I know there have been a pile of hot-box threads already, but I figured I'd post this both because it shows how to resolve some problems that others have encountered (temperature overshoots for example) and because the requirements and resulting solution were a bit different. My requirements in building a hot-box were as follows:
1. Temperature over-shoot no greater than 3 degrees F
2. Average temperature within 1 deg. of target after warmup
3. Temperature differences along the skis no worse than 5 degrees F at equilibrium
4. Accomodate 2 pairs of skis of up to 220 cm length (i.e. DH skis)
5. Accomodate 2 pairs of skis of up to 185 cm length while using no more than 6.5' of floor space
6. Be transportable in a vehicle with a 5' long cargo area
7. Be light and durable enough to carry and transport
8. Be capable of maintaining 140F while used outside in winter conditions (for example covered by a tarp on a condo deck).
(4), (5), and (6) can't be achieved by a single unitary box, so they require either an accordion-style expansion of some sort or a modular system. I went for the modular approach to keep things simple (no slides etc), so the basic structure of the box consists of a ~5' main box and 2 interchangeable extensions (for < 185 and < 220 cm skis respectively) that attach to it via spring-loaded latches. Here is a picture of all 3 parts (on the floor of my messy garage). The thermocouples at each end are for external monitoring of the ski tip and tail areas.
The internal structure consists of a lower bay with fans/heaters, and an upper bay with shelving supports for the skis. The dividing shelf between the bays extends almost the full length of the hot-box (including the extensions) such that there are no significant "stagnant pockets" of air:
The basic structure consists of 1/4" plywood framed with 2x2 furring strips. The strips are routed into L sections to save weight/space. The insulation is 1/2" polyiso (R-matte 3). There are hardwood (poplar and birch depending on application) stringers to add longitudinal stiffness to the top and bottom of the opening and also to serve as "feet" below the box.
The following picture shows the interior of the short extension, including both the continuation of the dividing shelf and the main thermocouple, which measures temperature just before the air "turns up" into the ski compartment (i.e. it measures the highest temperature that the skis could possibly see). The thermocouple sticks down through the mid-shelf, though you can't really see that here. The wood blocks on either side hold the clamps. The foam around the opening is 3/16" thick boot-fitting foam that I bought by the roll ages ago:
The following picture shows the lower bay of the main box without the dividing shelf, including both heaters and fans. The heaters are a pair of 350W Vulcan enclosure heaters connected in parallel. I chose to use two because the box isn't wide enough for a single 700W heater, and because this gave me the ability to disconnect one of the heaters if overshoot became a problem (it ultimately didn't). Note the quick-and-dirty plenum that concentrates the airflow through the heaters. The fans are Dayton 120x38 mm 120 VAC types, rated for ~40 cfm each:
At the total rated airflow of 120 cfm (~= 0.06 kg/sec pessimistically assuming dry air at 55C) and the maximum heater output of 700W, the expected delta-T across the heater is about 20F (i.e. the amount by which the air will increase in temperature from intake to exhaust when the heater is at full power). This means that I have to keep the heaters running at fairly low power to achieve my 5 degree uniformity goal. Assuming that half of all heat loss happens in the top chamber with the skis (losses in the bottom chamber don't count against my uniformity goal), this means I need to keep the heaters running at ~50% or less at steady-state. I could increase that by switching the fans out for higher-flow models, which I may do at some point. I could also increase it by humidifying the interior air (humid air has much higher specific heat than dry air, as anybody in Arizona will inform you ad nauseum - "but it's DRY heat!"). I would note here that high airflow is the single easiest way to deal with both overshoot issues (for example due to the thermal mass of the heaters) and uniformity issues. If in doubt, get the burliest, noisiest fans you can tolerate.
Heater control is handled by an Auber 2342P PID controller, contained in a metal enclosure on the right end of the main box (along with the terminal block for most of the 120 VAC circuitry). The Auber controller has a maximum temperature rating of 122F, so the electrical enclosure is both insulated from the rest of the hot-box interior and vented to the outside (see the second picture above). This cools it down just enough to stay within spec when the rest of the box is at 150F (though I never run it higher than ~130 with skis). The controller delivers negligible overshoot (1-2 deg F) and almost perfect temperature control even with default gains. I did initially have some trouble when I used a thermocouple with a metal cylinder around the probe, as the cylinder added too much thermal mass and slowed the thermocouple's response time to the point where the system oscillated by up to 20F. All was well after I switched to a bare-probe K-type thermocouple instead.
Both modelling and testing showed that the enclosure as-built could sustain a delta-T of ~70F over ambient at a steady-state power of 350W (recall that this is the most steady-state power I can allow and stay within my uniformity goal). This is enough when running indoors, but would not be enough for use on, say, a patio during winter. To do that I built a partial "exoskeleton" that wraps around the box and adds an additional layer of 3/4" polystyrene insulation. This is held together by gravity and interlocking bevelled surfaces (not shown). With this the 350W steady-state delta-T goes up to about 110F, which is all I realistically need:
Don't ask me how much this all cost. If I figured it all out then I might have to tell my spouse...