The process of research and development is the process of filling the pit, the hardware engineer has too many pits, both feet are not enough to step on, and it is probably not enough to add two hands.
Today to share the pit that was stepped on at the production line, and it is related to the buddy, I helped fill it in by the way.
Business trip to the production line support is the hardware engineer's commonplace, once I traveled on a business trip, solved the problem of my own project team after stealing in the customer office for half a day, suddenly brushed to the production line group, there are brothers in the project team brothers shouted people to borrow the string line, as if the serial mouth has a problem, just I have one in my hand, go, past to see.
After passing, I saw my brothers sitting at the small table in the corner of the workshop with sad faces, and what aroused their joy was not my arrival, but the string of threads in my hand, which seemed to be worth more than people.
After changing the string line in my hand, the problem still did not change, and the brother returned to a sad face. There is a brother of software next to him, with a calm face, I understand what he means: it has nothing to do with software, it is a hardware problem.
I understand the next pass, is the product modem module abnormality caused by the machine can not start, if you want to analyze the root cause of the problem of the modem, you need to analyze the log output through the serial port, but the ship broke the head wind, the partial serial port appeared abnormal, can not spit out the log normally, the first batch of trial production of the board 100% failed, there is no locking problem root cause, the back of thousands of trial production boards can not SMT patch, has been stuck in the production line, every minute of delay is the silver of white flowers.
No wonder the buddies frowned, I asked if there is a hardware problem or a software problem? Software colleagues speak first: it is not a software problem, the same software version is no problem with the machine at home. Looking at his confident face, I thought don't be proud too soon, who laughed and said that the end is not necessarily it.
The main problem at present is to solve the serial port exception.
Dude is a very experienced hardware engineer, I also remember that the line on the serial port line is very simple, as shown in the figure below, the serial port TX line is generally only a hardware switch, and then directly connected to the type C interface, how can such a simple line go wrong?
Dude took everything on the TX line apart, but the UART is still abnormal, I also wondered, cross-verification changed the computer, changed the serial line, the problem did not progress.
Are you sure everything on the line is torn apart? Dude is very familiar with his circuit, he said that everything that can be removed is removed, even the switch is removed, and now the UART is directly connected from the CPU to type c, and I wonder.
It is useless to talk more, look at the picture and speak.
I looked at the oscilloscope UART TX waveform, the waveform is a little close to the following figure, I was shocked: this charge and discharge is a bit like a capacitive load.
The software colleague stared at the oscilloscope next to him, and I exchanged glances with my buddies, all without saying a word.
Let the buddies fly on the motherboard to measure the UART waveform, and the motherboard measurement waveform is abnormal.
I don't believe that the wire will have such a large capacitive load, must find where the capacitor comes from, I said to disassemble! The TX data link is shown in the figure below, and the motherboard to connector is to the FPC to typec interface.
Remove the motherboard FPC waveform ok, UART log normal output, buddy eyes lit up, I said continue to disassemble!
After assembling, remove the small board connector, the waveform is OK, and the problem is locked on the small board!
Why is there a problem with the small board? This is a direct connection!
The brother opened the drawing again, after 120,000 careful inspections, in a hidden corner, saw a small capacitor in parallel on the UART TX, and the brother exploded!
depend on! Who's the capacitor in parallel with me here!
(Haha, I already have a number in my heart, and quietly persuaded the knife in the hands of the brothers)
Tear down this capacitor everything OK, it is this capacitor caused, fortunately I did not have the software at the beginning, otherwise the clown would be myself. At this time, the software colleagues did not know where to go to play.
Why is this capacitor? It turned out that when UART was working, TX would affect EMC, and colleagues in the brother department added a small capacitor here, so how should we finally solve this problem? For this problem, we and EMC colleagues do not have to struggle, because the UART is only used in trial production, not in mass production, and there are switches on the hardware that will disconnect this path, and everyone is happy.
There are many pits in the hardware, and there are more pits on the front line, in addition to a solid foundation, it also needs enough patience, courage and care, and luck. What pit will you step on?
To say more, like this kind of division of labor boundaries are relatively blurred area, conflict is normal, we should have an inclusive mentality to treat, maybe next time it is our turn to dig a pit for them, and EMC colleagues also have a good attitude, even sorry.