The red light on the receiver flashed briefly, then returned to normal.

Susan's fingers hovered over the keyboard as she stared at the screen. She retrieved the data from before and reviewed it at three times the speed. The signal waveform began to fluctuate slightly at the tenth second, followed by a brief disconnection fifteen seconds later, lasting 0.6 seconds. Then, the bit error rate jumped from 0.03% to 2.1%, before falling back down.

“No,” she said. “It’s not interference.”

Nana immediately accessed the log system and retrieved all transmission records from the past ten minutes. Twelve windows popped up side-by-side on the screen, each frame marked with a timestamp and signal quality score. She overlaid the data from the old and new systems for comparison; the lines almost overlapped.

Chen Hao walked over and stood behind the two of them, looking at the two parallel curves. "Does this mean... all our hard work was for nothing?"

"The performance improvement is less than 15 percent," Nana said. "It didn't meet our expectations."

“I don’t understand these numbers,” Chen Hao said, pointing to the map. “I just want to know, can someone ten kilometers away hear me?”

“No.” Susan switched to the remote feedback interface. “They tried to make a voice call three times. The first one ended after ‘hello,’ the second one sounded like a duck being strangled, and the third one didn’t connect at all.”

Carl stood up, supporting himself on the edge of the table, and walked to the main screen. His movements were still a little slow, but his eyes were clear. "We checked every step during assembly—power supply, modules, solder joints—everything was fine. The equipment passed its self-test too. So where's the problem?"

"Perhaps there's a problem with the testing environment." Chen Hao stroked his chin. "Too much wind? High humidity? Or is it the terrain?"

“Rule it out.” Nana pulled up the weather record. “The current wind speed is 3.2 meters per second and the relative humidity is 67 percent, both within the standard range. The terrain is open and flat with no obstructions.”

“That means the equipment is no good.” Chen Hao sighed. “The treasure we pieced together from dismantled antiques is actually just a pretty face.”

“Not entirely,” Susan shook her head. “At least it didn’t blow a fuse, and there was no self-restart protection. It works, but it’s not very effective.”

"So bad that it's negligible?" Chen Hao grinned. "Then you might as well let me shout it out; maybe it'll travel further."

“Theoretically, this shouldn’t be the case.” Karl stared at the heat map. “The antenna gain has been calculated three times, the modulation parameters have been set according to the optimal solution given by Nana, and the energy output is also within the safe threshold. All conditions are met, but the result is not as expected.”

"Science can be so frustrating sometimes." Chen Hao sat down, the chair making a dull thud. "You think everything is ready, but then the east wind just won't come. And when it does, it's just a gentle breeze, not enough to lift a kite."

"Should we try again?" Susan asked. "And try a different frequency?"

"No," Chen Hao waved his hand. "Don't change the settings yet. What we're most afraid of right now is making random changes, which will only make things worse. We need to know where the problem lies before we start."

“Then let’s start from the beginning.” Nana initiated a comprehensive diagnostic procedure, “Power module, signal generator, amplifier circuit, transmitting antenna, encoding logic, check the operating status item by item.”

“Okay.” Chen Hao nodded. “You report one item, and we’ll check one item. Whoever is in charge will speak.”

The first item is the power supply. The voltage is stable, the current is stable, and the power supply is normal.

The second item is the signal generator. The waveform generation is accurate, and the frequency lock is without offset.

The third item is the preamplifier circuit. The input-output ratio meets the design value, and the gain is up to standard.

The fourth item is the main amplifier module. The load is balanced, the temperature control is normal, and the power reduction mechanism has not been triggered.

The fifth item is the antenna array. The impedance matching is good, and the VSWR is within an acceptable range.

The sixth item is modulation and coding. The qpSK mode is running without errors, and the error correction algorithm has been successfully loaded.

After completing the seven checks, no obvious faults were found at the system level.

"That's strange." Chen Hao frowned. "Everything's normal, so why can't we get good results?"

“It might not be a problem with a single module,” Susan said. “It’s a problem with the overall team coordination. It’s like a team where each player can play individually, but when put together, they just can’t win the game.”

"That's called a lack of chemistry," Chen Hao said. "Or the coach is just too bad."

"Who are you implying?" Susan glanced at him.

"I'm implying to myself," Chen Hao said, shrugging. "I'm in charge of the coordination, so I'll take the blame."

Carl suddenly spoke up: "Is it possible that we overestimated the actual performance of the hardware? For example, the recycled capacitors may have met the specifications, but aging caused a response delay? Or the chip may be able to work, but its performance degrades at high frequencies?"

“Possibly.” Nana pulled up the component life assessment model. “Some components have been in service for more than eight years and have been in standby mode for a long time. There may be hidden losses after restarting.”

"So, they appear to be alive, but are actually paralyzed?" Chen Hao said.

“The analogy is apt,” Nana said.

“Our equipment is like a group of veterans working while sick; they shout slogans loudly, but their legs go weak when they charge into battle.” Chen Hao leaned back in his chair. “That’s pretty true.”

"What should we do?" Susan asked.

“What else can we do?” Chen Hao sat up straight. “Keep investigating. Since there’s nothing wrong on the surface, dig deeper. We need to see if something went wrong somewhere.”

"Where to begin?" Carl asked.

“Let’s work backward from the results.” Chen Hao pointed to the disconnection records on the screen. “Each signal failure occurred 30 seconds after transmission, it wasn’t random. This shows that the problem has a pattern, it’s not accidental.”

“I noticed that too.” Susan quickly flipped through the log. “The first anomaly occurred at 31 seconds, the second at 33 seconds, and the third at 29 seconds. They all occurred around 30 seconds.”

“It’s like some kind of cumulative effect,” Nana said. “It could be that the temperature rises, causing changes in material properties, or it could be that the power load gradually increases, causing a slight drop in voltage.”

“Then keep an eye on this period,” Chen Hao said. “Pull out the data from the first 35 seconds, break it down to the millisecond level, and see if anyone can’t hold on any longer.”

Nana immediately took action, breaking down the most critical time periods from the three tests into high-density waveforms. A dense array of lines began to scroll across the screen, each frame displaying the real-time status of each module.

Susan amplified the power supply output curve and discovered an extremely subtle downward trend—starting from the 25th second, the voltage slowly decreased by only 0.8%, which was within the normal fluctuation range, but happened to be synchronized with the signal attenuation.

“There’s something here,” she said.

Nana simultaneously checked other data streams and found that the current demand of the amplification module had increased slightly during the same period, while the internal resistance of the battery had also increased slightly.

“It’s not a single problem,” Nana said. “It’s a combination of factors. A slight decrease in power output, a slight increase in load, and lag caused by component aging ultimately lead to a decline in signal quality.”

"It's like ants moving house," Chen Hao said. "One ant can't move a mountain, but a hundred can."

“But we’ve already considered redundancy in our design,” Carl said. “Logically, this fluctuation shouldn’t cause a disconnection.”

“Unless,” Susan suddenly thought of something, “our redundancy calculations are based on ideal components, but in reality we’re using secondhand ones. There’s a thin layer of grease between theory and reality.”

"So the system thought it was safe," Chen Hao said. "But it was already standing on the edge of a cliff, and it was still grinning like an idiot."

The lab was silent for a few seconds.

"What do we do now?" Carl asked.

“There’s nothing we can do.” Chen Hao stood up, walked to the whiteboard, and picked up a pen. “We admit defeat, but we don’t give up. The equipment not meeting the target doesn’t mean we’re going in the wrong direction. It could be the materials that are holding us back, or it could be that we didn’t pay attention to the details. But none of these are reasons to give up.”

He wrote three words on the whiteboard: **Test again**.

"We're not in a rush to launch this time," he said. "We'll first figure out the true performance of all the components. We'll test the old and new ones separately, run them at full load individually, and see which ones can withstand the load and which ones will fail."

“We also need to check the connection points,” Susan added. “Poor solder joint contact can also cause momentary interruptions.”

“Yes.” Chen Hao nodded. “Every screw needs to be checked. I don’t expect to succeed on the first try, but I can’t afford to fail haphazardly.”

“I suggest establishing a test profile,” Nana said. “We can create an independent performance curve for each recycled component, which will help us avoid risky combinations when combining them later.”

“Okay,” Chen Hao said. “You’re in databases, so you’ll take the lead. Susan will be in charge of code-level monitoring, Carl will keep an eye on the hardware status, and I’ll coordinate the pace.”

"Aren't you going to sleep?" Susan asked.

"I can't sleep." Chen Hao chuckled. "I just said I'd get to the bottom of this, and now I'm going to lie down? That would be a joke."

"Aren't you afraid of collapsing from exhaustion?" Carl asked.

"Yes, I'm scared," Chen Hao said. "But I'm even more afraid of waking up tomorrow and finding out that I failed because I was lazy. If that happens, I won't even have the confidence to complain about the outdated equipment."

He put down his pen and looked at the flat signal curve on the main screen. "We're far behind, but we haven't lost yet."

“Export all the test records from the past 72 hours,” he told Susan.

Nana has begun connecting to the host port, the robotic arm is steadily inserting into the data slot, and the background progress bar is slowly advancing.

Susan opened the virtual interface and categorized the failed transmission segments.

Carl sat in a folding chair, holding a printed waveform comparison chart in his hand, his eyes never leaving the screen.

Chen Hao stood in front of the control panel, his fingers lightly tapping the table.

The signal curve remains stationary.

The only sound in the laboratory was the low hum of the machines running.

A new notification suddenly popped up on the terminal screen.