Damon's Laboratory Notes

Why are Engineering Tolerances important to design?

Sat, 26 Jul 2008 12:40:00 GMT

Anyone who has driven on the Jackie Robinson Parkway [Queens, NY] has already gained an appreciation of Tolerances. The posted speed limit is 40mph, but even that is too fast. Most drivers new to the parkway try to take it at 55 or 60. Big Mistake. There is a curve going to the Van Wyke or GCP where you can see all the marks made by cars as they side swipe the concrete and stone wall. The lane has a tight tolerance with respect to the width of a car and it makes it hard to navigate.

Many years ago I worked closely with a fellow who had a machine design shop. He designed machines for all sorts of industries. He was a mechanical engineer by training. He relied on my company to provide the electronics for his machines. I had an opportunity to ask him about his mechanical designs and he explained that the most important aspect (to him) was tolerancing. For instance, if you create a drawing for a one-off part used in the aerospace industry, it might be acceptable to specify a 0.1 mil tolerance on all your dimensions. Realize what is going to happen. When the machinist gets your drawing he might have to throw out the first 5 parts he makes because he had trouble holding that spec. Eventually he will learn how to make your part to that tight tolerance, possibly he might have to design his own fixtures to help him hold the part as he machines it. Eventually, with much waste, he will make your part. Your part will be expensive, but it may be less expensive than the engineering to realize a 'better' design.

A good mechanical design orients the parts so that they do not need to be manufactured with a tight tolerance. If you have ever assembled furniture from IKEA you might have noticed the large over sized screw holes to allow the part to 'float' with respect to the attached part thus alleviating the need for a tight tolerance.

One might think that a very good design then tolerates loose tolerances on all its parts. Such a design could be manufactured 'cheaper.' Well, if you remember back in the '70's how the Asian car manufacturers were advertising how well the parts of their car fit you think otherwise. Loose tolerances could give the perception of a cheaply made product to the consumer. You want your product to be manufactured inexpensively, but not appear cheap.

Tolerancing can take on a different aspect to the software professional. Back in 1981 John Postel coined the "Robustness Principle". In it he advocates that a browser should “Be conservative in what you do, be liberal in what you accept from others.” This holding of loose tolerances has resulted in a problem. See this link. For a browser the specified tolerance should be as tight as possible. Only in this manner can software manufacturers duplicate the performance of the browser over the range of web sites.

What conclusion can we make? Sometime a tight tolerance is good, sometimes its not good. But it is ALWAYS important to your design.

The Foreclosure Issue

Fri, 18 Jul 2008 10:45:00 GMT

Is anyone else besides me concerned about the proposals going through Congress these days. Seems to me like they are getting paid off. Why else would the politicians be trying to save these bad loans. They were bad loans the day they were wrote. No one writes a loan for 110% of equity without significant risk. They lost (the banks) their bet.

Lower costs for housing are pro family and good for EVERYONE.

You want to hear my proposal. Let the homeowners walk away from their loans. Thats right. Let the homeowners walk away and legislate a plan for the homeowners to regain their credit ratings. The banks made a bet when they lent the money, its not the place of the Fed or Congress, or the President to remove risk from the market. There is a risk/reward mechanism built into the market already The government is not welcome to give my money to the hedge funds and banks or to guarantee these loans (this is a disguised form of giving away my money.)

Secondly, if the government wants to stimulate the economy, do it like they have always done.Tax rebates, business incentives for investing, and lower energy costs. Yes, lower energy costs are possible, but thats another lab note.

The Devil Teaches me about being Anal

Mon, 26 May 2008 02:39:00 GMT

If you know me you know I am a very detailed engineer. In fact I was working with an engineer many years ago and this engineer told me I was 'anal' because this engineer thought I was overly concerned with getting my pcb layout just right. Could it be that wanting to get a design right is not a good thing? It did not bother me too much what this engineer actually thought but it did pique my curiosity about the meaning of this word, and about productivity, perfection and 'being anal.'

I have heard the term anal before. I have talked to many people about what they think the word means and have gotten many different answers. I got this definition from the Wikipedia:
--------------------
Conversationally, the term is often used to describe a person deemed to be overly obsessed with minor details. Its roots are said to be from the theories of Sigmund Freud.

The term is often used in a colloquial, derogatory sense to describe a person with such attention to detail that the obsession becomes an annoyance to others, and can be carried out to the detriment of the anal-retentive person
---------------------
It was the Devil* though that gave me the best definition of what is an 'anal person.' This is my working definition. He told me that someone who is anal is beyond perfectionism. Perfection is, or can be, productive. Anal is beyond that and is the realm of non-productivity. So for instance an anal person might spend inordinate amounts of time perfecting some area of documentation that does not benefit the actual purpose of that type of documentation.

It is a good thing then for a person to do their job and do it well as long as you are productive. The perfection exists in getting the design right, making your schedule, and attending to all the other details that make money for your employer. I have always endeavored to keep my eye on the 'big picture.' This means understanding all the pressures on the design and focusing my attention to the details that matter in the right places.

*The Devil appeared to me in the image of an ex-business partner, who we called "The Devil."

How to find a short on a Printed Circuit Board (PCB)

Thu, 27 Mar 2008 15:56:00 GMT

When I am troubleshooting new PCB's I have a unique technique for finding shorts on PCB's. I actually have to hand the credit for this technique to a business partner, Dr. Ed Kafrissen. He came up with a few non-traditional ways of doing business, and this is one of them.

So let me set up a scenario for you, and then show you Ed's solution. Suppose you have a client and you need to deliver 100 PCB's to him/her quickly. You have someone stuff the boards for you, now you have to verify they work. One of the first tests we would do is to apply power to the new PCB and bring the voltage up slowly. If the power supplies load down, then you got a problem to solve. You got a short on the PCB.

Ed came up with unique solution. We had this old humongous power supply lying around that was capable of suppling 30v at up to 5 amperes. This technique only works if the power supply, like ours, has an adjustment to limit the voltage and current. We would set the voltage to necessary voltage and the current all the way down. Ed would then hook up the PCB, which immediately loaded down the current limited supply. He would turn up the current on the supply and feel on the PCB for a hot spot! Invariably some IC would be stuffed in backwards that our visual inspection did not find, or we would get a PCB which was not completely etched. In those days we got horrendous PCB's.

Anyway, that was Ed's solution to quickly find shorts on a PCB. And it worked well as long as you were careful enough to not burn yourself! Yeow!

Simplifying the Programming, Concrete Example

Sat, 16 Feb 2008 02:53:00 GMT

I would like to offer a concrete example of simplifying the programming of a product. Years ago I was privileged to be given the opportunity to realize the programming and design of a home automation controller. This was the DHC Toscana home automation controller, still being manufactured by Leviton Mfg. Co.. When we first sat down with management, marketing, sales and all the other interested entities at that company I was literally buried in the number of features on this product. One of the other senior engineers and software guru's there, Dr. Michael Ostrovsky, helped me along and gave me a simple suggestion that cut my project in half!

To understand his suggestion I have to give you a brief technical description of the product. As you can see from the picture the device has a number of buttons and user interface elements to control lights, fan's, switches and other electrical devices you might find in a home. It also has a communications interface where it can receive remote commands from other interfaces devices in the home to control the same electrical devices. Therefore, an engineer might think he had two intertwined overlapping tasks to accomplish. First would be to control the electrical devices from the unit's front panel and transmit that command to other keypads in the home. The second task to control those same electrical devices from the commands received over the communications interface. The contention that could develop from controlling the same devices from several inputs is a classical problem. Mike suggested a better way. Mike was old school, Mike was from Russia. He knew how not to do the same task twice.

He suggested I split my project into two simpler tasks. The first was react to one of the front panel switches being pressed and to transmit the appropriate command indicating this on the communications interface. The second, separate task, was to receive a command over the communications interface and to turn on the light, fan, device or what not. To complete this scenario I had to let the device receive its' own commands, the one it just sent out! This way I only had to debug two separate smaller paths, that were more or less independent.

My 1 best secret to Great Software

Fri, 15 Feb 2008 14:09:00 GMT

Recently I have been thinking about the projects I have been involved with and what helped them become sucessful. I think I finally have something to report to you here in my laboratory notes as to MY 1 best secret to sucessful software.

My first thought was to look at the details of programming languages and techniques. I then thought about on a higher level how the architecture of the software was more important. Then I figured that even more important than architecture was the project management. I think I finally have a root cause to successful projects. Keep the project 'simple.' Now that word is a relative term. The exact nature of 'simplifying a project' depends on the skill of the software team. Second, it does not follow that a simple project produces a simple product. Read On.

I have developed a basis for this claim that this is the 1 most important ingredient to a successful project. As the number of requirements in a project grow, the software complexity grows exponentially (albeit this is a simplistic view) as the following chart shows.

The trick is then to define the project in simple terms yet design a dynamite product. What I want you to think about is where you want to focus your effort: Eliminating elements from the X axis of the chart above or troubleshooting program paths on the Y axis.

Nokia HF-6W defect, no sound

Tue, 15 Jan 2008 01:16:00 GMT

Nokia has some serious design problems... and they are not telling anyone! Let me explain.

About 2 years ago I did some shopping for a Bluetooth car speaker phone. I hit upon what I considered the most perfect speaker phone, the Nokia HF-6W. It cost me over $120, but it was worth it (while it lasted). It had a powerful 6w speaker and several microphones that tracked the sound source. It worked great as well.

About a year ago it died. I assumed it was a 'normal' failure. I returned the unit to Nokia returns and they sent me a new unit. This second one failed in the same manner. Further, after doing a search on the internet I found other people are experiencing the same failure of the Nokia HF-6W and here. Nokia's warrenty only covers the unit for one year. This is a design flaw. Nokia should admit it and do right by their customers.

The way the HF-6W speaker phone fails it that no sound comes from the speaker. The speaker is dead. Second, the unit will not reset itself. When the power button and volume up buttons are pressed simultaneously while the device is off, the unit is suppose to revert to its default state. The green and red LED's are suppose to blink as well. I cannot seem to get it to reset. The unit seems to work otherwise, bluetooth connects, etc.

Now my wifes unit, unit #3, failed the same way. Have you stumbled upon my post and are having the same problem?

[UPDATE 02/04/08]
Nokia replaced my HF-6w with their HF-33. It has a much smaller speaker and no built in retractable phone charger. The retail price is about 1/2 that of the HF-6w. I am accepting this replacement under the condition that it simply WORK.

The Deuce bus service Las Vegas

Sat, 12 Jan 2008 15:55:00 GMT

The "Deuce" is the bus that runs up and down the Las Vegas strip. It's an inexpensive way to get from downtown to anywhere on the strip. I have a story to tell about the Deuce. I had chance to use it several time while here at the CES show.

First you need to know that the name "Deuce" is a triple entandra. The bus is a double decker bus, and a trip costs $2. Some Las Vegas civil authority must have thought it would be cute to name it that. Second thing you need to know is that whenever the bus stops a pre-recorded announcement plays inside and outside the bus. It announces "Ride the Deuce, Have Fun. Only 2 dollars one way, 5 all day."

I want you to know about a couple inaccuracies in that announcement. The bus IS NOT $5 all day. It is $5 for 24 hours. Twice when I got back to my room I threw away the bus pass thinking it would be invalid the next day. So there, reading damon4.com just saved you money.

A second inaccuracy is tha the bus IS NOT fun. It IS pleasurable. It's clean. It is usually on time (they run every 20 minutes and run extra buses during heavy traffic). I must have rode that bus a dozen time waiting for the 'fun'. The most fun I had was watching people hit their heads as they sat in the lower level seat's on either far left/right side of the bus. There are low ceilings in these double decker buses, so watch your head!. The bus (and drivers) is a real pleasure to ride; what a bus should be.

The civil authorities in Vegas who run the bus service should rethink that 'fun' part of the message. I would also like them to consider adopting a different slogan. Instead of the "Deuce" call the bus service "A Nice Pair". So you would have the same double entandra PLUS a whole new set associations to work with!
Viva Las Vegas!

Ultra Wide Band (UWB) at CES

Fri, 11 Jan 2008 16:23:00 GMT

There were a number of companies showing Ultra Wide Band at the show. On display was silicon, modules, and finished products. Two applications for UWB are for pumping HDTV around the house and for wireless USB. With USB it can go the full 480 MB/s spec. With HDTV it is unclear to me if it can replace an HDMI 1.3 cable. That is rated at 10.8GB/s.

At my work we have application for UWB as a high speed cable replacement. I found at least one module vendor willing to talk to lower volume manufacturers.

More on Surface

Wed, 9 Jan 2008 01:00:00 GMT

In my previous note I said I do not understand the application for the Microsoft Surface. I want to explain a little more about it. They say the hardware is a 30" (at least screen). It could be mounted in any orientation. They use projected DLP. Also under the display are five infrared cameras as sensors. The cameras are necessary to pick up the multi-point input. The Microsoft sales person said it does not have to be one person inputing gestures, could be several. Applications we discussed were as a kiosk in a retail store where a customer could customize a snowboard virtually. We discussed how this could be used for video games or recreation. It could go into a board room as the board room table. I chimed in that it could be applicable for a draftsman, and the salesperson agreed.

Next I asked if this would appear on a tablet. The Microsoft guy could not see how since the hardware did not seem compatible. He did not rule out that some new touch screen interface could be invented, or admit if it already had been invented. When I mentioned the IPhone he immediately stopped me. Said it was apples and oranges. The IPhone touchscreen is a capacitive technology and has some limitations as to the number of inputs.

I stand by my last opinion. The surface computing seems to be gimmicky. If Microsoft pushes it, it will have some success. I don't see any market pull. I also think they need to further develop along these lines. With some different hardware the software might make a package more appealing. Who knows what could come along in the future?