Filling in a crawlspace?

Title sounds a little wacky. My house is divided by a large steel I-beam, and the crawlspace below (dirt floor) has one side about 4' deep while the other side is barely even a foot from the bottom of the joists. I've played with the idea of building some kind of mass heater for a while, and considered what it might take to actually support the stove above the crawlspace. This year I actually built a short, fat cob wall around my wood stove directly above the I-beam, but for a larger object, it would need to be continuously supported. I considered opening a hole in the floor, and building four walls down into the crawlspace and then infilling with rubble, and placing the stove on top, but then I thought, "why not just fill the whole thing?" Well, not the whole thing, but the half of the crawlspace that's already relatively shallow. There's no plumbing or electrical work there. I was just curious if anyone's ever done anything like this in an effort to gain more thermal mass and potentially install in-floor heating. I've also considered installing some kind of Japanese style hearth or Korean Ondol sytem.

My plan was basically to build a a nice thick stacked stone wall (broken concrete) directly under the I-beam, then tear up the wood floor in my living room, cut the joists away, host a wheel barrow party to fill in the empty space, and build up the layers as if you were doing a normal earth foundation and/or earth floor. The main issue I see is the pockets at the sill where the floor joists are currently resting. The stud walls are resting on the floor joists and rim joist, so removing the floor joists would require, at the least, fitting in another rim joist to replace them. I'd also be nervous about the earth floor resting directly against the rim joist, number one due to potential moisture, though that issue could be mitigated with some kind of membrane/vapor barrier, but number two do to the potential lateral thrust of the earth. Theoretically it would only be about 10" of earth/gravel/etc, since that's the height of the rim joist, and it wouldn't be shifting over time, but the house isn't old enough to have used a massive sill plate, but it's not young enough to have significant mechanical connections between the rim joist and foundation.

Thoughts? Or just a lot of work for not much return?

The only alternative I really see is I go buy a bunch of foam board and properly insulate the crawlspace (I've already sealed it as well as possible), and then try to pipe some warm air down in there from the wood stove so I can stop walking on cold floors all winter...

Wood touching dirt is really good at attracting termites, ants, and rots.

Without some type of floor structure you may have your walls start kicking out at the bottom over time, depending on how your structure is compared to where you are cutting out floor joists.

The foundation walls would be held in place by earthen loads on both sides. I believe the only wall that might start "kicking out" would be the stacked stone wall holding back the earth from the rest of the crawlspace. But I would build it at least a foot or so thick, and it would be less than 2 feet tall so the force exerted on it would not be significant.

There is currently no lateral thrust acting on the bottoms of the walls, and unless I built the earthen floor up above the stem wall there would continue to be none. the floor joists do not tie anything together, I'm not sure they're even nailed to the rim joist. There is still a concrete stem wall that rises above the below grade foundation. One of my specific questions is regarding the joists pockets, whether it's better to fill them in with some kind of stone, or just leave the floor joists ends in there and then extend the interior air barrier down to the top of the stem wall. I would likely install a vapor barrier and I would definitely install some form of insulation under the new slab floor.

There would also be no dirt in contact with wood in this application.

What I am describing is essentially taking this: http://epsindustry.org/sites/default/files/drawing3.jpg and converting it into this: http://www.finehomebuilding.com/assets/uploads/posts/15173/slab-frost-protected.jpg

Sounds like retro pouring a floating slab.

You will need to install a vapor barrier then foam insulation then another vapor barrier and then you could do the pour.
All exposed wood will need to be protected,you might experience condensation and the pour will give off water vapor as it cures.

As long as you address these issues prior to doing the pour it should work for what you are wanting to do.

So, any roof wieght or roof loads are directed straight down through an engineered system that contains the load in a plumb force, if there were parts missing the roof load could under some circumstances begin to spread your structure. This is why in old block shops from the 50s you will often see an all thread bolt every ten ft going from one wall to the other perpendicular to the ridge of the roof.

Just a reminder to watch for it that's all, once it moves it's difficult to move back into place.....

Kieran, are you talking about removing the I beam or cutting some floor joists and framing in double headers to support them so you can build a masonry heater?

You do have some structural items that need to be addressed (like Chadwick mentions) if you are going to remove a main structural component you have to be really careful.
This does not sound to me like part of your plan, but if it is, then you need to think like an engineer and consider the loads before starting any modifications.

If what you want to do is simply fill a portion of the crawl space with a floating slab, then double header the floor joist so you have a hole for a masonry heater to come up through.
There is less concern of wall spread but you do have to take care of the structural issues with the floor joist so the floor doesn't start to bow or flat out collapse.
You may need to add some floor joist support at the four corners of the hole you plan on cutting.

Note: just pouring a slab under a floor system does not add "thermal mass". Only if there is a way for heat to enter does it act as thermal mass, such as a thick floor for a green house or sun room.

If this pour is only going to support the weight of a masonry heater, it is not part of the thermal mass of the heater, that comes from the actual build of the heater not what is under it.

My understanding of your idea is that you want to pour a support structure that will allow the weight of a masonry type heat system.
If this isn't correct let me know and I can give you construction details of what you need to do for it to work without house failure.

Chadwick Holmes wrote:So, any roof wieght or roof loads are directed straight down through an engineered system that contains the load in a plumb force, if there were parts missing the roof load could under some circumstances begin to spread your structure. This is why in old block shops from the 50s you will often see an all thread bolt every ten ft going from one wall to the other perpendicular to the ridge of the roof.

Just a reminder to watch for it that's all, once it moves it's difficult to move back into place.....

I did not read the entire thread just looked at the pics but, this is not the failure path of most buildings. IRC code or a PE has the braced shear wall requirements that prevents buckling of shear walls under load. The biggest concern is the lower chord or collar tie spanned too far w/o a tributary break in deflection( s/b < 1") or ultimate failure of the roof. The biggest threat to buildings is not a static "plumb force" it is a dynamic bending moment.

I'd suggest hiring a pro with proven designs_builds verified before taking any more advice.

Kieran, I see you have been struggling with this while now feel free to email.

That is because most buildings' basic structure prevents it, I was conserned that cutting out structure can cause this not a failure of current.

Would you agree that removing the floor joists can have this potential, and would you not warn against it if a client was going to cut out multiple joists and leave the structure open to forces?

If you notice in the pics, the failure is due to removed structure and the forces on an incomplete structure.

Agree, do not want to cut out floor joist especially below a load bearing wall that has a footing unless reinforced properly. If you do the roof span has been broken and as I said above the first failure will be the roof not the walls. If you removed any adjacent wall bracing(sheathing, strapping etc) then the wall would buckle under shear as shown in your pics. If you remove a non-load bearing floor joist the decking will buckle under live load @ 200 PSF that s/b easy to deal with. Code describes the on center spacing, size, and clear span of floor joist, again to maintain deflection and prevent failure.

It is the roof failure in those picture that cause the walls to fail, if you look it is from the flattening of the roof pitch that the walls react. In fact I have arrows showing the force acting on the roof and into the walls only after the roof has reacted to force......

As the pitch of the roof flattens the top plate of the wall will be pushed out in the case of rafter removal, joist removal takes years but eventually the rim joist will fail and the walls will drift off the footing either into the house or out of the house, but they will eventually react to forces. However the roof is still structured so it will remain on pitch. This is simple structure physics, that I can prove in a structural model or a scale model.

So, any roof wieght or roof loads are directed straight down through an engineered system that contains the load in a plumb force, if there were parts missing the roof load could under some circumstances begin to spread your structure.

This statement is not entirely true, not all loads are down the worse are in bending from seismic, wind, shear, laterally. If braced wall requirements are correct or the most of the walls are sheathed with structurally rated sheathing at least 5/8 thick and attached to the proper stud spacing, and the max length of the wall bracing is correct or is broken by an internal wall, the lower chord or collar tie can fail in deflection and the walls should not buckle nor come off their footings. Most walls fail in shear not buckling from vertical loads. Your drawing is not showing dynamic lateral shear, it is only showing static vertical forces which is part of the equation but not the entire equation. If the rafters are connected to ceiling joist and seated on the walls, which is not always the case and code address, then removing a load bearing wall can occur if the ceiling joist and/or collar tie is capable of clear spanning. If it is not we can make them more capable by doublers or add intermittents, purlins, etc, that reduce deflection.

We do not design structures in a way that a roof failure would cause a total collapse of walls inward or outward, people to get hurt by collapsed walls, even in a seismic or 3 sec wind gust event they would partially fail in shear. Hurricanes or tornado's are a different story due to flying debri impacts or internal pressures. In general, PE/code does not allow a single catastrophic event such as partial failure of a roof to potential kill people. That is IF the structure was designed properly, that can be verified by a pro or by looking at code and seeing if the clear spans are accurate. We use a safety factor of two so there is some wiggle room if one knows where to find it.

Pudding as it were

Kieran Chapman wrote:Title sounds a little wacky. My house is divided by a large steel I-beam, and the crawlspace below (dirt floor) has one side about 4' deep while the other side is barely even a foot from the bottom of the joists. I've played with the idea of building some kind of mass heater for a while, and considered what it might take to actually support the stove above the crawlspace. This year I actually built a short, fat cob wall around my wood stove directly above the I-beam, but for a larger object, it would need to be continuously supported. I considered opening a hole in the floor, and building four walls down into the crawlspace and then infilling with rubble, and placing the stove on top, but then I thought, "why not just fill the whole thing?" Well, not the whole thing, but the half of the crawlspace that's already relatively shallow. There's no plumbing or electrical work there. I was just curious if anyone's ever done anything like this in an effort to gain more thermal mass and potentially install in-floor heating. I've also considered installing some kind of Japanese style hearth or Korean Ondol sytem.

My plan was basically to build a a nice thick stacked stone wall (broken concrete) directly under the I-beam, then tear up the wood floor in my living room, cut the joists away, host a wheel barrow party to fill in the empty space, and build up the layers as if you were doing a normal earth foundation and/or earth floor. The main issue I see is the pockets at the sill where the floor joists are currently resting. The stud walls are resting on the floor joists and rim joist, so removing the floor joists would require, at the least, fitting in another rim joist to replace them. I'd also be nervous about the earth floor resting directly against the rim joist, number one due to potential moisture, though that issue could be mitigated with some kind of membrane/vapor barrier, but number two do to the potential lateral thrust of the earth. Theoretically it would only be about 10" of earth/gravel/etc, since that's the height of the rim joist, and it wouldn't be shifting over time, but the house isn't old enough to have used a massive sill plate, but it's not young enough to have significant mechanical connections between the rim joist and foundation.

Thoughts? Or just a lot of work for not much return?

The only alternative I really see is I go buy a bunch of foam board and properly insulate the crawlspace (I've already sealed it as well as possible), and then try to pipe some warm air down in there from the wood stove so I can stop walking on cold floors all winter...

I'd need a diagram to clarify what I think you are wanting to do. If I understand without seeing it sounds like the steel I beam is acting as a footing to break the roof and floor joist spans, perhaps a wall is above the I beam I would gather. Your outer load bearing walls should have footings/CMUs/etc, you do not want to mess with nor the I-Beam I doubt needs anything below it. Add or remove any other walls is ok.

Your correct in the your biggest issue is the getting load from the walls to the foundation if you remove the floor joist. If the longest rim joist is seated on the foundation at it's ends and is capable of clear spanning (you find in IRC floor joist code you should be fine, doubling up the rim and cut floor joist would be good idea). If it is not seated make the connection since that will help prevent bending along with some lag screws you put into the doubled joist.

Unless there is a solar gain benefit through a window into this floor mass I'm not sure it is worth it. I'd use a mineral wool drain board or Roxul Comfortboar IS to insulate the rim and floor, it is rated for ground contact. Do not use foam/poly it is a microbial haven. Roxul will call out poly I do not use since I design breathable assemblies. Do not worry about earth pressures against the rim. Follow protocol for earthen floors, I don't do them.

The floor should make a good heat sink for the heater, better with some solar gain. There are easier ways to add mass where there is solar gains to floor structure and/or walls. I have a thread "Natural vs Mainstream Cost" that shows easy ways to add mass to drywall.

IRC (International Residential Code) : https://law.resource.org/pub/us/code/safety.html

IBC is commercial with much different requirements do not use.


Any questions shoot me an email or call.

While doing all the work entertain the thought of adding Hydronic Radiation, a small two element by Argo is 100% efficient or an air source water heater. You may only need 15-20,000 btu small wall mounted unit. Be carefull with zoning and solar gains to not over heat.

Chadwick, I admire your quest to understand structures better but you are getting off topic. We are not talking about removing ceiling joist or collar tie and the rafter/joist to wall joint does not pivot around a single pin as you have demoed. In most cases there is a birdsmouth seat that is toed nailed, or there can be ceiling hangers....Explaining the reactions to include decking/sheathing at this joint would take a lengthy post. You notice the guy in the video which is obviously a designer not engineer keeps pointing to a PE since he is not revealing alot of what is going on there. Seismic loads are largest at this upper joint, wind at the foundation. In a CMU to rafter tie this appears to be yes if you remove the collar tie or ceiling joist to vault you can see some local failing but not the entire walls, and yes the ridge would perhaps bend but not entirely from center of roof load.

Check out the model of a vaulted truss design I am working on now. This is a 50' clear span design, the lower chord and webbing takes most of the loads which are 25 PSF live load, 10 dead at the top chord, 10 dead at the bottom chord. The walls are kept from shear buckling with 5/8 sheathing. The lower sills sit on a slab in this case different than the OPs and are anchored with 4' OC 1/2" anchor bolts with large square washers. The steel studs are anchored to #5 rebar j ties in the footing-to-slab that takes vertical uplift loads and out of plane bending moments. I can guarantee none of my joints behave like those videos, and that is the same case with a rafter/purlin/collar design.

I suggest creating a new thread since all that is getting off topic.

What we are discussing (if I understood the OP correctly) is cutting out a non-load bearing floor area and filling with earth & aggregate. As I said this can be done with rim and/or joist doublers by providing a load path to the adjacent foundation seats and using rim joist that are approved for the span. Other ways to improve the area under the sill plates is to add more stub joist intermittently. At this joint the sill bolts accumulated act to resisted in bearing area of steel to keep the sill plates from kicking out or in or uplifting IF the washer is of adequate size, again this is not a pivot joint.

See the PIC below recommendation depends on the design I would need to see. I still say the cost of performance is not there unless there is a window with a high solar gain coefficient and perhaps some HR, find another way.

Kieran, seems like my wording can be confusing so here is a floor stick diagram and the solution if you care to take this path of cutting out floor joist to install an earth floor. The doubled cut back joist will exceed IRC 2015 allowable spans and use the spans for the doubled rims in code they will too and will also satisfy an inspector if a permit is required. Do check if they have any amendments or other codes. Permies, in this case we did not a PE we took a code path to prove the design and show it good.

Thats it for me folks, we are getting way too far off topic.

My home was built in the late seventies with progressive ideas, some worked out better than others. One that didn't work out was the slate floor with a crawl space below. The furnace was supposed to blow hot air into the space and warm the slate. It turns out that this connection took too much of the furnace's heating power away from the other vents and the slate never got very warm. The slate is lovely in the summer but quite cold in the winter.

I have thought of filling the space with a mass to be heated in a more efficient way. I would not want to mess with removing any structural elements though. I love the warmth of those radiant heat floors with the water pipes underneath but when a friend's relatively new system was ruined by pipes bursting I was turned off.

Is there a way to fill in your crawlspace without taking out structural elements at all?

I'm interested to hear if anyone has any experience with this kind of a retrofit project.

Hi all,

To anyone worried: no, I am not proposing cutting any portion of the I-beam. The I-beam runs down the center of the house, with floor joists running perpendicular to it.

I originally got the idea from this comment by Joe Woodall: https://permies.com/t/17088/earthen-floor/earthen-floor-wooden-floor#170037 Basically I want to remove a portion of the floor and the underlying joists, build some kind of containment wall in the crawlspace, up to the same height as the foundation wall, and then fill in that area with rubble.

The background for this is that I do not enjoy having a crawlspace for the most part, it's damp and dank and I've replaced multiple areas of the subfloor and even several joists due to the fiberglass batts that were installed for many years facilitating rot in between the joists. It was also very cold until recently when I insulated the rim joist. I am aware that I could "encapsulate" the crawlspace. Recently, I was thinking of building a mass heater of some kind, and rather than bulk up the wood framing I thought it might be a good idea to simply box a big hole in the floor, and form a "floating slab" as Bryant RedHawk called it just in the area where the masonry heater is. I was curious though, if instead of forming a slab just for the footprint of the heater, if I could build a raised earth floor for this entire room/section of the house.

The reasons for doing this would be that I could then create a large thermal bank in the floor. It is the northern half of the house and I am planning skylights on the southern roof slope which would shine down onto the slab. I would also like to experiment with running the mass heater's flue through the floor and letting the floor act as the mass. I have also considered simply installing a boiler that would run heated water through a hydronic floor system.

Chadwick, I had not considered the perspective that examines the continual small shifts over the lifetime of a building. You bring up an interesting point. I considered the floor joists to only be acting as support for the floor, but I can see how they could assist in holding the bottom of the building together. In my mind, because there are several walls perpendicular (two of which form the front and back ends of the house) to the exterior wall in my case, I would assume that these would be sufficient to hold the other wall in place. If not, another possibility would be to leave in 2 or 3 floor joists and double them and then pour the sand and gravel in around them, they could be wrapped or otherwise protected from direct soil contact. I'm not a big fan of that idea though.

Terry, it sounds like you're saying this plan is feasible and that you would recommend doubling the floor joists along the foundation wall so that the exterior stud wall has a sufficient footing.

The joists span are 2x10s that span from the top of the concrete stem wall to the I-beam in the middle of the house, about 12', so my thought was that I would cut back the floor joists back to the edge of the I-beam and the edge of the stem wall. Then I would double up the remaining joist stubs as per your recommendation Terry. Then I would build a containment wall for the earthen floor (in this case, I would build the wall directly underneath the I-beam, but it would not be supporting the beam. Then I would lay a vapor barrier down against the 3 foundation walls and the stone containment wall, and backfill behind this with stone, sand, etc. up to the top of the stem wall. At this point, the floor would be about 10" lower than the original floor because I would be at the bottom of where the floor joists used to be. The joist pockets would be exposed, but I would finish and insulate them as if they were just a part of the wall matrix. They will also be receiving a heavy dose of insulation from the outside as I am planning on super-insulating this north wall to the exterior when I extend the roof overhang to form a shed/porch. Does that all make sense?

Matu, I will certainly update if I ever decide to do this. (would be next year at the earliest) My house is a huge experiment for me right now: understanding natural building techniques and practices in the context of a conventionally constructed home. It is a place that has been brought back from the brink of ruin in an area of Detroit filled with sunken and abandoned houses, but also resilient and reliant people and buildings. Always much to explore and learn.

Terry, one thing I didn't quite follow was your sketch where it said "double rim joist" but then also "x4" Did you mean the rim joist would be doubled on all 4 sides of the house? That doesn't seem right.

Kieran, I seem to recall another thread of yours way back when discussing your crawl issues we showed old FG insulation to be an old bad choice builders made and are still making for floor insulation. Today we know better insulation's like mineral wool drainboard underfloor and against a stem will perform much better. That leaves the question of how well the vents are performing or if they s/b sealed, which is a topic of controversy even among building scientist since it is difficult to analyse convective air flows and mold growth. We face the same challenges in our walls/boundary conditions, the way we build them in the USA.

Cutting the floor out may not yield the results you are seeking since many earth materials including soils, aggregates, are just as capable at producing mold and rot, especially if plastic and/or foams/wood/steel are in the layers. It is a myth to think that is not the case and natural materials are incapable of it. Besides that there is no guarantee that the system you are contemplating will perform. It is reliant on some level of radiant heat reaching the floor, and since radiant heat can travel in any direction only a percentage will reach the floor. That and the fact that convective air loops coupled to a mass floor are not the most effective way to heat it, nor does air carry much heat in tubing to mass like water. Passive solar through a sky light will probably loose more btus/hr than the solar gains annually. The only defense is a double or triple pane low iron glass argon or better krypton filled with a low e-coating and high SHGC that allows the short solar waves in and blocks the long on the south side, low teens u-value, that will not be cheap. How effect that is will depend on IF the short waves reach the mass, alot of variables, again difficult to determine. More than likely at lease a couple of skylights, thermal holes, and with the mass north probably not that effective. If a south window typicality 32" off the floor were close to a mass floor the area by the window would solar gain then distribute to other colder mass like our body radiates heat to cold windows or floors. Just a heads up that this path may not be a good one without some mechanicals in this situation.

Other options that should be compared for cost of performance are the Argo electric boiler I mentioned tied to PV. In this case you leave the floor intact, use aluminum strips or a thin slab if you can tolerate a raised floor. See below. If you insulated your house well and air sealed, a small 15-20,000 btu/hr unit along with your wood stove to take the edge off in the cold Chicago winters may all you need. Get some quotes from a HVAC pro and ask them to run a new manual j or energy model to determine the new load. The model will also tell what u--value the sky lights need to be. Another VERY efficient design is an air source heat pump ductless minisplit tied to PV. Either one of those options could be installed for probably to same or less cost without the structural risk. Another, add hygroscopic mass to walls and ceilings if you have not already. These mechanical will couple to it. The cost of heat pumps tied to passive PV in most cases wins compared to deep energy retro-fits like this.

http://www.argoboiler.com/
http://www.radiantheatproducts.com/Thin_Slab_Heat.php

PEX_AL and properly installed connections do not have burst pressure issues if tested to the noted ASTM. Aluminum is highly conductive...
http://www.wattsradiant.com/products/radiantpexal/

A 2x10 can only span 11-8 @ 16 OC depending on species: http://www.wattsradiant.com/products/radiantpexal/

If exceeding that use a LVL/TGI I-beam. If not double it as I noted in my drawing and yes all four sides of the cut--out to resist both vertical and lateral loads(bending). Chad's drawings did a good job at showing some of those forces.

I would, if considering pouring around structure, look toward having a metal replacement made, something in square or rectangular cross section that has large flanges welded to either side that can be used to bolt through the rim joists. You can quadruple your strength at the same time as reducing the number of members crossing through your "filling" material.

Again this really should be engineered to suit your structure, no matter the solution used.


One would be surprised how much impact just slamming the door on a wall with no joists has, and the collective movement of that action over time. My house is poorly framed and I flex the wall when I shut the door, visibly moving the wall, if that wall were not contained top and bottom I would have over the last four years scooted the wall off the foundation in the local area around the door. Exterior doors wiegh from 10-30 lbs that is effectively a hammer that you are impacting the wall with.....

Chad, I agree. I grossly misunderstood the plan if it is to clear span any wall without a floor sill support. That is a no no. If I understood correctly there is just an opening for a small area to set a wood stove in, building up a 4th pony wall to hold some earth and aggregate, the other three walls will be the I beam and stems. No walls running over the cut-out/hole, no doors either.

Due to lack of a plan sketch here I would also agree hire local pro(s) or run the plan by some local building inspectors that inspect daily and see if it meet codes as I have outlined.

There are no walls running over the top of the area I'm describing. It is one side of the house ringed by three exterior walls, and one interior wall which rests on top of the I-beam. I am proposing removing the floor joists in this room and filling in the space with a slab of sorts. I agree with you that skylights may not be the best option and that convective air loops will not be sufficient to heat the mass, it's just an idea that I am currently exploring the feasibility of before figuring out the rest of the pros/cons.

Assuming that I could remove the floor joists in this given area, and "retro pour a floating slab" as Bryant RedHawk described it, I would have to decide where my floor would come up to. It seems easiest to have it come to the top of the stem wall, so it would be 10" lower than it currently is (the width of the joists), that way it would not be in direct contact with any wood. I'm wondering what the best way to treat the joist pockets would be. They would still have the floor joists stubs in them. Could I just install insulation and add blocking in between the joist stubs and then extend my wall (lath and plaster) down to meet the new floor?

Chadwick there are a couple of things I don't understand. 1. In a typical slab foundation, there are no floor joists. The walls do not bow out over time at the bottom, they are set on a sill plate which holds everything straight. WHy don't these bow out over time? Now granted, my house was not designed as a slab, so there are floor joists, but there's still a sill plate/rim joist I don't quite understand how this differs from the walls of a slab foundation house. 2. The floor joists in most houses all run the same way, which means that two walls (often the gable ends but in my case just two sides of my hip roof) are not held in place by floor joists, again they simply rest on a sill plate and are presumably also held in place by the walls set perpendicular to them. Why then is this the case that these walls do not suffer the issues you described? (I'm not talking about ceiling joists or any kind of tie beams either, just floor joists)

I'm wondering what the best way to treat the joist pockets would be. They would still have the floor joists stubs in them. Could I just install insulation and add blocking in between the joist stubs and then extend my wall (lath and plaster) down to meet the new floor?

I recommend Roxul or Thermafiber between joist 7 1/4" @ R-30 + 2" R-3 IS board @ R-8, whole = R-38 and fill all the floor open joist . Take some limecrete or hempcrete or other hygroscopic mass to the ground, no vapor barriers, foam products. Follow Roxuls "Split Insulation Guide" for underslabs for the new floor.

Again, the structure I designed exceeds code meaning it is "Engineered".

Good luck to you

Forgot to mention spray all mass upper surfaces with a natural silicone sand based sealer Siloxane. That will make impermeable to liquid water but 100%+ vapor permeable and breathable for fast drying, and chemically resistant. Use a mix with at least 40% solids and test it out, you should see water beads on the surface. No need for plastics.

Thanks terry, good tips. One aspect I didn't understand about the structure you designed though was doubling the rim joist. Since I would be cutting the floor joists back to the foundation wall, the stubs left behind on top of the foundation wall would be in the way of installing another rim joist. When I say rim joist I mean the "band joist" on top of the sill plate. I want to make sure the wall that rests on top of this is more than adequately supported, but I also don't want to fill the entire cavity on top of the foundation wall with wood or else I can't insulate. The slab I would be installing would be supported by a pony wall only on one side, so the other three sides are at the foundation. Does that make sense?

Thanks

Kieran, sorry for the confusion. I was saying double the new inner rim joist you terminate joist to IF you cannot meet the span tables in the IRC I posted. Other better options are TJI/LVLs you follow the MFG span tables. Two of the rims running the same direction as the wall rim/band, or 9O to the joist, s/b supported at both ends by the foundation for clear spans purposes. The other two can be hung. See my dwg. Other options if you cannot get the spans are piers that break the spans. There are many choices using code, use the one that makes the most sense. When I do not have all the dimensions or knowns (like wood species) I exceed code. You decide what is best.