Genuine discussions are good. People learn things from genuine discussions. Stating what you believe to be "real" then putting your fingers in your ears is not a discussion.zinsula said:
Suicide bombers agree on many things - what does that tell us? Perhaps that Dawkins' idea of memes may have an element of truth in it.zinsula said:
In a dictatorship things can be declared to be 'true' by an authority. In a democracy things can be declared to be 'true' by a majority. In capitalism things can be declared to be 'true' with money. Science, maths and engineering don't work like that. Proof by assertion, proof by popularity and proof by sales figures are not valid there.
So very true. And despite its crucial importance for all hardware advances, I still hear people, usually at the periphery of the field, expressing the notion that analog is obsolete. I had a speaker designer say to me over dinner at an AES convention in 1997 "So I guess you don't have anything to do now?"
This attitude is almost as prevalent as the misunderstandings of what "digital" actually is. The notion is often that something that switches is inherently digital. This reached total absurdity at Harman many years ago, when the ad copy writers were told that a switching power supply was a "digital" power supply (it used a Power Integrations IC).
This is not to disparage using the symbol domain (or numerical domain as some prefer) where it is appropriate. And as I remarked once, it's almost feasible to do useful work fresh out of school based on a curriculum heavy on "digital", whereas analog design takes quite a bit longer to reach a level of productive competence. And there are precious few teachers out there.
John, thanks for your reply. One more aspect that I would like to be touched is the role of output transistors Ft in functioning of GNFB mechanism.
By taking respective measures against dynamic overloading of any stage of an amp, one can get equal technical features at 20...20kHz range of two amps with various output transistors.
But, listening experience says another. Being built on the ground of the same schematic, the amp version with Ft=200MHz output transistors sounds better than the same amp with 20MHz flat h21 japaneese transistors. When we had a listening session with these two amps, people mentioned that 20MHz version sounded as a very good transistor amp, while a sound of the 200MHz amp was qualified as a good vintage tube amp sound, but with perfect bottom end. Passband of both amps was limited by input LP filter around 1 MHz.
I have realised that the subject of IMD is not at its final form yet.
Last edited:
There are enough discussions about this, hence no need to restart. Anyone interested can read upon, in this very thread.
I like the genuine suggestions of some here. They have enough credibility for me.
Wow Dick, I never expected a comment like that, maybe you should read the archives. I also love the implication that solid engineering just doesn't quite make "what sounds good". I sure glad the diagnostic value of data collected in medical instruments doesn't relay on these unmeasurable nuances of component behavior.
Zinsula - Don't you also hang out at a place with a long AD815 thread? Well, it does have more than 8 legs. The cast of characters you mention have also disagreed, or at least have made contradictory statements at times.
jcx - The JC-3 controversy can be put in a couple of very simple algebraic equations, no sims required. When confronted one side ran away.
Last edited:
Whose definition? I've always understood noise to mean anything that's unwanted (i.e. anything that's NOT the "signal"), whether it's random or not.
se
After entering our local university as a EE major, it was time to get approval from the assistant dean to approve our specific major.
I presented my documents and he became very angry with me because I wanted to focus on analog engineering. He refused to approve it and told me I could leave school or target digital courses to continue. He told me that he was doing me a favor!
I recall this happened around 1984.
I presented my documents and he became very angry with me because I wanted to focus on analog engineering. He refused to approve it and told me I could leave school or target digital courses to continue. He told me that he was doing me a favor!
I recall this happened around 1984.
Last edited:
equations?? - you must be joking - expect the audience I was addressing would admit any equation describes a circuit's operation?
I do hope that even those allergic to math but somehow doing circuit design would be more comfortable with schematics, sims
I provided source files so genuinely interested circuit designers could poke, modify - even show where I was wrong
I did move from the idealized to to more "realistic" modeling hoping to get some interaction, constructive argument rather the flat rejection of the analysis
now the same crowd is once again doing the "waving the bloody shirt" rhetorical thing about closed minded engineers
If precision is needed the term "stochastic process" and allied terms can be used, but the substitution "random noise" is what one usually sees.
I wrote a paper once where, early on, I made a distinction about uses of the term "noise", and said all noise discussed would be random noise unless it was stated otherwise. Then I noticed, after I submitted the document, that I'd never mentioned anything otherwise
But it was an unedited and unrefereed thing for SPIE. Beware, it has some errors, although the basic conclusions about charge preamp optimization are correct, qualititatively (and not too far off quantitatively).
After entering our local university as a EE major, it was time to get approval from the assistant dean to approve our specific major.
I presented my documents and he became very angry with me because I wanted to focus on analog engineering. He refused to approve it and told me I could leave school or target digital courses to continue. He told me that he was doing me a favor!
That's disgraceful, but it does underscore my remark about the dearth of good teachers.
E/M field theory has to be relearned for those that do high speed digital layout, and though seperate in many ways again high speed digital uses various analogue functions such as s-parameters for simulation.
Analogue is not totaly dead, but you do find that real analogue engineers are becoming fewer when you go to different establishments and they tend to be older, though where I am now has numerous analogue enginers(and acoustic ones) so its nice to know there are still some. Agree though digital is seen as the main stay of electronics, though with digital you do have to have a large software team in the background... Maybe its the sexier option at Uni these days, whereas analogue is seen as old fashioned.
Analogue is not totaly dead, but you do find that real analogue engineers are becoming fewer when you go to different establishments and they tend to be older, though where I am now has numerous analogue enginers(and acoustic ones) so its nice to know there are still some. Agree though digital is seen as the main stay of electronics, though with digital you do have to have a large software team in the background... Maybe its the sexier option at Uni these days, whereas analogue is seen as old fashioned.
I see some ground plane teachings, as well as stripline stuff and lead inductance lumps/understandings. But I really do not see much in the way of real understanding of E/M theory coming out of the Uni's beyond the simple math manipulations.
Actually, digital was pretty much viewed as sexier back in '78-79.
jn
I believe there is no fruit to bear along that line of reasoning. As Scott pointed out, typical sample lengths reduce errors well below the mud.
I haven't seen any equipment out there with "infinite" as a window size.
jn
Dick,
I am now at a computer rather than a phone so I could look at your capacitor ESR piece. (As I mentioned I had been taking some breaks.. about six in the foot and a few in the rib cage,)
The issue I was trying to get at is that to me the math shows that distortion must increase in any series capacitor circuit even well above cut off. Some of this is due to the tilt that is always present because of the still decreasing capacitive reactance. That is not done by an increase in the harmonics but rather a decrease in the fundamental.
As you show DF also can be modeled as another RC network in parallel that also decreases with rising frequency. So although there is no generation of new harmonics the distortion again rises due to the decrease of the fundamental.
Now there is still the issue of what out of band low frequencies do when confronted with both a theoretical capacitor and real capacitor forming a high pass filter.
That is where I think the C dv/dt term becomes more important. The most basic issue is that there is a baseline of vibration everywhere (.005 G is very still.) Now when you have a large DC voltage across a capacitor that has been oversized for bandwidth considerations the microphonic (dc/dt) energy may be of concern as it can become large enough to be a non harmonically relate signal large enough not to be masked by low level signals.
Then there is the issue of the microphonic effect even when no DC is present due to sub audio modulations such as record warp.
Now none of this is new. It is just not common knowledge. There are folks who laugh at isolation pads under preamps etc, not understanding that they may actually provide improvement for some circuits. (As would mounting considerations for the capacitors.)
I am now at a computer rather than a phone so I could look at your capacitor ESR piece. (As I mentioned I had been taking some breaks.. about six in the foot and a few in the rib cage,)
The issue I was trying to get at is that to me the math shows that distortion must increase in any series capacitor circuit even well above cut off. Some of this is due to the tilt that is always present because of the still decreasing capacitive reactance. That is not done by an increase in the harmonics but rather a decrease in the fundamental.
As you show DF also can be modeled as another RC network in parallel that also decreases with rising frequency. So although there is no generation of new harmonics the distortion again rises due to the decrease of the fundamental.
Now there is still the issue of what out of band low frequencies do when confronted with both a theoretical capacitor and real capacitor forming a high pass filter.
That is where I think the C dv/dt term becomes more important. The most basic issue is that there is a baseline of vibration everywhere (.005 G is very still.) Now when you have a large DC voltage across a capacitor that has been oversized for bandwidth considerations the microphonic (dc/dt) energy may be of concern as it can become large enough to be a non harmonically relate signal large enough not to be masked by low level signals.
Then there is the issue of the microphonic effect even when no DC is present due to sub audio modulations such as record warp.
Now none of this is new. It is just not common knowledge. There are folks who laugh at isolation pads under preamps etc, not understanding that they may actually provide improvement for some circuits. (As would mounting considerations for the capacitors.)
- Status
- Not open for further replies.