They are relevant, it is easy to measure and see (and hear) differences with SPDIF using mediocre transformers etc..
The relevance depends on which SPDIF receiver is used. Some are better at handling poor signals with jitter than others.
This is often overlooked on commercial DAC designs that I've seen. Its quite normal to see common-mode currents routed through sensitive audio circuits.
An electrostatic screen on the transformer doesn't reduce the common-mode currents, just provides the option to divert them elsewhere. Which may well not be useful in practice. How does a planar transformer help here - do please elucidate?
I was going to take issue with you on a different subject, but having found this, I don't think I'll bother.
None of you guys in this thread can be taken seriously if you just let stuff like this go by.
Take issue on any subject with me that you want but I don't have any WMD's. Please note that down.
Classic musical instruments are analog, violins are not made of wood without a reason. Some analog electronic devices like tube amps do sound better in wooden cases for whatever reason, not my cup of tea but for many tube guys a known issue. Even if it is non science and maybe thin ice to think/believe/experience wood sounds better it was demonstrated to me a few times and I was surprised.
I personally never use wood. I do use metal for known properties like shielding that align better with electronic stuff IMHO. But I am not the person that thinks some one elses truth is less true than my truth (certainly not with audio). Live and let live.
Classic musical instruments are analog, violins are not made of wood without a reason. Some analog electronic devices like tube amps do sound better in wooden cases for whatever reason, not my cup of tea but for many tube guys a known issue. Even if it is non science and maybe thin ice to think/believe/experience wood sounds better it was demonstrated to me a few times and I was surprised.
I personally never use wood. I do use metal for known properties like shielding that align better with electronic stuff IMHO. But I am not the person that thinks some one elses truth is less true than my truth (certainly not with audio). Live and let live.
Last edited:
You can approach audio technology (a result of applied science and engineering) either based on evidence, or based on faith.
Evidence can be discussed and debated, but trying to debate or challenge faith is rather pointless.
That is patently untrue. Have you listened to any contemporary popular music?
Any evidence to support this opinion, or is it just a subjective opinion?
OK, let me state it this way. I am sure that not all differences in sound quality can be measured in the ways we now know. More than once I built identical devices that only had 2 different caps and measured equal and had fully identical waveforms but sounded different (please not that I don't write "better"). I can put my head in the ground and think this is scientifically not true but my ears don't lie. So I accept it and use the PCB with the perceived different but to my ears more agreeable sound and continue with other stuff. Since I can not agree on the subject myself from a technical point of view I always build 2 identical devices and do the trick again.
It is not in the least important so if some "take issue" on me like if we are personal rivals or something equally imbecile: there are way more serious troubles in the world to think about or even fight for.
It is not in the least important so if some "take issue" on me like if we are personal rivals or something equally imbecile: there are way more serious troubles in the world to think about or even fight for.
Last edited:
If you wanted to approach it from a standard engineering / applied science point of view (we are not talking fundamental science here), you would have to acknowledge that yes, your ears (or rather your ears and brain together) do lie pretty often, and verify that what you hear really is due to actual differences in the sound. The accepted and widely used method for that is double-blind ABX listening tests.
Absolutely. But challenging your statements on a factual basis should not be seen as "fighting" - it is part of a normal dialogue.
The problem is that internet audio forums are full of all kinds of audio voodoo, misinformation and conflicting advice. Thus anything not backed up by evidence has to be taken with a rather large grain of salt
There's no evidence for 'sound' in this meaningless mantra 'actual differences in the sound'. The brains/ears combination always lies in the sense they together present a percept (audible sound) from what they're given, which is vibrations. So perhaps you meant to say 'actual differences in the vibrations' ?
There is indeed the confusing dual meanings of the word "sound" as both the physical "a vibration that propagates as a typically audible mechanical wave of pressure and displacement" and the physiological/psychological "perception of such waves by the brain", but I was assuming that the context made it clear that I was using the first meaning.
I'll make a mental note to ask for clarification next time I encounter you using the same confusing word in future 🙂
Deja Vous
http://www.diyaudio.com/forums/digi...asuring-square-wave-spdif-cable-possible.html
Planar transformers can be useful little beasts, you can build them directly onto your PCB if you have enough layers. But they are very good for the ability to minimise inter-winding capacitance and thus reduce noise coupling. Common mode noise can also be combated by both correct layout practice and further filtering such as common mode chokes and other methods. Probably a bit extreme for DIY, but do-able (and fun), planar transformers and associated filtering is used on numerous bus based digital Manchester encoded signalling...though more often space conscious designs and cost no problem. Regular transformers will do just as well again with whatever filtering is required. DIY designs do seem to shy away from filtering and EMC, but if worried about noise from one system to another then it is worth learning and implementing.
Most buses (SPDIF is simple point to point, maybe going 1.5m tops) now use isolating transformers (often with a CM choke built in) Ethernet and CAN Bus are two examples.
http://www.diyaudio.com/forums/digi...asuring-square-wave-spdif-cable-possible.html
Planar transformers can be useful little beasts, you can build them directly onto your PCB if you have enough layers. But they are very good for the ability to minimise inter-winding capacitance and thus reduce noise coupling. Common mode noise can also be combated by both correct layout practice and further filtering such as common mode chokes and other methods. Probably a bit extreme for DIY, but do-able (and fun), planar transformers and associated filtering is used on numerous bus based digital Manchester encoded signalling...though more often space conscious designs and cost no problem. Regular transformers will do just as well again with whatever filtering is required. DIY designs do seem to shy away from filtering and EMC, but if worried about noise from one system to another then it is worth learning and implementing.
Most buses (SPDIF is simple point to point, maybe going 1.5m tops) now use isolating transformers (often with a CM choke built in) Ethernet and CAN Bus are two examples.
Why would a trafo built into a PCB have lower interwinding capacitance than one wound on a core? ISTM that a planar trafo would be likely to have more capacitance, not less as it'll have more common surface area between windings when the windings are flat. Also aren't you pretty much stuck with FR4 as the insulator? Wikipedia says FR4's epsilon is 4.8, seems rather high when you want to reduce interwinding capacitance. As a reference there are trafos by Scientific Conversions that achieve capacitances of the order of 1pF - see here:
Scientific Conversion, Inc. - Transformers and Inductors
Not that I use these as they're a bit on the dear side.
Scientific Conversion, Inc. - Transformers and Inductors
Not that I use these as they're a bit on the dear side.
You build in layers of electrostatic screening, you can also reduce the number of turns. As I said it not really DIY or should I say it would be at the more extreme end of DIY. The main advantage for commercial design is cost of assembly and better performance in high vibrational situations.
Just thought it would be fun to add to the mix. You can use the technique to build transformers as well:
https://www.google.co.uk/search?q=p...et%2Fenglish%2Fsayfa1.asp%3Fid%3D1842;593;546
Again this is advantageous if you have no access or the skills to build your own, some transformer winding gear I have seen is like a cross between some medieval torture equipment's and a coked up spider. Doing a few variations on a small PCB doesn't cost the earth.
Its not for everyone, but if you have a few layers on your board it costs nothing to build the transformer in and just stick on a core, instead of trying to solder a quite heavy termally challenging component.
A lot of FR4 used these days is slightly lower than 4.8, the higher Tg ones used for lead free processes are around 4.3 (not the greatest difference) and this figure is used widely when doing basic simulation and trace impedance calculations.
Just thought it would be fun to add to the mix. You can use the technique to build transformers as well:
https://www.google.co.uk/search?q=p...et%2Fenglish%2Fsayfa1.asp%3Fid%3D1842;593;546
Again this is advantageous if you have no access or the skills to build your own, some transformer winding gear I have seen is like a cross between some medieval torture equipment's and a coked up spider. Doing a few variations on a small PCB doesn't cost the earth.
Its not for everyone, but if you have a few layers on your board it costs nothing to build the transformer in and just stick on a core, instead of trying to solder a quite heavy termally challenging component.
A lot of FR4 used these days is slightly lower than 4.8, the higher Tg ones used for lead free processes are around 4.3 (not the greatest difference) and this figure is used widely when doing basic simulation and trace impedance calculations.
Last edited:
Oups,
Is the Murata 78601/4MC pulse transformer 100uH good enough ?
I planed to put it between the spidf input of one of my two Subbu V3 and the pcb-spidf plug before casing. So finally near the input and not at the streamer output (a SB Duet)...
Is the Murata 78601/4MC pulse transformer 100uH good enough ?
I planed to put it between the spidf input of one of my two Subbu V3 and the pcb-spidf plug before casing. So finally near the input and not at the streamer output (a SB Duet)...
No. Neither is 78601/3C. Bad LF transfer characteristic, you can clearly see LF tilt.
In general: avoid low interwinding capacitance/high leakage inductance transformers. Transformers with a screen are particullary bad, screen is acting as additional capacitor, lowering interwinding capacitance and increasing leakage indutance. Any other ratio as 1:1 is No, No. If you want Murata, go for DA101C (through hole) or MC (SMD).
Eldam - the problem is the primary inductance of 100uH is too low. From memory I seem to recall 1mH or more is required. PE-65612NL is spec'd at 2.5mH.
My view differs from stormsonic's one - I prefer to sacrifice leakage inductance in favour of low interwinding capacitance as I'm much more bugged by CM noise than I am by jitter. But that's because I only use multibit DACs which, compared to S-D DACs, are relatively immune to jitter. If you prefer S-D then the equation changes. I agree though to avoid screened trafos and anything not 1:1.
My view differs from stormsonic's one - I prefer to sacrifice leakage inductance in favour of low interwinding capacitance as I'm much more bugged by CM noise than I am by jitter. But that's because I only use multibit DACs which, compared to S-D DACs, are relatively immune to jitter. If you prefer S-D then the equation changes. I agree though to avoid screened trafos and anything not 1:1.
Signal integrity and clean edges do matter for the SPDIF interface. It's true that the data is digital, but the clock is regenerated from the data stream. This is a known issue with SPDIF. The clock is actually data dependent, so there will be jitter intrinsic to the interface protocol itself. Improper connections and loading will make this worse. Light pipe is the best solution for this reason.
I remember seeing transformers in SPDIF designs, but I'm not sure exactly why they are used... is it to eliminate capacitors or impedance matching or what?
I remember seeing transformers in SPDIF designs, but I'm not sure exactly why they are used... is it to eliminate capacitors or impedance matching or what?
Agrees with Abraxalito, CM noise can be problem.
But, if S/PDIF connection is running in domestic enviroment with consumer S/PDIF voltage levels (1Vp-p unterminated and 0.5Vp-p terminated) and transformer is followed by differential S/PDIF buffer, then CM noise is less of a concern.
But, if S/PDIF connection is running in domestic enviroment with consumer S/PDIF voltage levels (1Vp-p unterminated and 0.5Vp-p terminated) and transformer is followed by differential S/PDIF buffer, then CM noise is less of a concern.
- Status
- Not open for further replies.