We presented conclusions from this study as a session entitled “Leveraging Existing Custom Fabrication Abilities in Small Affordable Houses” at the AIA Conference on Architecture in Orlando April 27th. Approximately 150-200 people were in attendance, and several engaged through questions and conversation at the end of the presentation. The presentation slides can be accessed here. While we have concluded the formal pursuit of this study, the issues explored are regularly relevant to our practice. Visit www.curb.cc to check for new projects engaging custom fabrication and affordable housing design.
Following successful communication of the design intent through available software plug-ins, we acquired quotes from two manufacturers within a regional vicinity. Facility A uses automated saws and jig tables in the facility and Facility B has no automated machinery.
Despite the significant difference in the two quotes, the difference may not be the result of automation. The Facility A designer said, “There are about 285 total sticks used to build these trusses and 145 are unique. I don’t see automation significantly decreasing the cost until they build robots to do it and then the robots are paid for. Experience has shown me though that even when they build robots, 2 years later your robot is obsolete and the newer, faster, more productive one will cost more. It is a cycle.” Despite the automation, he states that complexity increases cost and the true savings occurs for large projects with like trusses. For this project, he says, the difference is likely no more than $100.
An issue of economies of scale
Without having a larger pool of cost estimates, the cost difference between the two manufacturers could be attributed to economies of scale, which may turn out to be the primary barrier to the cost-effectiveness of complex designs. A third quote we had hoped to have in comparison was from a very large and fully-automated manufacturer in the region. Likely to offer the most significant cost-savings for a complex project, this facility currently has tremendously demand from builders of large and repetitive projects and it is not cost-effective for the facility to take on a project as small and specialized as our test design.
With economies of scale, manufacturers can produce more product in a shorter amount of time because the process is repetitive and eliminates certain steps like re-jigging a table. Thus, taking on a smaller project where re-jigging must occur between each truss incurs at least the opportunity cost of producing many more trusses with a more repetitive project in the same amount of time. While it is possible that technology may increase opportunity for affordable or market-rate projects, our assessment is that the motivation to create this opportunity is in many ways misaligned with that of the industry and with the profit-driven culture of the economy. Even if a cost savings is realized by a manufacturer utilizing new technology, the savings may not necessarily be passed on to the consumer.
Design software has advanced to the point that any person with a bit of training can build a simple model, populate it with building components, and click “calculate.” The simplicity and potential time savings – not to mention the ability to bypass a notably flawed process of communication and iteration between architect and truss designer – seems to hold incredible value which can be passed on to the consumer. As we have sought to incorporate the most accessible truss design software plug-ins available to the architect in our designs for this research, we found ourselves able to work with a manufacturer in this capacity only through great initial effort. Hindered by a lack of access to software training, architects remain dependent upon strong communication between architect – component designer to achieve unique or customized component conditions.
Companies like Autodesk make revenue by selling licenses to their software and are therefore motivated to innovate to compete with other software companies for the business of the architect. Conversely, truss design software providers like MiTek and Alpine are driven by their sales of metal connector plates. The software is secondary, and although there is noteworthy competition for software innovation among the plate suppliers, architects are outside the target market. Suppliers have little motivation to consider how the architect may utilize the software, which is evidenced by the fact that the software licenses are not available for general purchase. Alpine’s ModelMap Revit plug-in, which simply converts a model to a format accessible to the manufacturer’s software, must be requested and a 1-year license acquired before it can be used.
Alpine does offer a Revit toolbar geared specifically toward architects and contractors called VisionREZ which sells for $600/year per license. Bryan Randall of Alpine says he gets a few calls per year from architects requesting access to the IntelliView software, which includes engineering features and is only available to manufacturers. He asks them, “Do you sell trusses?” and when they say no, he apologizes and explains that he is unable to provide the software. Furthermore, there is no training material available for architects using the free ModelMap plug-in and arguably not enough training material available for the paid software.
Randall explained that it is his personal desire to see architects more involved and model communication more streamlined, but this would seem to require a structural shift in the connector plate business model. The software provided by MiTek, the bigger of the two competitors, has even less flexibility. While MiTek’s plug-in is easier to acquire by downloading directly from MiTek’s website, the software does not interact with Revit, a common BIM software used by architects. The plug-in makes a Revit file compatible with MiTek’s design and engineering software, but no function allows the model to be reintroduced to Revit later. This is a way of keeping the software closed; if you wish to interact with the structural model, you can observe it in MiTek’s Sapphire Mobile Viewer. Alpine seems less concerned with exclusivity; their software is open source, a strategy which allows innovation to occur more freely because developers can build off of the progress of others.
It seems that the interoperability between Revit and component design software is inevitable, but indifference and even resistance within the industry remains an immediate hindrance to opportunity for this technology to benefit the architect and thus, in many ways, the design and end consumer.
Prior to attending the BCMC show in October, we had designed a 576 SF one-bedroom house based on one of the schematic design studies shown in a previous post entitled “Ceiling Landscape.”
The roof is a standard gable, but the ceiling contains multiple planes sloping at various degrees and heights, creating a faceted surface.
Software Troubleshooting: BCMC Software Demos
We took the Revit file for this project with us to a MiTek Sapphire demo set up at the BCMC show but discovered that Sapphire could not recognize the ceiling planes the way we had modeled them – as a generic model without a family designation. We modeled it this way initially for the ease of creating planes of various slopes and rotations on multiple axes.
Next, we went to Alpine’s booth and requested a demo. Alpine is MiTek’s primary competitor for the sale of metal connector plates and as such a software rival as well. We had the same issue with Alpine, but the representative there took some time to remodel the planes as roof assemblies. The model did not import, and we were unable to troubleshoot further due to time constraints. We left with the expectation that if we could figure out how to model the complex geometry as a ceiling assembly, or as a generic model assigned to ceiling family attributes, the model should be able to import properly.
Software Troubleshooting: Local Truss Manufacturer
We reached out to three truss manufacturers in our region with the goal of acquiring cost estimates for our truss package using MiTek’s and Alpine’s software to communicate design intent. We had the most success while working with Anderson Truss (manufacturer) and who partners with Alpine (software provider).
We began with the same model we brought to the BCMC show except with the ceiling planes modeled as in-place components using the ceiling family category and parameters. This did not lead to a successful import. Following this, we sought the guidance of Alpine representatives and in waiting for their response, tried importing a model where the ceiling was modeled as a roof assembly using different settings than were used at the BCMC Show. This also did not import. (With each attempt, there remains the possibility that the export settings were not correct, but with access to instructions or support limited by either software or staff, we could not determine whether this was an issue.)
This diagram shows phases of communication between the architect (A), the components designer (CD), and the product representatives (PRx) who represent Alpine. The process took place over a 3 month time period, during which most of the time was spent waiting for a return call or email and periodically following up with Alpine representatives. Ultimately, we were able to receive a cost estimate based on DWG files and an RCP drawing.
After this experience, we discovered a more successful method of modeling and exporting through further trial-and-error. The only model type Alpine’s IntelliVIEW is currently able to accept is a ceiling created with the designated ceiling tool in Revit. Below is the simplified process through which this technology has the most value.
The result was the ability to import the engineered trusses back into our Revit model.
Due to the complexity of the ceiling pattern, we had been unable to successfully utilize the ceiling tool earlier in the process, even with the help of representatives at the BCMC Show. Without access to simple instructions for the process and representatives willing to invest in our project, what may be considered a straightforward process became laborious. The promise for this software to ease communication and make complex designs more accessible remains for those who can self-troubleshoot effectively. However, without much instruction or assistance from the primary software providers, the process will unlikely become mainstream. The next post will address the economic outlook of the connector plate industry and why this may be preventing mainstream accessibility to this technology.
The Disconnect: Communication Between Designer and Designer
For architects, it is odd to view component manufacturers as “designers” because they do not typically make decisions determining the quality of space, aesthetic, and form. However, the component manufacturing industry is filled with “designers,” and they pride themselves on the skills required to fill this role. Truss designers are responsible for preparing the truss shop drawings, which entails determining truss spacing and the exact locations of the truss members according to what is structurally necessary, often reflecting some knowledge of field assembly and concern for avoiding field issues. Essentially, these designers are given the sometimes complicated task of “filling the box” that architects create, as Sean Shields of the Structural Building Components Association explains.
Through the attendance of sessions and conversations with manufacturers at the Building Components Manufacturing Conference October 19-20, 2016 we learned that the architectural design information provided to components manufacturers for any given project often contains inaccuracies or is incomplete. As a result, truss designers understand their role to be not only to “fill the box” but to also accurately define the box boundary, which the architect might have designed for nominal lumber or misunderstood the actual location of a top plate required for their anticipated ceiling height, for example. We heard a surprising number of reports of manufacturers receiving napkin-sketch-equivalents to design from architects. Even more often, the architect has no interaction with the manufacturer whatsoever, and the interaction is solely between the contractor-manufacturer; or, the projects they receive don’t involve an architect at all.
Time loss occurs for truss designers when they must resolve inaccuracies or fill in incomplete information provided to them. As a result of increased design time, the cost can be affected. In absence of direct communication, the architects’ designs become inflexible toward potential simplification or material conservation strategies. To this point architects’ limited direct contact with the building components industry generally is notable. Beyond improved communication, our research points to architects’ opportunities for leveraging custom fabrication possible in the component industry – if architects develop knowledge of the industry and its processes.
The Disconnect: Goals and Mindsets
The logic of manufacturing is rooted in economies of scale, conflicting with the efforts inherent in the design of small, custom projects. Despite this logic, plant managers prefer business strategies involving personal relationships, not always adhering to the most “efficient” methods of production. For example, word of mouth is preferred to strategic marketing, and where excessive design time causes a fee change, the increase is generalized based on the designers’ experience with past projects, thus not highly calculated.
Leaders of the Metrics for Managers session at the 2016 BCMC explained that the tendency to “under-estimate complexity” is costly for manufacturers, suggesting that complexity should be avoided. Customers with overly-complex requests are known as “shop-stoppers” who hinder profitable operations. While an iterative design process is often desirable for the architect, it is unprofitable for the truss designer.
Sean Shields observed that, despite these general observations, the business tactics manufacturers take vary widely, and there are occasions, although rare, when truss designers and architects effectively work together with mutual benefit.
Through the advancement of automation and software capabilities, manufacturers boast an ability to produce unusual and complex truss designs otherwise inaccessible or unrealistically expensive. After visiting one facility with impressive capacities for automation, and another facility without any automated machinery, we see a great promise for design customization utilizing the speed and precision of automated saws and jig tables. However, whether the consumer or client is able to access this value remains to be seen. Although the time-savings is evident, the expenses for manufacturers to automate their facilities, for truss designers to design unusual or multiple unique trusses, and for framers to learn or accept new framing methods may prevent the value from reaching the clients of custom, architect-designed residences who might benefit from this technology..
The structural component system runs most efficiently and affordably with economies of scale. Despite the 6 second speed of a jig table realigning to a different truss, manufacturers presently place high value on the savings that occur with repetition. At the Metrics for Managers session at the 2016 BCMC, using templates and avoiding complex projects were suggested methods for conserving time. This suggestion seems to contradict with the values of smaller, minimally automated component manufacturers like Kevin Brown of Brown Truss Company, who considers every project a custom project. While the promise of customization and the desire for economies of scale appear to be in direct conflict, the accessibility of custom components in market-rate housing depends upon whether the values and processes of the plant manager and framer/installer can be aligned with those of the architect and client.
The “Metrics for Managers” session at the 2016 Building Component Manufacturing Conference provided insight to the complex issues plant managers grapple with when organizing people and processes in their facilities. Below are some of the concepts discussed.
“My primary job is to buy and sell lumber…I just package it funny.” | Simplified, the goal is to input lumber and output a different kind of lumber, one with higher value. The concept reminds us that manufacturing is driven by efficiency and not design.
“Guide your sales force to the things that make you money and away from what loses money.” | An example of this is recognizing good customers. Many customers are unreliable and ultimately cause wasted time for the truss designer. Calling customers frequently to confirm, recognizing signs of an uncommitted customer, and securing a repeat customer base are ways to focus a sales force.
“Everything works great when everything is working.” | This is the idea that, at maximum efficiency, supply equals demand, nothing is stored, and everything is delivered exactly on time. However, this is very rarely the case, and the difficulty balancing supply and demand can result in more than just shortage and surplus. The more time lumber sits in the facility, the more likely it is to change colors, warp or mold. Ideally, the lumber gets used right away to minimize loss. However, if the price of lumber is expected to increase in the future, storage may become more desirable.
“If you don’t need them tomorrow, I won’t build it today.” | The most difficult thing to predict is when a project is going to start, but getting as close as possible is important. The need to use lumber promptly and store as little product as possible prompts manufacturers to wait to produce until right before they can deliver. Lumber warping can occur in as little as 10 days of storage.
Mistakes Cost | The one element manufacturers don’t want to put pressure on is the productivity of truss designers. If a mistake can be traced to the manufacturer it can cost a lot to replace. Truss designers take on the most risk with multi-family housing projects because the demand is for quick production and there is pressure to shorten design time.
Many manufacturers will refuse to address issues which are outside of their legal responsibility, but the authors of this session suggest that the extra effort to anticipate field issues is beneficial for all parties. Stair locations are a frequent cause for callbacks, and strategies such as double-checking stair head-heights in plans and avoiding trusses over stairs when possible are recommended. Labeling drawings well including locating plumbing fixtures, recessed lights, range hoods, and vents can ensure proper placement and even double-check the truss designers’ truss spacing. In the areas where trusses are in danger of intersecting other objects, it is best to frame around rather than over as much as possible. The framers can always fill in, but fixing a misplaced truss can be very costly.
Framing the American Dream (FAD) is a study conducted by the Structural Building Components Association (SBCA) comparing component and stick framing through the construction of two identical houses. The study was originally conducted in 1995 and again in 2015. The SBCA collected and visualized data from its study to create a “primer” for educating framers and builders about the benefits of utilizing component industry products. Though it is intended as immediate marketing material for SBCA members, there is a lot to learn from the primer.
Time lapses of the 2015 study can be seen here:
On 29 September we visited Huskey Truss and Building Supply, a highly-automated building components manufacturer in Murfreesboro, Tennessee. According to Jackie Crutcher, the plant manager, Huskey’s motivation to automate its processes is driven primarily by anticipated return on investment. Other factors incentivizing these changes include Nashville’s booming development, which has them so busy they are not entertaining new customers, and considerable challenges in maintaining a reliable workforce. While clearly demonstrating the power of automation technologies, the visit also raised questions about the value of human labor in the construction process and whether or not such mechanisms, at their most advanced, can be harnessed in service of architectural expression or whether they will enforce their own logics that favor more of the same.
Each workstation at Huskey is equipped with a computer, and large screens displaying the components in production hang above many of the workstations. The plant employs around 110 people, but the labor turnover is significant and finding replacements can be difficult. As a result, Huskey employs automated machinery and assembly methods that simplify tasks or require fewer people to operate.
The company began its transition toward automation just a few years ago, and while most manufacturers have too few orders to justify the expense of the equipment, Huskey experienced a significant benefit right away. One saw which cost the company about $150,000 had a less than 9-month return on investment and requires a lower skill level to operate than the saw which previously accomplished the same task. As the machine cuts each piece of lumber, it makes additional cuts in the excess to create pre-cut floor truss members from what is often considered waste or would typically require an additional workstation to produce.
Having display screens throughout the facility helps to quality-control the production and helps Jackie to better manage the facility. The amount of linear feet of lumber cut and in production is updated live on multiple screens, so that he can check the productivity level from most locations within the facility. If the cut number falls too close to or behind the assembly volume, Jackie knows that something must be adjusted to maintain productivity at the assembly stations.
Huskey is one of the few manufacturers with consistent wall panel production, enough that panels account for about half of the plant production. Each panel or truss passes through multiple workstations before leaving the facility, enabling the correction of most errors prior to on-site delivery.