Author Archives: The Tarnz

DataPlex revamps PrefixSuffix.com

DataPlex enhanced PrefixSuffix.com, a major online English language reference site. DataPlex engineers added clear navigation, engaging content and revenue generating ads, converted a static word root list to a MySQL database, then greatly expanded it, while adding some “sticky” educational games.

These enhancement were designed to make the site more attractive to more visitors and to keep visitors on the site longer. One week after going live with the improvements, PrefixSuffix.com experienced a four-fold increase in traffic and and twelve-fold increase in revenue.

Free Internet Node, Part 3 – First Free Internet Service: A Website

The first service we will install is a web server because it is typically the main purpose for an Internet node. It is your “HTTP service” that establishes the network presence of a website and responds to browser requests. Fortunately, this is one of the easier services to set up.

You have many choices for HTTP servers ranging from free software to expensive software from reputable companies. Since we are doing this project on the cheap, let’s check out the free software.

Of all of the free software, the one with the greatest traction (almost two thirds of all websites) and ongoing support is the open source HTTP server project from Apache Software Foundation. Some readers may wonder – isn’t Apache’s server for Unix and Linux systems? Not quite. Historically, Apache’s best running version has been the one for Unix/Linux with the Windows version somewhat of an experimental poor cousin. However, years of development have yielded a fine Windows-based Server every bit as good as the original.

Go to the Apache’s website and read about their HTTP Server Project. You will see that there are essentially three versions of their HTTP server for Windows, versions 1 and 2.0 and 2.2 (ignore the notes for Linux and Unix as those are different operating systems). Apache HTTP Server 1.3.xx (‘xx’ is some number representing the current release) is an older release that is only now supported for bug fixes and security plugs. Apache HTTP Server 2.0.xx is a newer, more feature rich version, and Apache HTTP Server 2.2.x is a major new release with many added features.

While some techies might jump right into 2.2.x, I was a bit more conservative. 2.0 is a serious improvement over 1.3, and it has several years of stability and security bulletproofing, therefore, version 2.0 is the one that I recommend today. Besides, my notes that follow are all based on version 2.0, so if you decide to go with 2.2, you’ll have to translate my notes accordingly, and hopefully you won’t run into any serious conflicts.

At this point, I feel that I should add a **Warning**: This project is educational and is direct towards amateurs wanting to set up their own Internet node and learn the details about web and email servers and the like. If you are really concerned about reliability, uptime, providing services to clients, security, etc., then you might consider having your service hosted by a mission critical ISP, you know, the one with a data center with 24 x 7 support? Hey, don’t complain later that your site went down while you were vacationing in Jamaica! (Actually, there are tools for that — I will try and make a “vacation support” post at some point.)

Okay, we are at the point where we download and install the Apache 2.0.xx HTTP server. Go ahead and do that, run the install and check back here.

Once you have installed the server, you should have an entry in you Start Programs folder for activate the Apache HTTP server. You might also have an icon on your desktop. Either way, start the service. That’s it! You now have you own web server running! To try it out, go to your browser on your Internet node PC and type in “http://localhost” and see the Apache server’s default home page.

Of course, you will want to change the contents of that home page. So, let’s do that now but creating a simple HTML web page. Open up wordpad or some other Windows-based text editor and type in the following:

<html>
<title>My First Free Web Page!</title>
<body>
<p>This is my <em>very first web page</em> for this website.</p>
<p>Click <a href=”http://www.google.com/”>here</a> to go to Google.</p>
</body>
<html>

Hopefully I am not going overboard by throwing in some html language elements. If you already know HTML, then that is a good thing, and you will recognize the above example as being a crude minimum page where I define a header with a title which will appear in the title bar of a browser’s window and a body which contains some simple text and a hypertext link to the search engine Google. Knowing HTML at this point isn’t that important if you already have a website and are just going to port in over, or if you have a friend or employee who takes care of your HTML. Still, if you would like to learn more about HTML, there are numerous HTML references and resources on the web.

Save this file as “index.html” in the apache’s default folder located in the folder:

C:\Program Files\Apache Group\Apache2\htdocs

Now, when you type in “http://localhost/” into your browser window, you should get your new web page or an error if you typed it in wrong. By editing this file or by using a variety of HTML editing tools, you can transform this web page into a spiffy looking home page for yourself. In the next part, we will show you how to port over existing website files you may already have.

Okay, remember your IP address? If you go to another PC attached with access to the Internet and in a browser you type in your IP address in the format “http://209.175.22.103″ (where you substitute your four numbers), you should also get your sample website home page.

If you recently set up a domain and it has been 1 or 2 days, you can try entering the domain name into the browser (e.g. “http://www.yourdomain.com”) and see what you get back. If the domain name has not rippled through the DNS system yet or if your DNS record has been incorrectly set up, then you will get an error. Otherwise, you may get your sample website page. If so, then congratulations are in order! You have successfully set up a domain name and its corresponding website.

Let me point out a few rough edges here. First off, we are using the default Apache folder which is okay but not great in terms of organization and security. Also, we haven’t fine-tuned the server’s parameters for our site as we just left them at the defaults so certain server abilities are not available to us (yet). In the next several parts, I address these issues and show you how they can be resolved.

Next Up: Free Internet Node, Part 4 – Porting my Existing Website and DNS

Free Internet Node, Part 2 – Setting up the Internet Node PC

If the PC you have selected to be your “Internet Node” is not already hooked to the Internet, you should take the steps many other people have to get it permanently attached with a static IP address. Because of their reasonably high speed and low-cost, I recommend interfacing through either cable or DSL with either a cable modem or DSL modem, respectively.

I recommend Windows XP as your operating system since older versions of Windows simply do not have the reliability or the security, especially with the release of SP2 for Windows XP. Windows Vista is too new, and many experts suggest waiting until late 2008 before using a Windows Vista machine as a reliable server. The last thing you want is to have your Internet node compromised and shut down because a hacker used a vulnerability in your PC.

Some experts suggest running Internet services on Linux boxes (computers) which they consider more streamlined and less prone to attack, and that is okay if you have the experience and expertise. However, my purpose here is to show how to transform an “ugly duckling PC” into an “Internet node swan,” and Windows is what most PC users are familiar with.

I leave it up to you to decide if you need Windows XP Professional over XP Home. I use both, and they both seem to operate equivalently for our Internet purposes here. XP Pro is more tweakable and has additional features, but none of that is necessary for a typical low-bandwidth node.

The absolute best way to set up Windows XP is to restore the factory image of your PC harddrive from the Restore CDs that can with your PC. If you lean more towards being a PC expert, then you can clean up your PC, but please be thorough.

To clean up your PC, first remove any software you don’t want compromised, any financial spreadsheets, documents, etc. Then eliminate all extraneous software that you won’t be using through their Uninstall procedure or Add or Remove Programs in your PC’s Control Panel. Run ScanDisk, Defrag, and scan all of your drives with your antivirus software after making sure it’s virus list is up to date. You should also run anti-spyware software to make sure that your PC hasn’t already been compromised by hackers.

Once you have your Internet Node PC cleaned-up, you can verify your Internet speed to see if it is going to be able to handle the bandwidth you will require. A good rule of thumb is… If your bandwidth seems low, you might review some higher bandwidth services from your telephone or cable companies.

The next stop is to figure out what to do about your domain name, e.g. “yourdomain.com”. If you already have a domain name, then you will inform your current registrar about your IP address change, so that they can change your DNS Record. (If they cannot because they do not provide that service, you may have to switch registrars.)

What’s all that about the DNS Record? Well, the global Internic Registry used for decoding domain names into IP addresses does not, as some might assume, contain a direct pointer to your IP address but instead a pointer to a DNS Record that has the ability to direct the different services and sub-domains of your domain to unique IP addresses. Typically, your registrar maintains your DNS record for you and haven’t had to deal with it until now. Live and learn! Some registrars allow you to create your own DNS Record but will also do it for you through their customer support. If you want to learn how DNS Records work, it is certainly educational from a “how does the Internet work” perspective, but not a requirement for our task at hand.

If you do not have a domain name, now is the time to get one. First, choose a domain name you like and that you think will be easy for others to remember and use. Then, sign-up your new domain using one of the registrars that are authorized to provide domain names. We recommend assign a domain name that ends with the top level domain “.com” unless you have grand reasons for using another (e..g. non-profits typically use “.org”). If your domain is already taken, the registrar will let you know and offer alternatives. You can go with one of those or iterate and pick another domain name to try and register.

There are many low-cost registrars that will sign-up a new domain name for under $15 a year, so don’t get suckered into paying $25 or even $35 a year. The last registrar I used was aplus.net which charged me $11.90 for a one-year registrar of a “.com” domain including DNS service. Note if the registrar you use changes extra for DNS service (or even provides it – some registrars do not) as you will need it in order to have your registrar point your domain to your IP address. Normally, if they host your site at one of their own IP addresses which by the way is a huge profit margin for them – we are bypassing this typical configuration.

Okay, so you have registered or transferred your domain or changed your domain’s DNS Record to that you can . Now the bad news: It is going to take up to 48 hours for your change to propagate through the DNS System so, when someone, including yourself, types in “www.yourdomain.com” in a browser, it is directed to your new IP address. Do you now have to wait for a day or two until you can access your services?

Checking and Setting Up the “hosts.” File

The answer is a definitive “no,” and you can keep going in setting up your Internet node even which your domain name addressing is in transition. First off, you can simply type in your Internet Node’s IP address and service port number to get to a service. But don’t do this yet as we haven’t installed and activated any services yet – we are saving that for the next installment.

Also, you can edit a file on one of your local network PCs that tells your PC to go locally, even to itself, when it sees particular domain names. This local redirection will work also on the Internet node PC you are setting up. This file is called the “hosts” file (actually the file name is exactly “hosts.” with a period and a blank extension) and is located somewhere in one of your Windows folders, somewhere in “c:\windows\…”.

You should use Window’s find feature to locate this file – search for the file “hosts.” starting in “c:\windows\” and include subfolders. On my Windows XP Home PC, the file is in my “c:\windows\system32\drivers\etc\” folder. Open it in Notepad or Wordpad. Wordpad is preferred since it remembers previously opened files in case you have to reedit.

There should be many lines starting with “#” that explain in a semi-obtuse way the use of the hosts file. Basically, all you need to do here is add a couple of line in the following format:

127.0.0.1 yourdomain.com
127.0.0.1 www.yourdomain.com

The IP address “127.0.0.1″ is one that means “this PC” and reroutes the specific domains right back to the current PC without attempting to go out over the Internet and resolve the domain name, which, as we know, is in transition. So, at least for the next few days, we will be working with this redirection and you can continue to set up your Internet services and test them out. Then, once your domain’s URL is relocated, your Internet services will be all ready to go without delay.

If, in your hosts file, you already have a line:

127.0.0.1 localhost

that is fine and you can leave that line there as well.

While we are on this topic, let’s do a little security review. Any other entries, lines that begin with an IP address (without the ‘#’) may be a hijack by a virus, spyware or hacker, so if you see anything that looks suspicious, you might want to check it out and delete it. For example,

254.22.172.14 citibank.com

is very likely a hacker intercepting the citibank.com domain and redirecting it to a phishing site in order to steal your citibank online ID and passcode. If you see anything like this, your PC was at least at one point compromised and may still be, so rerun all of your antivirus and anti-spyware software first making sure you have the latest updates. Even better, use alternate software to evoke a wider gamut of detection.

Next Up: Free Internet Node, Part 3 – First Free Internet Service: A Website

Free Internet Node, Part 1 – Equipment and Location Evaluation

Okay, let’s get started. Remember, I am going to describe my actual approach and make more general comments to cover slightly different hardware and approaches. Please feel free to chime in by adding comments. (By the way, we have a few what-should-be-obvious rules for posting comments.)

The first step I took in setting up my own Internet node was to evaluate my potential equipment and locations. Depending on the amount of traffic I might expect, I would require different bandwidths and therefore would need to set up my equipment and choose my location accordingly.

Given my estimated low load of, say, under a hundred or so hits per hour, I decided to locate the node at my residence on its existing DSL line, noting that I could always scale upwards in case my traffic takes off or I add what become a popular website. I determined that I could get by with my existing 512 kilobits per second DSL modem and one of my old PCs as the Internet node. (DSL lines can actually go up to 1.5 megabits per second depending on your DSL service. You can test the speed of your current DSL connection.)

For reasons having to doing with noise and power consumption, I switched over later to a new eMachines mini-tower PC which cost me around $250 at Best Buy after all the rebates came back. Here’s another eMachines pricing sample along with some specifications directly from eMachines themselves.

Anyway, just make sure whatever PC you use is less than a couple of years old and, preferably, running XP Home or Pro. You can upgrade not too old PCs without any big hassles. Not to promote Windows XP too much, however you can optimize your security access, benefit from XP’s advanced features, and add hardware with minimal effort. (You can use older PCs running older versions of Windows such as Win98, but the end results will be sub par and your security will only be slightly better than nil.) The PC should have at least 20 gigabytes of available space, use an Intel-based processor running at least 2 GHz, and be very quiet if it is going to be within earshot. The latest version of an eMachines PC that I bought easily met this criteria.

What about Windows Vista? It may work well actually, but we did all of our work on Windows XP, and we haven’t found a compelling reason to upgrade. From what we hear, Windows Vista is a different animal when if comes to PC operating systems and there are incompatibilities with some software packages. All it would take is one small module not working to put a serious damper on things. Anyway, the incompatibilities with using Windows Vista may diminish greatly after Microsoft’s final release of Service Pack 1 (SP1) which is scheduled for early 2008 (a release candidate is available now). Anyway, I may post a Windows Vista report later since that is the operating system most easily accessible today. For now, downgrade to XP or just buy a used PC with Windows XP on eBay.

So, with my existing equipment of an older tower PC and a DSL line and modem all of which could already access the Internet, I encountered my first dilemma. The typical $30 DSL line to a residence (such as through SBC) requires a log in. Until your equipment logs in, you do not have an IP address. And should you have to log off (or be forced off) for any reason and then sign back in, you will have a different IP address.

Web servers, email servers and ftp servers are best located on a static IP address, a permanent address on the Internet that never changes. With the low cost residential DSL line with a dynamically allocated IP address, you would have to change some network parameters each time your IP address changes. (I am being vague about the “parameters” as that would be a digression, and I do not want to lose focus right now.) While that is possible, it is a technical pain, and, depending on the reliability of your DSL provider and your telephone lines, you might have to change the parameters several times a month. Also, your node will be down until your equipment logs back in and changes its parameters, maybe a good part of a day.

Fortunately, I found that I could upgrade to a static IP address for an extra $25 a month. I feel that paying such a small fee to not have to deal with changing IP addresses. If you disagree, you can explore the dynamic IP address route with either DSL or cable modems and let me know how it works out for you.

The change from a dynamically allocated to a statically allocated IP address took less than 48 hours. Since I use a LinkSys router, I did have to switch it over from PPPoE to Static IP Addressing. Note: This step is different for different routers but is typically achieved using your browser to access your router’s parameters at your local URL address http://192.168.1.1.

Thusly, I was able to evaluate my existing equipment and location and found both acceptable with the one change to switch over to a permanent IP address. I estimate this effort took about four hours of my time, most of which was dealing with the phone company and the router.

Next Up: Free Internet Node, Part 2 – Setting up the Internet Node PC

Free Internet Node, Introduction – How to Set Up a Free Internet Node & Web Server

Hello and Welcome! My name is “The Tarnz,” administrator and blogger ordinaire for DataPlex.com, so please bear with me as I get things bootstrapped and organized. Feel free to make comments and I will respond when appropriate. (Thank you!)

Over the past couple of years, I have set up a node on the Internet equipped with a bunch of free services and technologies including:

  • web server
  • ftp server
  • email server
  • database engine
  • forum engine
  • blog publishing engine
  • PERL, PHP and other languages
  • a bunch of utility software tools

As you may imagine, I have been through a lot dealing with the hardware and so many software packages, so I thought I would share the experience and occasional challenges in case anyone else is geeky enough to set up their own free Internet Node.

Okay, maybe the node wasn’t completely free for the value of own efforts, the cost of any new domain name registrations, and the cost of a low-end Windows XP Home Desktop PC and, optional but highly recommended, a reliable Uninterruptable Power Supply (UPS). But it sure beats paying $200 to $2,000 a year to an ISP for what can often be very bad hosting. Any problems and enhancements are solely my own.

Today the node has a fully functional multiple-domain web server, email server, ftp server, database and blogging engines along with many fully-implemented technologies such as Perl, PHP, AJAX, SQL and more.

Whoa, wait a minute! A half-dozen, fully-implemented Internet services, for free, on an eMachines running XP Home? Surely, you jest, Tarnz ol’ boy! That’s the domain of IBM blade servers and other vertical niches rack mount hardware costing thousands of dollars!

Not necessarily. In the past few years, several new technological developments have made it possible. Severe cost-costing competition has led eMachines (owned now by Gateway) to release surprisingly high quality PC boxes for only a few hundred bucks. The release of Apache 2 for Windows (yes, free) turns an XP Home PC into a screaming top notch web server.

But that’s just the beginning. I still had to set up the email and ftp servers, deal with security and spam, develop a backup and restore plan, and handle the everyday maintenance issues that come up. Even though it has been some work, I find myself being rewarded in a number of ways – for example, understanding better the Internet, setting up the servers exactly the way I want, paying much less for domain names, website and email support. I will get into those particulars that too.

Disclaimer: This project is educational and is direct towards amateurs wanting to set up their own Internet node and learn the details about web and email servers and the like. Of course, we cannot be responsible for any problems you encounter, although if you post comments, we will do our best to help. If you are really concerned about reliability, uptime, providing services to clients, security and all that, you should have your service hosted by a mission critical ISP.

Setting up your own Internet server, or Internet Node as I call it since in actuality the node’s PC has running on it several different types of servers, is not for everyone. High volume and mission critical applications should be run in professional 24 x 7 installations. But if you are a consultant like me, then hosting your own Internet Node is less expensive, educational, and not necessarily any less reliable than what I have been encountering with several ISPs.

Anyway, if you stick with me, you will come to understand what I went through, and maybe my experience with embolden you to set up your own node as well. Or, maybe it will make you run to your nearest ISP?! If you do set up your own node, you might then share with me your deviations, enhancements and clever insights to make me smarter about all this too.

Next Up: Free Internet Node, Part 1 – Equipment and Location Evaluation

DataPlex unveils PrincessMania.com

As part of its ongoing study of Internet sales and optimization, DataPlex established PrincessMania.com, an online store with a specific marketing theme. Powered by several technologies from Google and amazon.com, PrincessMania simplifies the selection of princess-theme gifts, books, DVDs, electronics and toys. To aid site administrators, it has associated with it a collection of easy to view sales and marketing reports.

Implementing State-of-the-Art Audience Response Systems

Volume 3, Number 4

“This game is simple,” says Bob Sagat on NBC’s 1 vs. 100 television game show. In this NBC show that airs in prime time on Friday nights in the United States, multiple choice questions are asked, and people are eliminated as they answer incorrectly.

In a “1 vs. 100″ game, a contestant is pitted against one hundred other people, known collective as the “mob,” and collects money for each mob member who gets eliminated. The contestant attempts to increase the pot and take home a sizable amount of money, either the accumulated total of the pot or a million dollars if all members of the mob are eliminated. If the contestant answers incorrectly, the remaining mob members who answered the last question correctly split the pot up to that point.

What may appear “simple” on screen is the result of the successful operation of a multitude of integrated systems in what has to be one of the most technically advanced game shows ever conceived. Behind the scenes is a major IT effort, controlled by a custom version of DataPlex’s state-of-the-art Audience Response System (ARS).

1 vs. 100 set under construction showing four rows of handsets

A view of two of the eight rows
of voting handsets during
construction of the 1 vs. 100
TV game show set (U.S.)

Mob members are placed into the various “pod” locations and often rearranged at the discretion of the director. Each pod has is own graphics display behind the player, microphone and voting handset. There are groups within the mob — lawyers, cheerleaders, janitors, kid geniuses — that are each tracked statistically. Our ARS system supplied through our client Quick Tally Interactive Systems for 1 vs. 100 has a player registration module that is used to set up the demographics and print badges with barcodes for each potential player. Once the mob members are in place in their pods, portable scanners are used to associate all mob members to their locations whose data is then processed by another one of our ARS modules.

Now, how simple is that?

The director likes to know which pods contain returning mob members (mob members who continue to answer questions correctly are carried forward into another game) and how well have they performed in the past. Often, there is a “reigning mob champion,” someone who has answered a significant number of questions correctly over several games. A sophisticated SQL database platform is used for managing the ARS data from multiple games. In a matter of seconds, the ARS data from previous games is processed along with the player location information of the current game, and a report is generated.

There are other IT processes, for example, if a mob wins, a list of the remaining mob members and their information is generated for the show’s accounting department. Also, post production uses demographic-based reports to show interesting factoids on the bottom of the television screen.

Not Your Father’s Audience Response System

A typical Audience Response System, also known as an ARS, collects votes and can generate a limited amount of graphs — bar charts of answer choice selection percentages, pie charts, and some cross reference displays. More recent systems can export directly or indirectly to Microsoft Office™ products, for example to PowerPoint for presentation purposes and Excel for further off-line number crunching. Reasonable stuff, actually.

DataPlex’s experience with audience studies started when in 1980 its early client ASI Market Research wanted to convert an analog dial system to a digital version.  The digital dial version was a huge success and is still used today for allowing audiences to evaluate movies, television shows and commercials before general release.  The audience feedback often had a significant impact on a show’s editing or whether a commerical was shown or not. In the 1990′s this system was expanded for remote voting outside of theaters and for supporting text-based answers by survey respondents.

In 1988, DataPlex started selling its DataPlex DataKeeper, a handheld mobile computing device that featured a world clock, professional time billing, a mileage logger,  and an easy-to-use database manager.  This database manager was used by several customers to conduct surveys where at the end of the survey period, the survey information from each device would be downloaded to a central PC, consolidated, and summary reports produced.  In the mid 1990′s, the DataKeeper received wireless communication capability where survey information and votes could be monitored and analyzed in real-time.  Around that time, several companies developed wireless ARS system packages specifically to handle surveys and votes in a localized region such as a meeting room or a conference hall.

The demands on ARS continue to grow. No longer are more demanding customers content with single location polling and simple summary bar charts. They want to know what’s behind the summary results. They want to compare ARS results across multiple sessions and locations and also correlate the results with those of their other systems.  They want to delve into the data and statistically pull out significant information that will help them improve their business.

On the high end, ARS are expanding above being merely a localized vote gathering tool. Now, it is asked to number-crunch, produce custom reports, feed other systems in real-time, and perform advanced digital and analog I/O (input/output) control.

Metadox ARS performing additional I/O

DataPlex Metadox ARS
Software driving ten individual
player video monitors for
Reader’s Digest’s Word Power
Challenge Championship

In another Metadox application supplied through DataPlex client Quick Tally Interactive Systems, the Reader’s Digest National Word Power Challenge, the application additionally controls ten video monitors with custom statistics for each of ten contestants. Metadox generates special graphics for each display, including showing each player’s name in large text and a red “X” when a player gets a question wrong. Different methods of revealing which players got what question correct or wrong keep the proceedings lively by preventing monotony.

Implementing a State-of-the-Art ARS

With the older ARS, you install the software from a CD that arrives in the mail or downloads from a website, configure it, and you are good to go. On these systems, as features are added, the configuration can get bogged down, and some of the new features may not work with your particular installation.

A state-of-the-art (SOTA) system has the pertinent features tested and qualified by an engineer as part of the process of delivering an exacting product. This step includes testing any hardware add-ons and writing and testing any custom programming code.

Once delivered, a SOTA ARS will require some additional fine-tuning because the installation is brand new. Fortunately, with technical support only a phone call away, most issues can be dealt with quickly.

In extensive applications that require much more customization than is typical, the ARS purveyors might recommend that a feasibility study be performed as a first step as a way to document the scope and detail a reasonable plan for development, lest the project run amok, which of course, is to no one’s advantage.

We are Engineers who Customize for Specific Applications

DataPlex’s niche is providing engineering and consulting services to meet exacting, custom requirements for existing or brand-new systems. We pride ourselves on providing quality products with appropriate levels of training and support. For more information on audience response and voting systems for your own needs, please visit our clients’ sites below.

For more information, please visit:

DataPlex to Develop Electronics for MRI’s M-20 Industrial Tester

DataPlex has been selected by Measurement Research, Inc. (MRI) to develop the realtime control electronics for MRI’s new M-20 Industrial Tester.  The M-20 electronics will receive sensor data and control several mechanical assemblies positionally driven by different types of motors.  DataPlex will design the custom electronics, develop the firmware and Windows application, and also manufacture the M-20 circuit boards.  DataPlex had previously developed and manufactured electronics for MRI’s M-10 and M-15 product lines of Industrial Torque Testers.

The Advantages and Pitfalls of Updating Old Electronic Designs

What is an Old Design?

By “old,” we mean systems that were designed more than a generation of technology ago. “Generations of technology” are usually indicated when the class of the components are superseded by a new one. Here are some examples:

  • Microprocessor controlled hardware replacing hard-wired circuits. (There are many advantages for doing so; the resulting design is often less expensive and more flexible.)
  • Digital analog to digital converters (ADCs) and digital to analog converters (DACs) replacing analog circuits. (Typically, a microcontroller site between ADCs and DACs to provide digital processing of the digital version of the signals.)
  • HCT Series CMOS superseding LS Series TTL for driver and interface circuits. (CMOS has faster switching, lower power and can be made smaller.)

A particular hardware circuit is comprised by different sections, and each section may be different numbers of generations old, so a weighted analysis can be used to explore trade-offs of a design upgrade.

For lack of any other measure, an electronic design can be considered old if it is more than ten years old. The electronics industry as a whole advances quickly, and few electronic components are used in the same way ten years later.

One useful question to ponder: If a qualified engineer were to design a circuit today for the same application, how would it be different? The answer helps to point us in the right direction. A new electronics design will utilize parts that are currently available and have a complete yet cost effective feature set. If the design were to be significantly different, then that is a strong argument for not perpetuating the approach of old design, although.

That said, an engineer might still choose to perpetuate an old design based on economic, compatibility and political factors. He may wish to use up existing supplies of old parts, or continue to provide products which meet with customer expectations. Something more technically modernbut significantly different may be perceived by customers as inferior or undesirable. Other technical reasons relate to support and repair and intellectual property issues.

How an Old Design is Updated – The Nine Step Process

Once the decision is made to update an old design, the update process is straightforward. Here is the general process we use at DataPlex:

Step 1
Locate or Develop a Specification of the Current Design

If it was not the first step to break inertia, this should be the easiest step. Old manuals and product datasheets are often useful references.

Step 2
Identify the Key Features of the Current Design

A bit more challenging, identifying key features requires an understanding of the products and, not only how it is used by customers, but also how is it not used. Features that are under-utilized but require significant ongoing support may need to be dropped.

Step 3
Develop a List of New Features, Ordered by Priority and Usefulness

Probably the most difficult step, choosing which features remain and what new features need to be developed has to take into account what are the expectations of the marketplace and weigh them against the up-front engineering expense, the per-unit production cost, and what alternatives provides and what competitors are doing.

Step 4
Research Components and Design Approaches

The design team researches components and explores overall design approaches that seem to optimize between cost, reliability, serviceability and an appropriate feature set. For economic and development schedule reasons, not all features will make the cut (hence the need to order them).

Step 5
Begin Design Approaches

Begin to develop in a top-down approach several different designs to see how then begin to flow. Often one design will leap ahead of the rest as a better or less expensive way. Pick it as the primary design.

Step 6
Review the Designs

Explore the designs with management and (very important) also with key customers to see how well their feature sets fit with their actual needs and requirements.

Step 7
Iterate the Design

At this point, a design team may iterate back to an earlier step should, with management and customer feedback, it become obvious that none of the design alternative sufficiently meet expectations.

Step 8
Start the Primary Design Phase

Once a design seems to be acceptable, the actual primary design phase begins. While it may feel as though the actual design work is a long time in coming, actually most of the difficult decisions have already been made, the complete function set is known, and it is just a matter of implementation. Backup designs can also be begun now or held off for a later time. Also, the design team should work with management to prepare contingency plans should something go wrong.

Step 9
Don’t be Afraid to Go Back

While the main design phase is underway, should any new information indicate that the results may not meet expectations, management should decide which earlier steps should be reviewed to make the necessary adjustments to bring the design back in line. In some occasions, an alternate design replaces the originally preferred design based on the new information. Within reason, it is much better to take some extra time to develop a product that is right for the marketplace rather than push throw with one that is known to be inferior.

When the Design is Complete

When the design is complete, we move on to prototyping and testing.

What Can Go Wrong?

Updating an old electronics design is still an electronics design process and can derail for any number of reasons. For updating a design, a design team must be sure to meet the important targets of feature set, backwards compatibility, production rates, delivery schedule and price points.

It is especially important to get one or more prototypes ready as soon as possible so that the targets can be evaluated by different groups such as management and key customers. These prototypes can also be used for showing to potential customers and demonstrating trade shows while production ramps up. At all times, be vigilant for discrepancies between the prototype and the marketplace’s expectations and acceptance and review Step 9 if necessary.

Should something go wrong — say, production doesn’t ramp up quickly enough to meet the delivery schedule – there is no need to panic. Management should have already developed contingencies such as using more expensive vendors but ones that can fill in the production schedule as necessary.

One area that is particular debilitating is so vast that is deserves its own section…

Intellectual Property Legal Issues

While we do not practice law or offer legal advice, this article would not be complete without a discussion about intellectual property, both developing it and being sure not to step on others.

If it comes to your attention that you are stepping on someone’s intellectual property, you might see what it will take to license it. If the patent assignee’s (owner) is not a direct competitor, you should find a warm reception since you will potentially be providing a new revenue stream.

Or, if you think the updated design improves on the current state-of-the-art for the product class in a novel and unobvious way, you might explore applying for a patent. We won’t tell you that patents are easy — they are not, but should you get one, you will have a government sanctioned monopoly on the invention disclosed in the patent for close to 20 years, hindering competitors and providing a cash flow from licensing.

Note Well: There are ongoing changes to both domestic and international patent law, and you need to be informed about the precise process you will have to follow. For example, as of this writing, a United States invention must have its patent filed before it is first offered for sale lest it jeopardizes its international validity.

In any case, should you feel that you are in such an industry or have a product that requires or steps on intellectual property, gather your design team and consult your patent lawyer.

What to Do if You Have an Old Design

Please consider DataPlex as your one-stop design team. Our products can be found in many different industries and chances our experience staff and proprietary design tools can get you a redesign in quick order. Also, with our being technical and having been successful in written what are now issued patents, we can help you document any new inventiveness so that you can more readily apply for intellectual property.

For more information, please visit:

Maintaining High Reliability for New Electronic Designs Intended for Harsh Environments

Volume 3, Number 2

The Issues of Harsh Environments

Designing a piece of electronic hardware such that it works reliably from the get-go is a challenge, especially if it has to work outdoors or in extreme environments where temperature, humidity, vibration and radiation are the enemies of electronics.

Unfortunately, while a growing number of new electronic designs are destined to be used outdoors, the art of designing for harsh environments is not typically part of an engineering school’s curriculum. As a result, most engineers make the erroneous assumption that designing for a harsh environment is very much like designing for hospitable one but with a greater emphasis on testing and improving the design empirically, addressing the observed failure modes. Although this approach is certainly along the recommended path, it should not be the only extra consideration; otherwise, the deployment could fail.

What is a Harsh Environment?

A quick definition of a harsh environment would be anywhere not “indoors,” although this definition is incomplete. Any condition of extremes relative to the human condition applies, so that includes temperature, humidity, atmosphere (including pressure), radiation and shock, whether indoors or not. While one can easy to see that sitting outside at the south pole with minus 50 degrees winds blowing at 80% humidity easily meets the requirement, it isn’t readily apparent to him that a handheld device that could be dropped five feet also meets the requirement.

A simple test is asking the question, “if this device were a human being (scaled up or down as appropriate) and subjected to the conditions of its environment — the highest and lowest temperatures, the amount of pressure, or amount of shock, would it be expected to survive?” A cellphone operates in a harsh environment, indoors, because it can be dropped, and the amount of force it can hit a tile floor would be terminal to an appropriately scaled human being. (It often is to the cellphone too depending on the height of the drop and the quality of its design.)

The Engineering Approach

Addressing the harsh environment begins before the design phase, in the R&D phase. Knowing that the design includes a harsh environment, an engineer sets the research parameters and, later, the product specifications accordingly. Paramount of concern to him are meeting the minimum and maximum storage and operating condition limits.

To maintain high-reliability for a new design, the engineering approach must:

  • list the extreme storage and operating conditions
  • set the design specifications accordingly
  • select components that meet the specifications and reject others
  • be mindful of circuit board physics, e.g. trace lengths, component mounting methods
  • explore conformal coating and heating and cooling elements, as appropriate
  • ensure that the fabrication drawing is consistent with the harsh environment design
  • select appropriate mounting hardware, cabling and enclosure
  • test the sensor inputs and software over the complete range of operating conditions

“Testing over the complete range of operating conditions” is easy to write, but could be very difficult in actual practice, owing to the exponential increase of possibilities over an increasing number of parameters and ranges. We have found that Monte Carlo Simulation and linear programming are two very useful aids for analyzing the effects of different parameter combinations.

Post-Design & Post-Installation

No design for harsh environment is complete without a plan for follow-up during the implementation or deployment phase. A well-designed circuit may allow additional or alternative components to allow for tweaking the final design after production. For example, in one DataPlex design, a standard delay line circuit slowed down so much in cold temperature that an alternate delay circuit was automatically switched in, and the switch point was controlled by a variable resistance that could be easily changed in the field.

Software is Affected Too

It may not be obvious at all, but software can be affected by harsh environments as well. We are thinking here less of the actual computing hardware that runs the software — that is still covered by the hardware related issues discussed above — and more of the way that the software has been designed and tested in a warm, cozy lab. Software that received sensor data such as values from analog to digital converters (ADCs) may find itself attempting to deal with extreme values that it did not encounter in the lab, and that usually leads to bugs and possible failure.

Taming Harshness

In this article, we explored the nature of harsh environments, what constitutes a quality engineering approach to design a piece of gear for those environments, and the importance of follow-up in order to make adjustments and preempt out-of-spec operation or catastrophic failures.

If you have a requirement for a harsh environment design, we invite you to consider the experience of DataPlex’s staff with difficult environmental conditions and how our refined approach to R&D and engineering can mitigate certain risks for brand-new designs. Please contact us to discuss your project.

For more information, please visit: