Digital Audio transformers

[tiroth]

All electronics has some very inexpensive (4/$1) pulse transformers that appear to be usable for digital audio transmitter isolation. I have only done preliminary testing, but the transformers I recieved pass a signal of 13MHz, so should be usable for 96kHz transmitters. YMMV. This is much cheaper than the $9/each that Digikey charges, although there may be capacitive coupling that reduces the common-mode rejection to an unacceptable level for some applications.

http://www.allelectronics.com/cgi-bin/category.cgi?category=search&item=PXR-3&type=store

 

[HarryHaller]

Digital transformers

These look like the Scotts that DigiKey sells. Tell me the primary inductance and I can tell you. These are good transformers if they are what I think they are. I designed pulse transformers for digital audio and was mentored by John Marshall that designed the Schott parts. Glad to discuss drive circuits and apps if interested.

H.H.

 

[tiroth]

Sorry, I sort of lost track of this post. I'm afraid I am not setup to make inductance measurements.

I'm currently using the interface schematics off of the Crystal application notes; I'd be very interested if there were alternative designs--although I must warn you that this will be a very one-sided "discussion". ^_^

 

[yves]

I came across a website (sorry, I lost the link ( ) which mentions
that a simple spdif transformer can be made by bifilar winding
magnet wire around a 1 inch ferrite diam core.

Any thoughts on how good these are, or how to test them?
I am thinking of a 13MHz square wave, if these can be
passed then it should be good enough, no?

thanks in advance
Yves

 

[tiroth]

Yves,

This appears to be the design of the input transformer on my Technics AC300 decoder, although it is quite tiny. IMHO this is fine for SPDIF, as isolation/short-circuit protection are the goals. However, I would think capacitive coupling would be far too high for AES/EBU.

13MHz square wave was what I used for testing...seems reasonable to me. I think I am out of my league though with regard to "real" testing.

 

[Jocko Homo]

Crystal diagrams

I'm not fond of their diagrams, or input circuit, period. The guy who designed that has no RF experience.

Build a good interface circuit, transformer isolated will do, that has LOW REFLECTIONS. Can't be done with any circuit that uses hysterisis.

And then hook up to their RX chip.

Jocko.

 

[HarryHaller]

Transformers

The transformer actually belongs at the transport/source end and you know it Jocko. What you need at the DAC end is a nice differential video opamp with the ground resistively decoupled. Something to buffer the cable from the DAC input circuit and reject common mode noise on the cable. something like the AD8#) from Analog Devices......

H.H.

 

[Jocko Homo]

1.) You don't always have control of the TX end.

2.) It doesn't hurt to have one there. If you do it right.

 

[HarryHaller]

Transformers

1. Nonsense, put the Transformer in a small box between the source and the digital cable with a short pigtail cable to the source/transport.

2. Yes..... and I can count the number of people in the digital audio business that know how to do it on one hand (including you and I) and still have enough fingers left to give you the rude hand gesture we all know and love. How many articles have you read on neutralizing pulse transformers for resistive termination. I think I have seen one, and in an RF journal at that! Face it, we are faced with "75 ohm RCA plugs", 50 ohm BNCs on 75 ohm cables, and BNC connectors with six inches of twisted pair between them and the load resistor in some DACs. And people wonder way digital audio sounds bad often times.........
Turn the ship upside down Unk.

H.H.

 

[lohk]

Here is a suggestion for a DIY SPDIF transformer. This word perfect to get full ground isolation, I am not really sure if it is a genuine high-end solution:

You will need a 6.3mm ferrite ring core (like Siemens B64290-K37-X830)
and 0.4mm isolated wire.
You can solder this littler transformer between two ends of RG59B coaxial wire without a box.

You need six windings for each side like this:

screen----- --------center
6w core 6w
center----- --------screen


Absolute phase is not significant (biphase coding).

5 windings to 7 windings could make a AES/EBU transformer.

regards
Klaus

 

[Jocko Homo]

Hairy Holler:

Are you giving me the "Italian salute"? Or just accusing me of knowing what I'm doing. I dunno 'bout that........I went to "bootcamp" at the Football Institute of NoTechnology.

The problem I see with your approach is that you cannot expect your brilliant transformer thingie to behave exactly as you want when you can't control what the idiots on the ends have done. ANd you know they are idiots. Some DACs don't even bother terminating the input. Or put some horrid low-pass filter on it. And then there is that wretched transformer Philips puts on evertything they build. Great for EMI, but too bad it can't pass a square wave. (You need to pass the 7th harmonic to make it look good.)

But never mind that........our buddy is going to hook up something to that crudbucket Crystal receiver, and when it sounds like doo-doo, he is going to blame his surplus transformer. The only approach that works is to minimize reflections (transformer or not), and you can't do that with a nasty Schmitt trigger input. So why don't you give him a nice ACTIVE circuit that he can shove in front of it.

And transformer too, if you want.

Put it on a plate, son. You'll enjoy it better.

Jocko

 

[Sandor]

Here is a suggestion for a DIY SPDIF transformer. This word perfect to get full ground isolation, I am not really sure if it is a genuine high-end solution:

You will need a 6.3mm ferrite ring core (like Siemens B64290-K37-X830)
and 0.4mm isolated wire.
You can solder this littler transformer between two ends of RG59B coaxial wire without a box.

You need six windings for each side like this:

screen----- --------center
6w core 6w
center----- --------screen


Absolute phase is not significant (biphase coding).

5 windings to 7 windings could make a AES/EBU transformer.

regards
Klaus

could you, please, provide more information about the construction of the transformer? The schema is not clear to me. Do you use a bifilar winding tecnique?
Regards.
Paul