During this circuit board test, I found that one circuit board could never be programmed. So I checked:
1. There are no issues with the power supply and ground.
2. Using the oscilloscope to measure whether the crystal oscillator is oscillating, I found a strange issue: a 24MHz sine wave appeared at the XOUT pin, but nothing at XIN? What could be the reason? I couldn’t find out.
So I had to replace the main chip. I used to struggle with QFP128, but now it’s quite easy to disassemble and assemble, thanks to the guidance of my colleague Xiao Wang (a little proud, you know I used to be afraid to solder even the most ordinary SMD components).
However, after replacing the chip, although I could program it, I measured the crystal oscillator again and still found nothing at XIN, but XOUT was present. It can be confirmed that the chip is working normally, so why can’t I measure the oscillation of the crystal oscillator? The circuit is correct, so I can only check if there is an issue with the oscilloscope.
I performed a self-calibration on the oscilloscope probe, and the 5V 1KHz square wave was normal. So where is the problem? Is it impossible to measure the crystal oscillator? That can’t be, I’ve measured it before.
I noticed that I was using the X1 setting on the probe, so I tried switching to the X10 setting and suddenly found the oscillation waveform. It’s strange why there was nothing at XIN with X1, but it appeared with X10?
Looking at the probe, I found on Tektronix’s website that it is related to the capacitance of the probe.
The characteristic parameters of the oscilloscope probe include:
Type
Cable Length
Attenuation
Bandwidth at -3 dB
System Input Resistance
Typical Input C
Max Voltage
Compensation Range
Read Out
ID/Gnd Ref.
Tip/Head Style
Corresponding:
P2200x10X/1X200MHz/6MHz 10MO/1MO16pF/95pF 300V/150V
Here, the Typical Input C is very important. At 1X, it is 95pF, and this capacitance affects the oscillation of the crystal oscillator. The matching capacitance for the crystal oscillator is 30pF, so it is normal that the waveform cannot be measured at XIN, and it does not affect the chip’s operation. If you want to measure whether it is oscillating, use a probe with a smaller capacitance, such as selecting the 10X setting.
Similarly, from the above analysis, it can be seen that, strictly speaking, the waveforms measured by the oscilloscope are distorted and not the actual values. Of course, for the crystal oscillator, regardless of whether capacitance is added and how much capacitance is added, it affects the shape quality of the waveform but does not affect the frequency. Therefore, the system will operate normally.
A simple issue, but still needs to be treated with caution!
