As in all these discussions, we have come back around to the simple statement that the primary source of nonlinearity in a capacitor is the known nonlinearity in the dielectric. This appears to be unrelated to DA performance. All well known for decades.
What we seem to have had is a measurement: 'the yellow car was faster' followed by an assumption: 'it went faster because it was yellow'. It then takes lots of effort for others to argue that colour is not a factor in car speed. Of course, there has been little research on this as no sensible person would think that colour is a factor. Hence the 'yellow is faster' merchants can declare that this has not been disproved, so it must be true.
What we seem to have had is a measurement: 'the yellow car was faster' followed by an assumption: 'it went faster because it was yellow'. It then takes lots of effort for others to argue that colour is not a factor in car speed. Of course, there has been little research on this as no sensible person would think that colour is a factor. Hence the 'yellow is faster' merchants can declare that this has not been disproved, so it must be true.
Um, you are aware that this is the Blowtorch thread, yes? Did the sign fall down again?
se
As Jan listed the components of a capacitor I am wondering if anyone has or even can look at what happens at all the junctions of dissimilar materials in a real capacitor. Sy seems to have the best background in materials science but I am not sure if this can explain any of the possible missing elements that could be missing in a capacitor model? Something as simple as the material used to weld the leads on or any other parasitic function that is happening at each junction of a different substrate? Is this where you would possibly find any real non-linear function that is being conjectured about for about a week now.
Just about any textbook on solid state physics. The whole field of NLO is based on this.
Hard for me to get my head around. Capacitors work by charge moving to/from or in/out of dielectric stuff. DA works by charge moving to/from or in/out of dielectric stuff. Now, the first kind of charge moving around is non linear. But the 2nd kind, which is the same kind of charge, and the same dielectric, is linear. Moreover, the two have nothing to do with each other.
Hmmm.
Jan
It very much depends on the dielectric. All of the external factors you mentioned can be minimized (and mainstream manufacturers have done so), but the dielectric has its own nonlinearities (very small for organic polymers) and contaminants (can be significant, especially in PET). Add in the mechanical aspects (serious issue for soft polymers like PTFE, much less so for hard polymers like polypropylene).
Compared to a tube, transistor, or (in the passive world) an inductor, capacitors are MUCH closer to ideal.
If I create a dipole by applying a field, that can look like moving a charge. Temporarily, of course.
Wait wait wait.
Ideally, capacitors don't work by charge moving to/from or in/out of a dielectric. Charge moves to/from/in/out of the conductive plates. Ideally, charge does not move in the dielectric.
With DA, charge moves into and out of the dielectric, or as SY mentioned, the dielectric may polarize to a certain degree. But ideally, there is no movement of charge in the dielectric.
se
Your "monster" H/K is based on which model, and what have you done to it, the poor thing?
yes.... that IS the question. If we can eliminate ceramics and stick to the polar and film...... I have taken apart a lot of brands and types to see the construction.... 99.9% of the visible lead termination difference is in the how the leads are attached. And, it varies a moderate amount.
Most leads (copper or copper-weld) are not soldered nor spot-welded (with exceptions, of course, like with PS caps). They are crimped to foil plate or the lead is laid against the plate material and molten aluminum metal is sprayed at the ends to bind the plates and lead to the plate material. This is done as a large group of caps at one time. The quality can vary a lot by this method.... the plasma gun is hand-held and literally sprayed back and forth... time, duration, all covered the same amount varies and is only under the control of the 'painter' doing it..... which when I watched, did not look all that consistent and uniform. But seem to pass muster.
One can tell if it is the contact/connection by measuring same dielectric made with end terminations done various ways..... soldered plates, metal sprayed, crimped. And then redo with changed lead material... copper, steel, copper covered steel.
I think you will find the same amount of thd as with a metal part connected to another metal part by same methods.
THx-RNMarsh
Last edited:
No current per se. Think of a bunch of dipoles oriented randomly, so there's no net field of any polarity. Apply a field of a given polarity. Some of those dipoles will align to the field. See?
se
Richard,
As someone who has actually done some metal spraying back in the day I can see so many ways that having someone do that by hand would give varying results. That would seem like a very good area to add automation to the process. In my application I used metal spraying to make tooling, a very difficult and expensive method with iffy results and lots of added problems that just weren't worth trying to improve. I sold my equipment to do that long ago. I do remember coming out looking like the Tin Man the first time I used that gun.
As someone who has actually done some metal spraying back in the day I can see so many ways that having someone do that by hand would give varying results. That would seem like a very good area to add automation to the process. In my application I used metal spraying to make tooling, a very difficult and expensive method with iffy results and lots of added problems that just weren't worth trying to improve. I sold my equipment to do that long ago. I do remember coming out looking like the Tin Man the first time I used that gun.
Its quite the contraption, too... and really loud... everyone using it or near it has to use hearing protection.
There are potentials for 'bad' connection if not used with the right combination of metals to have serious issues.... but cap makers have that worked out. But some foreign cheap made parts... you never know for sure what they are willing to do to cut costs.
So, one can just go thru one by one and check each and every known parasitic element. Almost like a process of elimination to see what has the largest affect or dominates over the others.
-Richard
There are potentials for 'bad' connection if not used with the right combination of metals to have serious issues.... but cap makers have that worked out. But some foreign cheap made parts... you never know for sure what they are willing to do to cut costs.
So, one can just go thru one by one and check each and every known parasitic element. Almost like a process of elimination to see what has the largest affect or dominates over the others.
-Richard
Last edited:
Ahh, that's a clear picture, thanks! But if there's no charge going into the dielectric, why is it called 'absorbtion'?
I'm not trying to make fun of anything - really interested to learn the ins and outs.
Jan
- Status
- Not open for further replies.