Internet commentary

Citation

Ellis, B. (2003), "Internet commentary", Circuit World, Vol. 29 No. 3. https://doi.org/10.1108/cw.2003.21729cag.001

Internet commentary

Keywords: Internet, Viruses

• •

  A cause of annoyance or anger; a bête noire [1]

How many times have I harked back to the same subject, viruses. At the time I am writing this, a new one, Bugbear, has been assailing us. I must admit that this one is cleverer than some of the older ones and is, perhaps, more mischievous. It is both a worm and a Trojan horse. One of its features is that if A's computer is infected, like all the other recent worms, it sends malicious E-mails with infected attachments to everyone in an Outlook address book, but it sends them as if it came from B, not A, where B is one of the addresses, to which the virus is not sent! So, C, receiving it, can never find out where it really came from but, if he has an anti-virus checker, he thinks it comes from B, who then receives all sorts of messages accusing him of dire acts or even threatening him, when he is not even infected! Of course, this extra traffic all conspires to further reduce the Internet bandwidth available.

We have also received a new kind of hoax, lately, many times over (the last one just today). The message reads,

"I am very sorry about this.A friend's Address Book has been infected by a virus and it was passed on to my computer. My Address Book, in turn, has been infected. The virus (called dbgmgr.exe) is not detected by Norton or McAfee anti-virus systems. The virus sits dormant for 14 days before damaging the system. It is sent automatically by messenger and by the address book, whether or not you sent E-mails to your contacts. Here is how to check for the virus and how to get rid of it...''.

It then gives precise instructions on how to delete the debug manager which is an integral part of every Windows system (and has a cute little Teddy Bear icon). It could have been a lot worse if the perpetrator of this sick joke had chosen a more essential Windows system file. As it is, there are probably thousands of users who no longer have a debug manager on their system. I know this hoax has been translated into many languages by well- thinking persons who have blocked the Internet backbone by sending it to all their addressees. How sad! By the way, if you have been taken in, you can easily restore the missing file from your Windows CD-ROM by starting in a DOS window, changing to the CD-ROM directory and using the EXTRACT command. Put it back in your base Windows directory, usually C:\windows or C:\winnt. You can find the syntax for your system, first, by typing "extract /?'', if you have never done this before. Better still, you can reinstall the whole of Windows over the existing one. I find this is a good thing to do, every few months, in any case, because it automatically restores the system to its optimum, without touching the precious registry. If you wish to make sure that no problems arise with your existing set-up, go to Start Run and type "regedit''. Click on Registry and make a backup of your existing registry files by clicking on Export Registry Files. You can then restore them by importing them again in the very unlikely event of something going wrong with your system. In an enterprise system, ask your administrator to do all this for you, of course.

I am going on strike! I refuse to tell you again how to prevent your Windows computer becoming infected with real viruses and related problems. I have done this so many times in the past that you, dear reader, must be fed up reading the same things, time and time again, just as much as I am writing them! Unsympathetically, if your computer is infected, you deserve everything you have received, in trumps!

Of course, one of the weaknesses of modern personal computing is Windows. If only we could change that for something better! It is true that Linux is more reliable, although a few viruses are beginning to attack it, as well. One disadvantage of Linux is that it is possible to find some of the weaknesses, although they are fewer than with Windows, by examining the source code. In the past, I have tried various dialects of Linux, including Corel and Caldera. I have never been able to configure these correctly to my systems and always something has not worked correctly. This may not mean that the operating system is faulty but more that drivers are missing or, worse, essential information cannot be found. I have tried to put a recent Red Hat version on my office computer, dual booting with Windows. This is not something for either the faint-hearted or the computer-illiterate to do. If you are lucky enough to have a straightforward installation, allowing you to boot into the KDE X-windows system, well and good, but this has never happened to me. In this case, after a couple of false starts with errors and dire warnings splashed across my screen, I thought I had succeeded. That is until I tried to log-in with either the administrator or a user password. It simply would not accept it, or any other, in graphics mode. I can do so in command line mode from the boot diskette and, from there, most things appear to work (except that I have not yet managed to configure my network card). Many E-mails and several hours of trials later, the Red Hat techie tells me that there is a problem with my graphics card (even though it is a very popular one, sold literally by the million and especially to corporate users) and could I try a different one? This is not the way that Windows will be supplanted. I understand that about 10 per cent of all PCs are now running Linux and that about half of these are dual-booting with Windows. Many of the 5 per cent that run exclusively on Linux are in dedicated systems, so Bill Gates is hardly quivering in his shoes, yet. My brief trials, at different times, have shown me that Linux has much to offer in stability and reliability but it can take off only if installation is trouble-free ands it works "out of the box''. Another example, is that I have not been able to find out how to configure my laser printer for double-sided printing. However, I must say that I appreciate the Internet access of Red Hat Linux, behind a good, configurable software firewall. This is the first time I have been able to do this on any of the systems I have tried and, being a Netscape user, I valued the Mozilla browser and E-mail client in it. If only I could get that darned password in with the X-windows set-up, I would be a happy bunny.

I was intending to look at the three types of simulation for my review section. A quick glance through what was available made me realise that this was a mammoth task that could never fit into one article. I, therefore, decided to forget the digital and realtime simulation and concentrate on analogue simulation. I have reviewed ECAD systems in the past and these generally cover everything from the schematic diagram through to the Gerber files, sometimes with the possibility of using in-system simulation or imported systems. Probably the most famous of these (for analogue circuits) is PSpice ( http://www.pspice.com). This latter used to be an independent company, which has unfortunately been taken over by an ECAD company, that has tied it in essentially to one of their products. This has discouraged competitive ECAD suppliers from supplying interfaces between their systems and PSpice. In my research for this article, I came across the site of a UK company offering independent analogue simulation with very decided advantages, not the least of which is price. However, from what I understand, it is analogue simulation with a difference, based on four independent softwares that do not use the Spice algorithms. This company has been around in the business for 20 years, so it cannot be considered as a Johnny- come-lately. I understand that some ECAD systems have two-way interfacing into their main software, called SpiceAge, so that integrated simulation is possible. I propose to make this a single-site review, because I am sufficiently taken with what I see on the site that I feel this company probably deserves to be better known. It can be found at some of these sites.

http://www.spiceage.com/

Do not be put off by the appearance of the Home Page; I know it is pretty hideous! The basic design looks as if it was drawn up about 5 years ago, but it has merit. It is fast to load, has no enormous graphics. It has a tiny JavaScript routine to show the date (why?) as the only piece of superfluous code, although I do not like the marquee-scrolled address just under it, either. It has all the meta descriptions and keywords for search engine spiders to find. It has a simple text and button menu system to advance further into the site. It also has an address for sales in other countries. Quite frankly, this page does not inspire confidence and I doubt whether many surfers would advance beyond this page – but I did!

The sub-page I propose visiting first is the SpiceAge one, second in the list; I will come back to the Spicycle one afterwards, for reasons that will become clear. This is the basic simulation package. This is only slightly informative but starts to whet the appetite, even if the background schematic shows such components as a beam tetrode (come on, be honest, how many of you know what a beam tetrode is!). However, the interest level breaks the top off the thermometer when I click on List of Features. If you compare this list with the corresponding lists of many other simulators, well, you can not – it seems to have everything you can find in all the other systems together and more besides. I became a little sceptical, at this stage, and I realised that there are probably not very full component libraries, although there are generic components, but more of this later. From here, I clicked on Analysing with SpiceAge and it was this that made me less sceptical, although the examples given are quite simple. Please allow me to make a brief comment on each one. The first one is a simple series resonant circuit with a resistance and the outputs are given as the reflection coefficient, the transient response ("ringing''), the internodal voltage vs. frequency response, the current amplitude and phase vs. frequency response, the response as a filter and the response to a square wave of lesser frequency to the resonant frequency. I find it a pity that these curves are not slightly larger on the screen, because the text is unreadable. So, I am slightly impressed at this stage. The second analysis is an unadorned family of curves showing the frequency and amplitude response of a simple 741 operational amplifier with different values of feedback resistor. This is followed by the frequency and phase response of a three-stage passive LC bandpass filter, which can, of course, take component tolerances into account. The next example is a fifth order low pass Chebyshev active filter, using the ubiquitous 741 operational amplifiers and it shows the frequency and phase response and the transient response. Then comes a single stage transistor amplifier in common emitter mode, where the input signal overloads it, showing cut-off and saturation, and a Fourier analysis of the resultant distortion. This example, in the accompanying text, describes one of the advantages of this system over its competitors, by judicious simplification of the models allowing faster simulation to be achieved. The next analysis is of a 555 astable oscillator and clearly shows why the first pulse is longer than the succeeding ones (as is known for most relaxation oscillators). From here, we go into something a bit more complex, a phase-locked loop using an XOR gate as the phase detector. You can see how the output comes into lock with the pilot signal. The maker claims that "Phase locked loops have always presented a challenge to simulation software because of the convergence problems associated with the control loop. Not so for SpiceAge.'' The next demonstration is of a diode mixer. The output waveform is as one would expect and the FFA shows no fundamentals, the plus and minus main components and the spurious frequencies and harmonics due to the non-linearity of the diodes. I would have loved to have seen the effect of diode tolerances! Two opposing pulse sources feeding the gate of a triac, with a variable firing angle demonstrate a fade-out of load current in the next case. This is followed by something I am very familiar with, the plotting of the mutual conductance of a KT88 beam tetrode, by stepping the grid voltage and sweeping the anode voltage. The non-linearity of the valve (tube) is easily seen. It also shows that the anode voltage has relatively small effect on the anode current above a certain threshold, varying between about 50 and 150V, according to the current. Straight back to my student days! The final case study is that of a high- power valve (tube) amplifier consisting of a two stage preamplifier, a long-tailed pair phase splitter and a four-valve power stage in parallel-push-pull, without negative feedback. The frequency response is sensibly linear from 10Hz (that must have been a darned good output transformer!) to 30 kHz. The transient response to a 20 kHz square wave shows the expected rounding off of the corners and the distortion due to the non-linearity of the valves. I would have loved to see the difference of adding some negative feedback through the transformer, although I suspect the model for this was not very sophisticated.

This shows that the simulator proper is able to cope with a wide variety of types of circuitry, so what are the other three softwares for? Let us go back now to the first one, Spicycle. This is a general draughting software and is available in two versions. The more important is designed mainly for drawing the schematic diagrams, designating the variables and interfacing it directly into SpiceAge. All the examples that I described above have their Spicycle schematic. This avoids the tedious task of inputting the data as node-to-node text: this is done for you. The second type adds what appears to be a rather crude and simple printed circuit draughting system. It does not seem to have an autorouter and I was not able to determine whether the distributed capacitance, inductance and resistance of the PCB could be brought back into the simulation. Wonderful if it could!

The third software is called Modelmaker. Its purpose is to allow you to create your own component models, quickly and simply. All you have to do is to fill in a form from the manufacturer's data sheets. It gives one example for a small signal NPN transistor asking for the Hfe, Icmax, ft, Cob, Rbb and Hoe. It is claimed that the resulting model works much faster than the corresponding Spice model, although it is implied, rather than stated, that the Spice model can also be used if wished.

The remaining software must be a great help to anyone who needs to design filters of any type (passive, active or digital). It is called, appropriately enough, SuperFilter. From the description, it enables designs to be made automatically, presumably from the desired characteristics as input, although information is a little sparse. It apparently draws the resultant response curves and it says the characteristics can be checked in SpiceAge. As the two softwares were written by different persons, it seems that such a check may use independent algorithms.

While writing the above paragraphs, I have downloaded the demo versions of Spicycle, SpiceAge and the manual. I propose breaking off the writing at this stage and I will tell you how I get on with it, in practice...

...OK, the installation and interfacing between the two modules were no problem. The software seems to add minimal data into the registry and nothing into the Windows directory, which is always a good thing. Most of the demo models used for the analyses I mentioned in an earlier paragraph, and others besides, are available as samples. I added to the valve amplifier the negative feedback that I wanted and I was able to obtain a substantially flat response to 100kHz, with less gain, of course. As I guessed earlier, the output transformer has an ideal characteristic and there does not seem to be an easy way of changing this, at least at first sight. I suspect this is possible with the ModelMaker module, which I did not download, as there is a demo with a loosely coupled mutual inductance. Some of the proposed demonstrations are quite complex and permit the introduction of distributed variables and parameters. Particularly impressive is the Fourier analysis of a filtered conductor being tested for EMC emissions (Plate 1). Another tricky one is the effect of mismatching two transmission lines at both ends, with the results of the reflections at both the Tx and Rx ends with a pulse input.

I then went into drawing a simple schematic. Although most of the operations are intuitive (I checked the DC response against temperature of a Wheatstone bridge with resistors of varying PTC and NTC values within a few minutes of my first try), a few of the operations are slightly less so. For example, there is a tool-bar row of symbols, including semiconductors, for drawing the schematic diagrams. This is good and simple to use with passive components (including putting in values, tolerances, temperature coefficients, operating temperature and so on). But you are stuck if you drag down, say, a transistor. There is no way to specify its characteristics. However, you can draw a transistor of the type you wish from a library, but I admit, in my brief excursion into this software, that I did not find out how to edit the characteristics. I am sure that this must be possible, as some of the components are quite generic in nature. Maybe ModelMaker can do the trick.

Plate 1 Simulation of EMC, with conducted emissions to a filtered power cable. The equivalent schematic is above, the signal from a transient at bottom left and a Fourier frequency analysis with a Hanning window at bottom right. (The two curves, reduced in size to fit, have been imported from the SpiceAge module.) Also seen is the Spicycle window layout with the toolbars and status line

Time has not permitted me to delve deeper into this demonstration version. I would have liked to have tried interfacing it with my Vutrax ECAD software (yes, I did find I have the interface for this). However, it is quite possible that this may not have been feasible as the limitations of the demo version are mainly around copying and saving (I must admit I prefer "time-bombing'', with or without something added to the output to reduce the usability, such as a file size limit.). I should perhaps mention that the downloaded manual is not really an instruction manual so much as a detailed catalogue of what Spicycle can do, but it is an interesting reading.

All in all, SpiceAge must represent one of the best values for money in the analogue (or mixed) simulation world. As much as anything can be stable in these volatile times, it would seem that the makers are here to stay, as they have been around for two decades and claim to have sold over 23,000 systems, which is no mean feat, in itself. One wonders, then, why they are so relatively unknown. I think they deserve better.

Brian EllisCyprus b_ellis@protonique.com

Note

Oxford Concise Electronic Dictionary definition:bugbear / / n.

1. 1.

   a cause of annoyance or anger; a bête noire,

2. 2.

   an object of baseless fear, and

3. 3.

   archaic a sort of hobgoblin or any other being invoked to intimidate children. [obsolete bug "bogey" + bear2]