SPDIF driver - 74HCU04 output impedance?

[Adrian]

Hi,

I'm looking at a simple capacitively coupled spdif driver using parallel 74hcu04s. In calculating the resistor values to get the correct ouput level and impedance is it worthwhile taking account of the 74hcu04 output imdedance - if so could someone suggest the value to use [can't find one on the relavent data sheets except a generic one from from Fairchild which suggests HC normal outputs are in the region of 40-50 ohms - http://www.fairchildsemi.com/an/AN/AN-393.pdf.]

Thanks.

 

[Jocko Homo]

40-50 ohms may be a bit high, but...............

Don't use parallel gates. Maybe if they are in separate packages, with ferrites to isolate them. Maybe.

But, yes, you do need to take into account the output Z of the chips. Something that a lot of alledged "high-end" designers never learned. Or if they did, neglected to take into account.

BTW.......I would measure it. A resistor and DVM will be enough to get you close.

Jocko

 

[Elso Kwak]

An opamp with feedback will have almost zero output impedance. Wait a minute Jocko, did I say f-e-e-d-b-a-c-k?

 

[Bricolo]

round 1!

 

[Adrian]

Thanks Jocko. Can you confirm what the problem with parallel gates is?

Adrian

 

[Elso Kwak]

right place?

round 1!

Bricolo, you are sure you are in the right place?
We don't have a gong here.

 

[Bricolo]

is it better now?

[IMGDEAD]https://www.lucinda.net/wbo/graphics/gong.jpg[/IMGDEAD]

 

[Adrian]

Meanwhile, back at the original question - any info on why not to use parallel gates would be appreciated.

 

[Jocko Homo]

I didn't see him mention feedback. Did you? And how much open-loop gain do you think that one has???? Enough to lower the Z to zero ohms??

Anyway........if you have 6 inverters in one chip drawing six times the current, then the di/dt should be 6 times higher. Only one ground pin on the IC.......all that current in phase through it.....................you fill in the missing pieces.

Jocko

 

[Guido Tent]

I didn't see him mention feedback. Did you? And how much open-loop gain do you think that one has???? Enough to lower the Z to zero ohms??

Anyway........if you have 6 inverters in one chip drawing six times the current, then the di/dt should be 6 times higher. Only one ground pin on the IC.......all that current in phase through it.....................you fill in the missing pieces.

Jocko

Hi

Jocko is right both in terms of Zout as well as on advising on paralleling inverters. Additional note: Zout is only defined at the transfer of 0 to 1, in all other states the output impedance is high.

cheers

 

[macboy]

I am still baffled as to why so many people continue to insist upon using logic gates to transmit or receive S/PDIF, which is fundamentally an analog signal (carrying digital information of course). Logic gates perform well at what they are designed for - driving 0 to +5 V signals into high impedances over short distances. They fall over backwards when asked to deal with S/PDIF. Sure, you can jump through hoops to try to make it work, but why? What we have is a high frequency signal that needs to be driven into a 75 ohm transmission line. This has been done thousands of times before in video transmisison. It is no fluke that S/PDIF has the same basic signal specs as consumer level composite video: 75 ohms, 0.7 Vrms, >6 MHz BW. An opamp based line driver circuit is an ideal solution to this problem, if designed and implemented properly. So you need to add another (-5 V) supply. Big whoop. Look at all the crap needed to make the logic gates "work".

 

[Elso Kwak]

I am still baffled as to why so many people continue to insist upon using logic gates to transmit or receive S/PDIF, which is fundamentally an analog signal (carrying digital information of course). Logic gates perform well at what they are designed for - driving 0 to +5 V signals into high impedances over short distances. They fall over backwards when asked to deal with S/PDIF. Sure, you can jump through hoops to try to make it work, but why? What we have is a high frequency signal that needs to be driven into a 75 ohm transmission line. This has been done thousands of times before in video transmisison. It is no fluke that S/PDIF has the same basic signal specs as consumer level composite video: 75 ohms, 0.7 Vrms, >6 MHz BW. An opamp based line driver circuit is an ideal solution to this problem, if designed and implemented properly. So you need to add another (-5 V) supply. Big whoop. Look at all the crap needed to make the logic gates "work".

Hi macboy, Well then you may like my interface...

 

[Jocko Homo]

I disagree. In a lot of cases, it is harder to add that extra -5V supply than it is to make an unbuffered '04 gate work. It really isn't that hard to get one to work close to what is needed. You just have to think as if you are making an "analog" signal.

Yes, it will work better if you use stuff intended for analog transmission. But there is lots of gear out there that is poorly designed, needs to be fixed, and only has a single +5 V supply. With no room or provision for anything else, including a decent +5 V supply.

Jocko

 

[Guido Tent]

Hi macboy, Well then you may like my interface...

Hi Elso,

Why do you leave those 2 gates in ? Get rid of them.....

It is my experience that additional RF transformers solve a lot of problems at the LF side of the interface.

I agree with Jocko's post: There is a vast difference between upgrading and designing new equipment. In the latter case we all would leave the SPDIF interface for what it is worth.

 

[macboy]

Hi macboy, Well then you may like my interface...

Nice.