Toslink input for CS8412?

Status
Not open for further replies.
Jump to Latest
#2
knucklehead23 said:
What is the best way to add a Toslink input to a CS8412 receiver? Simply connect the output of the TORX to RXN and RXP? Or do I still have to use the resistor/caps/transformer recommended in the Crystal data sheet?

Thanks in advance

Jan
Click to expand...

Jan

I'd suggest to go straight into the 8412. Does the TORX have a balanced output ?

I prefer to use balanced drive for the Crystal

regards
 
#4
No, and no......

Avoid TOSLINK at all costs. Find a way to add a coax digital out. Unless you don't care how it sounds.

Do not go into the RX chip directly.....use a balanced active circuit that does not use saturated logic of any kind.

Jocko
 
#6
fdegrove said:
Hi,

Shogun-san,



High time for you to show the horses' front...

Mayonara,😉
Click to expand...


Hi

I'd suggest to add a balanced output transormer after the Toslink receiver, add small series stopper resistors (100 ohm) from secundaries towards the 8412.

And decouple the receiver using series ferrite beads towards the seperate power supply

Jocko ?

regards
 
#7
knucklehead23 said:
What is the best way to add a Toslink input to a CS8412 receiver? Simply connect the output of the TORX to RXN and RXP? Or do I still have to use the resistor/caps/transformer recommended in the Crystal data sheet?
Click to expand...
Different views here, but why don't you start to connect as the datasheet? You can also check how I have done.

We have all different ears. Maybe this is sufficient for you?
 
#8
Toslink may very well work fine, as peranders says why don't you just connect it as the data sheet suggests?

It will work without the cap to ground and the cap to rxp i.e. direct connection to the receiver from rxp and directly to ground at rxn (I'd connect rxn to the ground pin at the toslink reciever).
You might want to add a small resistor 22 to 100 ohms in order to dampen potential overshot.

I've found that toslink receivers and transmitters are very sensitive to its power supply, make sure its lownoise and stable up to and above its range, ideally individually supplied.

Guido usually have very sound reasons for his suggestions but adding a transformer when using toslink escapes me.....? Guido?
 
#9
Guido usually have very sound reasons for his suggestions but adding a transformer when using toslink escapes me.....? Guido?
Click to expand...

Use RCA-SPDIF and a transformer or the balanced active circuit like Jocko advises, you won't regret it. This subject has been discussed several times. Please use the Search function on digital transformer.

Toshlink ( the correct name ) gives not as good results as RCA-SPDIF will give.
 
#11
A 8 said:

Guido usually have very sound reasons for his suggestions but adding a transformer when using toslink escapes me.....? Guido?
Click to expand...

Hi,

Using the balanced inputs, and driving them balanced as well, prevents spitting a nett amount of charge into the CS8412 (reducing data correlated jitter at the clock output pin 19).

In addition, the common mode RF generated by that same chip is not transfered to differential mode back into the SPDIF

regards
 
#12
Re: No, and no......

Jocko Homo said:
Avoid TOSLINK at all costs. Find a way to add a coax digital out. Unless you don't care how it sounds.

Do not go into the RX chip directly.....use a balanced active circuit that does not use saturated logic of any kind.

Jocko
Click to expand...

Jocko,

I have been thinking about your suggestion. What about a high
speed class A discrete dif pair degenerated to gain of 1 or 2
loaded with cascoded CCS for CMRR? Cap couple RXN/P to
collectors. I am inclined to run IP trans4mer before dif pair
for better isolation.
Any comments?

Cheers,

Terry
 
#13
Elso:

The '04/'86 works ok. I use it in places. I was refering to stuff like Schmitt triggers, which unfortunately already existsi nside that stupid Crystal chip. Putting something in line with it......an '04 will help.

All:

Why transformer?? May be trickier to implement, but will yield a balanced signal. And some CMRR, which may not be an issue if the stages are 3 cm apart.

Terry:

Yes, something like that will work. I use an arrangement like that on commercial products that I have designed.

Lots of problems with TOSLINK. It has slow rise/fal times that are not symetrical. There are optical interface issues that are too obtuse to discuss here. One being low bandwidth of the link.

Reflections occur in optical links, as well as copper ones. The distortion mechanisms that arise are obviously of a different breed, but they are handled in the same matter: impedance matching. Single mode fiber has other complex issues, such as mode hopping from excess light being reflected back to the source from links too short, i. e., under 1 km.

A TOSLINK won't work at those distances, but because it is a multi-mode link, it has its own set of waveform distortion problems.

A good coax, with excellent impedance matching, will outdo any fiber connection, unless a lot of care is taken. And coax can achieve isolation with an inexpensive transformer. Try the ones by Schott that Digi-key sells.

Jocko
 
#14
Jocko Homo said:

All:

Why transformer?? May be trickier to implement, but will yield a balanced signal. And some CMRR, which may not be an issue if the stages are 3 cm apart.

Terry:

Yes, something like that will work. I use an arrangement like that on commercial products that I have designed.

snip

A good coax, with excellent impedance matching, will outdo any fiber connection, unless a lot of care is taken. And coax can achieve isolation with an inexpensive transformer. Try the ones by Schott that Digi-key sells.

Jocko
Click to expand...

Thanks Jocko,

WRT dif pair IP, will give it a go, with & witout tr4mr. Have some Scientific
conversions on hand.

Cheers,

Terry
 
Status
Not open for further replies.
Top Bottom