同样还是Flash内容改变的问题。
在采取措施保证VDD不在没有彻底放电前就又加电后，Flash（程序）改变的概率降低了，但仍然存在。
终于找到TI的一个errata：http://focus.tij.co.jp/jp/lit/er/spmz506d/spmz506d.pdf描述了这个故障。虽然不是我使用的型号，但估计有着同样的问题：VDD和VDDC得有个加电的顺序。
Errata描述如下：
Flash corruption or device failure may occur at power on
Description:
There is a small risk of flash corruption or device failure on power up.
The issue can occur with
certain VDD and VDDC power sequences. The failure is not in the flash
memory itself but in the control
logic to the flash.
Workaround:
To eliminate the risk of flash corruption, two power-on requirements must
be met:
■ The ramp of both VDD and VDDC must begin below 0.2 V.
■ VDDC must reach at least 1.0 V before VDD rises above 1.5 V.
这个描述得比较清楚了。
回头再看LM3S2948的datasheet的第171页：Note: On the printed circuit board, use the
LDO output as the source of VDD25 input. Do not use an external regulator
to supply the voltage to
VDD25。完全就是TI（Luminary）在扇自己的耳光！让用户这样做，对于一个好的芯片是正常的，但对于LM3S，纯粹就是在把用户往火坑里面推啊！
这个LM3S芯片真是错误百出！ |
楼主
标题置顶
标题高亮
