OK, the transformer was there to provide isolation, so the input was actually balanced (ish) all along. So what you have done is to connect a source with one impedance and signal level to an input with a different impedance and expecting a different signal level via a cable which probably matches neither of them. This is meant to be an upgrade?
Lampizator has his own private view of the universe, and how electronics works.
Lampizator has his own private view of the universe, and how electronics works.
Do you have any specific recommendation on how I should connect sound card with the converter in balanced mode?
If you have an AES/EBU output then this expects a 110ohm balanced cable with a 110ohm termination. You need to buy or make the correct cable, and change the termination resistor from 75 to 110. Or leave as it is and accept that you have been lucky due to the robust nature of digital audio. Just don't claim that doing it wrong is somehow better than doing it right.
It is going to on a wrong and expensive direction generate by a incurable wrong understanding of things.
Ears tend to favour what your mind wants it to favour. Since you were expecting a change you found one, and the power of suggestion is so strong that it will overrule objective assessment.
Besides, there is no correlation between objective measurement and subjective perception - with most things in life, not just audio.
With that out of the way, would there be a difference in the two ways you set it up? Yes there would and it's not due to jitter. SPDIF is bits, but it's also a 3MHz radio wave. At those kind of frequencies the length and impedance of the line matters quite a bit and you may be changing the signal seen by the receiver quite significantly.
I can't say that either your initial method of converting to 75 ohm (which you haven't mentioned specifics of) or running a 110 ohm connection into a 75 ohm terminated load are optimal - and I can't ascertain which is worse. Definitely listening to it is not a good measurement, you will need to take a scope to the I2S lines and see what is happening. I can tell you it will be mangled.
Anyway, to do what you want to you will need to unbalance the line somewhere and terminate it properly. There's plenty of reading material online for you to look at, for example this:
Interfacing AES3 and S/PDIF
For a minute set the circuitry aside, and examine the differences between the two types of signal and termination. Since they are so vastly different, it would be naive to expect signal integrity at receiver if you connect one to the other directly. A suitable transformer would be the minimum recommended method, and would provide the full benefits of the balanced output from your (admittedly excellent) soundcard.
A few years ago I used exactly the resource that you mentioned (rane.com) for information on how to convert AES into SPDIF. Someone told me to purchase Lundahl transformer LL1574 to convert 110 Oms to 75 Ohms and I did it and used it. I connected this transformer to the output of the soundcard then resistor divider 220 and 100 Ohm just according to the schematics in the datasheet for the transformer. Please find it attached. Later I read the datasheet for SN75LBC176D (in the converter) and I found out that this chip can easily cope with AES/EBU voltage. Hence I removed the resistors and the transformer. Do you think I should not remove the transformer?
By the way I did comparison listening tests with and without transformer and resistors a few times in the past and I could not hear any difference in the sound. But I heard the difference for the better only when I changed spdif cabe to a twisted pair. This really made a difference to me.
That would be the correct transformer and attenuation for this application.
Twisted pair is not the correct impedance and will be causing some change in the received signal.
You may be able to hear a difference and prefer it, yet is is suboptimal in application. There may be some added distortion due to the instability of lock.
Digital audio, as said, is fairly resilient and can withstand quite a bit of foolishness without going to pot. But it can be audibly degraded by practices such as this, which would classify in the realm of voodoo. Listening and liking is one thing, but best practices are totally another.
You are free do select which you would like to follow. At the end of the day, it's only you listening. But please don't pass it off as advice or learned insight, because it is neither.
Twisted pair is not the correct impedance and will be causing some change in the received signal.
You may be able to hear a difference and prefer it, yet is is suboptimal in application. There may be some added distortion due to the instability of lock.
Digital audio, as said, is fairly resilient and can withstand quite a bit of foolishness without going to pot. But it can be audibly degraded by practices such as this, which would classify in the realm of voodoo. Listening and liking is one thing, but best practices are totally another.
You are free do select which you would like to follow. At the end of the day, it's only you listening. But please don't pass it off as advice or learned insight, because it is neither.
Can I ask you a few questions? Why did the designer of a sound card use AES/EBU instead of SPDIF? Can the SN75 LBC176D chip in the converter understand balanced connection? Then why on earth is it necessary to convert AES/EBU to SPDIF? Why not connect them in balanced mode? Why convert if we can go without it?
Regarding proper termination please note that 75 Ohms resistor is at the secondary winding, not the primary winding. In view of this is it really necessary to convert the impedance of the line to 75 Ohms by inserting a Lundahl transformer?
Take some time and read about signal integrity and eye diagram test.
Q: at what sample rate you made the tests?
Q: at what sample rate you made the tests?
The receiver chip may cope with both EBU and SPDIF signal levels. What you have wrong is that you are feeding an SPDIF impedance (75R) with a cable of unknown impedance fed from an EBU 110R source. This cannot be better than a correct connection. If you can genuinely hear a difference then this difference must be a degradation.
Adding a transformer does not convert the impedance of the line. This remains unknown, and so probably creating signal reflections. You really ought to do some learning about transmission lines before sharing your 'findings' and misleading others.
Adding a transformer does not convert the impedance of the line. This remains unknown, and so probably creating signal reflections. You really ought to do some learning about transmission lines before sharing your 'findings' and misleading others.
Seeing as part of your question was already answered by the two gentlemen above me, I'll take a stab at the other bit:
The fact is that your 'soundcard' is not that at all, but a piece of professional audio equipment that is designed to interface with something like an ADAT outboard unit or a Forsell converter (or two) in a studio environment and get as many channels into the computer as possible. It is an excellent way to use the highest level of equipment in the recording/monitoring chain.
These kind of things have very specific design needs and criteria, and don't care much for consumer needs. That is why they did not have S/PDIF in the first place.
The fact is that your 'soundcard' is not that at all, but a piece of professional audio equipment that is designed to interface with something like an ADAT outboard unit or a Forsell converter (or two) in a studio environment and get as many channels into the computer as possible. It is an excellent way to use the highest level of equipment in the recording/monitoring chain.
These kind of things have very specific design needs and criteria, and don't care much for consumer needs. That is why they did not have S/PDIF in the first place.
Can you give me any links to read about it? I mostly listen to 44.1kHz material only.
The recording industry favors 110ohm balanced AES. The broadcast industry favors 75ohm AES-3 which is very similar to SPDIF, in fact directly compatible in most cases. Both interface types are widely used and anyone in the business has to deal with both.
You convert between the two with simple balun transformers.
Canare Corp.: Impedance Transformers: Adapter Type(BCJ-XJ-TRC, BCJ-XP-TRC, BCJ-XJ-A10TRC)
Last edited:
- Status
- Not open for further replies.