eadbtxaawxyxzcbw

bogatin

's profile
image
Dean, Signal Integrity Academy

Eric is the Dean of the Signal Integrity Academy, www.beTheSignal.com, and an adjunct professor in the ECEE department at the University of Colorado, Boulder. He received his BS in Physics from MIT and PhD in Physics from the Univ of Az in Tucson. He can be reached at eric@beTheSignal.com.


bogatin

's contributions
  • 02.07.2014
  • Total loop inductance/length in 50 Ohm transmission lines: Rule of Thumb #6
  • MaleEngineer- yes, you are correct- it is 16.6 nH, not pF! And yes, rules of thumb are powerful tools to hav in your tool box to help you get to the appropriate answer faster. Be sure to read the Rules of Thumb 0 to see the right time and place for rules of thumb, over other tools. The first place to go for an answer should never be a field solver. http://www.edn.com/electronics-blogs/all-aboard-/4424571/Rules-of-Thumb--0--Using-Rules-of-Thumb-Wisely
  • 01.03.2014
  • Rule of Thumb #4: Skin depth of copper
  • Roy- Good comments. It is rather counter intuitive, but since the skin depth goes as SQRT(1/conductivity), when the conductivity is low, like copper oxide, the skin depth is really large. This means that in the very thin region of the surface of a copper trace that may be oxidized, the skin depth is huge compared to the geometrical thickness, so skin depth plays no role. This is not the case for ENIG. The plated Nickel's conductivity is about 2%-10% that of copper. It will have a larger skin depth than copper, but still small enough to influence the current path. This and the slightly higher permeability, about 2-5, means above 1 GHz, much of the current will see the higher resistivity of the nickel plating and the series losses can dramatically increase. Plating thick gold or silver on the outside will reduce the series resistance of the outer conductor and effectively hide the nickel plating. See this column I wrote about the work Yuriy Shlepnev and Scott McMorrow did on this topic: http://bethesignal.com/impact-from-electroless-nickel-on-microstrip-insertion-loss-and-propagation-delay/
  • 12.16.2013
  • The Bloch Wave effect
  • Hi MaleEngineer- The reason I brought in the concept of Bloch Waves to describe the "structural return loss" effect is partly because I am a physicist and learned about this effect when I first studied solid state physics and partly because I wanted to illustrate an important principle of science and engineering. I am always astonished how just a few principles can be applied across a wide spectrum of effects. The wave nature of voltage signals, electromagnetic fields, sound, even matter waves all behave exactly the same way. The mathematics that governs one type of wave applies to all the others, since they are all a result of a second order, linear differential equation. If you like the idea of structure return loss, then you're ready to understand band gaps in semiconductor crystals! If you understand band gaps in crystals, then you can understand the glass weave induce insertion loss suck-outs. Learn one principle, apply it to many effects.
  • 12.16.2013
  • The Bloch Wave effect
  • VectorForce- thanks so much for reporting this error. i don't know how I let this misspelling through. We still have 2 "No Myths Allowed" mugs to give away for errors found in our content. If you send me your mailing address off line, we'll send you one. (eric at bethesignal.com)
  • 12.08.2013
  • An important secret about transmission lines
  • I've taught over 10,000 engineers over the years. Probably 75% of them are fixated on the LC approximation to transmission lines and don't realize that a transmission line has its own individual set of important properties that dominate how signals interact. For example, while the characteristic impedance is sqrt(L/C), the importance of the characteristic impedance is that it is the instantaneous impedance a signal will see as it propagates. In a uniform line, there is one value of instantaneous impedance that "characterizes" the transmission line. It is not about sqrt(L/C). Engineers should downplay the n-section LC model and embrace the transmission line model.
  • 11.19.2013
  • Rule of Thumb #1: Bandwidth of a signal from its rise time
  • DaveE- thanks for sharing this derivation. What a coincidence, that a simple, intuitive estimate of the bandwidth of a signal should match the same value calculated based on the -3 dB point of a filter. Of course, in the case of this rule of thumb, we are really identifying the highest frequency component in the SIGNAL. There is no -3 dB cut off, there is a brick wall cut off in higher frequency components above the bandwidth. No other assumptions other than it sort of looks like a modified square wave. In the case of a filter, we are relating the -3 dB point of the filter and the 10-90 rise time of the transient signal that would come out if an ideal step went in. While these are not exactly the same thing, they hint at the underlying consistency of thinking of the world bilingually in both the frequency domain and the time domain. I think this is a very important skill for all scientists and engineers.
  • 10.28.2013
  • How to think in dB
  • Wow- you guys are really good at finding the tiniest little errors. I have to be on my toes when I write for such a sharp eyed group. First, to Don J- I had the numbers for the sound level for the Saturn V, but not specifically the numbers for the sound level of the Space Shuttle. I used these as two illustrations of similarly really loud sound sources. I suspect the Shuttle is s little quieter than the Saturn V, but it is also much more fragile. To Nidan.danny- you are correct that I should have asked in my teaser, is it an amplitude or a power. However, when describing the "voltage" of a signal, I think I got this one correct. When S21 is - 3dB, the ratio of the amplitude coming out to the amplitude going in is 10^(-3/20) = 70%. Nidan is very correct in pointing out the ratio of the powers coming out to coming in is 1/2, but the ratio of the amplitudes will go with the sqrt of the ratio of the powers: sqrt(1/2) = 70%. When you deal with S-parameters a lot, i find it very convenient to skip the intermediate step and think only of the ratio of the voltages which we can get directly from the dB with that factor of 20. You guys are going to force me to stay sharp.
  • 01.10.2011
  • FPGA Technology at the pointy end of the spear
  • I think the goal of many comm systems is to deliver a quality of service with no errors in the lifetime of the product. Of course, it is impossible for the physical layer to provide this, let alone, as you suggest, to test it. This is why there is an overhead in the data transmitted for error correction. A BER in the phy layer of 10^-12 in a 20 Gbps link will be an error once a minute or so. It would not work without error correction. I should have been more clear that I was referring to the system BER, not the phy layer BER. thanks for the clarification.