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

On Monday morning, the device group will connect the test bench ahead of schedule and carry out the twelve sets of compressed electrical pulse tests one by one.

For the next three days, Fang Xudong and Lao Sun took turns guarding the instrument, while Jiang Ming checked the current generated each time according to the device number, until the last set of parameters was completed.

On Wednesday evening, Fang Xudong pasted thirty-six device process cards on the blackboard, dividing them into four columns according to pulse amplitude, and then arranging each column into three rows according to the duration of action. The forward resistance, reverse resistance, and conversion loss were marked with red and blue pencils, respectively.

The three point-contact germanium diodes in the same group were taken from adjacent slices, the stylus was ground from the same batch of tungsten wire, and the clamping position of the spring bracket was also verified by gauge, but the data were very scattered.

The lowest group had a forward and reverse resistance ratio that differed by only 8 percent, while the highest group had a difference of 42 percent. Among several devices with acceptable conversion losses, there were also samples with significantly lower reverse resistance.

Wu Hanzhang stood in front of the blackboard, his pencil tracing the twelve sets of data line by line. He initially looked for patterns in the changes based on the amplitude of the electrical pulses, but when he reached the fourth column, he stepped back and rechecked the data from the end of the device number.

"These four groups have the highest dispersion, and the pulse amplitude and duration are different, but the germanium sheets are from the same crystal segment."

Fang Xudong pulled out the corresponding process card and spread it on the worktable. The germanium wafers used for the thirty-six devices were numbered sequentially, and four groups of high dispersion samples all fell in the middle of the fifth furnace crystal, near the tail.

Old Sun took out a micrometer and retested each of the sealed spare germanium wafers from the same batch. He filled in the thickness readings along the record sheet, and the maximum deviation was still within the allowable range of the slicing process.

He then disassembled the diamond wire saw's clamps and inspected the soft metal liner, guide rail clearance, and locking screws. No mechanical damage could be found on the cut surface to explain the 40% dispersion.

"If the thickness is aligned and the wire saw moves straight, then if the clamps wobble, all the pieces cut in the same stroke will be affected."

Jiang Ming took the resistivity curve of the fifth furnace from the folder, fixed it to the bottom edge of the blackboard with a glass strip, and circled the area where the abnormal germanium sheet was located on the curve.

This section of crystal is located within a 30-millimeter uniform region. The four-probe retest values ​​remain within ±2%, and the curve does not show any continuous decline due to the enrichment of impurities at the tail end.

Fang Xudong rearranged the device data, temporarily obscuring the pulse parameters and retaining only the germanium chip number, slice position, and forward and reverse resistance ratio. The previously scattered numbers were arranged into a slanted boundary according to their positions.

"The closer to the middle and end of the chip, the greater the difference between the three devices, but the average value on the same chip is still within the operating range."

Jiang Ming picked up a disassembled spring bracket, stacked the stylus pressure mark and the current-generating paper tape together, and the results of the three rounds of retests were close to each other, indicating that the pressure change was insufficient to cause the current system deviation.

The amplitude of the electrical pulse, the duration of action, the pressure of the stylus, the thickness of the germanium sheet, and the material resistance were removed from the list of uncertainties, leaving the remaining variables to be determined by the cut surface itself.

"Da Liu, select four pieces from the high aberration group and four pieces from the low aberration group, repolish them, and then place them together under a polarizing microscope."

Da Liu fixed the eight germanium sheets onto the stage one by one, first aligning them according to the cutting marks, then rotating the dial to find the matting position, while Fang Xudong noted the angle of each sheet.

The four discs in the low dispersion group dimmed when they were close to the same scale, and the angle difference was within the reading error range. However, the high dispersion group was divided into several extinction positions, which were separated from each other by three to five degrees.

Jiang Ming asked Da Liu to repeat the test, and then changed the order of the samples for a second test. The second reading still fell within the original range, and the direction of the deviation also corresponded to the position of the crystal slice.

Old Sun found the shape record of the fifth batch of crystals. The crystal column had slightly bent when it grew to the second half. The process card only marked the continuous surface growth lines, so the outer circle was still used for slicing.

The fixture locks in the shape of the crystal column, but the main axis of the internal lattice is deflected along the growth direction. The closer to that point of curvature, the greater the difference between the outer circle center line and the true crystal orientation.

Jiang Ming wrote the polarization readings on the right side of the blackboard and listed the carrier mobility relationships of different crystal orientations of germanium crystals next to them. The dispersion of several sets of devices finally fell into the same explanatory framework.

When the same contact formation conditions occur on different crystal planes, the carrier transport characteristics will change accordingly, and the forward and reverse resistance ratios and mixing conversion losses will also produce system differences.

Wu Hanzhang opened the 24 GHz pre-research plan, found the column for the paired diodes in the preamplifier stage, and tapped the allowable range of dispersion with a pencil.

"The preamplifier needs to use paired diodes. If the dispersion exceeds 10%, the equilibrium state cannot be maintained. What you've given me is 40%."

"The device fabrication process still needs to be continued, but the main reason for this round has regressed to material processing. Before cutting, the lattice principal axis must be determined, and then the cut can be made along the target crystal plane."

Jiang Ming wrote down four requirements on the back of the process card: orientation before cutting, crystal angle marking, turntable locking, and post-cutting verification, limiting the rectification to within one degree.

After he finished writing, he walked to the iron cabinet, first took out Qian Xuesen's black lecture notes, turned to the crystal structure annotation page, and then took out Lei Zhenbang's manuscript from the lower file box.

The red pencil mark in the manuscript is still clear: the purity of five nines must be combined with crystal orientation control, otherwise the performance difference of devices in the same batch will swallow up the progress of materials.

Fang Xudong looked at this sentence side by side with the material curve of the fifth furnace for a long time, then looked down and flipped to the first page of the manuscript to confirm the paper's age and the date it was written.

When Lei Zhenbang wrote his judgment, even stable high-purity germanium materials were hard to obtain domestically. Now, with the fifth batch bringing uniformity within ±2%, that reminder has finally taken effect on the thirty-six devices.

Old Sun squatted down next to the cutting table, slid his index finger along the diamond wire saw guide, touched the locking seat, and then compared the angle on the crystal column bending record.

"I can pinpoint the exact spot to cut down to a hair's breadth, but the crystal lattice is hidden inside, so I need to know which way to cut first."

While a polarizing microscope can identify the approximate extinction direction, the birefringence response of the sample itself, the surface finish, and the reading resolution limit further orientation.

Jiang Ming pressed eight polarized photographs next to the manuscript and drew error bands along the angle of the high dispersion group.

"Optical orientation is enough for us to find the problem, but the device matching requirement is within one degree, so we have to switch to X-ray diffraction."

Wu Hanzhang closed the preliminary research plan and asked, "Does the factory have this kind of equipment?"

"The Japanese-made X-ray diffractometer at the Institute of Physics is fully booked, and it will take at least two more months before it's our turn."

The weak pulses from Aksai have already compressed the pre-research time; two months is enough to delay the finalization of the mixer diode, and the subsequent receiver pre-amplifier assembly will also be delayed accordingly.

Jiang Ming took out a piece of graph paper and drew the relative positions of the X-ray tube, the single crystal sample, and the film: the X-ray passes through the central hole of the film and shines on the crystal, and the back reflection spots fall back onto the film.

"Waiting for readily available equipment is too late. As long as we can find a usable X-ray tube and a high-voltage source, we will set up a back-reflection Laue orientation device in 502."

Wu Hanzhang stared at the tube voltage column on the diagram, then glanced at the more than 40% device dispersion on the blackboard, and tucked the preliminary research plan into his canvas briefcase.

"I'll go check with Director Zhang about the equipment supply chain. Tonight, make a clear list of the pipe source, power supply, shielding, and film reading conditions. We'll find the items according to the list tomorrow morning."

Jiang Ming folded the graph paper and put it into a folder, while Fang Xudong gathered the polarization photos, copied the crystal position, extinction angle, and device dispersion into the last page of his notebook, and finally added a line.

Five nines solved the material purity problem, but a 3.8-degree directional deviation blocked the 24 gigahertz preamplifier.