Digital pulse transformer : Lundahl, Scientific Conversion, Murata or Newava ???

Status
Not open for further replies.
Jump to Latest
#41
SY said:
If someone has controlled (ears-only) listening tests showing that, within reason, the details of the shape of SPDIF waveforms have an effect on the audio, I'd be interested in that!
Click to expand...

See #31.
It's been a while but I remember we were with three guys comparing different SPDIF transformers.
There were audible differences, no doubt. The best "sounding" transformer turned out to show the best signal on the scope (rise time - square wave).
The DAC in question had a CS8412 receiver; I also remember that the DAC was pretty immune for the quality of the CD drive compared with other DAC's.
 
#42
Same experiences over here. I must say that most of my experiments were in the CS8412 and the much worse YM3623 (or whatever it was called) era. I recall having good results using Lundahl LL1566, Minicircuits types etc. Some of the standard outputs of certain CD players had non optimal waveforms and making that better was clearly audible and visible on an oscilloscope. Being without the right equipment to understand what was going on only trying and putting time in it was possible. So swapping transformers was part of the modding hobby.

It never occurred to me at that time that you could do without transformers. I really thought they were absolutely necessary as every device had them. I had a laugh when Lukasz Fikusz (Lampizator) showed some oscilloscope views of transports and DACs connected directly. After reading the info in the following link I decided that one transformer in the chain was more than enough 😉

http://lampizator.eu/LAMPIZATOR/TRANSPORT/CD_transport_DIY.html

You know how the human mind can work. After bad experiences with several optical cables and transmitter/receivers I never bothered again using those. Who knows they can be better dan coaxial nowadays.
 
Last edited:
#48
Pulse transformers (one in source, the other one inside the DAC) and an interconnecting coax cable should be considered as one tuned circuit when it comes to S\PDIF transmission.

In other words, two Muratas can sound really good with interconnecting coax cable A, but if you change the interconnecting coax cable A LENGTH only, Newavas may sound better. Scientific Conversion ones may sound great with coax cable B….

Best, by far, is not to spend too much money on pulse transformers, but to use the CRO and tune either cable type / cable length, or try different pulse transformers AND different pulse transformers’ loading resistors…. The pulse transformer inside the source is easily overlooked; however it is as important as the one inside the DAC.

Boky
 
#49
I would like to see some waveforms both analogue and digital, so that any effects that the digital waveform may have on the analogue output can be defined, IMO this would be the only way to determine if the sahpe of the digital wave has any effect on the analogue output...
Sorry if I am being to objective for some, but outsuide of standard signal integrity issues with digital signal transmission I cannot see how the output could be affected so much, again I believe to analogue effects are being ascribed to digital data, which is going to be either 1 or 0.
As to the SPDIF interface it is not a tuned circuit as such, again this is RF terminology, you are trying to minimise impedance discontinoutities, by having the signal see one impedance value as it travels down the PCB trace, through a connector via the wire etc etc.
 
#50
That would be good to see...... haven't ever seen anyone post that exact link - there have been many posts though showing scope waveforms of various qualities.

I think many consider of spdif in terms of a series of 1s and 0s. But I've come to think of it as an analogue square wave and as such is prey to all the things that could affect any analogue signal. Things like rise time, ringing, slew rate of chips etc etc etc. I've spent quite a while thinking about rf attenuators and how they could possibly lead to the improvement in sound that I have observed in several set ups - I think that the SPDIF Tx we are talking about here can affect the signal in similar ways. These points above are all I can theorise about.

Sorry to mention rf-attenuators, as I think that the objectivist crowd will now descend here to rubbish my post (crisis in sector G - send a team immediately!).

Fran (who is not an EE - let fire!)
 
Last edited:
#52
An interesting discussion but it still leaves me puzzled.
If
Better transformers = better SPDIF square waves (we've seen some waveforms)
And
Better waveforms = lower jitter (listening tests)
Then
Why doesn't someone simply measure the IIS clock coming out of the receiver chip? Does a better waveform = less jitter out? Is there a threshold? That might start down the path to telling us something. MCLK, LRCLK, whatever.

The next step might be to verify what various amounts of jitter do to the analog outputs of different DAC chips. There are only a few receiver chips used in high end DACs, right? Anyone have a jitter testing rig?
 
#53
Pano said:
Better waveforms = lower jitter (listening tests)
Click to expand...

How do we know this is so? Not that better waveforms lead to lower jitter (that's uncontentious) but that we can actually hear the results of lower jitter isolated from effects of lower common-mode noise?

Useful listening tests would need to be for one variable only while others are controlled would they not? But until there's some kind of a measurement for CM noise, how can that be controlled? As far as I'm aware the only listening tests that have been carried out on jitter told us fairly convincingly its not audible at the levels typically found on SPDIF receivers and DACs. Admittedly this was (if I recall correctly) not tested with correlated jitter which may make a significant difference and is normally by far the dominant component of jitter with SPDIF.
 
#54
Yes of course, but I wanted to be methodical!

If there really isn't much difference in the clock jitter with various transfos and waveforms, would it even matter?
If there is, then the good, the bad and the ugly clocks could be identified and make a nice basis for a listening test.
 
#57
We like white papers here's a real good one for AES from SC http://www.scientificonversion.com/AES1998.pdf
IMO CM noise and grounding is a much bigger issue than eye balling the signal fidelity of the data stream esp for 75 ohm commercial single ended apps. see 110 ohm professional apps for better data interface solutions. (reason> all re-clocking receivers use a Schmidt trigger at the interface but common mode noise is injected directly on the recovered "clean" clock on unbalanced 75 ohm interfaces .
 
Last edited:
Status
Not open for further replies.
Top Bottom