Management of Wabtec Embedded Software Team

Wabtec’s Winnipeg plant is actually in Oakbank, about 30 km (18 mi) outside of Winnipeg.  Oakbank is a small community of a few thousand people, surrounded by farmland, and, to the northwest, hills with rock quarries.

This was originally a small Winnipeg company called iders (pronounced by most people as I-D-ers), who had been pioneers in the development of PIN pads for Canadian banks and retailers, and later successfully landed the contract to develop the Conviron CMP4000 environmental growth chamber system, among many projects of varying sizes and complexity.   In the 2001 to 2008 time frame, they also manufactured the NxtPhase TESLA DC Isolation Module, which Horst Koelzow and I had developed.

One of the industries that iders broke into in the early 2000s was electronics for the railroad industry.  Their experience in communications in general, and computer networking in particular, led them over several years, to create what’s now known as the GoLINC ACC (Auxiliary Communications Cage), which has the AAR standard for communications on railroad locomotives.

iders was acquired by GE Transportation in 2016.  In 2018, GE Transportation itself was divested from GE, merging with Wabtec.  The Oakbank plant was still referred to as “GETW”  (GE Transportation Winnipeg) until early 2020, when the migration from GE’s IT systems to Wabtec’s IT systems was mostly complete.

I was hired as the Embedded Software Engineering Manager for GETW in August 2018, with 11 direct reports.  A few days after starting, I was advised that I had 3 hires to make – which, for historical reasons, were each assigned to separate hiring managers, and were each managed by separate recruiters.  The hiring took some time to implement, as we worked out the projected seniority of the hires and reviewed dozens of applicants… but the big thing is that I was busy in the interim.  We made the hires in January 2020.

There was a lot of development going on at GETW.  Some was new, some was remedial work, but it was all very intense.

When the COVID-19 pandemic hit, we were quite fortunate to be able to work effectively from our homes, because Wabtec has quite a good internal communications infrastructure.  There were issues, mind you – we had trouble accessing the Checkmarx static analysis scanner – but we worked around that using a Raspberry Pi that I put on my desk at the office, until the proper routing could be arranged through the VPN.

The pandemic was not as serious in Manitoba as in many other parts of the world.  When things settled down in mid May, we carefully moved back to the office, taking appropriate precautions.

Unfortunately, as with many industries around the world, the locomotive industry was hurt by the economic downturn that accompanied COVID-19.  As a result, my employment with Wabtec came to an end in June.  Many others were let go at the same time, and apparently there will be a long term strategy to consolidate Wabtec’s electronics manufacturing into other locations, and downsizing the office footprint, saving considerable money in the process.

Challenged More and More: DR2 and TESLA on the Same Platform

We did focus group studies, discussed with customers, went to trade shows and conferences, and finally, settled on the development of two platforms for the power utility market: a TMS320C32-based platform for the relay, which we would call the DR2 (DR1 would have been the units developed using TMS320C30 at the UofM), and a TMS320C44-based platform for the recorder.

The ‘C3x family is fast and inexpensive.  The ‘C44 family is well connected and easily networked.  No comparable family at the time gave us both.

Management deemed this too expensive and risky, so we were told to go back to the drawing board and do it with only one platform.  We gasped… sighed… and started over.

I can’t say that it was wrong.  The development was successful… the product met market demand… and, with updates and changes, still in production today.

The core Main Processor Board developed as part of that effort, was in active production and use until being finally retired in 2018.  Not bad!

Work on GPS, Heads-Up Displays, Vision-Based Vehicle Speed Measurement

After years of very long hours and hard work, I was relieved when Vansco offered me the position of Chief Research Officer.  Vansco built me a lab in a quiet area of the plant, gave me resources and ideas, and asked me to play.

This only lasted a short time, as both my personal priorities and Vansco’s priorities changed, and within a year I was back in APT – although, no longer as General Manager, which was fine with me.

Global Positioning System

At our CEO Dave Sokol’s suggestion, I dove into GPS positioning, real-time kinematics (RTK corrections), and their potential use in agriculture.  I obtained some early GPS receivers and experimented with them.  At that time, they were very expensive!  I was amazed at their accuracy and reliability, even then.

Heads-Up Displays

I also toyed with heads-up displays.  We anticipated that this would be in demand in agricultural equipment soon.  I examined the existing state-of-the-art, and projected where we might take the technology.

Implement Velocity Measurement

One of the challenges in agriculture is the accurate measurement of ground speed.  This is required to accurately control seed and material application rates, and measure harvest efficiency.  Unlike road vehicles, you cannot just measure wheel rotation, because considerable slippage is seen all the time.  The state-of-the-art at that time was to use a RADAR sensor aimed at the ground at an angle, measuring the Doppler shift in the returned signal.  Taking into account the trigonometry of the angle, you can accurately measure velocity with respect to the ground.   However, on smooth surfaces, such as puddles and some roads, this system will experience almost-total reflection away, and no Doppler shift can be measured.

Our solution was to use a colour camera, watching the terrain below the vehicle.  By correlating the images in real-time, the 2-D displacement between the images can be estimated, and velocity can be accurately calculated.  I obtained a patent on this innovation.

Appointed General Manager of the APT Division of Vansco

With the departure of the Conviron business, Ed and Terry Van Humbeck asked me to take on a new venture – creating the APT Division and leading it into the power utility sector.

We had been doing collaborative work with Dr. Glenn Swift and colleagues at the University of Manitoba for years, but the designs were theirs, and we found them to be difficult to build.  They didn’t necessarily design for production.  This would take our learning from those efforts, and create entirely new products.

I hand-picked my dream team, and they gave me all the persons I asked for, except for one  – my friend Filipe, who was very busy working in the Vansco PC division.  Vansco built us a new office area in the strip mall at 1253 Clarence, next to the Wire Harness Division, and we got started.

For the next 4 years, I was Chief Architect, Hardware Designer, Low Level Firmware Developer, IRIG-B Processor Developer, Product Development Manager, General Manager, and reported to the board meeting monthly.  It was pretty crazy – kind of hazy now, honestly, as I worked far too hard during this time.  But, did we accomplish a lot!

Promoted to Engineering Manager, Industrial Systems

I clamoured for a promotion, and finally got it!

James White and I were named Engineering Managers.  James was given half the team for Embedded Systems Engineering, and I was given the other half for Industrial Systems Engineering.  Our offices were side-by-side in the building at 1305 Clarence.  We worked very closely together, often reviewing each others’ work, and seconding our team members back and forth.  We both always interviewed any candidate for either group.  And, for existing historical projects, James and I sometimes would cross over as well.