Chapter 8 - Starting to lose their grip on the place

Spanning the 3rd Revolution

The 3rd revolution is better known as the "Digital" revolution 

In this chapter we are going to examine the gulf between the Industrial and the digital revolution, and how the Group handled their transformation because of it. The first revolution was mankind's harnessing of agriculture. The move from a primarily agricultural society to a society that embraced machinery and automation, which replaced much of handcrafted products and goods offered before, was the second revolution. The third is the digital revolution. This ushered in what has become known as the information age. The roots of this phase actually started back towards the later part of the 19th century but did not reach critical mass until the invention of the transistor, and then the integrated circuit. The Grass Valley Group came into being right as the transistor was making major inroads into electronic circuitry. Because of Dr. Hare and Bill Rorden's early mastery of the transistor, the Grass Valley Group never offered a product that used vacuum tubes. Another electronic juggernaut, that made the same early commitment to solid state electronics was Sony, which entered the United States shortly after the Group got started.

Vertical Integration

Even though the Group was now under new management it continued its rise in the ranks of the high end television production equipment industry. This rise manifested itself by the number of employees and real estate the company occupied. While today some companies can grow financially large with few added employees, witness the web economy, back in the 60s and 70s that usually was not even a remote possibility. To grow a company like Grass Valley, they needed to put more people to work to produce more product as sales increased. Either by contracting some of the work out to others, or by hiring people to do the work "in house." Grass Valley evolved using the second approach. It was a highly Vertical Integrated company as it handled almost all of its required stages of production. Early on it even handled all of its own sales and marketing.

It made sense that the Group took this approach when it did, for a number of reasons. One reason was timing. When the company started in the electronic manufacturing business the equipment itself, and the components that made up that equipment were much simpler to create and assemble. A basic machine shop, some assembly stations, or a rudimentary assembly line where circuit boards could be stuffed with parts and soldered, and installed in a case or enclosure, and finally some test equipment to make sure the assembled product was working correctly.

When the Group first started shipping product they did not even have to manufacture or etch circuit boards. As mentioned earlier they used boards that had rows of continuous traces that they would manually cut to organize the components into the required circuits. None of the electronic components back then had any complex mounting requirements. Frames and enclosures where constructed using sheet metal, steel ribbing, and tubing. This approach to enclosures and cases actually continued into the 90s at the company long after circuit boards and control panels had evolved to be quite complex.

Another reason why many smaller companies that grew up around the Grass Valley Group are still very vertically integrated is that they manufacture what is referred to as "low volume, high mix." Historically equipment aimed at broadcasters and other users of high end television equipment was quite expensive. As such it had a limited market. Also most of the gear these folks buy are not stand alone pieces of gear, as a camera or microphone might be, but they are used in conjunction with other equipment that comprise a system.

That system is composed of a mix of gear that is installed to allow the users to perform a specific workflow. That is, a series of tasks performed both sequentially and in parallel to deliver video programming to the viewer. A station that does a lot of news, will have an entirely different workflow than a movie channel. Sports channels will work differently and have people doing different tasks than a channel that does talk shows.

Thus a piece of equipment in this industry usually has to be able to be vastly configurable to meet these varied demands. Often the model X video switcher needs to be produced for one customer that needs 20 video sources fed to it, while another needs 100. As we will see video and audio does not come in one flavor, and today they number more than Baskin-Robbins has flavors of ice cream.

Vertical Integration makes sense when supply chains are not stable, like right after WWII, and before just in time overnight shipping came into being. Transportation is much more robust today. When the Group started roads from all directions into the area were 30 miles of two lane road, through hills and mountains. Even the main road from the town of Grass Valley out to the Bitney Springs site had a one lane bridge. Often large trucks headed to the site had to make a dozen mile detour and approach from the other direction.

So for the first 30 years of its existence vertical integration made sense for the Group. The Grass Valley company today still does much of its own manufacturing, just not in the Grass Valley area anymore, as we will see.

Dan Castles, who was the Groups president in the mid-90s said that when he got involved with the company it looked like a postwar company, as it was heavily vertically integrated. The first time Dan drove up highway 49 in the mid-80s he could not believe a 1000 person company could be up in that part of the world. Dan was the VP of Tektronix's Video Division. At the time the Group had 1200 people working for it worldwide, 900 in grass Valley.

Printed circuit board layout - before computers and CAD

More art than engineering

An eye for detail and a steady hand


Being an outpost in the Sierra foothills, the Grass Valley Group evolved doing as much as possible for themselves. From printed circuit board fabrication (PCB), stuffing parts into those PCBs, machining frames and panels, to testing the final product, originally the Group did it all.

Early on, the Group took raw material and parts in, and through machining, etching, and assembly produced finished products for sale. As we have mentioned, at the volumes we are talking it would not have been very economical to job any of that out in the early years of the company. At one time the Group had about 100 people working multiple shifts in its machine shop. Switcher control panels had hundreds of cutouts, mostly done by hand, for all the switches and other controls that would end up in them.

GVG Machine Shop employees, day and night shifts


Machining hardware into finished products was a multi-step process. Raw metal stock generally arrived as sheet metal or metal bars that underwent the process of being formed into the desired part by a lathe or a milling machine, or both. Drilling operations were common on many parts. Both the milling and lathe processes are similar, but the simple explanation is that with a lathe the part being shaped is rotated while the cutting tool is stationary. With a milling machine the piece being worked on is stationary while the cutting tool rotates. Often individual parts were then manually de-burred, welded to other parts to make a larger assembly, sanded, and then painted or etched/anodized.

Anodizing is an electrochemical process that converts a metal surface into a durable, corrosion-resistant, anodic oxide finish. Aluminum is ideally suited to anodizing, although other nonferrous metals, such as magnesium and titanium, also can be anodized.

Even back then not everything was done in house. Any parts that needed to be painted were sent to outside vendors. Raw sheet metal, which generally formed the outer enclosure needed to undergo punch, bending, additional shearing, drill and tap operations. In addition these parts would often undergo etch/anodizing or nickel plate processes. Nickel plating provided a combination of corrosion and wear resistance. It can add brightness, and luster. It also adds adhesion properties for subsequent coating layers, which is why nickel is often used as an 'under coat' for other coatings, such as chromium.

So the Group, by the 80s was doing almost all of its own machining, and thus most non-off-the-shelf parts where created up on a ridge northwest of town, that is the Bitney Springs site. When you make your own parts you not only have to worry about keeping the assembly line stocked, but also need to create replacement parts for product already out in the field. As such the Group divided the parts they used into three categories with the highest grade, parts that had some cosmetic appeal, such as operator panels, were individually packaged in foam. The next level parts were individually wrapped, and the lowest level might be packaged in paper and plastic, and simply piled together in a parts bin.

The Group in the early 80s went from assembly lines where a single installer might install all the components on a particular board, or completely assemble the enclosure for a product, to what they called flow Lines. Here each assembler only installs a couple dozen components, and it then moves on to the next assembler. The Group hired an assembly engineer, Larry Hoy, to implement that. The company invested in machines that pre-bent and cut leads for the components that were now just dropped into holes in the circuit board by assemblers. This led to higher productivity, but for the person doing the assembly the task became repetitive and there was a loss of pride as a complete assembled board was not a particular assembler's anymore.

Flow line assembly, where parts are "stuffed" into circuit boards, wave soldered, and then those boards are cleaned of all the solder resin residue.

Printed circuit boards


The PCB operations grew to a point that in early 1982 the mother company, Tektronix, gave the Group an order for 3,200 PCBs for a new product line being launched up in Oregon.

To build a complete system you must wire all the sub-components together. Hence the building of wire harnesses. The flow line concept was helpful when it came to constructing harnesses, the bundles of cables that interconnect circuit boards, connector panels and control panels. The ones in a video switcher could take 180 hours to build. If a single installer built the entire harness, by the time they were done the person might have forgotten the initial steps.

Finally you need to be able to test and configure the final product

As we will see later on, as the company continued to change hands, and merged with operations from other companies, some that made competing products, the Grass Valley and Nevada City engineering and manufacturing fortress began to lose out to others spread around the globe. That was not always the case. Shortly after the Group was acquired the third time, by French electronic giant Thomson, a television router division they had in Salt Lake City was brought to Grass Valley. But in later years the bulk of the work migrated away from the area. By 2015, and with the fifth acquisition, this time Belden Cable, the last of the manufacturing was moved to Montreal. We will see later how this saga played out as owner after owner applied successive management philosophies.

But just as Thomson had their Salt Lake City router operations moved to Grass Valley, they also tried to have as much as possible made in Vietnam. The problem is this is only economical if you make many of something, not a few, like Grass Valley was doing. If you are building offshore a few Engineering Change Orders (ECO) that crop up can throw any savings right out the window.

The thought here is that many manufactured goods come out of China today, and since Vietnam is a neighbor, although neither are friendly to each other today, the same cost savings might be had. There are a couple reasons this did not work. One we already mentioned. But the other difference is that the Chinese have the infrastructure to iterate many times until they get a manufacturing sample and then copy it millions of times. Hard design is a small part of that process. Again, you cannot afford this in low volume/high mix manufacturing.

While process is important in any set of steps to take raw parts and materials and turn it into a finished shippable product, the task gets much more intricate as more of the finished product is siloed together. Birney Dayton, who had a number of hats at the Group, director of engineering was one of them, and later spearheaded the Groups only real spin off, NVISION, was an early apostle of process, and the aspect of who was responsible for it every step along the way.

Complicated electronic equipment, which is what most of the Group's products were, created many opportunities to get things wrong in their design and construction. Not only did the engineers need to create solid, stable, and robust designs, in both hardware (PCBs, frames, control panels) and the software that even back then drove much of the hardware, but then manufacturing has to be able to replicate the design and see that it is built so that one example of a product is indicative of all the others like it.

At each juncture, the chance for missteps is many. It starts right after the design phase. Even if the engineer nails the circuitry needed to accomplish a given function, how is that design, usually in the form of a schematic, conveyed to the people who need to document the design. While considered quaint today, schematics and other parts diagrams used to be considered a required ancillary part of any product package. Thirty years ago schematics were considered essential as most professional television facilities expected to be able to repair the equipment they bought.

At one time technicians at a facility were expected to troubleshoot problems down to a particular component on a board and replace it. Over time PCBs became so intricate that they could not be repaired in the field as the owners of that equipment could not justify investing in the expensive equipment necessary to accomplish that. So troubleshooting was just to determine which PCB was defective. Today many product's parts and subsystems are so intertwined and constructed as to not allow anyone but the manufacturer or authorized repair agent to successfully deconstruct, repair, and re-assemble a broken product.

So even though schematics are useless now to the end user, they are still vital to the manufacturer. Just as the product has become too perplexing for outsiders to cope with, it has become increasingly complicated for those who build them. So back to the schematic. The blueprint of the designer's solution. It is a collection of data points that must be managed and tracked. The manufacturing engineer must control each piece of data, that there is one point of entry, one master record. They must control who can change that record or piece of data. Inevitably new designs will have lots of changes as problems crop up in produced products, and they are corrected. In the 80s, before software like AutoCAD became ubiquitous, schematics were usually drawn by drafters. The engineer would hand draw their circuits, pass them off to the drafter, who would return the "formal" schematic to the engineer for sign off.

Today, with all the advance design software available engineering will create the schematic, it won't be scratched on a piece of paper and handed off to a drafter. Today the schematic capture tool has the parts data in it and creates a Bill Of Material (BOM). No human hand actually touches the BOM. It has hooks to add mechanical pieces to the board. Now PCB layout software adds location and orientation data. If the layout software catches potential errors: such as traces that are too long, or the need for more capacitors, engineering has to correct that, not someone else. Today only the original engineer should touch the schematic.

The manufacturing department programs the pick and place machines, which includes the order of assembly. Big parts might be found to be in the way of small parts. Again new data, that must become part of the same master record. An important part of the auditing in this manufacturing system is who owns each data entry in the record. Mistakes happen when many can change a record and thus no one owns that piece of data. Errors can quickly creep in.

While greatly automated today as you can see, imagine in the 80s when the process lived on paper schematics and checklists, and not in database records as they do today. What automation was available then lived as islands of data, and not networked together so one database can control the whole process.

Today the biggest hurdle is that the folks overseeing the corporate IT structure often insist on being the overlords for the process and manufacturing engineers that are held responsible for the final result.

A big goal for a company is to have as many early examples as possible of a product be fixable enough to be salable. In the 80s that might mean only 50% of first run boards. Today it is expected to be near 100%.

In addition, as we have mentioned, the Group, as with many other vendors of high end video equipment, both in the area and worldwide, must be able to cope with low volume, high mix products. When you were doing low volume, and high mix you need your own manufacturing, otherwise it is too expensive. This means those companies need an economic manufacturing quantity of one. They need to make one of something and make money at the same time. This is accomplished by arranging their manufacturing line to a place where changes to the line for a different product, besides maybe changing out a parts wheel, is as little a task as possible.

As we will see shortly through the 80s and into the 90s the vertical integration process worked fairly well for the Group. But not always.

Hence: You need a facility that was known as Bitney Springs!