Creator 2.0中的PSoC 5需要用于USB的XTAL时钟

我想这里有一点小小的帮助。
问题的根源在于IMO对于PSoC 5是不准确的。因此，当从IMO计时时，USB将失败认证。它可能“工作”，但不会顺从。我们在2版本中添加了对该siutOutlook的设计检查。
简单的解决方案是在设计中添加一个晶体，并设置XTAL作为USB时钟。设计检查指示你这样做。但是，我们忘记了另一个可行的解决方案：使用外部时钟信号。应该把一个好的时钟放在GPIO上，并用它作为USB时钟。该设计是附加的，显示如何命名信号并在时钟编辑器中使用它。可悲的是，设计检查，我们错误地禁止使用XTAL作为IMO源，并使用该USB。我们会在下一个版本中解决这个问题。
但是，还有更多。在最后一篇文章中，帕夫洛芬巧妙地展示了一种绕过虚假错误的方法。代替使用IMO作为USB时钟源，他通过PLL运行IMO，然后运行到USB。这闪避了错误条件，让你在没有水晶的情况下运行USB（虽然你仍然无法运行内部的IMO时钟）。干得好，帕夫洛文，我们没想到那个！当然，这对你在PLL上运行有点限制，但在你的设计中可能不是问题。
--马克。
时钟编辑
1.6兆字节



     以下为原文
  I think that a  little sum-up would help here...
    The root of the problem is that the IMO is not accurate on PSoC 5 as we would like. As a result USB, when clocked from the IMO, would fail certification. It might "work" but would not be compliant. We added a design check for that siutation in the 2.0 release.
    The easy fix is to add a crystal to the design and set up XTAL as the USB clock. The design check directs you to do just that. But, we forgot another viable solution; using an external clock signal. It should be possible to put a good clock onto a GPIO and use that as the USB clock. The design is attached, showing how to name the signal and use it in the Clock editor. Sadly, the design check we put in erroneously prohibits using XTAL as the IMO source and using that for USB. We'll fix that in the next release.
    But, there's more. In the last post pavloven has cunningly shown a way around the bogus error. Instead of using IMO as the USB clock source, he runs the IMO through the PLL and then to USB. This dodges the error condition and lets you run USB without a crystal (you still cannot run it off the internal IMO clock though). Nice job, pavloven, we'd not thought of that one! Of course, this does put a bit of a restriction on what you run the PLL at but that may not be a problem in your designs.
    -- Mark.

•   Clock Editor.bmp 1.6 MB
  
  

帕维伦伟大的工作！



     以下为原文
  Pavleon Great job !

要回答我自己关于电容器的问题：这取决于：一个需要查找数据表的晶体使用，然后计算电容器。当使用公式Cult=（（C1*C2）/（C1+C2））/CSTRAY时，假设Cstray是10PF（PCB的5PF，PSOC5输入的5PF，根据数据表），12PF和22PF值不是那么远。
我昨天在050号上焊接了我的水晶。我使用的是一个负载电容20pF，我焊接22pF电容器（上面的公式给出20pF）。焊接0603个部分是一个真正的痛苦，我需要多个尝试来得到它的权利：
我的测试用例是把晶体频率除以256，并用这个时钟驱动一个16位PWM块。这给出了一个不错的1.4秒周期，并且从一开始就开始工作。
之后，我尝试了USPFS HID示例项目，它自然不起作用…最后的原因是代码指定了一个3.3伏的电源，但是这个项目在某种程度上被设置为5伏特（以及电路板）。在改变代码设置一个5伏电源，一切都很好。
由于我使用的是12MHz晶体而不是推荐的24MHz的晶体，所以我从上面使用PavoLvin技巧：使用晶体作为PLL的源，并从中产生所需的48 MHz频率。然后可以使用PLL来驱动USB时钟。唯一的缺点是，PLL不能用于创建另一个频率，这在某些情况下可能是需要的。



     以下为原文
  To answer my own question about the capacitors: it depends :) One needs to look up the data sheet for the crystal to be used, and then calculate the capacitors from that. When using the formula Cload=((C1*C2)/(C1+C2))/Cstray, and assume that Cstray is 10pF (5pF for the PCB, and 5pF for the PSoC5 inputs, according to the data sheet), the 12pF and 22pF values are not that far off.
    I soldered my crystals on the -050 yesterday. The one I used was specified with a load capacitance of 20pF, and I soldered 22pF capacitors (the formula above gives 20pF). Soldering the 0603 parts was a real pain, and I needed multiple tries to get it right :(
    My testcase was to divide the crystal frequency by 256, and using this clock to drive a 16bit PWM block. This gives a nice 1.4 second period, and worked right from the start.
    After that, I tried the USBFS-HID example project, which naturally didn't work... The reason in the end was that the code specifies a 3.3 volt supply, but the project was somehow set to 5 volts (and the board too). After changing the code to set a 5 volt supply, everything worked fine.
    Since I was using a 12MHz crystal instead of the recommended 24MHz one, I did use pavlovens trick from above: use the crystal as source for the PLL, and create the needed 48MHz frequency from that. Then one can use the PLL to drive the USB clock. Hte only disadvantage of this is that the PLL then cannot be used to create another frequency, which might be needed in some cases.