User talk:Brews ohare/Archive 1

Archive 1

Refs?
Brews, your contributions would be a whole lot more useful if we could verify them by checking your sources. Can you please add refs to say where your info is from? Saying "is known to be" is not useful, even if that's what it said before, especially if you change what it's known to be. Let me know if you need any help figuring out how to do citations. Dicklyon 05:33, 5 November 2007 (UTC)


 * I'm happy for the feedback. Do you have some specific cases in mind? I thought I was being more careful about this following the discussion fo Early effect.


 * Most recently, I was reacting to small-signal model. Dicklyon 15:13, 5 November 2007 (UTC)


 * I believe the Jaeger-Blalock reference covers my contributions here. There may be some other issues that I inherited. If you point them out, I'll fix them.


 * Perhaps it does, but since the footnote was on one particular item, it wasn't clear if this source would apply to the rest of the stuff you were changing. And there was no URL to make it easy to check.  So I added one; I haven't looked closely yet, but it appears to mostly support the section, though with quite a different set of symbols and formulas.  See how I used GBS to find a URL, and used cite book to get a footnote in standard format.  I left out the equation and page, however; we could probably find a way to put those back in if needed; see Template:cite book. Dicklyon 16:01, 5 November 2007 (UTC)


 * Thanks for the help. I found that clicking on the ISBN in the reference list provides a choice of sites to look up the book, so I had decided that was adequate. However, a standard format is the way to go.


 * By the way, I find that using math formula /math seems to choose a font size that cannot be controlled: sometimes small sometimes way too large, by caprice. Can you advise? Thanks much.


 * It depends both on your preference settings and on how complex the equation is. It sets it as inline html when it can, which makes it too small; when it can't, it makes an image, which is big.  I agree it seems annoying and capricious. Dicklyon 16:01, 5 November 2007 (UTC)

Diode small signal model
Hi Brews. Why did you move the small-signal diode model section from the diode modelling article back to the small-signal model article? In the small-signal model article its just a random example, but it adds to the completeness of the diode modelling article. -Roger 01:31, 10 November 2007 (UTC)


 * Hi Roger: Thanks for the note. My thinking was that without the diode example, the small-signal article lacked any illustration of how small-signal models were done, and it seemed a bit odd to leave the small-signal article dangling with the large-signal diode equations.


 * In contrast, the diode modeling article doesn't really need a small-signal section, which opens a new topic altogether. I did add a reference to this small-signal example in the modelling article.


 * If the diode modelling chapter is to be extended, a different way to go would be to add treatment of some other diodes, like the Zener or Shottky diode, in much the same fashion as already done for pn-junction. What do you think about that??

SPICE article changes
Hi Brews. I'm working on updates for the SPICE article. One thing, though: A while back I flushed a ton on external, commercial simulator links which were basically advertising. Now the only simulators mentioned are the big, historical ones: HSPICE and PSPICE. You just opened a can of worms by adding a reference to LTSPICE -- which is a fine simulator, but if that goes in, National Instruments, Tina-TI, and everyone else will add their links and I'd just as soon avoid the mess. LTSPICE is commercial; it helps Linear Technology sell their chips, as do a number of other give-away SPICEs.

Also, the device model section you added doesn't have any context and is, I think, not directly related to spice itself (although it's true enough). Perhaps it belongs in the semiconductor device modeling article, as it relates to compact modeling. YoungGeezer 18:19, 10 November 2007 (UTC)


 * I guess I can live with your decision, although it would be helpful to readers to be able to find the freebie version.


 * There is a point to indicating that SPICE has little value without the compact models. A reader may very well not realize this point about SPICE.

More SPICE article changes
Hi Brews: Please see the discussion page on the SPICE article. Thanks. YoungGeezer 18:06, 12 November 2007 (UTC)

Template:Electronic structure methods
This template is not intended to be an all-inclusive list of methods but rather one that points to the key methods, which often link to other related methods. I have therefore reverted you recent additions as I do not think they are key methods. If we added all the other methods that had articles the list would be very long. Bduke (talk) 22:27, 16 November 2007 (UTC)

License tagging for Image:MOS Capacitor.png
Thanks for uploading Image:MOS Capacitor.png. You don't seem to have indicated the license status of the image. Wikipedia uses a set of image copyright tags to indicate this information; to add a tag to the image, select the appropriate tag from this list, click on this link, then click "Edit this page" and add the tag to the image's description. If there doesn't seem to be a suitable tag, the image is probably not appropriate for use on Wikipedia.

For help in choosing the correct tag, or for any other questions, leave a message on Media copyright questions. Thank you for your cooperation. --ImageTaggingBot (talk) 19:08, 27 November 2007 (UTC)

Errors in expressions
Hi Brews. I think there are some errors in your expressions in the common base article. Your definition $$R_{in} = h_{11}= \begin{matrix} \frac{v_{in}}{i_{in}}\end{matrix} \Big|_{v_{out}=0} $$ I believe is only true in the unilateral (h12 = 0) case (see Admittance_parameters). If you assume Rl = Rs = ∞ then the expression shouldn't involve ro and the exact expressions should be Rin = re = 1/gm. Can you double check this? Thanks. -Roger 21:05, 4 December 2007 (UTC)
 * Hi Roger: Thanks for looking at this. Here's how I arrived at my result for h_11:

Although it is possible to solve the thing using coupled equations, it is easier to solve by making the h-equivalent and the hybrid pi agree for two different sets of terminations. If the circuits agree for these two sets, then they agree for every case. The terminations that make for the easiest math are those that cause one or the other of the two dependent sources in the h-equivalent to disappear. One case is V_2 = 0 (short-circuiting the output) to make the back voltage zero and driving with source I_1, and the second case is setting I_1=0 (open circuiting the input) making A_i I_1 = 0, and driving the output with voltage V_2.

Looking at the h-equivalent in Fig. 3, the back voltage vanishes if V_2=0. So I set V_2=0, which means short-circuiting the output. In that situation, from the h-equivalent, h_11 = V_1/I_1.

Then I go to the hybrid pi model and short the output (collector) to ground, to obtain the same situation used with the h-equivalent. Applying I_1 to the emitter, I find V_1, the emitter voltage. This voltage is above ground by the voltage drop across r_O, which has its collector node at ground because the collector is shorted to ground to set V_2 = 0. The current in r_O is I_1-(g_m+1/r_PI)V_1. Solving V_1 =(I_1-(g_m+1/r_PI)V_1) r_O for V_1, I find h_22 = V_1/I_1 = r_O//r_E.

Hope this makes some sense to you. Another way to check things is to avoid the math and set up SPICE with the hybrid pi and with the h-equivalent. For arbitrary values of output voltage on the collector and input current at the emitter, both circuits are the same if they show the same input voltage at the emitter and output current through the voltage source at the collector.

I confess I have not actually done that. However, if your opinion is that there is an error, I'll go ahead and do it, assuming you agree that is a valid test. Brews ohare (talk) 21:41, 4 December 2007 (UTC)
 * I've done the SPICE simulations and everything looks OK Brews ohare (talk) 01:36, 5 December 2007 (UTC)


 * I get the same result h11=(re//ro), but thats just the input impedance in the unilateral case. In the general case the input impedance would depend on re, ro and RL. If ro = ∞ or Rl = ∞ then the result is just re=1/gm. Basically my only issue was that Rin =/= h11, similarly for the other parameters unless you call them the short circuit current gain, input impedance, etc. When I worked on the Characteristics section I deliberately assumed ro=∞ to avoid that problem. Its unavoidable for Rout in the common collector circuit where you have to put the expression in terms of the source resistance Rs (also for Rin, but you can fix that by combining RE and RL).
 * Also the expression for Rin should be the same in both h and g-parameters. -Roger (talk) 02:28, 5 December 2007 (UTC)
 * I wonder if there is a semantical issue here? The subject of the "general case" for instance? Maybe the "general case" refers to the case with a general driver (with source impedance, say R_S) and a load (say R_L)? My approach to the general case has been: do the simple case (tabulated case) first, to find an h- or g- equivalent with ideal terminations, and then tack the "general" terminations onto that. That leads to the same results as doing everything at once, but it's a bit easier to handle. Do we agree this approach will work?? Brews ohare (talk) 03:32, 5 December 2007 (UTC)


 * I agree that finding the two-port parameters for an amplifier (or any linear two-port) is the most general approach. The issue I have is with the definition of input/output impedance and gains. For the y-parameters, for example, the input admittance is found from Admittance_parameters, it isn't just y11. For unilateral networks the input/output impedances don't depend on the terminations, but they do for bilateral networks.
 * Try repeating the example you did in Common_base to find the input current for a common collector using the same method. If you ignore the load connected to the emitter (RE) then your input current will be wrong. -Roger (talk) 03:56, 5 December 2007 (UTC)
 * Hi Roger: I'm going to take a little time to revise all this. I think I haven't been clear about what is being done. Hope you will have a little patience while I struggle with the presentation. Brews ohare (talk) 04:11, 5 December 2007 (UTC)

Hi Brews. It looks better now, but I have a few more suggestions. I think using R11 and R22 in the tables is a bit confusing and having the labels "current gain", "voltage gain", etc. in the first column may not be strictly accurate. How about if we combine the first two tables and make them just about g and h parameters (and possibly others)? Then in the following discussion we can just make the unilateral approximation (and refer to Admittance_parameters) to derive the driving point and transfer functions. This way we'd avoid too many special cases and carrying forward approximations. Also I think "Back voltage gain" is usually called "reverse voltage gain" or "reverse transmission" in the general case.

Are you familiar with the symbolic circuit analysis program SapWin? It can derive pretty much any of the expressions we'd need in symbolic form and would be very useful if we wanted to include parasitics in the transistor models. -Roger (talk) 20:37, 5 December 2007 (UTC)
 * Hi Roger:
 * I'm temporarily exhausted at this point, so I'll have to take a rest before considering another big revision. It seems that the unilateral approximations are OK only in some cases, so it might be hard to avoid some special cases. I'm OK with including the terminology 'reverse voltage gain". I chose "backward" because the symbol "B" seemed a logical choice for backward where "A" is chosen for the forward gain.
 * I was unaware of SapWin. I'll take a look. Brews ohare (talk) 21:45, 5 December 2007 (UTC)
 * Roger: Have you any comments on latest version of Common base? I believe the proposal to refer to Admittance_parameters really requires writing an article like this for each of the possible two-ports. Presently only z-parameter and y-parameter have pages of their own apart from the two-port article. Each type has its own equations for relating driving point and transfer functions. Brews ohare (talk) 19:31, 6 December 2007 (UTC)
 * Its coming along nicely, thanks for incorporating my suggestions. I think there might be too many tables and figure though. Its probably going to be overwhelming for someone not too familiar with this stuff. Maybe we could put a section that uses the simpler model (without ro) to kind of ease the transition to the more complete model. We could eventually follow that up with better model with parasitic capacitance and discuss the bandwidth, etc.


 * When I get some free time I'll try to tabulate the expressions for the various two-port parameters. That said, are you sure the two-port approach is best for this article? We could make it a lot more concise if we based the discussions directly on the hybrid-pi model and just list the two-parameters for reference. -Roger (talk) 20:11, 6 December 2007 (UTC)


 * I agree it's more complicated than most users want. I suspect a slight amplification of the latest "Summary" paragraph would do the trick. Maybe there still is a place for these two-port examples for those readers that wnat to continue past the summary? Brews ohare 21:46, 6 December 2007 (UTC)
 * Roger: Do you think the revised article for common gate looks like a better way to go? Brews ohare (talk) 18:31, 9 December 2007 (UTC)
 * Hi Brews. Yes, I think thats a better approach. Having the two-port parameters as well will be useful, but I think its better to focus directly on the model. One thing - why didn't you directly define Rin as (what you called) Ramp? You don't need to take the source impedance into consideration when finding the input impedance since its the impedance looking directly into the amplifiers input terminal.
 * I also think we should tabulate things as a summary and to make it easy to compare the different amplifiers. I should have some free time soon, so I'll try to help you out. -Roger (talk) 04:55, 10 December 2007 (UTC)
 * Hi Roger: The source impedance plays a different role for the voltage as signal than for the current as signal. Somehow that has to be brought out, don't you think? Brews ohare (talk) 15:31, 10 December 2007 (UTC).
 * Yes, but the input impedance should still be independent of the source resistance and whether its a current or voltage amplifier. -Roger (talk) 15:50, 10 December 2007 (UTC)
 * Hi Roger: I've done a tabulation for common gate. Take a look, please. Brews ohare (talk) 16:31, 10 December 2007 (UTC)

Loading effect
Hi Brews. It seems like we're duplicating some stuff in the current divider and voltage divider articles regarding the loading effect. Perhaps we should move it to its own article, or move all the stuff from one article to the other and just link to it. What do you think? -Roger (talk) 17:46, 21 January 2008 (UTC)

Hi Roger: Glad to see you back. This is a trade-off I think: the cases are a little different, and having several examples might be helpful to a novice, who probably is the audience after all. The downside is it takes more storage space: is that a cost that has to be considered? Brews ohare (talk) 18:51, 21 January 2008 (UTC)

It looks like the creation of a new article on loading might be necessary if we were to include loading effects of feedback two-ports on feedback amplifiers? Brews ohare (talk) 17:59, 22 January 2008 (UTC)


 * Hi Brews. Storage space isn't a concern and if anything we'd be saving space by eliminating redundancies. I didn't think the section on loading effects would evolve the way it has, plus if we add more examples it'll end up dominating the divider articles. So I'm thinking it deserves its own article.
 * What do you mean by the loading effect of feedback two-ports? Do you mean calculating the input impedance (e.g. with Blackman's formula or the EET)and accounting for that? -Roger (talk) 18:10, 22 January 2008 (UTC)
 * Hi Roger: I'm referring to the approach to feedback networks in negative feedback amplifiers used by Gray and Meyer, and Sedra and Smith: they replace the feedback network (a T-section of resistors) with a two-port, plug the two-port into the amplifier circuit, and then set the two-port dependent sources to zero. The remaining R_11 and R_22 of the two port result in loading of the gain. Brews ohare (talk) 18:23, 22 January 2008 (UTC)
 * Ah, you mean where they idealize the feedback network by moving the m11/m22 terms to the amplifier. We'd probably have to write an article on the two-port feedback amplifier approach before we got into that.
 * I never did like that method, too many approximations and not as elegant at the EET or the asym. model. -Roger (talk) 18:33, 22 January 2008 (UTC)

Asym. model article sections
The recent reordering you did to the sections in the Asymptotic gain model‎ article seem a bit strange (e.g. the examples are scattered). Is this a work in progress? -Roger (talk) 18:36, 22 January 2008 (UTC)
 * Please take another look; a number of changes were made. Brews ohare (talk) 18:46, 22 January 2008 (UTC)

Note in feedback amplifier article
I'm still not getting what you mean by "...main amplifier may not be a two port". Do you mean its non-linear? -Roger (talk) 02:07, 24 January 2008 (UTC)

Hi Roger:
 * I'm not sure what to do about this. The reference to Jaeger goes into detail. One cause of the main amplifier not being a two port is that there is an implied fifth terminal besides the main two input terminals and the main two output terminals. The example Jaeger treats is an output bipolar transistor, where one output terminal is the collector, the other the emitter and the port condition is therefore not satisfied. Of course the third terminal in this case is the base, but it can be connected into the amplifier in a way that makes it non-obvious that there is a third terminal. I'll think more about this.

Brews ohare (talk) 06:16, 24 January 2008 (UTC)
 * I have added a section called Negative feedback amplifier that spells out the problems. I've left in a reference to Jaeger where an example is discussed in great detail.

Brews ohare (talk) 20:16, 25 January 2008 (UTC)

"Other methods" in return ratio article
Hi Brews. A few things about your last change to the section. I think your focus is too much on SPICE, its more general than that.


 * If a small-signal circuit is available, the above steps can be directly implemented in a SPICE simulation using the small-signal model directly...

I don't think its worth pointing that SPICE can be used, but only rather when it can't be used. Also I think people rarely start off with a small signal model of whatever active device they're using, where they have direct access to the VCCS source within the device. Usually people use SPICEs built in BJT/MOS (nonlinear) models, in which case its pretty much impossible to perform the steps outlined in the article.


 * ...avoiding some tedious algebra, but forgoing the convenience of a formula.

I think this issue is beyond just avoiding algebra, its also a practical issue when you have to measure loop gain in a lab. -Roger (talk) 20:06, 28 January 2008 (UTC)


 * Hi Roger:
 * I made some more changes before reading your comment. I see your focus is on the lab, and maybe on simulations where device models are not available. However, that is not everything, of course. If a small-signal model is available, as it is in an introduction to the theory for example, it is very straightforward to follow the recipe as given. However, many might not think of actually putting the small-signal circuit into SPICE themselves, because they don't usually do that.


 * It seems to me that a good way to go about this is to have a separate article showing the equivalence of a Rosenstark approach to the simple-minded small-signal approach. That is non-trivial and non-introductory too.

Brews ohare (talk) 20:19, 28 January 2008 (UTC)


 * Well I'm not really more focused on lab stuff (I often prefer SPICE), but, again, the issue is more general than just a SPICE one.
 * If there are small-signal models already available and the users have direct access to the controlled sources, then of course they can follow the steps outline, there's probably no need to point this out. Manually finding and using your own small-signal equiv. circuit might be useful as a basic check, but I still don't think its a practical enough to point out. -Roger (talk) 20:35, 28 January 2008 (UTC)
 * I see no harm in pointing it out. Anyone who finds that obvious can just sniff at it. For students it can be useful. Brews ohare (talk) 20:40, 28 January 2008 (UTC)

Reference section in Metre
Would you please explain the odd layout of the Reference section in the Metre article? Some references are indented more than others for on apparent reason. —Preceding unsigned comment added by Gerry Ashton (talk • contribs) 22:06, 13 February 2008 (UTC)
 * I can't claim much understanding here. It looks to me like the indent just identifies NIST related articles. Brews ohare (talk) 22:09, 13 February 2008 (UTC)

Feldkonstante
Hi, the Germans adopted the term de:Feldkonstante for ε0 and μ0. I am considering resurrecting elektrische Feldkonstante, but I do not write German very well. It is a pity that there is no corresponding construction in English, but I do not think that the German term should be in the lead of the article. I have requested a move of the French article to Constante électrique (it needs an administrator because of previous history). /Pieter Kuiper (talk) 23:04, 15 February 2008 (UTC)
 * I moved it to a footnote. I've done a number of other revisions on this page. Please take a look.

Brews ohare (talk) 23:35, 15 February 2008 (UTC)

User page
Hey Brews. Nice job on the user page, and thanks for the nod. Its been a pleasure editing with you these past few months.

I really like how some of your figures came out. What software are you using to make them (e.g. )? -Roger (talk) 22:52, 21 February 2008 (UTC)


 * Hi Roger: This figure was made using Excel 2007 and pasting shapes from Insert Shapes. After its composed I copy it and paste it as .PNG using Paste Special. Then I copy the .png and save it in PAINT, where I access it from Wikipedia. It's all a bit kludgy, I'm afraid. Brews ohare (talk) 23:41, 21 February 2008 (UTC)

Two discussions relevent to you
Hello! I've introduced two discussions which you ought to be aware of. One is c versus c0, on Wikipedia talk:WikiProject Physics. The other is in the discussion page of free space. I promise that it's just a coincidence, I don't have anything in general against your editing, and I haven't been purposefully seeking out things you've done to complain about. But just thought you should know about both of those. Best wishes!! --Steve (talk) 01:19, 1 March 2008 (UTC)

Metre
Hi Brews, Metre is spelled as such in the article - by consensus I think. 82.20.28.142 (talk) 22:35, 3 March 2008 (UTC)

Lorentz Force
Hello! You added back in the sentence, "In addition, if the charge experiences acceleration, for example, if forced into a curved trajectory by some external agency, it emits radiation that causes braking of its motion.". This is the radiation reaction force. Why isn't the link which I added sufficient? Thanks!! --Steve (talk) 18:05, 10 March 2008 (UTC)

Dielectric
Hi! I noticed Dielectric, and I thought that you might have views on that section. You might also want to have a look at Displacement current. /Pieter Kuiper (talk) 10:16, 29 March 2008 (UTC)

Free space
Your change to the definition of free space is wholly at odds with the one before it. It only makes sense if you presuppose that you are talking about electromagnetic radiation and simultaneously claim ignorance of such radiation as possibly propagating through it. The analogy to absolute 0 in temperature is misleading. You cited ad nauseum about the unrelated issues surrounding vacuums, yet cited nothing aside from linear supposition in your definition. Free space is its own idealized form of a vacuum. Going from:

"Free space simply means that there is no material or other physical phenomenon present except the phenomenon under consideration"

to

"The concept of free space is an abstraction from nature, a baseline or reference state, that is unattainable in practice, like the absolute zero of temperature"

..is exactly the difference between "nothing" and "something", as it is clear to me you wish to bring in the modern ideas concerning things in a vacuum into the definition, however, by definition, they are not "at issue". That which was *at issue*, the permittivity and permeability of free space, you dropped.

Jok2000 (talk) 22:04, 16 April 2008 (UTC)

Just wanted to say, nice job on editing the article. J. D. Redding 07:51, 17 April 2008 (UTC)

By the way, I have a couple of those Tesla 10,000,000,000 dinar notes. I feel sorry for the guy who got his picture printed on the 10 dinar note. Also thanks to Reddi for putting back the permittivity & permeability of free space. Saved a few dinars worth of effort. Vacuum should go with "vacuum", though. Jok2000 (talk) 14:08, 17 April 2008 (UTC)

The Faraday Paradox
Brews, I think that your problem lies in not being able to see where the vXB term fits into the flux rule. We are agreed that the partial time derivative Maxwell-Faraday law cannot explain motionally induced EMF. We are further agreed that vXB is needed to explain motionally induced EMF. Hence we are agreed that the Lorentz force is needed to supplement the modern (Heaviside) Maxwell's equations in order to cater for motionally induced EMF.

What Feynman has failed to see is the mathematical link between the full total time derivative flux rule (Faraday's law) and the vXB effect. That link was explained in a paper in the Toth Maatian review, January 1984.

A total time derivative splits into a partial time derivative and a convective term of the form (v.grad)B. That convective term is in fact the curl of vXB.

Hence, if we use the full expression for E, as per Maxwell's equation (D) in his original eight, and take the curl, we will end up with a total time derivative version of the Maxwell-Faraday law.

The problem nowadays is that the electric field term E = F/q is only considered to apply to stationary points in space. Hence we never write the expression E = vXB for the component of the electric field that is induced by motion.

This is what has caused all the confusion. George Smyth XI (talk) 07:08, 20 April 2008 (UTC)


 * Regaring the Lorentz force and Faraday's law, once again, I think that you now have the correct physical picture in your mind.


 * But unfortunately I can't give you an informed view on the maths in relation to the integral versions. For veracity of the maths, you might be better to seek the opinion of somebody like Steve Johnson who is an applied maths professor and sometimes contributes to these pages.


 * I'm not planning on altering your section. The only advice that I would give would be to leave it for a few days. Read it again and stand back and ask yourself if this is the clearest way to express the physical point that you are trying to make.


 * If you are absolutely sure that your maths is 100% correct as regards integrating the vXB effect into the integral versions, then I suppose that there is no harm in having it there. George Smyth XI (talk) 06:54, 24 April 2008 (UTC)

Your diagrams
You've drawn some great diagrams, kudos for those. Would you consider learning to use Inkscape, so that you contribute your diagrams in SVG format (vector graphics - making it easier for others to build on your contributions)? Also, would you be able to use the proper multiplication symbol (×) instead of the letter x when multiplying in your diagrams? Thanks! All the best, Stannered (talk) 19:46, 1 May 2008 (UTC)

Moving magnet and conductor problem
I've rewritten the intro to moving magnet and conductor problem. Having done that, I'm left feeling that the use of A is a much better approach than the E and B method, because E and B introduce the false dichotomy between electric and magnetic fields, mainly due to lumping the solenoidal and conservative E-fields together. These two components of E-field should be kept separate, as they have different origins and different behavior under switching of frames of reference. A cleaner formulation would leave out B-field and refer instead to A-field, while E--field would refer only to the conservative E--field. Of course, the gauge-transformation freedom makes you think that the potentials aren't "real", but one could argue that the E- and B-fields aren't real either - it's only forces and currents that matter. Differently put, we'd have the E-field redefined as only the conservative part and redefine the B-field as B = −∂ t A + v ×  curl A. Would that fly better? (I'm not suggesting Wiki be rewritten like this, just thinking out loud.) Brews ohare (talk) 13:11, 2 May 2008 (UTC)


 * Brews, the most important point that you have made here is the fact that E has got two components. And just as you say, these two components are quite different in nature. One is a radial conservative field with a source, and the other is a curl affair. Most people are unaware of this split.


 * From a purely hydrodynamical analogy, if we treat A as fluid momentum, then (partial)dA/dt becomes a force. It will have a radial component and a tangential component. The radial force will be your conservative inverse square law E term. The tangential force will remain in the form (partial)dA/dt and be associated with vorticity. Curl A will then of course be vorticity. You could view B as a vorticity.


 * That should help you realize the relationship between the two E's and B.


 * Gauge freedom only exists if you don't know the physical significance of A. Once you know what A means, then there is no gauge freedom.


 * It's a bit like the acceleration of a car. It can't tell us the exact speed. But if we know the exact speed then we are no longer at liberty to guess the speed.


 * I personally don't mind using E and B, because in all the scenarios which you are talking about in EM, we know that E is always referring to (partial)dA/dt. When have you ever been using an E in EM induction with the Coulomb force in mind?


 * On balance, my answer to your question would be 50/50. David Tombe (talk) 18:30, 2 May 2008 (UTC)

Centripetal Force
Brews, your derivation of centripetal force on the centripetal force page was correct in principle. The physics was correct. The symbolism was perhaps a little overcumbersome. The underlying principles are simply a vector diagram showing that the centripetal force is radially inwards in the limit. Then using a bit of trigonometry we establish that v = rω and hence acceleration equals v^2/r or rω^2.

And just as you say, there is no other force when viewed from this Cartesian perspective. We are looking at a radial direction which is rotating.

But you cannot overlook the issue of 'inertia'. Inertia is the tendency of a body (in the Cartesian frame) to carry on in a straight line motion unless acted upon by a force.

In the circular motion case, that singular force is the centripetal force.

But 'inertia' itself becomes centrifugal force when we view its effect radially.

Do the exact same vector diagram again, but this time consider a straight line path referenced to the same origin. No centripetal force. No circular motion.

This time the vector diagram leads to an outward pointing acceleration of the same magnitude.

Hence centrifugal force is implicit in inertia.

Centripetal force in circular motion is merely cancelling out with inertia, or in the case of polar coordinates, it is cancelling out with centrifugal force.

The physics is exactly the same, irrespective of which frame of reference we do the calculations from. David Tombe (talk) 07:35, 3 May 2008 (UTC)

Centrifugal Force
Brews, There are a number of points to clear up.

(1) The edit war has been principally about the fact that these people here have been trying to deny that a centrifuge works BECAUSE it is rotating. They have been arguing that centrifugal force occurs when we view something from a rotating frame of reference. The two concepts are quite diffent.

I have been trying to insert the cause and effect aspect into the main article, but it gets deleted instantly every time.

These people think that a rotating centrifuge is equivalent to observing a stationary centrifuge from a rotating frame whose axis is on the axis of the centrifuge.

Clearly you can see that this is nonsense. A centrifuge will not be made to work in that manner. There is no equivalence principle involved in all of this.

However, there is a group here that are trying to promote the equivalence principle, and they are quite wrong. They are denying the age old Bucket argument.

(2) Orbital theory. Polar coordinates show up both the centrifugal force and the Coriolis force. However, Kepler's law of areal velocity eliminates the Coriolis force term. No Coriolis force is involved in the gravitational field.

We are then left with a radially inward gravity force and a radially outward centrifugal force which is absolutely real.

(3)Action - Reaction. In the real scenario, when actual rotation occurs, we get a radial centrifugal force, which as regards the issue of action-reaction, behaves exactly like gravity. However in the purely fictitious situation in which we observe a stationary object from a rotating frame, any effects are only fictitious and Newton's third law will be totally irrelevant. The group that are controlling this article are focused exclusively on the latter scenario. David Tombe (talk) 04:14, 11 May 2008 (UTC)

Science debating club time is over
You're right. I'll stop. It's frustrating; over and over again, I kept expecting him to have an "aha!" moment any moment, but he's had this described to him in every conceivable way, to no avail. And it's stopping us from writing the encyclopedia. Have you seen my suggestions at the bottom of Talk:Centrifugal force? -- The Anome (talk) 13:15, 22 May 2008 (UTC)
 * Thanks for bringing your suggestions to my attention. They are very nice ideas to improve the article. I left a note on the talk page. Brews ohare (talk) 13:20, 22 May 2008 (UTC)