Rivets

 All about rivets

Well not "ALL" about rivets but the rivets used on the Clipper. Here's a reference for most types of rivets.

Solid Rivets

I've found that the majority of the solid rivets used on the Clipper are of the type AN455AD-4-X. 

"AN455" is a specification for the size and shape of the rivet head and is commonly referred to as a "Brazier Head". For any give diameter of the rivet, the Brazier Head is slightly wider than a AN470 "Universal Head" rivet. This wider head helps distribute the load on the soft aluminum.

"AD" refers to the type of the material used. AD means 2117-T4 aluminum. These are sometimes referred to as "structural strength" rivets. They are made with 2117-T4 are identified by the a small visible dimple in the center of the head of the rivet. 

"A" type rivets (ie AN455A-4-X) are made with 1100 aluminum which is nearly pure and relatively soft. "A" type rivets have no dimple, the head is visibly smooth. An important physical property of rivets is "sheer strength"; AD rivets made from 2117-T4 have rough three times the sheer strength of A rivets made from 1100 aluminum.

Another item of note for "A" vs "AD" rivets is that "A" rivets used with 2024 aluminum can apparently lead to galvanic corrosion and some aircraft manufacturers (e.g. Boeing) recommend against using "A" rivets with 2024 aluminum (2117 aluminum rivets are fine). Looking at a galvanic series chart it appears that aluminum alloys have a broad range of potentials and this leads to the possibility of corrosion, but it also appears significantly less than other problem areas that we encounter in the vintage trailer world (e.g. steel vs aluminum).

"-4" is the diameter of the shank in multiple of 1/32", so a "-4" means 4/32" or 1/8".

"-X" is the length of the rivet shaft. This is dependent upon the thickness of the material stack you are trying to fasten together. The rule of thumb is that the length of the rivet should be the thickness of the material + 1.5 times the diameter of the rivet shaft (ie material thickness + 3/16" for a -4 rivet). This added length becomes the "work head" (the part of the rivet that gets smashed/mushroomed) when the rivet is installed. You can buy rivets of a standard length and cut them to the correct size before using them with a special tool.

When repairing the Clipper I'd like to use solid rivets of the type AN456AD-5 (note the 456 vs 455 and the -5 vs -4). These are known as "Modified Brazier Head" rivets. Typically when you remove and replace a solid rivet you need to drill out the hole to the next size up, in our case I'll be going from -4 (1/8") to -5 (5/32") diameter rivets. The function of the "modified" head is that a AN456-5 head is roughly the same size as AN455-4 head so that when you replace the rivet with the larger size the head will appear the same (ie you won't notice the newly replaced rivets). 

HOWEVER AN456AD rivets are hard to find so you're stuck using either AN456A (ie softer, non structural) rivets or AN470AD-5 structural universal rivets. The head size is roughly the same and it's difficult for the untrained eye to tell the difference (the head has a slightly taller crown). In my opinion its easier to spot where a repair has been done with 456 "A" rivets than 470 "AD" rivets because you typically view them head-on (the crown height is not apparent) and the "A" rivets won't have a dimple and that's fairly obvious.

Blind ("Pop") Rivets

Blind, or "pop" rivets are what most people think of when someone talks about using rivets (either that or images of 1920's hot riveting skyscraper ironwork). The are called "blind" because they are inserted from one side, hence the user doesn't need access or is "blind" to the other side. They are commonly called "pop" rivets because most are set by pulling a mandrel that separates from the rivet with a "pop" action and/or noise.

There are a huge number of blind rivets available. Most have "open" ends when the mandrel pops out so they will let air and water pass through. This makes them less useful for exterior applications. However there are some types that are "sealed" or have "closed" ends.

Another type of blind rivet that is popular for aluminum trailer repair is known as an "Olympic style" or shaved head rivet. These rivets work like a typical blind rivet but the mandrel breaks off proud of the head. The mandrel can then be "shaved" down to resemble a solid rivet head. This allows panels to be repaired or replaced with ready access to the back side of the panel.

Blinds rivets can fasten items very securely but they are not considered "structural".

Tubular Rivets

Tubular rivets look a lot like solid rivets but the end is hollowed out. They are set using a pinching mechanism (vs rivet gun) with a die (aka "set") on both ends vs a set on one end with a buck on the "work end". Tubular rivets are not structural but they do fasten metal pieces together well and can be set very thin. They can also be used to form a pivot point in light duty mechanical parts.

For the Clipper the tubular rivets are used in the Hehr "Standard" windows. They are used to fasten the frame together and also as a hinge pin in the window "link levers" that are used to open and prop open the windows.

"Tinners" Rivets

The wheel wells are constructed with galvanized steel not aluminum. The steel is galvanized, ie coated with zinc, to prevent corrosion of the steel. My research led me to the use of galvanized "Tinners" rivets to assemble the wheel wells.

A "Tinner" is a tinsmith. Tinsmiths  make and repair things made of light sheet metal, hence the name tinner's rivet. Tinners rivets come in a variety of materials and are sized based on the approximate weight of 1000 rivets. For example approximately 800 1.25 tinners rivets come in 1 lb of rivets and have a shank diameter that ranges slightly smaller than 1/8". These are known as 1.25 tinners rivets because 1000 would weigh about 1.25 lbs.

I used 1.25 tinners rivets from Hanson Rivets to fabricate the wheel wells. I actually purchased them through Grainger because it was cheapest. I used a flat set in my rivet gun and a standard bucking bar on the "work end". I found they install very much like typical solid rivets.

Subfloor Material

Advantech Subfloor

Just a short post to discuss my downselect of subfloor material to Advantech by Huber. I special ordered it via Home Depot for about $60/sheet (including CA fees and minimum order costs). This is actually much cheaper than my 2nd choice which was marine grade plywood (which I used in the Boles Aero), and I won't need to use penetrating epoxy as a sealer.

I was considering the following options:

1. Coosa Bluewater 26 - suuuppppeeerrr expensive composite that won't rot
2. Marine grade plywood - expensive but less prone to rot than normal plywood
3. Advantech or other engineered subfloor

After much deliberation I chose Advantech. It's an engineered subfloor specifically designed to withstand soaking moisture during the construction of a home. The first time I read about it I was repelled because OSB is notorious among the trailer restoration crowd and Advantech looks like typical OSB. But after some investigation I found that it's being adopted by vintage trailer restorers, particularly when they don't want to spend $300+ per sheet for Coosa board.

I had some initial difficulty finding it but contacted Huber Engineered Woods and they recommended that I special order through my local lumber store. It's not a normal stock item because local contractors won't spend the extra money to use it. Generally only large corporate builders use it because they recognize the cost tradeoffs associated with rework.

In any case it took about 3 weeks to get it delivered to the local Home Depot, but it's now in my garage waiting to be installed. It's the nicest and most uniform sheet good I've ever unloaded. Hopefully it lives up to it's promises.

Wheel Wells

Wheel Well Fit Check

The wheel wells need to be placed on the frame and under the subfloor to prevent water thrown up from the road from rotting the subfloor. I was excited about installing the subfloor when I re-remembered this step. So I stepped back and went about fabricating new wheel wells.

The wheel wells for my Clipper were missing (along with most of the interior walls, etc. Based on internet posts I believe the original wheel wells were constructed from simple plywood boxes. The only remnants of those boxes was the outer aluminum wall that is riveted to the inside of the exterior trailer skin.

Remnant of Original Wheel Well

I had decided a while ago that I'd fabricate some new wheel wells. I really didn't want to use plywood boxes that would rot out fairly quickly because replacing them would mean tearing down the entire trailer.

I considered several alternatives but ended up designing a simple wheel well and ordering the metal pieces from SendCutSend. I'd considered purchasing sheet metal and cutting it myself, but after pricing in shipping, handling and waste it was going to be about the same price.

I ended up using 0.036" G30 galvanized steel which falls somewhere between 22 and 20 gauge. That's pretty stout for wheel wells and "the internet" said it was a much better choice than using aluminum because it's stronger. The G30 designation indicates how much zinc is deposited for corrosion protection (G30 is on the low end). I would have preferred G90 grade which has more zinc deposited for rust prevention but it was not available when I ordered. Later I'll talk to some additional efforts I made to prevent corrosion.

Laser Cut Parts from SendCutSend

Each radius of the wheel well is made of 3 identical cut pieces. The sidewalls for each wheel well are 2 identical cut pieces (6 and 4 total for the 2 wheel wells). I've found that ordering a quantity of 4 or more from SendCutSend is the "knee in the curve" for pricing so when possible I make design choices that allow for part commonality. In this case the wheel well is slightly taller than required (3" above the max axle loading deflection per Dexter) so that I could use a common design for the radius pieces.

I used my 36" straight sheet metal bending brake to make the appropriate bends. Of course I misbent a couple bends on the first one and had to hammer them back (which is do-able with steel) and rebend it.

Parts Bent into Shape For Assembly

For assembly I wanted to make encourage corrosion prevention. I considered a few options but settled on using galvanized steel "Tinners" rivets. These are flat head rivets used by "Tinners", ie sheet metal fabricators. I'd never heard of them before but they are available from Hanson's rivets and their distributors. I could find very little on the internet about their use. The sizing is odd compared to other rivets: they are sized according to the weight of 1000 rivets. I used a 1 1/4 lb rivet which is just under 1/8" diameter (Hanson Rivet part number TRSL01.25) which gives about 800 rivets per pound.

Small number of 1 1/4lb (1.25) Tinners Rivets

Drilling and setting these rivets is almost identical to the aluminum ones except they have a flat head vs a "universal" or "Brazier" head. I used a flat rivet set on my rivet gun. The lack of a domed head to center the gun was a little challenging but not much of a problem - fortunately no one will be able to see the few occurrences where the gun skipped off the rivet head.

As is my practice with the aluminum skin panels, I used Sikaflex 221 between the panel seams. So the process is clamp; drill, cleco, drill, repeat; disassemble; spread Sikaflex 221; reassemble with clecos; rivet removing clecos as you go. The Sikaflex makes it a little messy where it squeezes out but it's easy to clean up.

Wheel Well Assembly (1 of 4)

Wheel Well Assembly (2 of 4)

Wheel Well Assembly (3 of 4)

Wheel Well Assembly (4 of 4)

Once assembly was complete I sealed the inside seams with a liberal application of Sikaflex 221. Then I sprayed the inside with Rustoleum Professional Grade Rubberized Undercoating Spray. It took a can for each wheel well to accomplish the 2 recommended coats (wear a respirator!) I chose the Rustoleum product based on some YouTube comparison videos. It's not as thick as I'd imagined and I'm a little concerned about durability. I was hoping for something similar to bed liner. 

I also sprayed the outside of one sidewall on each wheel well. The outside of the sidewall will sandwich up against the inside of the trailer's aluminum skin and be riveted to it (it replaces the aluminum box side in the image towards the top of this post). The zinc coating of the galvanized steel should prevent dissimilar metal corrosion between the wheel well and the trailer wall but I thought an insulating coat of rubber wouldn't hurt.

Wheel Wells Drying after Undercoating

One last note. One of the sidewalls of each wheel well is not flanged. That sidewall will be riveted to the inside wall of the trailer. I left the sidewall long and plan to trim it to the shape of the exterior wheel well after I wove the shell onto the new frame. I did trace the old wheel well onto the new sidewall, but I'm not 100% confident of the final fit of the shell on the trailer so I wanted to leave myself some wiggle room.

Dry fit of the new Wheel Wells

Overall I was very happy with the way this subproject worked out. It was fairly straight-forward, not too time consuming, and reasonably priced (about $250 for materials).

Outrigger Design

 

Finite Element Analysis of Perforated Outrigger at ~2.5x Nominal Load

I've been doing a little design work in a tool called FreeCAD. As the name implies it's a free computer aided design package that can be used to design and analyze solid objects. Originally I started using it because I needed to fabricate some Hehr Standard window link levers (the levers used to open windows) for my Clipper. I wanted to user a laser cutting service to make some stainless steel levers and they needed the information in a digital file format. FreeCAD (and many other programs) can generate this file and there was a large amount of training available on YouTube.

In any case I had some experience in FreeCAD before I started the trailer frame design process.

I did some rough paper sketches of the trailer frame I wanted to design. I then did a more refined drawing in PowerPoint (!). I then broke down the parts needed and did some iterations to modularize the design and create common components. I know from experience that the unit cost of parts drops pretty quickly if you order more than one so common components would significantly reduce my costs.

In the case of the outriggers I need 14 total all with the same basic "outer mold line". 4 are used to bookend the wheel wells. These need to be a "solid" design because they close off the belly of the trailer from the exterior. The other 10 can be solid or they can be perforated to save weight or serve as a pass-through for things like electrical conduit, plumbing or HVAC return air.

I made measurements of the original aluminum cross-members so that I could duplicate their profile. Using these measurements I created a sketch in FreeCAD.

FreeCAD 2D Sketch of Outrigger

From the sketch it's fairly easy to create a 3D solid model in FreeCAD. And when you have a 3D solid model you can use FreeCAD's Finite Element Analysis workbench to analyze the design. It's easy (but a little tedious) to change materials or material thickness to determine how they might respond to different "loads", ie the amount of force being applied.

I focused on using mild steel (vs something like aluminum) because it's relatively cheap, easy to weld and performs well in the real world. I iterated some material thicknesses with very demanding loads (up to 2.5x max design load applied to the outside end of the outrigger) to arrive at a conservative material selection.

FEM of Solid Web Outrigger for Wheelwells

Based on the design of the solid outrigger I then moved on to create what I'm calling a "perforated" outrigger, ie one with holes in it to remove excess weight. Weight is a key concern since I'm transitioning from a super lightweight frame design to a more study, but heavier design in steel.

I ended up iterating the basic design 3 times. First was just an oblong hole, next was a series of 3 holes, then the final design is simply a large cutaway. The outrigger is actually designed in 3 parts: a top flange; the center "web"; and the bottom flange. I ran FEM on just the web during the iterative process as I was really primarily interested in the stress and deformation at the outer end of the outrigger under load.

Initial Iteration of Perforated Outrigger

Iteration 2 of Perforated Outrigger

Iteration 3 of Perforated Outrigger - ~2.5x Max Load at Outside

After settling on a web perforation design I then moved on to model the entire outrigger with both top and bottom flanges.

Based on the analysis I semi-finalized the design and have ordered a set of laser cut parts to fabricate 4 perforated outriggers. Why not all 10? Well I'm using a "tab and slot" design approach that I've never used before so I thought it would make sense to do a "preproduction" run before I spent an even bigger hunk of money.

I'll follow up with a post on the tab and slot design approach as applied to the outriggers.

Laser Cut Templates Uploaded in SendCutSend.com

Frame Layout

Conceptual Frame Layout

I'm planning to replace the original "Pipe Frame" with a more traditional "Ladder Frame". This is a major commitment and undertaking and probably the most serious challenge I'll face with this renovation project.

There are a few things driving me to entirely replace the frame:

• fresh and grey water storage
• space available to route the plumbing
• the original pipe frame is not structurally robust and prone to failure

First the plumbing. There really is no good location for suitably-sized fresh and grey water tanks. To me 20 gallons would be the bare minimum to support multiday off-grid trips. The primary driver is showering: we need to be able to shower at least every other day if not every day. Without this it's simply aluminum tent camping. That's fine for some people but I'll have trouble getting Becky to agree to something like a 2-week trip to Yellowstone/Tetons/Glacier if she can't have ready access to a shower.

Lots of people have renovated their Clippers with large water tanks placed under benches, etc. I've tried diligently but I can't figure out how to do this and maintain a safe weight distribution while also keeping the trailer tongue weight under 500lb (a typical max for midsized SUVs). As the water drains from the freshwater tank to the grey tank the weight distribution shifts and tongue weight shifts.

Placing the tanks close to the axle in the center of the trailer  with a low center of gravity is the optimal location. The pipe frame has some space for this but it would limit fresh and grey water capacity to about 15 gallons. The pipe running down the center of the frame really limits the options for underfloor tanks of any size.

Another consideration is routing of the plumbing. The central pipe has limited clearance above and below it so plumbing can easily run from one side to the other. This is particularly true of the 1.5" diameter grey waste dump valve and the large water fill line. Both need to be accessible on the "street" side but one of them would need to run to the "curb" side past the center pipe frame.

It makes my head hurt. I really wanted to make this as simple as possible.

Another consideration is the soundness of the original design. Pipe frames were phased out of Silver Streaks and Airstream in the early 1950s for good reason. While they are exceptionally light, they are rotationally flexy when compared to ladder frames. Additionally the single central tube limits the weight carrying capability when compared to a dual frame rail design.

As a side note I found it interesting to see that the advertising brochure of the day featured a "4 inch 11 Gage Steel Seamless Shelby Tube Design". A little online research revealed that "Seamless Shelby Tube" is a reference to a particular manufacturer and the type of tube was used in the aircraft industry (as well as bicycles). Yet another tie-in to the post-WWII aircraft ancestry of many of these trailer designs and fabrication techniques.

Cutaway From 1950s Sliver Streak Clipper Brochure

I plan to keep the weight down as much as possible but a large water tank immediately adds 200+ pounds. Add a few batteries for boondocking and it starts to add up.

Central Backbone 4" (0.0125" wall thickness) Round Steel Tube, aka "Pipe"

Coupled with this is the approach taken on the Clippers and pipe-frame Airstreams: they used dissimilar metals in the frame. The crossmembers are made from 0.062" aluminum that is bent into a C-channel shape. Two C-channels are sistered back to back to form a single crossmember. These cross members are then riveted to steel tabs that were welded to the central 4" steel pipe (which on my Clipper is 0.125" wall thickness - quite thin). The aluminum to steel connection is prone to failure through galvanic corrosion and simple mechanical failure.

Aluminum Frame Crossmember

Aluminum Cross Member Riveted to Steel Tabs Welded to Central Pipe

The other interesting design element is the axle attachment. There are two steel C-channels running parallel to the central pipe and the axle attaches to these as it would any other ladder-type frame. But these C-channels are attached to the aluminum cross-members which provides another failure-prone connection. Decatur had several of these fasteners missing or rusted out when I purchased it and they needed to be replaced before I could tow it home.

Axle Attachment to Frame

Steel Axle Supports Riveted to Aluminum Cross Member (Note self tapping screw repair)

Most Clipper and pipe-frame Airstream renovations I've seen have had to address frame failures. Often the aluminum cross members are replaced with steel ones that are welded to the central pipe. This seems to be a reasonable fix although it can rapidly increase the weight as steel is 3x more dense than aluminum. It also doesn't address the plumbing issues I mentioned above.

In the end most of the more thorough renovations I've seen have gone the route of a frame replacement. Most of these replacements utilize a ladder frame design with external cues that nod to the original pipe frame, for example a pipe extending from the back even though it isn't the central spine of the frame.

I plan to take this same approach.

At a top level the frame will feature rectangular steel tubing as the main ladder elements. My analysis indicates that 0.120" 4"x2" rectangular tubes will be more than sufficient.

A simple comparative analysis of the frame rails can be done by comparing estimated beading of rectangular tubing to the round tube used originally. For a given size of tube, say 4" diameter like the original, a rectangular tube 4" tall with the same wall thickness will bend less than a round tube of 4" under the same load. Analytically this is because beam deflection is proportional to 1/(moment of inertia) and the moment of inertia of a rectangular tube is higher than a round tube hence the beam deflection is lower for the rectangular tube. Given that I'll be using 2 beams, each of which resist bending better than the single round tube, I'm confident this approach is sound. I'll also note that I did a simple finite element model of a single rectangular beam loaded at 4x the rated axle weight (3500lb) and it indicated negligible beam deflection.

Simplified FEM of Single 4"x2" 11gauge Rectangular steel Frame Rail

The cross members will primarily be fabricated from 12 gauge mild steel. I say fabricated because I intend to use a "tab and slot" design approach to build up the crossmembers. This approach gives me complete control of the crossmember webbing design meaning I can remove material to cut weight but also run computer analysis to make sure I'm not impacting necessary strength. I'll do a post with more detail on this as I've just completed the outrigger design and ordered the initial set of laser-cut parts that I'll be welding up.

The other feature of the crossmembers will be different frame depth sizing. I've found some 30 gallon water/grey tanks that should work but they are 7" tall. So Decatur is going to have a 2 inch "baby bump" around the axle area. The belly pan is going to drop down a little to accommodate the tanks. Of course the frame crossmembers in that area will also need to be deeper to accommodate the bump out. I'm planning on 8 inch deep crossmember in that area vs the normal 6 inch belly pan depth.

A few crossmembers will be made from angle iron (probably 1.5"). The ones in the very front and rear where they really only support the subfloor are strong candidates.

The tongue will use a 32 degree couple vs the traditional 50 degree coupler. This fits the design better and will lessen the apparent deviation from the pipe-frame underpinnings. It will also provide an additional foot of tongue that will allow me to mount the spare tire between the propane tank and the trailer body front end.

Finally I do plan to maintain the protruding pipe from the rear of the trailer. I have in mind that I'll weld a 2" hitch receiver inside the pipe to allow for bicycle carriage.

As I mentioned I've ordered the first set of outrigger parts so I'm off to the races.

Water Heater Tradeoffs

A primary driver for purchasing the Decatur and renovating it was the ability to have a toilet and shower inside the trailer. Our Boles Aero has a slide-out porta-potti for emergencies but no provision for a shower. We have a portable on-demand shower water heater and a shower tent, but that arrangement is not optimal. For one it means you'll be draining the shower "grey water" and most campground technically don't allow it (although I see this arrangement used frequently in National Forest campground).

If you're going to have a shower then you really need a water heater. Cold water shower might be an occasional delight but not something I want to endure every time I feel the need to shower. So in this post I'll focus on the water-heater trade-off's I've considered. I'll address the toilet trade-offs in another post.

The first trade is a water heater with a tank vs a tankless, or "on-demand" water heater. I believe we'll be using the Clipper more in locations such as National Parks where hook-ups are unavailable or limited. So fresh water will be limited to what we're carrying and we'll also need to capture our grey water. After reading a lot of forum posts from people who've used both I've decided that a water heater with a tank will probably be best for Decatur. We'll need to watch our water usage carefully so the waste associated with an on-demand system isn't practical, plus limited hot water will encourage shorter showers conserving water as a result.

Given we'll likely be off-grid much of the time it's obvious that we'll need to rely on propane as the primary fuel source. The real question is whether we also want electric heating to supplement propane when we do have it available. As you'll see all the of the options include the ability to have supplemental electrical heating as an option, so that decision can be deferred.

Tank materials vary between water heaters. Some use a porcelain-lined tank and others use an aluminum tank. The porcelain-lined tanks require an anode that needs to be replaced periodically and they also weigh more.

Water heaters are primarily delineated based on tank capacity running from about 3 gallons up to 16 gallons. Based on forum posts and some online articles it seems that an efficient shower requires on the order of 5 gallons of water. If you take drastic measures you can use a lot less, but I'll assume 5 gallons. If about 1/2 of that water is hot and we need to provide 2 showers in sequence then a tank capacity in the 5 gallon range is necessary.

In the past there were water heaters that relied on a pilot light. You'd need to go outside the trailer and light it when you got to camp. Now they all appear to have electronic ignition (aka "direct spark ignition"): there is a switch mounted inside the trailer and it (along with the water heater thermostat) initiates the ignition of the burner. 

A few of the models from Dometic include a thermostatic mixing value that combines cold and hot water based on a user-selected temperature setting. If you crank up the temp on the water heater then this can effectively provide more hot water, albeit at a slightly lower temp. The literature also says there is the added benefit that hotter tank water kills pathogens better. In reality though you can get the same advantages by installing a shower faucet with an integrated thermostatic mixing valve. There are a large number of inexpensive shower faucets with mixing valves so this is a much more economical approach.

How fast the water heater can "recover" from being emptied is known as the "recovery rate". This depends on the tank type, water heater burner size and also the design of the heater. Faster is generally considered better but it may not be a real issue if you don't have a lot of people lined up to take showers.

The physical dimensions of the water heaters varies a bit however within each size/capacity category they are very close in dimensions.

Here are the water heaters I'm considering:

Dometic WH-6GA/WH-6GEA (gas only/gas-electric)

• 6 gallon capacity
• electric option WH-6GEA
• Aluminum tank (no anode)
• 17 lbs empty weight
• TBD/TBD recovery rate (gallons/hour - gas/electric)
• 12.88" H x 12.88" W x 19.5" D
• Thermostatic mixing valve available ("EXT" model and also upgrade kit)

Dometic WH-6GEA

Suburban SW4D/SW4DE

• 4 gallon capacity
• electric option SW6DE
• Porcelain-lined steel tank (anode required)
• 29 lbs empty weight
• 7.6/6.1 recovery rate (gallons/hour - gas/electric)
• 12.75" H x 12.75" W x 16.125" D

Suburban SW6D/SW6DE

• 6 gallon capacity
• electric option SW6DE
• Porcelain-lined steel tank (anode required)
• 32.9 lbs empty weight
• 10.1/6.1 recovery rate (gallons/hour - gas/electric)
• 12.75" H x 12.75" W x 19.19" D

There are a couple other interesting options that I've read about: the Propex Malaga 5/5E; and the Truma Combi.

The Propex Malaga is used primarily in European caravans. Propex manufactures very popular small propane furnaces that are widely installed in camper van conversions. I have one of the HS2000 furnaces installed in my Boles Aero. The Malaga is a small capacity water heater (13 liters/3.4 gallons) that is correspondingly small in size. An attractive feature is that it doesn't require the large sidewall cutout that the Dometic and Suburban models require. Rather it's installed from inside the RV with a relatively small exhaust vent through the sidewall. However it doesn't seem to have any sales or support base in the US but it appears it can be ordered.

Propex Malaga 5

The Truma Combi is a combination furnace/water heater. They are quite popular in European caravans and are also being installed by US RV manufacturers as well. The water heater capacity is small (10 liters, 2.6 gallons) and the physical dimensions are larger than other water heaters but that's to be expected since a furnace is integrated. I imagine it also is more efficient.

Truma Combi

The major downsize of the Truma Combi is that in the US it is not available for DIYers. It's also quite expensive, much more than a furnace and water heater purchased separately. There is a work-around for these downsides: there are Chinese knock-offs available from several sources including AliExpress. I priced several in the $1300 range which is similar to the cost of a separate water heater-furnace pair. The problem would be lack of support.

As of the writing of this post I'm leaning towards the Dometic WH-6GEA. It's a bit lighter but more importantly it requires less maintenance (no anode replacement) than the Suburban units and the prices I've found are comparable.

Furnace Tradeoffs

 I'm sorting through the options for the furnace in Decatur. I think the primary requirements are:

• roughly 20k BTU heating capacity (I've read 1k BTU per foot is recommended)
• propane fueled
• needs to fit into ~13 inch wide space, height is less critical
• minimized external appearance (no external door - means it must be installed from inside)

I used a Propex HS2000 in my Boles Aero but the Boles has a very small interior. The Propex is only 6 kBTU but it's a real easy install and has no obvious external appearance (it exhausts through the floor vs the wall). Truthfully I've never had to use the Propex unit while camping so maybe my 20 kBTU requirement is overkill but I think it'll provide peace of mind for any shoulder-season camping we might do.

Here's a list of options with pros/cons that I've come up with:

Suburban NT20-SEQ, 19 kBTU

Pros

• Small size: 9.5" H x 9.5" W x 22.5" D
• Relatively light weight (22 lbs)
• Easy install
• Probably the cheapest of these alternatives

Cons

• Single duct and/or direct vent
• Larger external vent (2 holes for Suburban models vs 1 for Dometic)
• Internally visible (intake) panel required
• Can't easily upsize later if we find that 20 kBTU is insufficient

Suburban NT-20SEQ, Direct Vent

Dometic DFSD20121, 18 kBTU, small size cabinet

• Pros
• Small size: 7" H x 12" W x 20" D
• Can be installed vertically as well as horizontally
• Relatively light weight (20.8 lbs)
• Small external vent (single vent hole vs 2 for Suburban models)
• 2 (min) + ducts (for front and back of Decatur)
• Vent location selection easily accommodates horiz or vert installation
• No internally visible (intake) panel required
• Cons
• Can't easily upsize later if we find that 18 kBTU is insufficient
• Pricy and seems to have limited availability (note: 20111 models readily available and cheap but require large external door).

Dometic DFSD20212, Small Case, Small Vent

Suburban SF-20FQ, 19 kBTU medium sized cabinet

Pros

• 2 (min) + ducts (for front and back of Decatur)
• Vent location selection location not as good as Dometic
• No internally visible (intake) panel required
• Medium size furnace cabinet size could be swapped for unit with up to 35 kBTU if needed later

Cons

• Larger cabinet (7.5" H x 17" W x 20" D) size requires vertical installation
• Added cost of vertical installation kit
• Larger external vent (2 holes for Suburban models vs 1 for Dometic)
• 29 lbs
• Less expensive than medium sized Dometic unit

Suburban SF-FQ Series Furnace, Medium Cabinet

Dometic DFMD 20121, 18 kBTU medium size cabinet

Pros

• Small external vent (single vent hole vs 2 for Suburban models)
• 2 (min) + ducts (for front and back of Decatur)
• Vent location selection easily accommodates horiz or vert installation
• No internally visible (intake) panel required
• Medium size furnace cabinet size could be swapped for unit with up to 35 kBTU if needed later

Cons

• Larger cabinet (7" H x 16.5" W x 20" D) size requires vertical installation
• 35 lbs
• Pricy

Dometic DFMD20121, Medium Cabinet, Small Vent

Each of these options has some installation requirements that need to be considered before a final selection can be made. For example the cabinet dimensions need to accommodate side and top clearances depending on the orientation (vertical or horizontal) of the furnace.

Right now I plan to install the furnace in a small closet between the kitchen area and the bathroom. I think it will be about 13" wide. This won't provide much clearance area on the sides so a vertical orientation may be required for even the small Dometic 20121. Given these constraints either the Suburban NT-20SEQ or Dometic 20121 are leading contenders.

The NT-20SEQ is attractive because it's small and cheap. It also has the least cabinet interior mounting restrictions (5/8" clearance from the top and sides). It also has a duct/vent on the left side that could be plumbed into the bathroom to assist in drying. The downside is the front direct discharge which isn't attractive and will force a fixed cabinet door (but would simplify ducting installation).

The Dometic 20121 is attractive because of the ducting options and the small exterior vent. It also doesn't require a front vent. But it will require a lot more cabinet real estate and may not be possible given the installation requirements: 3" required clearance on each side plus the 7" width of the furnace may be more than is available. The Dometic may also be a lot more expensive and possibly unavailable.

Before I can lock down a decision I'll need to get a bit further down the road and nail down the floorplan and cabinet dimensions. Given what I know it appears the NT-20SEQ is the leading candidate.