Scott,
Electrical current is defined as dq/dT. Or the change in charge per unit of time. Electron movement is different. The movement of an electron through a conducter is much slower than the movement of charge. It is called the electron drift. But you knew all that.
The article in question attempted to explain passive inter modular distortion in terms of circuit theory. In doing so it used terms imprecisely to convey the technical content. It should be clear that circuit theory pure resistance, capacitance and inductance are lineàr and as such should not contribute to the non linearity that causes distortions.
However as there are no perfect components it should be clear that a non ohmic resistor passing a two tone signal will also create sum and difference components. The article pointed out that even surface scratches can do this for RF signals due to the increased current density from skin effect causing localized heating. As there is a thermal coefficient to cable resistance that causes a non ohmic current induced change.
Dimitri asked if I was not mistaking skin effect issues for something else.
Now if you go to lower frequencies it is not skin effects but flaws in the conductor.
If you stop limiting yourself to circuit theory models of conduction you can see other effects
As I thought it was clear what was being proposed before and since you still think the issue is open here is the simplest analogy I can make. Think of a small wave in the ocean passing a rock or two. It must flow around them and after a distance will recombine. But as there was a slightly longer path the wavefront is now distorted. Now if a large wave hits the same rocks it will go over and around them so the result is less waveform distortion.
Electrical current is defined as dq/dT. Or the change in charge per unit of time. Electron movement is different. The movement of an electron through a conducter is much slower than the movement of charge. It is called the electron drift. But you knew all that.
The article in question attempted to explain passive inter modular distortion in terms of circuit theory. In doing so it used terms imprecisely to convey the technical content. It should be clear that circuit theory pure resistance, capacitance and inductance are lineàr and as such should not contribute to the non linearity that causes distortions.
However as there are no perfect components it should be clear that a non ohmic resistor passing a two tone signal will also create sum and difference components. The article pointed out that even surface scratches can do this for RF signals due to the increased current density from skin effect causing localized heating. As there is a thermal coefficient to cable resistance that causes a non ohmic current induced change.
Dimitri asked if I was not mistaking skin effect issues for something else.
Now if you go to lower frequencies it is not skin effects but flaws in the conductor.
If you stop limiting yourself to circuit theory models of conduction you can see other effects
As I thought it was clear what was being proposed before and since you still think the issue is open here is the simplest analogy I can make. Think of a small wave in the ocean passing a rock or two. It must flow around them and after a distance will recombine. But as there was a slightly longer path the wavefront is now distorted. Now if a large wave hits the same rocks it will go over and around them so the result is less waveform distortion.
Last edited:
Yes, imprecise use of terms is often a good way to convey technical content.simon7000 said:
Circuit theory has no model of conduction. It assumes that some other model of conduction delivers components to which circuit theory can be applied. Circuit theory makes no assumptions about conductor linearity.simon7000 said:
The electrons are the charge, again you are missing a lot of things and really do need to go back and review the nature of EM wave propagation. There are no signal charges propagating at c as opposed to the free electrons in the conductor.
Water flow and valve analogies should be banned IMNSHO.
Charge movement is not the same as electron movement. Now where you get the bit about c is out of left field.
Did you get the bit about 1 MHz waves propagating in a solid piece of copper at 400m/S?
Are you guys perhaps debating something akin to current flow ?
I.E....
When I got my little bit of education (USNavy~1962 and junior-college
~1970) current flowed - to + (electron flow), but now it is taught + to -
(hole flow)......
It seems to change (not unlike whether or not eggs are good for you) ...
You have Math, Physics and then Engineering concepts. A lot of things get simplified when going into design/production or you would never get 'there' in a life time. I have some observations relating these three when everyone is finished with bashing each others models of reality.
[and after my air conditioner gets fixed... my brain is getting fried here]
THx-RNMarsh
[and after my air conditioner gets fixed... my brain is getting fried here]
THx-RNMarsh
Ok Scott I will let you have it your way. So when the electrons flow out of the wire at some point the wire will eventually run out of electrons. So you need to replace the wires every so often. . That is clearly why cables are directional. You have to reverse them every so often to replace the electrons.
(That's is nutty enough might just show up elsewhere as a good idea!)
Next we can talk about AC and maybe even the wave equation for non homogenous media.
. That is clearly why cables are directional. You have to reverse them every so often to replace the electrons. (That's is nutty enough might just show up elsewhere as a good idea!)
Next we can talk about AC and maybe even the wave equation for non homogenous media.
You have my sympathy. If you were closer I would lend you my portable wàter cooled unit.
Have to agree on the water flow valve issue. IMSWHO electron flow is a closed loop system with loses to heat etc. Water is viewed as a only a portion of its loop . Water gets recycled but on a long some parts out of view way . Electrons need the loop closed to work . Using the water/ valve analogy only show half the picture and leads to very wrong conclusions . It misses the big picture so to speak. Heaviside had some interesting thing to say about that.
Have to agree on the water flow valve issue. IMSWHO electron flow is a closed loop system with loses to heat etc. Water is viewed as a only a portion of its loop . Water gets recycled but on a long some parts out of view way . Electrons need the loop closed to work . Using the water/ valve analogy only show half the picture and leads to very wrong conclusions . It misses the big picture so to speak. Heaviside had some interesting thing to say about that.
Waves in water do not require valves or flow. Scott's comment is just not related to the post.
Then the Heaviside reference comes in to play with the wave and what it is riding on would it not ?
Bateman capacitor articles
For those of you interested in 'Capacitor Sound' or lack thereof, I have posted Cyril Bateman's landmark article series on my personal website.
Thanks to Cyril and the Editor at Electronics World for making this possible!
Enjoy!
Jan
For those of you interested in 'Capacitor Sound' or lack thereof, I have posted Cyril Bateman's landmark article series on my personal website.
Thanks to Cyril and the Editor at Electronics World for making this possible!
Enjoy!
Jan
The wave functions of the electrons in a conductor are spread out over large distances, waves in water and EM waves are not a good analogy it breaks down very fast because the physics are very different. Propose all you want, and let me be more specific, the large and small waves and large or small rocks it the way, or maybe as I have seen before the rough shores "surface defects" are not good analogies for large or small currents (or signals) in a wire.
Take the below as stated, to make sense the waves being distorted are on the order of the size of the rocks. In a wire this inplies giga or even tera Hz which don't propagate inside the wire. Small wave, large wave, rock what do those mean relate them to potential wells, field strength, etc. in a meaningful way. You can't.
I'm sorry I brought this up again, I really just wanted some (any) reference, where this is an important issue outside of imaginative audiophilia.
Last edited:
Wow, thanks Jan, Cyril and EWorld!
Here are the "straight forward" recommendations:
Cyril Bateman:
Capacitor Choice.
For the lowest distortion I still prefer PS, however from my measurements, it proved almost impossible to distinguish between an extended foil/PS and a similarly made foil/PP capacitor, apart from small increases in second harmonic, measured for the PP versions. Both types are easily available from mainstream distributors in values up to 10 nF.
For small, low distortion capacitors up to 10 nF, my personal choices would be C0G ceramic, perhaps also including discs up to N750, extended foil/PS or extended foil/PP, with the leadout wires soldered to the electrodes.
For capacitances up to 10 nF, low distortion, small, low cost capacitors are easily available, so I would avoid using metallised PET capacitors for such values. For capacitance values above 10 nF the near perfect C0G, foil/PS and foil/PP types are not easily available.
For 100 nF capacitance we find the lowest distortions are generated by choosing either C0G multilayer ceramic, metallised film
Polyphenylene Sulphide (PPS) or double metallised film electrodes with Polypropylene (PP) film.
Much better film capacitors were listed in my last article but at 1 uF, a metallised PET capacitor provides the economic choice. For the lowest possible distortion, especially with increased signal drive or DC bias, the better quality film capacitor styles shown in figure 1 and recommended in my last article, should be used.
Figure 1
100 mF choice.
Provided the AC voltage developed across the capacitor at the lowest audio frequencies is 1 volt or less and no significant DC bias is used, a double Bi-polar series pair provides an economic solution. When higher AC signal voltages, especially combined with significant DC bias, must be applied, the metallised PET combination produces less distortion. It costs eight times more and takes five times more PCB area than the double Bi-polar.
For the least practical distortion, an assembly of metallised polyphenylene Sulphide capacitors might be feasible. It needs double the board area and is five times more expensive than the PET assembly.
For small AC signals with modest DC bias and for supply rail decoupling, I choose the double Bi-polar 63 volt solution.
10 uF choice.
We have three possibilities. A double Bi-polar using two 22 uF 50/63 volt Bi-polar electrolytics, a 10 uF metallised PET or an assembly of three 3.3 uF PPS capacitors. The lowest cost solution for use with signal voltages less than 1 volt and no significant bias, is a double Bi-polar series pair.
A 10 uF MMK metallised PET takes the same PCB area and distorts less with DC bias.
The PPS capacitor assembly ensures lower distortion, especially when used with increased AC signals or DC bias voltage. However it occupies more board area and is expensive.
An assembly of Polypropylene capacitors, as used in the DC bias network, would provide the lowest possible distortion but
requires a five times larger board area and is most expensive.
For small AC signals and modest DC bias, I choose the 10 uF MMK metallised PET capacitor.
http://www.linearaudio.nl/linearaudio.nl/images/pdf/Bateman_EW_10_2002_mar2003_10_nF_and_less.pdf
http://www.linearaudio.nl/linearaud..._12_2002_mar2003_1uF_electrolytic_or_film.pdf
http://www.linearaudio.nl/linearaud..._10_to_100uF_caps_and_100_Hz_measurements.pdf
PS Was there any follow up to these articles? Replication, extension ...
Any testing of 'newer' caps (post 2003)?
Last edited:
Still a problem at lots of sites including big magazines. Some extended ascii fonts include the most common greek characters as a convienience, but there is no standard. As of a few years ago Postscript printers still had a "hack" (at least a hack IMO) so that printing was always OK. They had an explicit switch to symbol font to print the mu (micro) no matter what the running text font was. So you write an article and print it out locally and all looks OK, but through transmission elsewhere the characters above 127 ascii can get lost or substituted. Since it's Adobe I would imagine sticking with pdf and display postscript exclusively would always work?
Articles with "m" instead or the wrong character for "Omega" (Ohms) are everywhere.
Last edited:
I have a question for Jn, Scott or someone smart enough to answer. Since I am designing and getting closer to producing the enclosures I will use for my speaker and they will be made with a polyurethane rigid material how do you properly do the grounding in a non metallic enclosure? Do you just run the star ground to the heatsinks and use that as the single grounding point or is there a preferred method to do the grounding scheme? Thanks for any answers.
Steven
ps. This will be a self powered speaker with all components internal to it. Crossover, any eq, power supply and power amplifiers.
Steven
ps. This will be a self powered speaker with all components internal to it. Crossover, any eq, power supply and power amplifiers.
Last edited:
jn is the best guy for this but I figure not having the conductive shield you need to go back and consider all the loops/shared connections and twist together and ground things to minimize the undesirable coupling. In my experience star grounding is useful in minimizing the direct voltage drops due to current including the circulating current in bypass capacitors, but you need caution when dealing with higher and higher frequency components because the mode of coupling changes from simple voltage drops to field induced voltages/currents.
Last edited:
- Status
- Not open for further replies.