I started with a ship full of max-level veterans, and I'm just a complete noob.

The next morning, three high-purity alumina tubes were placed on the inspection rack in turn.

Lin Lanying followed the acceptance procedure in the meeting minutes and handed the vernier caliper to Fang Xudong, asking him to measure the inner diameter from the pipe opening, the center, and the tail end.

The dimensions of the old quartz tubes were laid out to the side, and the measured values ​​quickly revealed the first difference: the inner diameter of the new tubes was generally too small. If the original quartz boats were to be used, the expansion after heating could cause them to jam the furnace tubes, and even the movement of the molten zone would be affected.

Old Sun picked up one of them and examined its inner wall under the light; the opening was slightly elliptical.

He marked the thicker side with a pencil, then took the quartz boat outline and re-marked the dimensions.

"The sides of the boat should be tapered a bit, and the push rod joints also need to be shaved. Silicon material is less resistant to being stuck in the furnace than germanium material."

Lin Lanying clipped the material specifications of the alumina tube into the process booklet, and wrote "1500 degrees Celsius, ventilation and heat preservation" on the pre-firing test page.

"First, use an empty tube to raise the temperature to 1,500 degrees Celsius, observe the inner wall and the deposits in the exhaust gas, find out the amount of aluminum released at high temperature, and then decide which tube to put into furnace No. 3."

Aluminum in silicon is an acceptor impurity, and even if the amount released is limited, it can still change resistivity and junction performance.

Jiang Ming then adjusted the modification sequence, scheduling the pre-firing test before the mechanical installation.

After half a day of screening, the tube with the smallest change in inner diameter was selected as the main furnace tube, another was kept as a spare, and the third tube was transferred to the cutting test due to its large ellipticity at the tube opening.

The brief item "replacing the furnace tubes" in the meeting minutes has been broken down into three hurdles on the workbench: size, thermal expansion, and contamination.

When the silicon carbide rods were installed in the furnace, new problems quickly emerged: the rod ends did not fit evenly against the metal clamps, and after the plates were tightened, only two contact marks were left on the edges.

Silicon carbide is both conductive and brittle. If the pressure on the chuck is slightly off, it may break. Old Sun can only remove the rod body, slowly repair the end face with a triangular file, and then use fine sandpaper to grind away the file marks.

He inserted a layer of copper foil into the inside of the tablet press, using the softness and fit of the copper foil to compensate for the small undulations on the end face. He also had Fang Xudong measure the room temperature resistance of both ends and each section of the rod.

Fang Xudong pressed the multimeter probes to the numbered positions, and Da Liu read them out one by one. The record sheet quickly listed the values ​​at the center, end, and chuck connection of the rod.

The difference between the two silicon carbide rods was within the allowable range, but the resistance at the connection point was higher than expected. After repeated disassembly and reassembly, the readings shifted.

Looking at the records, Jiang Ming was already on guard, but for now, he could only assemble the structure and then use the high-temperature curve to determine the extent of the impact of the movement.

The tungsten-rhenium thermocouples coordinated by Wang Shouwu arrived in the afternoon. The calibration certificate was attached to the wooden box. The wire diameter was much thinner than the temperature sensing element used before, and the opening of the alumina protective sleeve also needed to be re-ground.

Old Sun clamped the sleeve into the cork liner and gradually widened the opening of the tube with a self-sharpening scraper. He kept the force in his hands within the material's tolerance range, fearing that small cracks would spread along the tube wall at high temperatures.

Da Liu then disassembled the original amplifier and readjusted the preamp gain according to the tungsten-rhenium thermocouple output. After the thermoelectric potential decreased, the contact potential and amplifier zero drift would occupy a larger proportion.

He replaced all the terminals with materials from the same batch, fixed the low-temperature terminals in an insulated box containing a thermometer, and then used a calibration source to check the amplification factor level by level.

Jiang Mingshou stood by the recording table, repeatedly checking the thermocouple output and paper tape readings until the errors of each range from room temperature to simulated high temperature signals were all within the experimental range.

After the air pipe ventilation was increased to 1,500 degrees Celsius, a very thin layer of deposit appeared on the inner wall of the furnace tube. Lin Lanying scraped a sample for preliminary inspection. The traces of aluminum release were within a controllable range, but further pre-firing and protective airflow were still required.

The first no-load power-on after the mechanical modification was completed was scheduled for the morning of the following day. All personnel retreated outside the shielding line according to the high-temperature test procedure. After Da Liu turned on the main switch, the power was gradually increased.

As the silicon carbide rod changes from dark to orange-red, heat gathers along the furnace towards the center. When the temperature curve on the paper tape crosses 800 degrees Celsius, the old PID parameters can still maintain the heating rate within the set range.

Fang Xudong recorded the voltage, current, thermoelectric potential, and furnace outer wall temperature at each temperature zone. The curve from 800 to 1000 degrees Celsius looked smooth, but everyone's attention was always focused on the higher temperature range.

After crossing 1100 degrees, the ink line on the paper tape began to swing left and right. Fang Xudong used a timer to compare the adjacent peaks. The period was about 40 seconds, and the amplitude was close to plus or minus half a degree.

"The low-frequency oscillation has appeared, and the same noise is no longer visible on the thermocouple signal. The control loop has started to correct back and forth according to the furnace temperature."

Jiang Ming instructed Da Liu to reduce the heating rate, while recording the delay between each power adjustment and the temperature response. The oscillation on the paper tape temporarily narrowed, but then widened again after the temperature range continued to rise.

At temperatures above 1300 degrees Celsius, the thermal inertia of the furnace increases significantly, and the resistance and current of the silicon carbide rod no longer change regularly along the low-temperature range, with the old parameters gradually losing their binding force.

When the temperature approaches 1420 degrees, the ink line fluctuates by ±1.2 degrees on both sides of the set value. Each time the integral phase pulls back the temperature, the next deviation appears from the opposite direction.

Jiang Ming switched to manual control, dividing the input power into smaller increments. Only after the furnace temperature returned to a stable range did he allow Da Liu to check the thermocouples, amplifiers, and control circuits in turn.

The thermocouple output remained clean, the amplifier zero drift was within the calibration range, and the low-frequency fluctuations on the paper tape corresponded to the current changes at the power supply end, gradually pointing to the heating element as the problem.

"First, cool it down. After the rod has cooled down, remeasure the resistance of each section, paying special attention to the connection ends."

Da Liu gradually reduced the power according to the cooling procedure. The furnace body returned to the operating temperature after a long period of heat dissipation. Lao Sun removed the heat insulation plate, and Fang Xudong then retested the previously numbered position.

The change in the resistance at the center of the rod is within the range after the material has been cycled at temperature, but the resistance at the two contact ends has increased by about 15 percent compared to the first measurement, and one end still moves after being repeatedly pressed.

Old Sun squatted down next to the furnace, took off his heat-insulating gloves, and used his bare fingers to feel along the seam between the silicon carbide rod and the pressing sheet. His fingertips quickly touched an area that was biased by the force.

"Jiang, the contact resistance here is active."

He disassembled the sheet and showed everyone the contact marks left on the copper foil. The single-sided spring sheet shifted outward after expanding at high temperature, leaving only the inner section as the actual conductive area.

"When the rod gets hot, it expands, and the force of the spring can't keep up. The contact surface gets bigger and smaller, and the resistance in the circuit also changes."

Fang Xudong marked the direction of force on the original structural diagram. Under constant voltage power supply, the change in contact resistance will directly change the current passing through the rod, and the input power will naturally drift accordingly.

Jiang Ming re-laid out the temperature paper tape, current records, and contact resistance data, aligning the same time points one by one. About half of the fluctuations could be found in the changes in end resistance.

The remaining part is related to the temperature range of the rod. When the temperature rises and falls to the same temperature, the resistance values ​​are still different, which is obviously due to the nonlinearity and hysteresis of silicon carbide itself.

When Huang Kun arrived at the furnace room, the paper tape was covering half of the worktable. After looking at the markings of ±1.2 degrees, he looked back at the No. 3 furnace, which was still dissipating heat.

"It's a little better than the manual control we used to have."

The old furnace at the Institute of Physics relies on manual adjustment and can usually only maintain a temperature range of ±2 to ±3 degrees. ±1 to ±2 degrees has already crossed a threshold, but zone purification and single crystal growth require a narrower temperature range.

Wu Hanzhang took the cigarette from behind his ear, lit it, took a puff, and then moved it outside the door, speaking decisively.

That's not enough.

Jiang Ming did not rush to adjust the PID. He divided the sources of fluctuation into two columns: random drift at the contact end is a mechanical problem, while the change of silicon carbide resistance with temperature is a control problem. Mixing the two together would only result in incorrect parameters.

"First, modify the clamp to suppress the random changes, then use the new rod curve to perform piecewise step response. The order cannot be reversed."

Old Sun had already taken the old pressure plate and drawn the double-sided force structure on the paper. The original single spring sheet was changed to a double-sided pressing, and the contact surface was also prepared to be increased from one layer of copper foil to multiple layers.

Jiang Ming added the direction of force on the clamp to the kraft paper notebook, and plotted the temperature, resistance and input power of the silicon carbide rod on the same coordinate system.

The positive and negative 1 and 2 marks on the paper tape reminded him that the controller, which had been validated and matured in the germanium era, had just entered the silicon temperature zone and had become a system that could barely function.

At the bottom of the page, he wrote down the direction of psychic search, high-temperature control system, and Soviet cybernetics system.