This next paragraph sums up the problem pretty accurately.
the next solution assumes the Verror is in the power cord. If Verror was due to the power cord impedances, then it may provide a small improvement.
But Verror is not due to the power cord. It is explained by D.Joffe and if you believe him, then Verror is in the source to receiver interconnects. The solution below is attempting to sort the wrong problem (see fig3 in the pdf attachment).
Tom now jumps to the far better solution: Adopt balanced impedance interconnections. All the pros use it. It works and works very well.
But that balanced impedance solution leaves all the unbalanced interconnects without any effective guidance to attenuate the interference.
The real solution for all the unbalanced systems is to attenuate the Verror using the guidance given by D.Joffe
See fig5 showing the two added resistors HBRR & HBRL
These two resistors in a two channel amplifier attenuate the interference (loop current) that generates Verror.
Last edited:
I think that George Soong is working hard to improve audio in the way he deems best, and I am sure he spends much of his time and other resources to accomplish the goals he set for himself. This is not to say that I am thinking along the same lines, or that I even agree on the goals, but that is only fair enough.
The sort of technical discussion I like most, is where these differences in philosophy are brought the the fore, with arguments flying in all directions. Like chicken fleeing from the ax.
Disagreement is part of that process. As soon as people put a personal dimension into it, I often cannot resist the temptation to respond in kind. Which I did just now, against my own will and in a moment of weakness, for which I apologize.
The sort of technical discussion I like most, is where these differences in philosophy are brought the the fore, with arguments flying in all directions. Like chicken fleeing from the ax.
Disagreement is part of that process. As soon as people put a personal dimension into it, I often cannot resist the temptation to respond in kind. Which I did just now, against my own will and in a moment of weakness, for which I apologize.
Last edited:
I think my blog is not to reveal technology, but rather to just make aware the effort and direction we wish to accomplish. In the process, there are people more specialised in different areas where hopefully I can somehow bring into the overall process when goals and experience complement. Basically this initiative started because I believe if the proper technical aspects are addressed, total cost of high fidelity audio systems can be affordable. Making money really takes a back seat. Tom's designs are quite reasonable if you consider the quantities and efforts he puts into his products, and I appreciate his work very much. Lots of my posts in this thread really is trying to find out where there is a match in views, objectives, and interest. Obviously there seems quite a gap.
If you go and review different products from the same company, you will generally find that sonic signatures vary quite drastically, you know that these deviate from high fidelity quite drastically as well. The basic obstacle is that lots of people are in for the money but do not have the accumulated experience and knowledge; bit and pieces are all over the place which should work quite nicely these different approaches are well integrated, but getting these kinds of people to work together is not easy. I know some local people here that spend there life saving trying to do this, for me, this is not a good idea either. So my own approach is just keep the project going without creating a financial burden, and operate however it best works.
To people whom do not appreciate the efforts other put forth, I think you will never get to work with the great people whom do. You really miss the fun of the process. Engineering gets boring when one gets into this rut.
If you go and review different products from the same company, you will generally find that sonic signatures vary quite drastically, you know that these deviate from high fidelity quite drastically as well. The basic obstacle is that lots of people are in for the money but do not have the accumulated experience and knowledge; bit and pieces are all over the place which should work quite nicely these different approaches are well integrated, but getting these kinds of people to work together is not easy. I know some local people here that spend there life saving trying to do this, for me, this is not a good idea either. So my own approach is just keep the project going without creating a financial burden, and operate however it best works.
To people whom do not appreciate the efforts other put forth, I think you will never get to work with the great people whom do. You really miss the fun of the process. Engineering gets boring when one gets into this rut.
Now now Mr. Shurv. I will not have you disparaging the fair ladies of Ipswich, Suffolk where I lived happily during my last years in civilisation.
But I will admit that Ipswich virgins are rare .. as are Danish & Californian ones. That's why good transformers are $$$.
Just for the record, having measured some 10 transformers with Mark Johnson's Quasimodo jig, I found leakage inductances varying between 30uH (a power transformer) and 20mH (a 2x18V/0.39A model).
http://www.diyaudio.com/forums/powe...rmer-snubber-using-quasimodo-test-jig-65.html
The Excel table with the results is attached to the post #643 at the above link.
Regards
I will need to add Ipswich to my travels.
Looks like a good time could be had, if I get there before you lot.
Looks like a good time could be had, if I get there before you lot.
Browsing through the attached document, fig. 5 is very similar to what we are doing. I did verify that it does make an audible difference, and slight modification can be carefully applied by reviewing the circuit and considering the layout in detail.
I think it should be single-ended or Bal IN (many DACs output bal V or bal I).
And single ended or Bal OUT. To cover all bases (which belong to us)
Oh, oh
And it should be low / no gain because so many of us never get beyond 9'o'clock on the vol. 😱
... with a switchable 'hi gain' option for annoyingly recorded, not-compressed music like ... Dire Straights, Brothers in Arms
... oh and maybe a tone control tap ... 😉
Cheers,
(and sorry for the more work!)
Jeff
PS you've been mumbling forawhile ...
It only took me 15 months to realize that a signal generator and a resistor could do the same job with the same accuracy. The lab set-up is called "Quasimodo ExtraLight" and it's shown in (QM post #561). The measurement is particularly simple when you use an old school signal generator whose square wave amplitude is variable, i.e., not a cheapie model with a fixed TTL-level square wave output.
By the way, I've measured a couple toroids whose secondary leakage inductance was below 10 microhenries. Were they the secondaries you'd expect to have low leakage inductance (few turns, large wire diameter)? Yes they were.
Actually, for the single-ended case, I lumped all error terms in series with the input signal into Verror. These error terms include a bunch of stuff. Power cord induced error is just a fraction (of unknown magnitude).
In the differential case, I lumped all common-mode errors into Verror. Again, the error terms caused by the power cord are only a fraction of this total error.
Tom
When I'm done, I can provide enough puzzle pieces for you to accomplish that. You'll have to work out the volume control and input switching. That can be as simple as a potentiometer or as complicated as a bunch of remote controlled relays, attenuators, and the like.
Tom
Please consider adding option for "tone control"
I know, it seems highly un-hi-fi, but really, I don't have my young whippersnapper hearing any longer (I was tested two years ago). A small, up 2 to 4 db shelving boost would be most welcome. I'm thinking starting at about 1.5K or so but perhaps it could be made quasi parametric by allowing the user to wire in a switch with decreasing caps paralleled so as to allow some modification of the turnover point.
Now that most CDs and downloads have been digitally scrubbed to remove tape hiss from old analogue tape masters, how "purist" is it to have the straight-line pre today?
I know, it seems highly un-hi-fi, but really, I don't have my young whippersnapper hearing any longer (I was tested two years ago). A small, up 2 to 4 db shelving boost would be most welcome. I'm thinking starting at about 1.5K or so but perhaps it could be made quasi parametric by allowing the user to wire in a switch with decreasing caps paralleled so as to allow some modification of the turnover point.
Now that most CDs and downloads have been digitally scrubbed to remove tape hiss from old analogue tape masters, how "purist" is it to have the straight-line pre today?
If you use a media player rather than CD player you can add that shelving in the digital domain very easily of course!
You're welcome. Nothing I haven't been pointing out for years elsewhere in regards to deficiencies in home audio gain structure conventions, though I think this is the most detailed small book, er, post I've composed on the subject.
The doc Andrew provided contains a couple sims for error voltage, actually. The limitation I found with sims is there's so much dynamic range in the mains system variables they don't really indicate anything not already yielded by more handwaving sorts of analysis.
Mr. Joffe appears to be trying very hard to reinvent the difference amplifier. You may find greater familiarity with that circuit helpful.
Thank you both DNi and Mark for the data points! Most numbers I came across when I last looked some years ago were in the 10 to 30uH range so I was sweeping 1 to 100uH in my analysis. Clearly that could go a bit higher---the common use of low VA EIs in sources might mean sources tend to have more leakage inductance bounce than power amps.
Yup. Be cautious with shelving filters, though, as they introduce a durable 180 degree phase shift which nearly always in my experience increases phase error even whilst providing magnitude correction. Usually overlapping low Q peaking (or dipping) biquads produce better correction.
I hear ya. I'm not able to support tone controls right now, but may be able to support those functions at a later date. This is actually the first request I've had for tone controls. Unfortunately, I have many projects and limited time, so I have to focus my efforts on circuits that are more likely to keep business going and hopefully support business growth as well.
You could probably find a combined volume control / tone control circuit somewhere. I'd look for Baxandall's circuits. There are several examples in Douglas Self's "Small-Signal Audio Design" book.
As Bill mentioned, you could also use the graphic equalizer in your media player (assuming you're using a digital source). For example CMD-ALT-E in iTunes (or Window -> Equalizer from the menu) will bring up an equalizer. Works pretty well...
Tom
Last edited:
Depending on the filter order... The shelving high-pass filter commonly used in tone controls is a zero followed by a pole. That would give a phase "speed bump" of no more than 90 degrees (if the pole and zero are decades apart). For 2-4 dB of correction, the phase bop will be rather low. This for a first order filter (common Baxandall). As the filter order increases, the phase "speed bump" increases in magnitude as well.
Tom