After another busy workweek where the project was on hold, I finally had time this weekend to return to the dry box design. This time, I decided to make some structural and wiring improvements based on the previous airflow and integration issues.
The biggest change is that I moved all the wiring—sensor lines, fan power, and PTC connections – to route through the bottom of the sealed box, leading into a newly designed lower housing compartment beneath the container.
Inside the main dry box, I mounted two centrifugal fans directly onto the inner walls to help guide internal airflow. In the center, I fixed the SONOFF TH Elite temperature and humidity sensor, which will continue serving as the main control input. All wiring is routed through sealed holes at the base using waterproof grommets, feeding neatly into the lower chamber.
The interior still includes a support frame for the filament spool, but this time the frame is directly screwed into the bottom of the sealed container. This provides much better mechanical stability, especially under heat and vibration. The bottom part of the support frame also serves as a mount for both the PTC heater and the cooling fan.
I finished designing and 3D printing the updated components, assembled everything quickly, and successfully built a new working prototype that’s now ready for the next round of testing.
To verify the temperature difference between the PTC outlet and the TH Elite sensor, I once again used a SONOFF SNZB-02LD temperature and humidity sensor. I mounted the main unit on top of the filament spool holder and positioned its external metal temperature probe directly at the PTC outlet.
This setup allows me to directly compare the readings from the SNZB-02LD probe
and the TH Elite sensor in the center of the dry box. It should give me a clearer understanding of the thermal distribution during operation.
When I powered on the system and began heating, I noticed a significant temperature difference between the SNZB-02LD probe at the PTC outlet and the TH Elite sensor in the center of the box. When the SNZB-02LD reported 60°C at the outlet, the TH Elite still showed just over 30°C. Even as the outlet temperature rose to 75°C, the center temperature remained below 40°C.
That’s when I realized something important: the heat produced by the PTC needs time to propagate through the enclosure. Simply keeping the PTC running doesn’t immediately warm the rest of the space – it just builds up heat at the outlet.
So I came up with a new idea: instead of heating continuously, I should try using an intermittent heating strategy – turning off the PTC when the outlet reaches 70°C, and letting the residual heat spread throughout the enclosure. When the outlet temperature drops to 60°C, I can turn it back on again. Repeating this cycle should gradually raise the overall temperature without overheating the hotspot.
Since the TH Elite’s sensor is currently placed in the center of the enclosure, I can’t use its built-in auto mode to control heating based on the PTC outlet temperature.
Instead, I’ve decided to manually operate the heating logic – by monitoring the SNZB-02LD probe at the outlet, and switching the TH Elite on and off via the eWeLink app.
Admittedly, this method is a bit clumsy, but it allows me to test the heating strategy without needing to redesign or rewire the prototype. If the approach proves effective, I can then revise the model tomorrow for a more precise and automated version.
By observing the SNZB-02LD readout, I turned off the PTC heater once the outlet temperature reached 70°C. The temperature at the PTC outlet continued to rise briefly, peaking around 72°C, before it began to slowly fall. Once it dropped to 60°C, I manually turned the PTC back on. The temperature would continue to drop slightly to around 58–59°C before it began rising again. Every time the temperature reached 70°C, I repeated this on/off cycle manually.
Although the outlet temperature cycled steadily between 58°C and 72°C,
the TH Elite sensor at the center of the box showed a consistent upward trend – increasing by 2–3°C with each cycle. After repeating this process more than a dozen times, I was able to raise the center temperature to around 50°C.
If I had used my previous approach of continuously heating until the center hit 50°C, the outlet temperature might have reached 80°C or even 90°C, which would very likely deform the structural parts near the PTC outlet.
At that point, I noticed the center temperature stopped increasing significantly,
indicating that the system likely reached a thermal equilibrium. This result confirms that the staged heating strategy works. Tomorrow, I plan to relocate the TH Elite sensor to the PTC outlet, so I can use its built-in auto mode to control the on/off logic automatically. That way, I can run hundreds or even thousands of cycles to observe the long-term effect of this intermittent heating strategy.
Leave a Reply