Nov 6 2012
Takt time – Early work at Junkers in Germany
In the TPS Only discussion group on LinkedIn, Joachim Knuf provided the following information, to which I have added a few illustrations:
“Junkers was building an aircraft for infantry support in 1917/18, the J4. After building airplanes in a handcraft mode for some years, this was the first attempt to meet larger demands. Junkers was known for not simply designing a product but the production process along with it. Clearly very lean and practiced by Toyota.”
The founder of the company was Hugo Junkers (pronounced ‘yoonkerz). He died in 1935 and was in no way affiliated with the Nazis, who took away his company and later besmirched his name by claiming association with it. Because the company still bore his name, it is linked in the minds of World War II forced laborers with their experience.
“The J4 was build in modules, incorporating components delivered from outside vendors, incorporating fixtures, templates and gauges for economic benefit. Subsequently, the F13, build in 1919/20 as the world’s first civilian passenger plane, was designed from the start with modular manufacturing in mind. This plane was built for customers around the world, with production numbers as high as 60/month.”
The only surviving J4, at the Canada Aviation and Space Museum.
Junkers F13 in 1925.
“Production was organized to an overall completion schedule. As a result, completion of modules had to be structured and synchronized. Final assembly was organized into six phases, performed at specific fixtures. Highly specialized work teams had a set amount of time to complete their fixture-based task, then moved on to the next fixture to repeat the job, followed by another specialized team (what we think of as a caravan system these days). The increment was the ‘progression interval’ (Fortschrittszeit). Airplanes/modules stayed in place. The result was a finished plane every 9 hours. This approach was shared with Junkers facilities working in other industries.
By 1926 this system was developed to the point that subassemblies could be produced off the main assembly and connected to it with moving lines that moved at set intervals. These intervals were then referred to as ‘Takte’ (plural, ‘takt intervals’). With the new W33/34 (first East-West Atlantic crossing in 1928), there was interest in the US to produce the plane in license, in preparation of which Junkers developed a complete production plan to allow large-scale production, identifying the most economical methods. At that point Junkers had 40% of the international market share. Some years later, Lufthansa orders for the new model Ju 52/3m required the further refinement of the ‘takt method’ (Taktverfahren), incorporating new technology and equipment.”
Junkers W34 at the Canada Aviation and Space Museum.
“After 1933 (and the nationalization of Junkers by the Nazis) this allowed the mass production of airplanes in serial assembly. To produce the required numbers of planes, eventually also using forced labor, Junkers began constructing large subassemblies in decentral locations within 20 miles of the main assembly facility, delivered just in time. Major subassemblies then moved down a ‘takt avenue’ (Taktstrasse) from station to station, remaining a uniform, prescribed ‘takt duration’ (Taktdauer) in each, creating the Junkers ‘Airplane High Volume Series Production to the Minute’ (Flugzeug-Grossreihenfertigung auf die Minute). Changes in takt were used to adjust production volume to demand. Continuous improvement was an integral aspect of this system (which also certified workers on their self-inspection skills).”
A 2010 video entitled Fischertechnik Taktstrasse mit Sortierung depicts a “Taktstrasse” as a transfer line.
November 7, 2012 @ 8:57 am
Comment in the TPS Only discussion group on LinkedIn:
November 7, 2012 @ 9:04 am
What the developers of TPS did with takt time was innovative. The point of this discussion is not to deny it but just to trace their starting point. I enjoy it because it is answering some questions I have had in the back of my mind for a long time.
It is remarkable that Toyota took ideas from the aircraft rather than the car industry. One key difference between the two is volume. In the early postwar years, the US was already producing cars by the millions every year, while the Japanese car industry was barely doing it in the thousands. It had neither a market that could absorb larger quantities nor the capital needed to buy the kind of equipment that GM or Ford used.
On the other hand, the aircraft industry makes much smaller volumes. The total number of B17s built over the life of the product is 12,500; for the Junkers Ju-88. about 16,000. The Mitsubishi Aircraft plant in Nagoya had been building planes at the rate of 1,000 to 2,000 per year.
So, if I put myself in the position of the Toyota people at the time, it would make more sense to learn production techniques from the aircraft people, especially considering that they were local and available. In The Japanese Automobile Industry (http://bit.ly/SLb6FM), Michael Cusumano confirms that “…directly after World War II, Kiichiro [Toyoda] recruited more design engineers from the aircraft industry…” (p.120).
November 7, 2012 @ 9:07 am
Comment in the TPS Only discussion group on LinkedIn:
November 7, 2012 @ 9:08 am
Comment in the TPS Only discussion group on LinkedIn:
November 7, 2012 @ 9:09 am
I don’t think anybody is questioning the filiation between US Industrial Engineering and TPS.
The one, unimpeachable evidence of a German connection is the word “takt” itself. It is not in the index of my copy of Maynard’s IE handbook (3rd edition), which goes straight from “tachometer” to “task.” And we now know that the Junkers Takt system was adopted in a Mitsubishi Aircraft plant in Nagoya. And we also know that Toyota hired aircraft engineers after the war. The case is getting pretty solid…
Incidentally, within IE, I think TPS owes more to the Gilbreths than to Taylor. The Gilbreths were observing operations in order to improve them, while Taylor was timing workers to prevent them from “soldiering” — that is, colluding to curtail output. This is clear if you compare the compilation of Gilbreth films available on line with the writings of Taylor.
November 7, 2012 @ 9:12 am
Comment in the TPS Only discussion group on LinkedIn:
November 7, 2012 @ 9:14 am
Comment in the TPS Only discussion group on LinkedIn:
November 7, 2012 @ 9:17 am
Comment in the TPS Only discussion group on LinkedIn:
November 7, 2012 @ 9:19 am
@Joachim — I first heard the term “caravan style” from my Japanese mentor Kei Abe, to designate the mode of operation of a U-shaped cell in which each operator loops through all the tasks. That is, of course, what happens when you only have one operator. If you have ten stations, then the operator will walk and do what needs to be done at stations 1 through 10 in succession. When volume rises, the takt times becomes shorter and you need two operators. You can have the operators follow each other through stations 1 to 10 — which Kei Abe called the caravan style but is now more often called the rabbit-chase system — or you can route operator 1 through stations 1, 2, 3, 9 and 10 and operator 2 through stations 4, 5, 6, 7 and 8 — which he called the baton-touch style.
The caravan style provides better job satisfaction to the operators but requires them to be fully multiskilled, and, experimentally, breaks down when you try it with more than two operators, because they all end up queuing behind the slowest one.
From your description, I assume that the Junkers Taktsystem involved the same team of operators following, say, one fuselage through multiple stations from start to finish of its assembly process, and that you had teams follow each other like groups of golfers through 18 holes.
In this case, the main difference with fixed-station assembly is that each station along the way only needs the materials, fixtures, jigs, tools and instructions for one single operation, and that, for each operation, these resources only need to be available at one station. Therefore, you need fewer resources, and the stations are simpler, because they don’t have to be changed over for every operation.
On the other hand, with the caravan system as you describe it, each team of operators must have all the skills needed to assemble the fuselage from start to finish, in the same amount of time at each operation, or else the slowest team will slow down all others.
I have seen this system work in the early phases of pilot production and ramp-up on a new line, and effectively support its debugging/improvement. At that time you can have teams of highly trained technicians. As it carries a unit of product all the way through, a team immediately sees the consequences of local problems on the whole process and can act on them. As volume goes up, however, you must involve hundreds of operators, and it becomes increasingly difficult to have teams fully cross-trained on all operations and to choreograph the work so that they are not in each other’s way.
Junkers may have been able to make it work with the takt times in days measured in days of the aircraft industry, but I don’t know of any car manufacturer doing it with takt times in minutes and seconds.
Lean is from Toyota, not Ford, and not 15th-century Venice boat builders | Michel Baudin's Blog
March 29, 2013 @ 11:37 am
[…] Early work at Junkers in Germany […]
Immer synchron im Takt
July 29, 2013 @ 1:41 pm
[…] Takt time – Early work at Junkers in Germany […]
Pitfalls of Takt Times | AllAboutLean.com
September 29, 2015 @ 4:01 am
[…] Knuf and Michel Baudin for this important piece of history – for more see the highly interesting post 1, post 2, post 3, and post 4 on Michel Baudin’s […]