.

Making sure that an idle input stays on one side is crucial, of course.
This is quite easy when you use a differential receiver. Are you
committed to a 74HCU04 or are you free to use whatever chips you like?

Cheers
Stefan
 
J

Jessie Dickens

Jan 1, 1970
0
I can use whatever chips I would like. Is there something that might work
better?

Jessie
 
S

Stefan Heinzmann

Jan 1, 1970
0
Jessie said:
I can use whatever chips I would like. Is there something that might work
better?

With a differential receiver you can bias the positive and negative
input separately, thereby defining the idle polarity. It is also easy to
control the hysteresis by using a small amount of positive feedback.
This should kill oscillations.

You may use an RS422/RS485 receiver. These have some hysteresis built in
already. They also often have a defined output state when the inputs are
open, so you may save the biasing resistors. Such things are available
in small packages, if you want. See the MAX3280E for an example. Note
that RS422 receivers are also quite resilient against damage by ESD or
overvoltage.

Cheers
Stefan
 

Similar threads

D
Replies
2
Views
837
Dirk Bruere at NeoPax
D
D
Replies
1
Views
934
Archimedes' Lever
A
D
Replies
6
Views
5K
Winfield Hill
W
D
Replies
4
Views
2K
JosephKK
J
Top