波长记录不够准确

您好PeeJay，这听起来像驱动程序错误，然后建议增加步长，对吧？
这通常是由扫描返回的波长值阵列引起的，这些波长值不足以使插值到等距点。
确认这是一种简单的方法是禁用此插值。
这是通过以下命令：* hp816x_returnEquidistantData（ViSession ihandle，ViBoolean equalSpacedDatapoints）; *测量可能会在没有错误的情况下完成，但这些点不会完全均匀分布。
原因可能是使用非常小的步骤，如0.1pm，并且增加步骤可以避免问题。
这取决于激光模型。
对于某些激光模型，也可能存在模式跳跃，特别是在扫描超过50nm或使用低功率时。
而激光本身可能会有问题。
可调谐激光器的型号是什么？
我的猜测是，这与探测器没有直接关系，但是当插入这些数据时会发生错误。
问候，迈克



     以下为原文

  Hello PeeJay,

This sounds like the driver error that then suggests increasing the step size, right?
This is generally caused by the sweep returning an array of wavelength values that are not linear enough for the interpolation to equidistant points. An easy way to confirm that is to disable this interpolation. That is with this command:
*hp816x_returnEquidistantData(ViSession ihandle, ViBoolean equallySpacedDatapoints);*
It's likely that the measurement will then complete without error but the points won't be exactly evenly spaced.

The cause could be using very small steps, like 0.1pm, and increasing the steps may avoid the problem. This depends on the laser model.
Also with some laser models, there may be mode hops, especially if sweeping over more than 50nm or using low power.
And the laser itself could have a problem. What is the model of the tunable laser?

My guess is that this is not directly related to the detectors but the error occurs when these data are being interpolated.

Regards,
Mike

你好迈克，谢谢你的启示。
我目前正在开发用于CWDM和DWDM盒模块的被动测试的常规软件。
在DWDM测试期间，扫描在1520-1570nm，0.005nm处进行。
这个似乎没有问题。
对于CWDM，我试图在0.2nm扫描从1511到1639。
在我的参考例程中，一切似乎都没问题，除非我碰到探测器3.顺便回答你的问题：我使用的激光模型是Agilent 81640A。
是的，错误表明我应该增加步长。
你的假设可能是正确的。
我用于测试目的的TLS似乎有随机出现的“激光重新对齐”问题。
我将尝试在今天的测试中实现禁用等间距数据点命令。
我想澄清一下：假设该功能被禁用，那么我可以预期我的结果数组可能不相等，对吧？
我的意思是，如果我的波长为1024个阵列大小，则等效功率阵列的大小可能不同，对吗？
最好的问候，PJ



     以下为原文

  Hello Mike,

Thanks for the enlightenment. I am currently developing a routine software for passive test of both CWDM and DWDM box modules. During DWDM testing, sweeping is done from 1520-1570nm at 0.005nm. There seem to be no issues with this one. For the CWDM, I was trying to sweep from 1511 to 1639 at 0.2nm. During my reference routine, everything seems to be okay except when I hit the detector 3. 

By the way to answer your questions: The laser model I was using is Agilent 81640A. And yes, the error suggest that I should increase the step size. 

Your assumption is possibly right. The TLS that I used for testing purposes seems to have 'laser re-alignment' issues that appears randomly.

I will try to implement disabling the equally spaced data points command in my testing today. I would just like to clarify: suppose that the feature is disabled, then I could expect that my resulting arrays may not be equal, right? I mean if I have 1024 array size for my wavelength the equivalent power array may not be of the same size, correct?

Best regards,
PJ

嗨PJ，是的，你应该得到相同的数组大小。
激光器将为功率计提供相同数量的触发，并返回相同的记录波长值。
不同之处在于驱动程序软件不会插入点以提供均匀的间距。
基本上你会看到原始数据。
由于在扫描范围很广的情况下使用相对较大的0.2nm间距时会看到这一点，因此看起来这个范围背后是问题所在。
激光似乎变得不稳定，可能在扫描结束时。
（您可以从没有插值的原始波长数据中检查出来。）长扫描比没有模式跳跃的50nm范围要大得多，并且还强制激光器以相当低的输出功率运行。
所以你可以通过稍早停止扫描来获得一些稳定性，比如1620或1625nm。
如果您偶尔收到“重新签名”消息，那么如果您之前没有，则可以尝试执行此操作。
这是激光菜单中的一项功能，它运行大约45分钟的广泛校准程序。
如果激光在某种程度上受到干扰（例如在运输过程中），它可以提供帮助。
但如果这不能解决问题，那么经常这样做是没有用的。
如果你运行这个，那么应该首先预热激光，然后在完成之前不要打扰。
否则这些消息可能是衰老的迹象。
81640A是一款相当古老的型号，可能会磨损。
祝你好运！
麦克风



     以下为原文

  Hi PJ,
Yes, you should get the same array size. The laser will give the same number of triggers to the power meter, and return the same set of logged wavelength values. The difference is that the the driver software will not interpolate the points to give even spacing. Essentially you're seeing the raw data then.

Since you're seeing this when using a relatively large 0.2nm spacing when sweeping a very wide range, it seems that this range is behind the problem. The laser seems to be getting unstable, probably towards the end of the sweep. (You can probably check that from the raw wavelength dat that you get without the interpolation.) The long sweep is quite a bit more than the 50nm range specified without mode-hops and also forces the laser to run at rather low output power. So you might gain some stability by stopping the sweep a little earlier, like 1620 or 1625nm.

If you're occasionally getting the "realign" message, then you could try to do this once if you haven't before. That is a function in the laser menu that runs an extensive alignment routine that takes about 45 minutes. It can help if the laser has somehow been disturbed, like during shipping. But if that doesn't solve the problem, then it isn't useful to do it regularly.
If you run this, then the laser should be warmed up first and then not disturbed until it is finished.
And otherwise these messages can be signs of aging. The 81640A is a fairly old model and might be wearing out.

Good luck!
Mike

嗨迈克，谢谢。
似乎我用于此测试的TLS存在一些问题。
就在今天，我尝试重新对齐但是我遇到了一些错误，迫使我取消了这个过程。
无论如何，我用于软件开发的单元已经用了很长时间才能用于扫描应用程序。
我在没有等距数据点函数的情况下附加了样本轨迹，用于从1520-1570以0.005nm步进扫描。
这是用于DWDM测试的。
我只是在1524nm区域观察到一点噪音。
这种波动可能导致波长记录问题。
我还观察到痕迹似乎没有对极化的变化做出反应。
根据我通常的观察，LHP，LVP，L45和RHC（Mueller-Matrix方法）的痕迹有点远。
无论如何，我的主要问题已经给出答案。
非常感谢您的及时回复。
最好的问候，彼得



     以下为原文

  Hi Mike,

Thanks. It seems that the TLS that I am using for this test has some problem. Just today, I tried the re-align but I got stuck with some error forcing me to cancel the process. Anyway, the unit I have been using for SW development has not been used quite some time for sweeping applications. 

I attached a sample trace without the equally-spaced datapoint function for a sweep from 1520-1570 at 0.005nm steps. This is for a DWDM test. I just observed a bit of noise at 1524nm region. This kind of fluctuation maybe causing that wavelength logging issue. I also observed that the trace seems not responding to changes in polarization. Per my usual observation the traces for LHP, LVP, L45, and RHC (Mueller-Matrix method) are a bit far enough. 

Anyway, my main problem is already given answer. Great thanks for your prompt response.

Best regards,
Peter

附件

• 1xNFT_Ref_Data_noESDP.doc342.5 KB