(C) 2005 Hank Wallace
You design electronic devices. That’s how you are wired. Your customers or company superiors come to you with problems, and you come back with solutions. Your palette is a million electronic components and techniques, your brush is your brain, your painting is the schematic, code listing, copper artwork, or mechanical drawing. You live to create and you are good at it.
Have I hit the mark? Thought so. When someone describes a problem to me, I see the circuit design spontaneously forming in my mind. Algorithms. Hysteresis. Digital to analog and back. Pulse width modulation. Amplifiers, filters and gates. People like us are despised by telepaths. (Don’t know many telepaths, do you?)
However, there is a problem with your solution, and it’s not noise or thermal runaway. It’s the fact that you might, just might, be infringing on someone’s patent. Ever worry about that? What a nightmare it is when the fax machine clicks on and out rolls a nastygram from Slime, Ooze and Guck, Attorneys at law.
In my freelance work, I deal with a lot of patentable products, and help my customers prepare patent documents so they can protect their intellectual property. That process forces me to spend time digging through the patent database, trying to determine as best I can (as an engineer, not a lawyer), how our methods mesh with prior art and the patent space. Not a simple task, and one fraught with ambiguity. As an engineer, I of course despise ambiguity, and I expect you do as well.
Now this ambiguity has many causes. Patent attorneys intentionally word patent applications as vaguely as possible, in an attempt to gain the broadest coverage possible for their clients. Add to that the sheer size of the patent space, not only in the U.S. but worldwide. It is impossible for even a small army of humans to thoroughly explore the patent space for matches in technique, letter and spirit, not to mention the endless possibilities that some minor prior art would invalidate the patent.
That said I still like to blame the patent whole mess on the lawyers!
The net result is that patents must be interpreted not alone (except possibly where your design exactly matches a described technique), but together with all other issued patents and the entire history of mankind. As they say, after the revolution, the first people with their backs against the wall will be the lawyers.
Much commentary has been written on the patent system, and I’m not here to regurgitate that. A simple internet search will load you with reading material for a week. However, years of scanning patents have brought to my attention several choice examples that lay claim to the most basic of methods in our art, and it is some of those I seek to entertain you with. After all, not only is the patent system the spawn of attorneys, it is run by the government, so there has to be some unintentional sadistic humor buried not too far below the surface. The best gags are perpetrated by people who don’t know they are being funny.
The patents that follow may be active or expired. Perhaps they were litigated out of existence. Fair enough. But at one time, they were issued and valid. Let’s see what’s out there.
Which of you has ever used a current source? Have you ever connected a resistor to a voltage source then to a fixed voltage device, like an LED? You have created a current source. Let’s say you want to vary the intensity of the LED among numerous levels, called amplitude modulation. Before you read the word “before” at the start of this sentence, your brain was screaming “DAC! DAC! DAC!”. Right. A voltage output DAC with a series resistor connected to an LED. A digital current source!
But WAIT! You can’t do that! It’s patented!
U.S. Patent 4940930 Digitally controlled current source
“A current source is controlled by the output signal from a digital to analog converter. The output signal from the digital to analog converter is applied to an amplifier unit. The output of the amplifier unit controls current through a pass transistor element, the pass transistor element current being the current applied to the load impedance. A feedback signal is generated by a differential amplifier in response to the current applied to the load impedance. The output signal from the difference amplifier is applied to an input terminal of the digital to analog converter with a polarity resulting in a change in the output signal of the amplifier unit which compensates for any change in the current through the load impedance. A voltage level changing element is included in the pass transistor control terminal to interrupt the current to the load impedance in the event of a degradation in the power supply.”
Isn’t that incredible? Had no one ever, ever, connected a DAC to a current source prior to 1990? Shoot, many DACs ARE current sources! I was using them in 1981, weren’t you? This patent details a closed loop current source, of course, so perhaps that was the innovation that the patent examiner considered novel, the addition of an opamp!
But this patent also refers to a current limit that is sensitive to the input power. Perhaps that’s the novelty? Certainly that was a unique idea in 1990!
A Patently Dim Bulb
Since we are running an LED, you might point out that a linear current source wastes power. For an LED that’s no big deal, but if we were running an incandescent lamp, the power waste would be significant. Before you read the word “before” at the start of this sentence, your brain was screaming “pulse width modulation!”. Right. Use a transistor, one of those newfangled power MOSFETs, to switch the lamp’s supply on and off at a variable duty cycle to change the amount of light coming out of the lamp. I did that with LEDs when I was learning about digital logic in the 1970’s. Didn’t you?
But WAIT! You can’t do that! It’s patented!
U.S. Patent 5113120 Dimmer circuit
“A pulse width modulated dimmer for use with incandescent lamps is powered by a D.C. source that has full on/off capability with a substantially full range of light level control. The dimmer is designed to replace a switch in a lamp circuit in which the lamp is connected on one side to ground and on the other to the switch and then to the power source. The dimmer is designed to operate with a 10 to 30 volt D.C. source. It is controlled in both the on-off mode and in the brightness mode by an N channel field effect transistor switched in response to the voltage level of any remote variable voltage source, such as a potentiometer, photoresistor, phototransistor or digitally controlled voltage source. The lamp is powered through the field effect transistor which is in turn controlled by a comparator which compares the user or automatically variable control voltage level with a triangle wave ramp signal to control the on-off duty cycle and thus the brightness levels of the lamp. When the user or automatically variable voltage source is adjusted high, the field effect transistor is off and the lamp is off. If the user controlled voltage is set very low, the field effect transistor is fully on and the lamp is operated at the D.C. source voltage less the small voltage drop across the field effect transistor. As the user variable voltage is adjusted between the “off” and “on” conditions, the duty cycle of the field effect transistor will decrease as the variable voltage is increased, thus dimming the lamps.”
When I read such a patent abstract, I simultaneously laugh and weep. It is quite entertaining to watch, say my kids. How could any self respecting patent examiner authorize such a patent? Have these people never designed a real circuit?
Certainly you have stored important data in your embedded system designs. If you have some data that is critical to system operation, you take precautions to ensure that the data cannot be corrupted, for example, by storing it in multiple memory spaces. A simple checksum or CRC shows quickly whether each copy has been changed, either by a software bug or some hardware defect. I did not catch you screaming this time, but I’m sure that solution was on the tip of your tongue a couple sentences back.
But WAIT! You can’t do that! It’s patented!
U.S. Patent 4802117 Method of preserving data storage in a postal meter
“A nonvolatile memory bank of an electronic memory and retrieval system is partitioned into at least three memory blocks, each block capable of storing an accounting program. The accounting program data is stored in duplicate with one copy of the data in one memory block and a second copy of the data in a second memory block. The remaining memory blocks of the partitioned memory bank are held in reserve status. The integrity of the data is preserved by comparing one data copy with the other data copy. When a difference between the two data copies is detected, a determination is made as to which data copy is the correct data copy. The data in the memory block having the correct data is duplicated into a reserve memory block, forming a new set of data in a new memory block, which is then substituted for the memory block having the incorrect set of data.”
So, if you have partitioned your memory into three or more blocks, and you make multiple copies of the data and check it, and fix the errors, you are a filthy patent infringer. Bailiff, take him away! I suspect I could find some computer science manual from the 1950’s that advocated the same process, but using vacuum tube flip-flops.
My favorite ridiculous patent is one on numerical integration. That’s right, the addition of a bunch of numbers. I think if the ancient Egyptians were still around, they would really be hot, especially their land surveyors. They used numerical integration to figure the areas of oddly shaped plots of land. That was what, about 5,000 years ago?
U.S. Patent 4240149 Measuring system
“There is provided a method for measuring a parameter whose value is calculable as an integral with time of a function of at least one variable so as to make possible the use of a microprocessor when measurements of such variables are desired. The variable is sampled periodically at a frequency asynchronous with the wave form which the variable follows and then there is periodically calculated the value of the function by summing the values for the function determined by the samples taken during the time between the calculations, thus determining the desired integral.”
Oh, but excuse me! The patent clearly states that a microprocessor is used to compute the value of the integral (that is, add up a bunch of numbers). I was not aware that ancient mathematical methods could be patented when rendered in software running on a micro. My mistake. I wonder what other Egyptian methods could be patented? Did the pyramids have an internet enabled home control system? Let’s have an IP seance with King Tut.
You cannot make me believe that no one before 1980 used a microprocessor and A/D converter to compute an integral, that the technique was not part of the prior art.
An acquaintance recently made me aware of the most hilarious patent I have seen. Apparently filed to illuminate the weaknesses of the patent system itself, here’s the abstract:
U.S. Patent 6368227 Method of swinging on a swing
“A method of swing(ing) on a swing is disclosed, in which a user positioned on a standard swing suspended by two chains from a substantially horizontal tree branch induces side to side motion by pulling alternately on one chain and then the other.”
When I saw the title page and the abstract, I thought it was a hoax. Then I searched the database myself and sure enough, it’s there. Let me warn you not to eat or drink while reading this patent, as I was, because there is a small but significant chance of choking while laughing. My favorite line is, “Young children often need help to climb onto a swing, and may need a push (sometimes even an ‘underdog’ push) to begin swinging.” God help us.
A slight bit of hope is provided by the reexamination of the patent which cancelled all the claims in July, 2003. For over a year I was a swinging patent infringer, but now I can again swing freely.
The Point Please
So, how do we work in this confusing, frivolously over-patented world? Publications on intellectual property identify two methods for protecting your company’s IP, patents and trade secrets. I suspect that engineers’ designs infringe many patents every year, but no one knows because the infringements are under the radar.
Perhaps the patented technique is used in an industry other than that familiar to the patent holder. Perhaps the infringer is insect sized in relation to the patent holder. Perhaps the method is buried so deep in the product that it would take a great feat of reverse engineering to identify the crime. Perhaps no one really cares.
I think a big reason we don’t see more patent litigation on the small issues, like that sly use of numerical integration in your last project (I’m talking to you), is that patent holders don’t relish blowing any more money on attorney’s fees. After all, a patent can be viewed as your sole right to continue laying out big money to keep playing the patent game. Many patents make zero money (or less) for the patent holder, and defending a loser patent is not a winning proposition.
However, every day you use techniques that are patented, and you are at risk. The risk is a hideous one because you cannot quantify it. Searching the patent database and concluding that your bright idea is not present is like trying to prove a negative — impossible. That’s where most of us shrug, stop writing checks to attorneys, and take our chances.
I wish I had some warm and fuzzy conclusion here, but I don’t. Fortunately, the patent database is available for free online at www.uspto.gov, and if you have a specific innovative idea you can search for it pretty quickly. (Let me also recommend PatentMax search software, which is a great tool and quite snappy. See www.patentmax.com.) But patents cover the space like patties in a cow pasture, some fresh, some decayed, some blended into the grass, and you never know when you are going to mistakenly step in one.
Hank Wallace is the owner of Atlantic Quality Design, Inc., a consulting firm located in Fincastle, Virginia. He has experience in many areas of embedded software and hardware development, and system design. See www.aqdi.com for more information.