QS-30 Nixie Clock_Mainboard

Continuing from the previous article 

The board was made in two steps. Today, I soldered both boards. After a brief debugging, I found that all the mistakes that should be made in the PCB were made. It was too difficult for me. 1. The pins of the transistor package in the boost circuit are reversed. 2. The mos tube on hand is a little longer than the package in the picture, which affects the circuit behind it. 3. The voltage of the boost circuit has increased, but the mos tube is full of fragrance. No details yet. Debugging 4. I found that the DC sockets I had on hand were not universally packaged and could not be plugged in. 5. The most serious problem was that the pin header interfaces of the upper display board and lower control board, two groups on the left and right, and one group of IOs were reversed for me. too difficult! 6. The through hole of the QS30 socket package is narrow. If it is a new socket, it is easy to solder. For example, when I dismantled the socket, I had a hard time without a solder extractor at hand. I couldn't insert it. I used a soldering iron for a long time to get it in. I will definitely change it later and never use other people’s packages for free.

Updated at 20:25 on August 1, 2020! After that, I wrote the main line for this project and the Display project. Depending on my mood, I took the time to weld two boards today and found some bugs for the boards. ┭┮﹏┭┮ I found that I didn’t add a switch, and then I found the reason for the aroma of the mos tube: two If the proportional resistor is improperly selected and the voltage is increased to 230, it may be overloaded. It is more likely that the 8550 package of the mos tube gate discharge circuit is wrong again. After flying wires and a simple wave of operations, the next step is to wait for procurement and coding. It is worth mentioning that I used a 0.76 OLED for parameter setting when designing. I will not say whether I want to keep it forever, but the program is prepared according to it to compile

Updated on August 8, 2020! !

This week I took the time to weld and air-condition to try out several modules. The ones that have been adjusted so far include:

1. Pulsator button

2. Clock chip (only used the test program, haven’t read the chip manual yet)

3. Buzzer IO, boost circuit control IO, neon bubble IO

4. Basic functions of the display panel (only test display, no display content control)

Not yet tuned:

1.OLED convulsion

2. Infrared IO (obviously the previous program was very useful!)

Below are pictures and videos:

It is worth mentioning that tonight I experienced a capacitor explosion for the first time in my life! Can you believe it? I have never blown up a capacitor after playing for so long. I have only burned mos tubes, transistors, and chips. This time, the tantalum capacitor clearly paid attention to the polarity, but somehow it overturned. I only saw the BOM after plugging it in! The capacitor turned red and then carbonized, turning completely black. I don't know if I connected it wrongly.

After that it's simple coding work, which may take a long time. . .

Updated on August 16, 2020! Today's mental journey is quite complicated. I took out the program in the morning and started copying and pasting. I decided to use the previous VFD program (for the sake of laziness, I just wanted to change the IO and display characters).

Suddenly I found that the OLED had not been adjusted yet, and I was about to die with him. First take out the 32-bit board and use the process to light up one piece. After several ten minutes of debugging, ok! It lights up successfully, then stacks this OLED on the previous OLED, and then there are still 8 lines. Confirm again whether the io is correct and whether the io is alive, and power on!

It turns out there is no problem with the program, 555

Dry off the original OLED and tear off the film, oh! It turned out to be cracked. . . . . .

Then after lunch, I learned that I was going to have a physical examination. It took four hours by car and half an hour for the physical examination. The afternoon was stolen.

I came back in the evening and found that the SD2405 program was not working. It turned out that the clock logic in another project was like this. It is read once when the computer is turned on, and then the whole process is simulated with a timer. When the time is modified, it is written back to the chip. I was wondering, why does it do this?

First of all, the program does not encapsulate the modification year and the like, as well as the alarm clock and the like. I went online to look for wheels, but couldn't find them so I had to make them myself. Looking at the manual, it seemed very simple, but the configuration of almost every register of the domestic chip was given as an example.

Then I discovered while making the wheel that the reading time of this thing seems not to be too fast? Reading three five bytes will make it float, just three alone will do. After finishing it and revising it for a while, I found that reading three times in a row didn’t work anymore. I also suspected that it was a problem with the analog IIC, and after checking for a long time, I found that there was nothing wrong. Alas, I simply used what I wrote before, but the problem remains the same.

So I was ready to give up this chip and use DS3231. There was a slot for 3231 reserved on the board. Because there was only one disassembled SD2405, I reserved the slot for 3231. I didn't expect that it would come in handy.

After changing the brain, it feels more comfortable. The previous program also used 3231, so there are less changes to make, hehe.

After some modifications.

It actually looks better to the naked eye, but the equipment is too weak to achieve the best results.

problem dicovered

1.What topology is used for the mini360 buck module? The power supply from the buck terminal can actually be boosted to the input terminal to power the 34063. Although the combined efficiency of the two boosts must be low.

2. Is SD2405 easy to use? I was too good at it, after all, I only worked on it for a few hours, so I crushed it to pieces in the end.

To be improved

1. The inductor of the boost circuit has serious whistling, and the circuit needs to be optimized ( see experts online ).

2. No switch drawn

3. No test LED drawn

4. The OLED power supply should have a switch, either soft or hard.

5. This pulsator switch is different from ordinary buttons. You may need to change the anti-bounce time. Now it will be accidentally touched.

6. One of the two interconnected cables is drawn backwards.

7. The four positioning holes are too small

8. The enable pin of 74hc595 should not be directly connected to ground. If IO is added, it can be used to control the brightness.

The parts not yet debugged

1. External sensors (photoresistor, temperature, humidity...)

2. Alarm clock function, flash reading and writing, infrared control logic, automatic on and off, automatic dimming

3. Nixie tube protection program (to prevent poisoning)

Through this project, I discovered again that I am not as patient as before when making things. As we age, we have less and less free time and more and more things to do. There used to be a whole chunk of time, but now it's hard. This should be the last project in the short term. After all, I have to study again and the future is uncertain.