Homeless Man Built Walls From Old Refrigerator Doors — Each One Already Insulated and Sealed

[00:00:00]You've seen a refrigerator door a thousand times. You've grabbed the handle, let the cold air hit you, closed it without thinking. You've never once looked at that door and thought about what's actually inside it. Virgil Doss had. He'd spent 9 years repairing commercial refrigeration units for a company called Klein and Roth out of Dayton, Ohio. The kind of place nobody's ever heard of that keeps every restaurant and grocery store's walk-in cooler running. 9 years crawling behind reachin displays, pulling apart residential side by sides, hauling dead chest freezers out of restaurant kitchens at 2 in the morning because someone's produce inventory was at risk.

[00:00:36]9 years knowing exactly what a refrigerator door was made of, what was inside it, what it was designed to do.

[00:00:42]The design hadn't changed much in 40 years. steel skin on the outside, plastic liner on the inside, and in between injected polyurethane foam, closed cell, dense, locked to both surfaces. The foam goes in liquid. It expands. It fills every corner. It hardens and bonds and never moves. An average residential refrigerator door comes out of that process with foam insulation rated around R six to R8 per inch, which is among the highest insulation values you can get from any rigid material. Add the gasket, that magnetic rubber seal running around all four edges, and you've got a panel that maintains a sealed thermally resistant barrier every hour of every day for decades. Virgil Doss knew all of this.

[00:01:25]By the winter of 2023, he was sleeping under a highway overpass in Evansville, Indiana. The overpass was on Lynch Road near where it crosses over the CSX freight line that cuts through the south side of the city. Two concrete support walls ran parallel, maybe 14 ft apart with the road deck overhead. In summer, it was tolerable. In winter, the wind came through that gap like something with intent. There was no way to block it with what most people built out there. Tarps pulled flat, cardboard stacked flat against a wall, sleeping bags rated to temperatures that didn't account for wind cutting straight through a 6-in gap in a tarp seam.

[00:02:00]Virgil had tried the standard setups. He knew the physics of why they failed. A flat tarp with no airspace behind it conducts cold directly through the surface. Cardboard works until it gets wet. And on Lynch Road, it always got wet. He was 44, medium build, a little heavy through the chest from years of physical work. Short brown hair that had been growing out unevenly since he'd stopped being able to get it cut regularly. Going gray fast at the sides in a way that made him look older than he was. He had a long scar on his right forearm from a sheet metal edge in 2019, pale and straight that he used to think made him look tougher than he felt. He always wore the same green canvas Carheart jacket, not the insulated version, just the shell with a piece of black zip tie replacing the broken zipper pull on the left chest pocket.

[00:02:47]He'd been homeless since April of 2022.

[00:02:50]The path was the same sequence that gets people into that situation so fast they don't see it coming until it's already done. Klein and Roth lost their three biggest commercial accounts when a national refrigeration chain expanded into the Dayton market and undercut their service contracts. The company folded in November 2021. Virgil had $6,400 in savings. He spent four months job hunting, driving across the tri-state area for interviews, burning through gas. A back injury in January 2022 put him out for 6 weeks. By April, the math had run out and the lease on his apartment on Wayne Avenue was gone.

[00:03:26]He stayed in his truck for a while. Then the truck's transmission went in August.

[00:03:31]He pushed it to the side of a parking lot off Fourth Street, and that was the last he saw of it. The overpass became his spot in October 2022, heading into his second winter outside. The first winter he'd spent moving around, shelters when beds were available, other encampments when they weren't. The shelters weren't always workable. The one on Lincoln Avenue had a 90-day limit, which he'd used. The drop-in on Oakill Road closed at 9. By October 2022, he'd started thinking about staying in one place, building something that held heat because moving around in winter was burning more calories than he could replace. The problem with the overpass, and he'd thought about this systematically, the way he thought about everything, was that the wind had two open ends to come through, east end and west end. If you closed one end, well, the wind just funneled harder through the other. The standard material for closing an end was whatever you could carry. A tarp, plywood scraps, salvage lumber. But plywood is heavy. Tarps move. Lumber means a frame. And frames take time and hardware. The refrigerator doors came from a conversation with a guy named Pete Carver, who ran a scrap hauling operation out of a lot on Green River Road. Virgil had done odd jobs for Pete through that fall, loading, moving, sorting, and Pete paid cash and asked few questions. Pete hauled commercial kitchen equipment from restaurant closures and business liquidations. The stuff he couldn't sell for parts he'd break down and take to the metal recycler. Refrigerators and freezers came through regularly. It was the door that caught Virgil's attention, not the unit itself. He was loading the back of Pete's flatbed one afternoon in November 2022 when a commercial reachin cooler, the tall glass door kind you see in convenience stores, came apart at the hinge. The door dropped and landed flat on the asphalt. Virgil picked up his end of it and felt the weight, maybe 30 lb for a 5-ft panel. He knocked on the face with his knuckle. Solid, dense. He looked at the gasket running around the edges, the magnetic rubber seal, still intact, still flexible despite the unit being maybe 12 years old. He set his end down and stood there for a second. A 5-ft tall, two foot wide panel, two steel skins, foam core injected under pressure, bonded to both surfaces, fully cured, the gasket on all four edges. It was by any reasonable measure a pre-built insulated wall panel. It had been engineered to maintain a 35° temperature differential for the entire life of the appliance. It was designed to seal. It was designed to resist moisture. And Pete was going to send it to the metal recycler for 11 cents a pound. Virgil asked if he could have it.

[00:06:11]Pete said, "Sure, have whatever you want. I'm not getting anything for any of this." He got four more doors over the next 3 weeks. Two came from a restaurant walk-in cooler Pete hauled out of a closed Mexican place on Riverside Drive. The kind of thick flat walk-in panels with no glass, just solid insulated steel faces, about 4 in thick, significantly heavier than the Reachin doors, but with proportionally higher R value. One came from a broken upright residential freezer. One was a standard sideby-side refrigerator door, lighter and narrower, but still usable. He had them staged in the back corner of Pete's lot with Pete's permission. Covered with a tarp, while he figured out the problem of how to connect them. The connection problem was real. Refrigerator doors aren't designed to be structural panels.

[00:06:57]They have hinges on one edge and a handle on the other. If you stand them up and lean them against each other, they'll work fine as long as nothing pushes on them. but they won't stay square and they won't form a tight seam where they meet. The gasket runs around the perimeter of each door to seal against the frame of the refrigerator.

[00:07:13]If you butt two doors edge to edge, gasket to gasket, you actually get a pretty reasonable seal. The magnetic strips in both gaskets will pull toward each other slightly, but you need them to stay square and upright under lateral pressure from wind. He worked this out at Pete's lot over two evenings using material he found in the yard. The structural answer was angle iron, L-shaped steel bar that he could bolt through the door skins at the edges and then run a section across the seam between two doors, holding them in plane with each other. Pete had scrap angle iron in the yard. He had a drill. He had nuts and bolts and a coffee can. He had an adjustable wrench. The first two-door test panel, a walk-in door and the residential freezer door bolted together with angle iron along the seam, stood on its own when he propped it against the support wall and didn't flex when he pushed on it from the outside. The gaskets at the vertical seam weren't perfectly mated, but they were close enough that the gap was maybe a/4 in, which he could deal with. He did some math on the insulation. The walk-in panels were 4 in thick, foam core, which put them at R28 to R32. The residential doors were closer to 2 1/2 in, which was around R16 to R20, depending on foam density. Even the thinner residential doors were performing at a level that matched or exceeded what most people put in an exterior wall in a new house. He also didn't have to build them. They were already built. They were already sealed, already structural, already moisture resistant. The foam was bonded in place and wasn't going anywhere. The gaskets had outlasted the appliances they'd come from. All he was doing was changing what the panels were used to seal. The build took three weekends in December 2022. The plan was to close the east end of the overpass gap, the upwind end, because the prevailing winter wind in Evansville came from the northwest and hit that side first. He wasn't going to try to close both ends. Close one end well, then build a secondary inner wall to create a dead air buffer zone, and the west end could stay partially open for ventilation without killing the warmth. He transported the panels from Pete's lot to Lynch Road using a two-heel furniture dolly he borrowed from Pete. Three trips, two of the walk-in panels were heavy enough that he needed a second person for the final lift. And he recruited a man named Gary Mets, who'd been camping at the overpass since September, and who helped in exchange for getting to sleep on the warmer side of whatever Virgil built.

[00:09:33]Gary was 58, had been outside for going on 3 years, and was skeptical of the whole project in a way that was honest rather than cruel. He said it was a lot of work and he didn't see how refrigerator doors were going to keep them warm better than the tarp setup they already had. He said this while helping carry the panels. He kept helping anyway. The east end went up first. He stood three panels side by side, walk-in door, residential sidebyside door, residential freezer door with angle iron bridging the seams.

[00:10:02]He ran a horizontal piece of angle iron along the bottom edge, bent into an L-shape that sat flat on the concrete and kept the panel assembly from sliding forward. At the top, he drilled through the panel edges and ran a piece of salvaged threaded rod into anchoring holes he chiseled into the concrete with a masonry bit he found in Pete's yard.

[00:10:20]It wasn't elegant. It held. The total width of the three panel east wall was just under 7 ft. The gap was 14 ft across. The remaining seven-foot opening he left as a pass through for now, planning to deal with it with a tarp because the structural solution for the second half needed more hardware than he had yet. The second component was the inner buffer wall set back 4 ft from the east panel assembly toward the center of the overpass. This was where the physics changed from simple insulation to something more useful. A wall with R20 insulation is good. A wall with R20 insulation plus a 4ft dead air space behind it enclosed by a second insulated surface is dramatically better. The air between the two walls can only move by convection and the distance is large enough that convective loops don't develop efficiently. The air essentially sits there still acting as additional insulation on top of what the panels themselves provide. Cold air can't reach the sleeping space without first crossing the outer panel then crossing 4 ft of still air then crossing the inner panel. Each crossing costs it temperature. He built the inner wall from two more refrigerator doors. The remaining walk-in panel and a second residential unit Pete had come across.

[00:11:32]Same angle iron connection method. This wall didn't need to be anchored as heavily because it wasn't taking direct wind load. He ran it between two existing pipes that ran vertically along the concrete support wall and zip tied the panel edges to the pipes at three heights. It wasn't structural in the same way. It just needed to stand still.

[00:11:51]Gary watched him setting the inner wall and said it was starting to look like something. Virgil said it wasn't done yet. He checked the thermometer for the first time on the night of January 4th, 2023. He bought a basic digital thermometer with a remote probe from Dollar Tree on Green River Road. $2 ran on a single AA battery. Had a wire sensor he could thread through a gap to read the inside temperature while keeping the display outside or the other way around. Outside temperature at Lynch Road at 11 that night, according to his thermometer hanging on one of the vertical pipes outside the panel wall, 19 degrees Fahrenheit. Inside temperature in the space behind both panel walls where he and Gary were sleeping, 41° F, 22° of differential with no heat source, no heater, no candle, nothing producing warmth except two adult bodies. The body heat somewhere between 250 and 300 BTUs per hour per person at rest. A number Virgil knew from his refrigeration work because walk-in cooler calculations always factored in occupancy heat load was being trapped by the panel assembly instead of radiating out into the open air under the overpass. 22° on a 19° F night. 41° F is not warm, but it's not 19° F. The difference between those two numbers when you're sleeping on concrete with a sleeping bag rated at 20° F is the difference between a functional night and a dangerous one. Gary said 38° would have killed him. He meant the old tarp setup. Tarp setups on wind exposed concrete don't hold body heat. The tarp itself is an R0.1 barrier and the wind punches right through gaps in it. Gary had spent the previous two winters waking up multiple times per night just to move enough to stay warm. On January 4th, Gary slept 5 hours straight. He said he hadn't done that since September. The assembly got better when Virgil fixed the gap problem. The 7-ft open section of the east wall, the part he'd temporarily covered with a tarp, got two more panels in late January when Pete pulled another restaurant kitchens worth of equipment and gave Virgil first pick before the crusher truck came. He got a commercial underounter refrigerator with a solid door he could remove and a narrow chest freezer lid that wasn't the right shape but could be worked into the framing with additional angle iron at an angle. More importantly, he fixed the seam gaps. The vertical seams between panels where two gaskets met edgeto edge had a/4in average gap that leaked cold air when the wind was strong enough to create pressure differential between the outside and inside of the east wall. He noticed this on a particularly windy night in early January when the temperature inside dropped 3° over 4 hours despite both of them being awake and generating heat. The gaskets were magnetic, but they weren't attracted to each other at zero contact. They needed to be touching to pull together. His solution was foam backer rod, which is a flexible foam rope sold in hardware stores for cocking window frames. He found a partial roll at Ace Hardware on First Avenue in the cocking section and asked the guy working if he could have the partial roll since it wasn't sellable in that condition. The guy said, "Take it." He pressed the foam rope into each seam on the inside face of the east wall, compressing it so the two gaskets on either side were slightly pushed toward each other. The magnetic strips in both gaskets then had enough contact to pull together and seal. The temperature differential went up to 28° F on the next cold night. outside 17 degrees Fahrenheit. He also addressed the floor. Cold concrete is a major heat sink. The earth below it has enormous thermal mass and pulls heat down through ground conduction about 25 times faster than still air loses it. He'd known this from his refrigeration work where walk-in cooler floor insulation was always a separate specification from the wall panels. He built a platform from wooden pallets sourced from the loading dock behind a Dollar General on Covert Avenue. two layers of pallets, one rotated 90 degrees from the other, so the deck boards bridged the crossmembers. On top of that, he laid sleeping pads, two foam camping pads from Goodwill on Washington Avenue, $4 total, and the sleeping bags sat on the pads. The pallet air gaps added R1 to R1.5 per inch of height, and two pallet layers came to about 6 in of raised platform. Not a lot, but combined with the foam pads, it was enough to meaningfully slow the ground conduction.

[00:16:08]By the end of January 2023, the cold night temperature differential was holding at 26 to 30° F consistently, which meant on the coldest nights Evansville saw that winter. A stretch in late January, where the overnight low hit 8° F, the sleeping space inside was sitting at 34° F to 38° F. Still cold.

[00:16:31]Cold enough to need good sleeping gear, but survivable in a way that pure exposure was not. The thing that other people at the overpass noticed was not the temperature differential. It was the fact that the walls looked permanent.

[00:16:43]Refrigerator doors are not the material of improvised shelters. Tarps look temporary. Cardboard looks temporary.

[00:16:50]Plywood looks like someone tried.

[00:16:52]Refrigerator doors standing upright, bolted together with the factory steel faces outward. That looked like something that wasn't going anywhere. A woman named Cass, who'd been sleeping at the west end of the overpass for most of the winter, walked over one afternoon in February and looked at the east wall from the outside for a long time without saying anything. Then she asked if it ever got warm inside. Virgil said warm was relative. She said she meant warmer than out here. He said yes, significantly. Cass asked if she could sleep on the inside sometime when it was really bad out. He said he'd think about it. He thought about it and said yes. He set up a third sleeping position on the pallet platform, which involved sourcing one more pallet from behind the Ruler Foods on Lynch Road and fitting it into the remaining floor space. The platform now occupied almost the entire footprint of the enclosed space, which was roughly 8 ft x 6 ft, tight, but manageable for three people sleeping parallel. On the night of February 3rd, which was the coldest night of the entire 2022 to23 winter in Evansville, the National Weather Service station at Evansville Regional Airport recorded a low of 7° F.

[00:17:58]Three people slept inside the refrigerator door walls at Lynch Road.

[00:18:02]Virgil's thermometer read 36° F at the coldest point, around 4 in the morning, 29° warmer than the air outside. There are things that a refrigerator door does that most improvised wall materials can't. The polyurethane foam inside is closed cell, meaning each bubble of gas is sealed within its own membrane and can't communicate with adjacent bubbles.

[00:18:23]Open cell foam, the kind in a sofa cushion, allows moisture to migrate through it because the bubbles are interconnected. Closed cell foam doesn't. The moisture resistance of refrigerator door insulation is nearly absolute. Rain hits the steel face and runs off. Condensation forms on the cold outer surface and runs down. The foam inside stays dry. The dry foam maintains its R value. This doesn't seem important until you've watched cardboard insulation absorb moisture over 2 weeks in February and collapse into something with no more thermal resistance than wet paper. The gasket, the magnetic rubber seal along the perimeter, is a piece of engineering that most people don't think about, but that took decades of development to get right. The PVC outer layer is flexible from 20° F down to aroundus 20° F before it gets brittle.

[00:19:11]Inside the PVC is a flexible foam core that adds its own minor insulation value. And embedded in the foam is a continuous magnetic strip that exerts a consistent sealing force across the entire perimeter. When Virgil pressed two doors face toface with their gaskets in contact, the combined magnetic attraction held the seam closed with a force that wind couldn't open. He tested this by pushing on the outside of the connected panels with his full weight.

[00:19:35]They didn't separate. The magnetic seal between gaskets held. He'd seen the same principle demonstrated in every commercial walk-in cooler he'd ever worked on. The door gasket is what separates a cooler that holds temperature from one that runs constantly because cold air is leaking out constantly. The gasket does more thermal work per dollar of material than almost any other component in a refrigeration system. He wasn't repurposing it. He was just using it for something it was already designed to do.

[00:20:02]Pete Carver came out to Lynch Road in early March to see what Virgil had built. He drove the flatbed out on a Saturday afternoon and parked on the shoulder and walked down to the overpass on foot. He stood in front of the east wall for a while. He walked around to the outside and looked at the angle iron connections. He noted that Virgil had added a horizontal steel channel along the top of the panel assembly. Pete recognized it as part of a shelving system from a restaurant kitchen that was acting as a cap to prevent wind from getting behind the tops of the doors. He opened and closed the gap in the inner wall that served as a doorway to the sleeping space. Virgil had rigged one of the walk-in panels on a simple piano hinge set, so it swung inward. Pete said he'd built walk-in coolers and restaurants that were less well put together than this. Virgil said the restaurants probably didn't have to work with whatever happened to come through on a flatbed. Pete said, "Fair enough."

[00:20:53]He said there was another load coming in the following week from a grocery store remodel on Green River Road. old display cases, reach in coolers, the full walk-in from the back. Pete said he'd pull the doors before the crusher truck came if Virgil wanted to think about what he'd do with more of them. Virgil said he'd think about it. What he thought about was the west end. The west end of the overpass was where the wind exited. He'd prioritized the east end because blocking the incoming wind was the highest value intervention. But with good panels, he could also reduce what was happening on the west side. He wasn't trying to build a sealed room.

[00:21:26]Sealed rooms without ventilation are dangerous for exactly the same reason sealed cars are. CO2 buildup, moisture accumulation, condensation on every surface. He needed air exchange, but he could control how much. The western panel assembly, which went up in the second week of March using four doors from the grocery store remodel, was designed with a gap. Two panels on the left, two panels on the right, a 14-in opening in the center at breathing height. The opening was covered with a heavy moving blanket. the dense quilted kind which he'd found in a furniture store dumpster on First Avenue that could be pushed aside to open or fall back to cover. In practice, the moving blanket cut maybe 60 to 70% of the wind that came through from the east while still allowing enough air circulation to prevent moisture and CO2 buildup. The temperature differential on the first cold night after the west panels went up jumped to 34° F. That was the best it ever got. Outside 14° F. He didn't have a way to measure relative humidity inside, but he could tell the moisture situation was better because the inside of the outer panels wasn't condensing as heavily. Condensation on the inside face of the outer panel was his indicator that warm, humid air from the sleeping space was reaching a cold surface. The moving blanket in the western gap was letting enough dry outside air cycle through to dilute the humidity. Gary noticed the difference in the first week of March and said the air inside felt different. dryer. He said it was the first time he'd woken up in the morning without the sleeping bag feeling damp.

[00:22:55]Virgil said that was the ventilation gap working the way it was supposed to. Gary said he'd figured it was magic. Virgil said it wasn't magic. It was just airflow management. Gary said same thing. The interaction with the outreach worker happened in April. A woman named Sandra Hollands who worked for Evansville's homeless services coordination office out of a building on Sycamore Street had been doing a Lynch road check as part of her regular circuit. She had visited in the fall when the overpass was a tarp situation and had noted it in her case files as a high-risisk cold weather site. She came back in April expecting to find the same thing. She found refrigerator doors. She stood outside the east wall for a while before she said anything. She'd been doing this work for 6 years and had seen a lot of improvised shelters. She said she'd never seen anything like this.

[00:23:41]Virgil was sitting on an overturned milk crate outside the east wall, drinking coffee from a thermos he kept filled at the gas station on Lynch Road when he had money. He said the basic physics were straightforward. Sandra said the physics weren't what she was reacting to. She took a photograph of the exterior wall for her case file. She took a photograph of the interior with the pallet platform and the three sleeping positions and the thermometer mounted on the inner panel wall at head height. She wrote down his name and the temperature readings he gave her from the past two months. She said she was going to put him on the priority list for the transitional housing program the county ran out of a complex on Lincoln Avenue. He said he'd been on a list before. She said this list was different. It had a specific timeline.

[00:24:24]He said okay. She asked if he built the whole thing himself. He said mostly.

[00:24:29]Gary helped with the heavy panels. She asked if he'd worked in construction. He said refrigeration service. 9 years. She wrote that down, too. The priority housing timeline Sandra described was 10 to 14 weeks. Virgil moved into a studio apartment in the transitional housing complex on Lincoln Avenue on June 22nd, 2230, which was 14 weeks and 2 days after Sandra's April visit. The apartment was on the second floor, faced west, had a window that didn't seal properly, and let a draft in along the bottom edge. He noticed it the first night. He knew exactly what it was. He asked the building manager about it and the manager said the windows were old and they'd been meaning to get to it. He bought foam weather strip tape from the hardware store on Washington Avenue and fixed it himself. Cost him $3.80 for the roll. Took 4 minutes. The refrigerator door assembly at Lynch Road stayed standing after he left. He built it into the concrete well enough that dismantling it would have taken tools and time, and he didn't see the point.

[00:25:29]Cass was still sleeping there. Gary was still there. Two other people had moved into the space on the east side of the inner wall. A man named Thomas and a younger guy whose name Virgil never learned who' started using the space sometime in February when word had gotten around that there was a warm spot under the overpass on Lynch Road. Virgil left the thermometer mounted on the inner wall. He left a note on a piece of cardboard tucked behind it explaining the ventilation gap in the west wall and why it needed to stay open at least partially. that if someone blocked it completely to try to get warmer, moisture would start building up inside and the sleeping bags would get wet and they'd be colder at 4 in the morning than they would have been with the gap open. He wrote it out in plain language, the way he'd explain it to a customer who didn't have any background in refrigeration.

[00:26:14]He wrote, "Keep the blanket there, but don't seal the gap completely. The gap is doing work. Trust it." He tacked the note to the inner panel wall with a bolt he threaded through a pre-existing screw hole in the door skin. The last thing he said to Gary before he left was that he should check on the angle iron connections every couple of weeks, especially the ones holding the bottom channel to the concrete. He said if any of them were loosening up, Pete Carver could probably come by and tighten them if Gary called him. Gary said he didn't have a phone. Virgil said he'd leave Pete's number on the note with the ventilation instructions. Gary said the whole thing was going to fall apart without Virgil there to maintain it.

[00:26:50]Virgil said probably not. He said it was steel and foam. Steel and foam didn't go anywhere on their own. Gary was quiet for a second, then he said he'd keep an eye on it. The east wall was still standing in October 2023, which was the last time Sandra Hollands documented it on a site visit. She noted in her file that the gasket seals on the panel seams were intact, that the pallet platform was still in place with additions.

[00:27:15]Someone had added a fourth layer of pallets to one section to create a slightly elevated sleeping position and had covered the new section with a piece of indooroutdoor carpet that had come from somewhere. She noted that the thermometer was still mounted on the inner wall. She noted that the note was still there. Refrigerator doors are thrown away every day. Every appliance recycler in the country pulls them before the units go to the crusher because the doors have to be removed for legal reasons. The Federal Refrigerator Safety Act requires them to be disassembled. So no child can get trapped inside an abandoned unit. The doors get piled in yards, stripped of their hardware, and eventually scrapped for the steel. The foam goes with it.

[00:27:54]The gasket goes with it. Decades of engineering designed to maintain temperature differentials and conditions that never change. All of it goes to the crusher at 11 cents a pound. What Virgil Doss did was not complicated. He changed what the doors were pointed at. Instead of keeping cold air inside a refrigerator, they kept cold air outside a shelter. The physics are identical.

[00:28:17]The material doesn't care about the application. The closed cell polyurethane foam inside a residential refrigerator door has no idea it isn't a walk-in cooler panel. The magnetic gasket sealing two adjacent doors together has no idea it isn't sealing a door to a frame. It just seals. There are things that require expertise. There are things that require materials. There are things that require money and time and tools and connections. And then there are things that just require someone who already knows what something is made of and is willing to think about what else that thing could be used for.

[00:28:49]Virgil Doss spent 9 years learning exactly what was inside a refrigerator door. He used it.