Wow, it's been a long time since I last posted to this blog. Much has changed with me and with our world. I seem to have some new thoughts I want to share so I'll add them here, as they certainly build on many of the ideas I've previously posted.

First here are a few links to web-based archives of some of the work I've been producing instead of posting new blog posts.

An article I wrote for Independent School magazine, house organ of NAIS, about the 1% and independent schools. https://www.nais.org/magazine/independent-school/fall-2012/our-1-problem/

YouTube: Videos made for various tasks/interests. https://www.youtube.com/@ottercreekltd/videos

SketchUp: Many of my SketchUp models. https://3dwarehouse.sketchup.com/search/?q=fred%20bartels&searchTab=collection

Thingiverse: Many of 3D printable designs. https://www.thingiverse.com/fredbartels/designs

Mastadon: I left Twitter in November, 2022, where I still have thousands of posts with the user name @FredBartels. My Mastodon posts. https://mstdn.social/@fredbartels

I think that's pretty much it. Much that I created and shared to Google+ and NING is now gone. 

I finally bit the bullet and decided to prototype a curved green roof. I set myself a very limited budget, and the resulting prototype is an experiment, not an actual roof structure. It is a testbed for trying out some ideas.

The underlying structure.

Early prototyping.

The waterproof membrane. 

The parts.

The mesh framework for holding the soil.

Some final reflections about the practicum project.

• Creating complexly curved roofs is really hard. There are a lot of issues that need to be addressed and very little information, technology or practice available for guidance. As a result I've had to come up with solutions to many challenges on my own. I hope, as I move forward, to work more with others also interested in exploring this area.
• Hempcrete is a building material with much potential, especially in the natural building community. This article in nauhaus lays out the arguments for hempcrete quite persuasively. I hope someone in the Yestermorrow teaching staff will do a project with hempcrete in the near future.
• Yestermorrow has been a supportive community for my explorations. Not as helpful as I had hoped, but I suspect that's because my ideas are pretty far "out there." Sadly for me, I didn't find my architectural tribe at Yestermorrow; perhaps because at this point I'm a tribe of one. I did learn a lot from all the classes I took, and really appreciate the knowledge and caring of the instructors and staff. 
• Maybe a way forward is to combine traditional rectilinear architecture with curvaceous triangular architecture. Some interesting possibilities there.

  

• Thanks to Bill Hulstrunk for his quiet support as my practicum advisor. Thanks also to Kate Stephenson for all her support with presenting virtually. Finally, special thanks to Ed Lowans for helping me understand the importance of context... and how architects think.
• A special thanks to Tom Simon and Emily Payton at Hempfully Green for their ongong support and advocacy of hemp as a sustainable building material. They also very kindly supplied me with materials for my hemp wall experiment. 
• If you would like a copy of the SketchUp model of the studio just drop me a line and I'll be happy to send along the file.

The sequence of images of the model I showed during my practicum presentation. The final image is a detail of the roof/wall connection.

Here is the video I made for my practicum presentation. I followed the video with a brief "walkthrough" of my SketchUp model of the organically curved-roof studio.

A couple of renders of the completed model.

Added the hard mache layer to cover and smooth out the triangulated roof decking. 

Two images below illustrate solar gain issues. The first image has winter solstice shadows. The sun penetrates deeply into the structure. The second image has summer solstice shadows. The sun barely enters the building. Both images were generated using the same time of day.

A render made later. Note the roof vents.

Finished the roof framing and the triangular roof decking pieces. A big job. Nice to have it done. Work from this point forward should be much more interesting.

In the embedded video I go through some of the steps needed to create a triangular roofing board and supporting rafters for the organically-curved roof I'm "building" for my Yestermorrow practicum project. Please note that the boards I call joists are actually rafters. (My bad, but I think you will get the idea.)

Also please see the two images below the video. I've finished the hub placement. There is an end in sight!

Another day of plowing through making irregular triangles and their support beams. If the tediousness of doing this forces me to come up with a better solution than it will have been worth it. Otherwise... not so sure. 

A couple of images showing the progress.

Grinding out the roof structure. I'm getting faster but it is going to take three or four days to complete the whole roof. After building the model I should be cured of any desire to build a real world version.

Also added a couple of windows since the previous post. All in all I'm making good progress toward getting this thing done on time.

Working on the final model for my project. Making progress but it's slow tedious work. Glad this is nearing an end so I can switch to more viable designs. There must be an easier way to do organically curved roofs but I'm not seeing it yet. Probably something involving molds and lots of technology. 

The two images below show the framework for the roof overhangs.

I made much progress over the past week or so. Following is a sequential photo essay that captures the "big moments".

Two inch diameter dowel pieces attached to the top of each post with Timberlock screws.

Previously created frame deconstructed for parts. It would be really nice to have the hub joist hangers available as a custom part like this.

Assembled hubs placed over the 2" dowel pieces. Turns out that tightening the bolts that hold the joist hangers in place is a bit of a challenge with the dowel in the center of the tube segment. This can be solved by temporarily removing the dowel pieces, one at a time, after putting together the triangular frame.

First beam/joists hung between the post hubs. I'm using 3/4" plywood with a depth of 12".

First triangle completed. The beams are 12" deep at the posts, 8" at the eaves. I jury rigged a jig to help do the length and angle measurements. It worked but I'd build something better and easier for an actual production.

The three triangles of the prototype completed. When I ran out of plywood I started raiding my scrap wood pile.

Fascia details. This, for a change, was actually easy to do as measurements can be taken off the existing framework. The cuts to match the adjacent fascia boards would be a bit of a challenge.

A sample cross-brace to reduce the span for the covering plywood triangles. Worked well but took more time than I would like.

An experiment using 2x4 pieces to "float" the roofing triangles above the framing. This would be fairly easy to do and could be used to increase the roof thickness for a higher R value.

Finally, a device I created to "measure" the irregular triangular shapes of the the roofing framework so as to be able to cut matching triangles from plywood sheets. I was originally planning on filling in between the joists with hempcrete, but now believe that there is likely to be too much movement for this approach to be viable. It is possible to "box" in parts of the eave triangles by cutting a triangle that will rest on top of the hempcrete wall on the in side. This "box" could be filled with dense pack cellulose. It was in trying to figure out how to easily measure the shape of the bottom of this "box" that I came up with the idea for the triangle shape reproducer.

Again, lots accomplished. The prototype is essentially finished but serves as a great test bed for further experiments. 

Lots of progress. Just needed some cooler weather to make working bearable. I learned a great deal about how to build better forms. The approach I used here is too complicated, and has other issues, but I now see how to easily make simple good formwork. I'll elaborate more in a future post.

I'm quite pleased with the look of the wall. Love the curves.

Some shots of the formwork I used for this prototype. The first three pictures show how I set the window boxes on a few inches of hempcrete, before putting up the wall forms.

Next a couple of shots of the formwork filled with hempcrete.

Very hot today so I only worked for a few hours, but they were productive. I added the insulation between the 2x4s of the sill plate. Foam board for the straight runs and expanding poured foam around the posts. Then I built a couple of plywood boxes for the window openings. (Ohh, for a good table saw.)

I really don't know what the tricks are for doing things like this so I just work along slowly and try to learn as much as possible as I go. Tomorrow I should be able to start mixing and adding the hempcrete.

I'm starting to wonder what I'm going to do with this thing when I'm done. Probably keep it around for a while and then deconstruct it. It's all a learning exercise.

A few pictures of today's work.

It was a day of two steps forward, one step back. Lots of rethinking caused by the reality of working with real materials at full scale. In general though, it was a productive day in which I discovered I could avoid some complicated steps that turned out not to be needed. For instance, there is no need to cut the tops of the posts into circular shapes to match the pipes that will sit on them. I'd thought I needed to do this so that the rafter bottoms wouldn't hit the posts, but realized that I could just trim back the rafter bottoms as needed to clear the posts.

Another thing I was reminded of today is the importance of making jigs for any repetitive process. It probably helped that Bruce Beeken, the furniture maker who gave Yestermorrow's summer lecture number two, stressed the importance of jigs in his work.

Here is a picture of a simple jig I made to help build the formwork jig to hold the hardboard wall forms in place.

Then here is the wall form holding jig, made with the help of the first jig.

Finally, the form holding jig moved up the post to near where it will be attached.

Did some work on the wall prototype today. I completed the wooden parts of the sill plate. The hempcrete will go on top of this. The work went pretty well. I first made a jig to help figure out the curves.

Then using the jig, and a 12" by 26" piece of hardboard representing a straight wall segment, I layed out a wall section with three posts on a piece of 3/4" plywood. Made a few mistakes but nothing major; the usual sort of thing when building a prototype. Tomorrow I'll add the truncated posts and the sill insulation, then build the form for the hempcrete, which will include a window opening. It's fun to be building at full scale, even it's only a prototype.

The mache-covered wood shed has been drying out in the sun for the past few days. Yesterday I noticed some shrinkage on a couple of the sides. As the mache dried it pulled back from the edge an eighth to a quarter of an inch. This may actually be an opportunity as I've been wondering how to handle the edge condition. I used some Liquid Nails adhesive to bridge the space between the mache edge and the plywood edge. After I've painted the roof, and its had a chance to weather, it will be interesting to see how the adhesive edges hold up compared to the mache edges. In either case the solution is not ideal and I'm still looking for a better approach.

Here are a few images of the roof and the shrinkage.

I've been working on a truss system to replace the ceiling rafters. The advantages to using a truss instead of a solid beam are that less wood is needed (in theory) and there is less thermal bridging. In order to get an R-60 dense-pack cellulose roof a thickness of around 18" is needed. For R-80 24" is needed. For those kinds of roof thicknesses a truss especially makes sense. Here are a few images of my latest truss designs.

A sample of Stuc-o-Flex "Waterway" rainscreen drainage mat arrived in the mail yesterday. It is the material I hope to use between the triangular plywood roof sheathing and the smoothing mache layer. 

I mixed up a small batch of mache and applied it to the top of the mat. The mache will take a few days to dry but after working with "Waterway" a bit I think it will work quite well. I love this mixture of very high tech and very simple technology.

OK, that was quick. The advantage of being on vacation and having some time to play.

I mixed up a batch of smooth mache using the ammonium sulfate cellulose; about a gallon in quantity. Then I used a plastering trowel to smooth the material instead of my hands. What a difference. This was a enjoyable process that produced a beautiful result. Very encouraging!

I added a second layer of mache to the test roof yesterday. This was a smooth mache mix which -after it dries in a week or so- will be touched up with joint compound and then painted. Although the layer turned out well enough it was frustrating to build. Partly this was because I was mixing mache in small quantities (about a gallon at a time) and partly because I didn't start with a good technique for spreading the mache, especially around the edges. I used my hands again to push and mold the material, a process that is both very time consuming and leaves behind marks of the hand working. To use this material at a larger scale it will be necessary to involve hand tools (like trowels) to spread and mold the mache. My next experiment will be in this area.

I'm still working out how to handle edges. In this case I will sand/shave off the uneven mache along the edges once the material has dried. Definitely looking for a better solution.

A couple of images of the roof with the second layer.

A double-shell roof. Latest ideas and a request for help.

Coming up with a method for venting an organically-curved roof was a challenge, but thanks to the help of John Unger Murphy and Bill Hulstrunk I believe I have a solution. In essence the roof consists of a double shell. The inner shell is composed of plywood triangles, which are attached to the framing, and covered on the outside with a vapor permeable "paintable" membrane like Tremco ExoAir 220. The outer shell is composed of hard paper mache covered with a paintable waterproof membrane such as roof mastic. In between the two layers would be a drainage mat -perhaps the Stuc-o-Flex WaterWay Rainscreen 9120 - which would allow moist air to flow up to the dome peaks to vent via small cupola like structures. The rainscreen would also allow the mache to "float" free of the plywood, helping to prevent cracks. That's the theory anyway. If you see problems with this approach please share your concerns!

Here are a couple of sketches that may help you visualize the double-shell approach.

There are two components for which I could use some help. These are the waterproof roof membrane and something that could be added to the mache to help it dry faster (the mix I'm using now takes about a week to dry). Bill Hulstrunk suggested I try duct mastic for the mache mix and I'll give that whirl in the next few weeks. Any other suggestions are much appreciated!  

Worked on my first large-scale use of hard-mache to smooth a roofing surface composed of irregular triangles. Used about 5 gallons of mache to put a first coat on the small structure that covers my firewood. 

It went well, and the new surface is definitely more rounded and smooth. A second less-rough coat of "finish" mache will go on in a few days; after this coat has a chance to dry.

The following two photos of a test piece provide some sense of how the finish coat will go over the rough coat, and then on top of everything will be a layer of waterproof mastic.

Wall framing on a structure with an organically curved roof is going to be a little different. I've played around with lots of ideas. This post reflects my latest thinking. Rather than attempt to explain in writing I've created a short video which is embedded below.

Insulating the foundation of a building is important in order to prevent the concrete from acting as a thermal bridge. Othen this is done by applying foam board to the outside of the foundation's concrete walls. With a curved foundation wall I was curious how this might work. Yesterday I did a little test. I scored one side of a piece of 2" foam board and was then able to bend it without a problem. This should work very well with my foundation wall, which consists of short curves and longer straight runs.

I'm considering using Hemcrete® for the walls of the organically-curved structure I'm designing for my Yestermorrow practicum project. I learned about Hemcrete from Thomas Simon (of Hempfully Green) during Robert Riversong's Hygro-Thermal Engineering course at Yestermorrow last March. Thomas brought some samples of Hemcrete and shared them with the other students in the course during breaks. Recently he very kindly provided me with some materials so I could mix up and experiment with my own batch.

I built a simple curved form yesterday using some scrap wood. The curved sides were formed with brown fiber board, which can easily handle the gentle curves of my design.

Yestermorrow is a wonderfully creative and quirky design/build school I've been attending since last September. I'm in their Sustainable Building and Design certificate program and have just one weekend class left to complete my coursework. I've spent over six weeks in residence at the school since starting in September and have a pretty good sense of Yestermorrow's culture. My background is in education so while I've been absorbing all kinds of great information about design and building I've also been observing and thinking about the pedagogy used to convey that information. If I have any design strength it is in the use (and teaching of) SketchUp. Some of my SketchUp work can be found here and here.

In my three week sustainable design core class I was the only student who made significant use of SketchUp. I wrote about this experience in a previous blog post. There was an interesting discussion about SketchUp at the end of the my project's crit session. During this discussion it became clear that the Yestermorrow teaching community is struggling with whether to continue teaching traditional drafting techniques or bite the bullet and move to SketchUp. John Connell, Yestermorrow's founder, participated in that discussion and mentioned that he was going to teach a sustainable prefab design course in April in which SketchUp use would be required. I attended that course last week (see previous post) and want to share some thoughts about how SketchUp was taught/used.

Most students in the class arrived with very little SketchUp experience. There was an optional Sunday evening session for students who wanted help getting a handle on the program. The intuitive nature of SketchUp's interface, especially the Push/Pull tool, allows students to start using the program quickly, but without a solid understanding of good SketchUp practices a quick start can often lead to signifcant modeling problems down the road. That dynamic very much characterized the use of SketchUp for many in the class. There was much more frustration than there needed to be, and the students' house designs could have developed much further than they did.

So the question is, how to quickly develop good SketchUp use practices?

Here are some suggestions.

1. Require students read SketchUp for Dummies by Aiden Chopra before the class. Aiden is an architect and SketchUp employee. His book is by far the best introduction and reference to SketchUp that I've found.

2. Have a scanner in the classroom so that floor plan drafts done on paper can be scanned and imported into SketchUp. 

3. Teach students how to use groups and components as early as possible. All significant elements of models should be grouped as soon as they are created.

4. Show students how to grab site information from Google Earth as soon as possible.

5. Have students set up and memorize (or learn the default) keyboard shortcuts for at least the Move, Orbit, Push/Pull, Line and Hand tools. 

6. Require all students to have a mouse.

7. Teach students to pay attention to the hints that SketchUp is constantly providing.

8. Teach students how to effectively use the Dimensions box.

9. Provide a SketchUp house model made with best practices. This model could include basic elements which students could copy into their models.

In the following video I build a simple house model making use of some of the practices listed above. If students master the techniques I present in the video they would tend to have a much more positive experience with SketchUp, and teachers could focus more on design concepts and less on SketchUp. (Note: This video was made in one take and a number of the illustrated techniques could be better explained in shorter better prepared presentations.)

I had the distinct pleasure of attending a course on sustainable prefab housing last week at Yestermorrow. The course was taught by architects John Connell (who founded Yestermorrow) and Giocondo Susini. One couldn't ask for a more thoughtful, kind, knowledgeable and entertaining pair of teachers. John Connell is massively connected into the architectural and building community in the Mad River Valley and surrounding environs. These connections benefited the class via the field trips and speakers John was able to arrange.

My classmates were a delightful mix of folks hailing from as far away as Austin, Texas and having a wide range of design/build experience. Everyone was in good spirits and we all shared many ideas through the week as our projects developed.

I worked on a possible retirement house design based on modules offered by Huntington Homes, a family-owned company whose factory we visited early in the week. Below are a couple of renderingsof the ranch-style home model I developed that would use four HH modules for the main structure. The garage and screened porch would most likely be stick built on site. Larry Roux, a HH homes sales manager, was kind enough to provide a preliminary quote for the building, which I'm pleased to say was within our price range.

HH is producing a very well constructed module with a 12" thick double 2x4 wall using dense-pack cellulose for insulation. This wall assembly struck me as a good compromise solution to a super-insulated wall, reaching an R-value of around 40.

We also had a great visit to the Connor Homes factory. Connor produces all the components for its higher-end houses and ships them out to the building site as contractors are ready for them. They produce really beautiful versions of the region's tradtional vernacular architecture.

In addition to the field trips we had guest lectures from representatives of structural insulated panel (SIP) maker Vantem and the famous (at least to fans of This Old House) Benson Wood timberframe/prefab company. Both lectures were extremely informative.

The class used SketchUp right from the start for all our design work. Students were required to bring a computer to the class capable of running SketchUp. I applaud Yestermorrow for taking this step, and hope that they apply it to the core class for the Sustainable Design and Building certificate program. There are still some rough edges in the approach Yestermorrow is taking to using SketchUp (more on that in my next post) but it is definitely the right way to go in a world with increasingly powerful digital tools.

If you are considering building a home you could do yourself a big favor by attending a future meeting of this course.