27 Feb 2025
by Roger Arrick
Arrick Robotics is a company I started in 1987 with $400. I designed PC-based motor control systems, positioners and mobile robots - all manufactured in Euless, Texas.
See the Arrick Robotics Archive for software and manuals.
While operating my first company, Arrick Computer Products, it became obvious that consumer product manufacturing in the US was on its way out. Cheap computer products from overseas flooded in. This idea was sold to us as 'free trade'. Products coming from Taiwan had a retail price less than my parts cost so there was no way to break-even much less make a profit.
So, I set out to completely change careers from computers to robotics. I wanted to continue product design and manufacturing. My reasoning was that industrial robotics would have higher margins than consumer computer products, plus, building robots had been a hobby for years. Initial investment to start Arrick Robotics was $1000 I made by designing a passive exercise machine. This venture was like all the rest, I did all the product design, hardware, software, manuals, even wrote the accounting software.
My first robotics product design was a robot arm for factory automation called SA4. It was quite ambitious with multiprocessor motor controllers. This was side-line tinkering while I continued to battle the computer market full-time.
In 1987 I got the chance to sell my computer company and break-even - this was scary but gave me a clean slate for the robotics venture. I continued to work for the buyer for 2 years developing computer products while working on the robotics product line at home.
During development of the robot arm I decided to package the motor control portion into a stand-alone product for use with personal computers. This was the MD2 Stepper Motor system. I wrote software drivers in C, QB, VB and Pascal, along with a decent user's manual.
The very first MD2 systems were sold for telescope positioning through astronomy magazines. The MD2 was also sold through ads in NASA Tech Brief magazine and some other industrial control magazines - this was all before the internet.
During this early time we also did contract electronics assembly for various companies including TI-EX, DemoTeller and American Microsystems to supplement revenue. We stuffed circuit boards, built cable assemblies and assembled various electronic products. At one time the number of employees doing assembly work was around 20.
I built all of the first products myself at home in 1987 but then added part-time employees. In 1991 we moved to a real office.
When the internet came along around 1995 I registered the ROBOTICS.COM domain name, completely dropped all of the expensive magazine ads, and went completely on-line. This was pretty daring at the time but my confidence was high that the internet was going be a big deal.
From there the product line grew to include linear and rotary positioning systems compatible with the MD2 motor system. These were used for computer controlled laboratory and factory automation. Applications included dispensing adhesives, filling test tubes, pushing buttons on phones, and scanning TV screens for dead pixels. MD2 systems were used at NASA for test machines, Disney for floats, and MIT for self-driving cars.
After some success with automation products I got the freedom to design mobile robots for the hobbyist market with many used in school labs. I was also able to experiment with walking robots that looked like geckos but none of those were brought to market due to a lack of resources.
In the mid-90s I began designing synthesizers which lead to Synthesizers.com. I slowly let synthesizers take over robotics but kept both companies together until the sale in 2018. The automation product line lasted more than 30 years.
The MD2 motor driver system was the first product I released in 1987. It connects two stepper motors to a personal computer through the parallel printer port. Typical applications are automating machinery such as glue dispensers, pick-and-place machines, and telescopes. Software was provided that let the user move motors in real-time or record movements and replay them. Subroutine libraries are included in C, VBWIN, VBDOS, QB and Pascal for making custom applications.
I made all the first production batches by hand at home but then hired part-time employees to help. The circuit boards for the very first MD2 batch was delayed and I had a couple of customers who were really anxious to get their product so I hand-wired the first units on perf board. This was extremely labor intensive. The MD2 product line sold for more than 25 years and there are some still in operation to this day, mostly factories. Overall production was in the thousands but they were not serialized so I don't know the exact number.
The very first units, named 'MD2', was a 4-phase unipolar L/R driver circuit and two 50 in/oz NEMA 23 motors running at 1 amp each. The next generation named 'MD2A' and 'MD2B' were 4-phase unipolar chopper designs using discrete chips. A 20khz oscillator turned on each motor phase by driving it to ground through an IRF640 MOSFET, then a comparator monitors the current and turns it off when the threshold is meet. Random oscillation sometimes caused a buzzing noise which was fixed by 4 .1uf capacitors soldered to the comparator on the bottom of the PCB. The MD2A was 50 in/oz 1 amp motors and the MD2B was 150 in/oz 2.9 amp motors. There was an MD2C for a short while which was a 300 in/oz 3.4 amp NEMA 34 motor.
The MD2 transformers were 25.2V from Radio Shack. I use to go store-to-store and buy all all their stock but eventually found out how to buy direct 100 at a time.
The MD2 software used the PC's parallel printer port as a general purpose digital I/O port. Four bits controlled the 4 phases of each motor. All of the translation was done in software, and speed was regulated by software timing loops. This worked pretty well until Windows came out which blocked access to the port and stole processing time in an unpredictable way making motion erratic. This was ultimately solved with the C4 controller which offloaded the real-time processing of steps to a microcontroller that connected to the PC via a serial or USB port.
See the Arrick Robotics Archive for software and manuals.
Soon after the MD2 motor system started taking off I designed some linear and rotary positioning mechanics. These were unusual in that they were made entirely of sheet-metal, no machined plates. This made them very inexpensive yet capable of better than .01" positioning accuracy and repeatability. Each axis could be driven by an MD2 motor system or any NEMA 23 motor. The systems were very modular and could be stacked together to create multi-axis positioners such as this 3-axis robotic workcell shown in the picture.
Trilobot is the first mobile robot I designed as a product to sell. Profits from the success of the MD2 and positioning products made this possible.
The first Trilobot version had 3 wheels and a shape that reminded me of a Trilobite. The processor was an 8051-like CPU running code in assembly language to control a single DC drive motor with a geared stepper steering system. The mechanics were an experiment in suspension with the drive wheel mounted to a hinged ring supported by opposing springs. Sensors included a Polaroid sonor range finder, digital compass, light, temperature and whiskers around the base for obstacle detection. The first customer was the robotics group at the University of Texas at Arlington called ARRI.
The second version of Trilobot was a complete redesign with dual-wheel differential front-wheel drive using optical encoders and a swivel wheel in the back. It had a multi-processing controller with a keypad and display. All of it was written in 8051 assembly language.
The number of sensors is amazing including a precision digital compass, ultrasonic range finder, directional light sensing, temperature, IR in and out, microphone, speaker output for crude speech creation and much more. Many of the sensors were on the head which could turn to any direction. The robot could input RC pulse data and had 8 servo motor outputs. A removable battery pack used standard D-cells or rechargable NiCAD batteries. Trilobot also had a gripper capable of picking up objects like a soda can or tennis ball.
We customized Trilobots for use with a school program distributed by Paxton/Patterson.
The Trilobot name has been used since then by other companies not related to Arrick Robotics.
See the Arrick Robotics Archive for software and manuals.
ARobot is a low-cost programmable mobile robot I designed for hobbyist use. The brains of ARobot included a PIC coprocessor to control the drive motor, steering and sensors. The main controller was the Basic Stamp created by Parallax. ARobot was provided in kit form but the circuit board was pre-assembled.
At one point ARobot got on the front cover of the Edmunds Scientific catalog. Since I had seen this catalog as a kid it was quite an honor. ARobot was also the focus of my book - Robot Building for Dummies by Wiley (see below).
See the Arrick Robotics Archive for software and manuals.
In 2002 Wiley & Sons asked me to write a 'Robot Building For Dummies' book for their popular series and I agreed. The book takes the reader through building the ARobot mobile robot and adding devices such as a compass and speech.
Overall this experience was difficult because of the rigid format, and wasn't worth it financially due to a busy schedule running 4 companies - Maketeck, Publishing, Synthesizers.com.
T-Glide is a design I created to utilize standard industrial T-slot framing to create linear slides for automated machines. Systems can be any number of axes, each with a travel distance from 1cm to many meters. The system includes brackets to connect motors, idlers, carriages with bearings, pulley reducers, limit switches and cable carriers. Each axis can be driven with an MD2 motor system or with any NEMA 23 motor. The timing pulley construction provides an accuracy and repeatability of .25mm (.01"). An online app provided instant quotes for any custom size.
In 1990 Ron McDaniel of Soft-Tec Systems in Fort Worth, Texas wrote software for controlling a telescope using an MD2 I modified for that use. The system was called Sky Probe 1000 and it came with a custom handheld controller we built. Ads were run in Sky & Telescope magazine.
Mini-Mover was my attempt at making a low-cost stepper motor system for hobbyists. It was a simple circuit board with a transformer and motor with a cable. It was a failure because the hobby market is very cost sensitive.
My first attempt at a robot product was the SA4 robot arm for use in factories. It was very ambitious and beyond my funding resources so it never got produced. I designed a fancy multi-processing system with Z80 CPUs to control the stepper motor for each axis. All the software was written in Z80 assembly language. The mechanical system was ultimately going to require precision gear reducers that I couldn't afford. Here is what's left of the first prototype I built by hand.
PD5 was a portable disk drive enclosure and cabling system I made for use with MEII Pharmacy computers which I built for them in the mid-90s. I made it into a regular product to sell through computer dealers and catalog dealers such as PC Tools.
Our first and ultimately our largest dealer was Personal Computing Tools run by Leon Hamner in Los Gatos California. Leon became a good friend and taught me a lot about marketing. The biggest thing was learning my low product pricing strategy was wrong and that prices needed to reflect their value to the customer more than my cost. PC Tools produced a nice catalog in the millions and we were well-placed there, even on the front or back cover sometimes.
Circuit Specialists was another catalog vendor that sold the MD2 and positioners. Edmund Scientifics carried our ARobot and it made the front cover which was very exciting.
This first catalog was around 1989. I did it in MS Publisher for Windows V1.0 on a 386 with 16 megs of memory. The second is 1994. The third catalog is 1996 and included the Trilobot along with cabinets from my other company called Maketeck. I did that one in Corel Draw.
From 1987 through 1991 Arrick Robotics was run out of our home. It grew from a bedroom, into the garage, and took over the living room. At one time we had more than 5 employees there. I was concerned about all of the cars parked at the house but the neighbors never complained.
In 1991 we rented an office in Euless. Strangely, this building was the same as my previous company (Arrick Computer Products) and had recently become available. It was about 3000 square feet. Here's a picture of it in 2002 after the Synthesizers.com sign was added. When we left in 2004 the neighbor Hurst Metallurgical took over the space.
In 2004 we moved to Tyler in East Texas about 100 miles away from Dallas/Fort Worth. The free-standing metal building was 6000 square feet. A few of the employees moved with us and we filled new positions from local university engineering programs. This was known as County Rd 2335 but we had the name changed.
In 2015 I added 1500sf to the back of the building.
I had another business card while working at home in the early days but can't find one to scan. Later years used the ArrickRobotics.com domain name but I can't find one of those either.
When the internet began to get popular around 1995 I registered the domain name ROBOTICS.COM. For more than 10 years I used it for the company but also for publishing general information about robotics. One of the features was the RoboMenu which provided a place for hobby robot builders to show off their creations. Later the RoboMenu was merged into Robots.net.
The early websites can be seen on the WayBack Machine internet archive. Here are image captures from 1996 and 2005.
Here's an archive with product manuals and software for most products. I've kept a sample of many products.
Resources used for Arrick Robotics slowly moved to Synthesizers.com over the years as my focus changed. Arrick Robotics was sold along with Synthesizers.com in 2018. The products were maintained for a while but were eventually dropped. Some products are likely still in operation. The remaining inventory was put in a dumpster with some items saved in my barn archive.
Arrick Robotics had a good history of 30+ years, employed hundreds of people, and made a decent living for my family.