8753C输出轨迹有噪音

你是否将源连接到R通道进行这些测试？
你在使用测试装置吗？
必须将源连接到R通道才能使锁相环工作。



     以下为原文

  are you connecting the source to the R channel for these tests?  Are you using a test set?  You must connect the source to the R channel for the phase lock loop to work.

非常感谢你的答复，。
仪器连接正确，ALMOST工作正常。
我所说的是源本身（在我展示的例子中）似乎比我预期的具有更多的相位噪声。
因此，仪器未通过基线噪声测试（设置为3 GHz的CW，获取线路的统计数据）。
当带宽设置为3KHz时，A / R走线的变化接近0.02 dB（相对于规格0.006 dB）仅R输入具有变化的噪声电平，但偶尔会超过1 dB。
。
当我将接收器带宽设置为窄时，它非常安静。
A / R标准偏差可能下降到0.005 dB。显然，这可能是接收器或源问题。
所以我查看了频谱分析仪上的信号源，看到了我在附图中看到的内容......它有很宽的带宽和幅度变化。
我认为这是不正确的？
（即，来源应该更纯净）。
假设是这种情况并且它没有正常运行，我还能在哪里看？
所以你问的具体答案是肯定的，源是连接到R输入，而不是运行开环。
我在线路上有一个分路器，因此频谱分析仪可以查看信号源。
我相信我会将我所看到的几乎锁相的特征描述出来，但会出现意想不到的高相位噪声。
再次感谢您的回复。编辑：michaelh310于2015年5月4日下午10:42



     以下为原文

  Thank you so much for reply,.

The instrument is connected properly and ALMOST works perfectly.   What I'm saying is the source itself (in the example I showed) seems to have more phase noise than I would have expected.   As a consequence, the instrument fails the baseline noise test (set to CW at 3 GHz, take stats of the line).   When the bandwidth is set to 3KHz, the variation in the A/R trace is approaching 0.02 dB (vs spec 0.006 dB)   The R input alone has a varying noise level, but occasionally exceeds 1 dB.  .   When I set the receiver bandwidth to narrow, it is very quiet.  A/R std deviation drops to perhaps 0.005 dB

Obviously this could be either a receiver or a source problem.   So I looked at the source on a spectrum analyzer and saw what I showed in the attached photo....it had a wide bandwidth and amplitude variation.

I assume this is incorrect?  (ie, the source should be purer).   Assuming this is the case and it's not operating normally, where else might I look??

So the specific answer to you question is yes, the source was connected to the R input, not running open loop.  I had a splitter in the line so that the spectrum analyzer could look at the source.  I believe I would characterize what I saw as almost phase locked, but with unexpected high phase noise.

Thanks again for your reply.

Edited by: michaelh310 on May 4, 2015 10:42 PM

鉴于您从源到R通道正确连接，您看到的情况可能与30-60 MHz合成器的某些降级有关，该合成器为脉冲发生器供电，脉冲发生器又驱动采样器。
来自IF的信号进入锁相环。
合成器的相位噪声或锁相参考的降级可能导致这种情况。
由于随着频率的增加性能变差，它更多地指向合成器中的问题而不是其他地方，因为它被采样器谐波数随频率上升而多变。
如果在Yig或Cavity振荡器中出现问题，则问题更可能在频率上保持一致。
查看从合成板到脉冲发生器的信号。



     以下为原文

  Given you have it connected properly from the source to the R channel, the condition you see is likely related to some degradation of the 30-60 MHz sythesizer that feeds the pulse generator which in turn drives the samplers.  The signal from the IF goes to the phase lock loop.  Degradation of the phase noise of the synthesizer or the phase lock reference could cause this.  Since the peformance is worse as the frequency increases it points more to a problem in the synth than other places as it gets mulitplied by the sampler harmonic number which goes up with frequency. If it was some problem in the Yig or Cavity oscillator, the problem would more likely be consistent across frequecncy.  Look to the signal from they synth board to the pulse generator.

太好了。
我会深入研究它。
谢谢你的指导。
该仪器可以按原样用于我的目的，但我希望它完全正确！
到目前为止，我一直在使用8754，这是一种能力的复兴。
我会及时向大家发布。



     以下为原文

  OK great.  I will dig in to it.  Thank you for the guidance.

The instrument is usable for my purposes as-is, but I would love to make it exactly right!   I have been using an 8754 up until now, and this is a revalation in capability.

I will keep you posted.

嗨，Joel博士......所以我一直在研究这个问题。
我一直在研究腔体振荡器的输出。
我希望看到3.8 GHz +/- 10 MHz的CW信号。
我所看到的是一个大约每秒一次在3.801 GHz和3.796 GHz之间反弹的信号。
当我第一次开机时，几秒钟稳定在3.796，然后在3.801稳定几秒钟，然后开始弹跳。
换句话说，它移动约5 MHz当我按下预设时，它在3.801稳定几秒钟然后开始弹跳。
这些测量值来自腔体振荡器的输出，并且A3模块的输出断开，因此它们是“开环”这是正常的吗？
或者这表明腔振荡器存在一些问题？
功率输出电平正常，约+10 dBm。
我看到一根电缆，四根线连接到腔体振荡器。
我认为两个是权力和地面，但另外两个是什么？
正如我之前提到的，该单元的性能几乎+正确。
我没有看到源中相位噪声的规格，但我认为我的规格有点偏差。
它是锁相的，但并不像我想象的那样干净。
正如您在我的第一篇文章中的频谱分析仪屏幕截图中所看到的，边带约为30 dB。
也许这是正常的，我正在追逐一个没问题？
由于频率的弹跳是如此可预测，我猜它是设计的一部分，但由于我没有在相关帖子中看到任何提及它，我想我最好问一下？
可能与刺激避免过程有关？
我会提到在CW和扫描模式下都会发生弹跳。
如果它是刺激避免的一部分，我可能只是在扫描模式中预期它？
再次，非常感谢您的帮助。
编辑：michaelh310于2015年5月16日下午4:05编辑：michaelh310于2015年5月16日下午4:09



     以下为原文

  Hi, Dr. Joel...

So I have been continuing to look into the issue.

I have been looking at the output of the cavity oscillator.   I expected to see a CW signal at 3.8 GHz +/- 10 MHz.   What I see instead is a signal that bounces between 3.801 GHz and 3.796 GHz about once a second.   When I first power on the unit, for a few seconds it's stable at 3.796, then stable for a few seconds at 3.801, then starts bouncing.  In other words, it shifts by about 5 MHz

When I press preset, it's stable at 3.801  for a few seconds and then starts bouncing.

These measurements are taken from the output of the cavity oscillator, and with the output of the A3 module disconnected, so they are "open loop"

Is this normal?  Or is this indicative of some problem with the cavity oscillator?

Power output level is normal, about +10 dBm.  

I see a cable with four wire connected to the cavity oscillator.   I presume two are power and ground, but what are the other two?

As I mentioned earlier, the performance of the unit is +almost+ correct.   I haven't seen a spec for the phase noise in the source, but I presume I'm out of spec a bit.  It's phase locked, but not a clean as I would have guessed.   As you can see in the spectrum analyzer screenshot in my first post, the sidebands are about 30 dB down.  Perhaps this is normal and I'm chasing a non-problem?

Since the bouncing of the frequency is so predictable, I'm guessing it's part of the design, but since I haven't seen any mention of it in related posts, I thought I'd better ask? Possibly related to the spur avoidance process?

I will mention that the bouncing occurs both in CW and swept mode.  If it were part of the spur avoidance, I might have expected it only in swept mode??


Again, thanks so much for help.

Edited by: michaelh310 on May 16, 2015 4:05 PM

Edited by: michaelh310 on May 16, 2015 4:09 PM

自从撰写上述文章以来，我更多地研究了系统，并相信我理解发生了什么。
在我看来，这是设计的正常部分，而不是问题的指示。
我认为5 MHz移位确实是支线避免系统的一部分。
它仍然在CW模式下看到，但每次扫描只能看一次。
它在扫描期间在较高频率下稳定。
额外的两根线可能用于频移变容二极管。
我对么？
正常？
如果是这样，真正的问题是源是否在规范内运行，或者我是否还需要寻找问题。
谢谢



     以下为原文

  Since writing the above post, I have looked into the system more and believe I understand what's happening.  Looks to me like this is a normal part of the design, not an indication of problems.

I think the 5 MHz shift is indeed part of the spur avoidance system.   It is still seen in CW mode but only once per sweep.  It is stable at the higher frequency during the sweep.   The extra two wires are probably for the frequency shift varactor.

Am I correct?  Normal?   

And if so, the real question is whether the source is operating within spec or if I still need to be hunting a problem.

Thanks

是的，你是对的。
5 MHz偏移发生在混频器的高阶乘积与接收器的频率交叉并引起杂散迹线的频率处。
就在此之前，我们将LO和Cavity 5 MHz单向捕捉跳过rx频率的杂散，然后当测量超过rx频率时将它们跳回。
我在1985年7月3日发现了这个问题，花了大约9个月的时间围绕那个特定的问题进行设计（毁掉了我的老板7月4日）。
如果这些信号中的任何一个的驱动器有噪声，则信号源会产生噪声，但由于频率为1：1，因此噪声不应随RF频率而变化。
然而，由于本地振荡器为30-60 MHz，预计噪声会上升（相位噪声），因此我们将乘以100倍以达到3 GHz。
对于每2x乘法，相位噪声跳跃6 dB。
因此，有大约7次倍频达到3 GHz，因此相位噪声相对于30 MHz应该上升42 dB。



     以下为原文

  Yup, you're right.  The 5 MHz offset occurs at frequencies where the higher-order products of the mixer cross the frequencies of the receivers and cause spurious traces.  Just before that happens, we snap the LO and Cavity 5 MHz one way to jump the spur across the rx frequencies, and then jump them back when the measurement is past the rx frequencies when they cross. I found that issue on July 3, 1985 and it took about 9 months to design around that particular issue (ruined my boss's 4th of July).

If the drivers to either of those signals is noisy, the source will be noisy, but since it is an 1:1 with frequency, the noise should not change with RF frequency.  

However, the noise is expected to go up (phase noise) as the local oscillator is 30-60 MHz, so we multiply up by 100 times to get to 3 GHz.  For each 2x multiply, the phase noise jumps by 6 dB. So there are about 7 doublings to get to 3 GHz so the phase noise should go up by 42 dB relative to 30 MHz.

嗨，乔尔博士和其他人......抱歉是痛苦，但我的痛苦正在增长上面讨论的8753C突然变得更糟。
我已经接受了源中的高相位噪声，因为它没有干扰我的需求，但现在它根本不会锁定。
有了外部资源，一切都很好。
Cavity osc的输出看起来很好，正好在3800 MHz，大约+10 dBm，良好的纯度，偶然的偏差-5 MHz，如预期的那样。
但我的第一个症状是无法在6 GHz范围内获得锁相，现在无法锁定3 GHz范围而测试48失败。
源调谐测试显示频率超出范围（如预期的100 MHz以上）。
来自Source的输出级别看起来不错，但我猜测YTO或其他组件已老化。
我记得你说过Cavity必须在10 MHz以内的频率，所以我猜测腔体频率和YTO调谐范围之间存在特定的关系。
可能是YTO超出了测试48可以处理的范围吗？
我所说的是，我的印象是实际上没有任何东西被破坏，但不能再正常工作。
我看到YTO的小电路板上有一个用硅胶密封的电位器，用于固定YTO ...有可能需要调整吗？如果有的话，程序是什么？
（不是源组件顶部的可访问的锅）我是否有可能尝试将腔设置为3800以外的某个频率？
或者是否有人可以修复源程序集？
或者我应该放弃并找到另一个8753？
我真的，真的很想让它重新投入使用。
我同时使用我的旧款8754A，它不是那么有趣，也不是我需要的所有东西。
再次感谢并提前获得任何帮助



     以下为原文

  Hi again, Dr. Joel and others...sorry to be a pain, but my pain is growing

The 8753C discussed above has gotten suddenly worse.    I had accepted the high phase noise in the source as not interfering with my needs, but now it won't lock at all.

With an external source, all is fine.

The output of the Cavity osc looks good, right at 3800 MHz, about +10 dBm, good purity, occasional deviations -5 MHz as expected.

But my first symptom was being unable to get phase lock in the 6 GHz range, now can't get lock in the 3 GHz range and test 48 fails.  Source tune test shows frequencies way out of range (like 100 MHz above expected).   Output levels from the Source look OK, but I'm guessing that the YTO or other components have aged.   I recall you saying that the Cavity must be on frequency within 10 MHz, so I'm guessing that there is a specific relationship between the Cavity frequency and the YTO tuning range.  Could it be the YTO has drifted beyond the range that test 48 can handle?   What I'm saying is that I have the impression that nothing is actually broken, but not longer working correctly.

I see the YTO has a potentiometer sealed with silicone on the little board that holds the YTO...is there a possibility that this needs adjustment and if so what is the procedure? (not the accessible pot on top of the source assembly)  Is there a possibility that I can try setting the cavity to some frequency other than 3800?    or is there someone who can fix the source assembly?  or should I give up and find another 8753?

I'd really, really love to get this back in operation.   I'm using my old 8754A in the meanwhile and it isn't nearly as much fun nor does it do all the things I need.

Thanks again and in advance for any help

进一步调查显示更多。
我将频谱分析仪直接连接到YIG的输出。
在Source Tune模式下设置为1.5 GHz，输出看起来很恶心。
频率非常粗略地在区域（1.5 + 3.8 GHz）但不完全正确。
此外，电平很低（约-20 dBm）并且有很多（几MHz）频率抖动/漂移。
它看起来像YTO是坏人，但我将为10针电缆构建一根延长线，将YTO连接到其余的源，以探测YTO引脚上的信号，确保没有一些不良电压。
任何提示/建议？
Arrrrgh ...



     以下为原文

  Further investigation shows more.   I connected the spectrum analyzer right to the output of the YIG directly.   Setting to 1.5 GHz in Source Tune mode, the output looks sick.  Frequency is very roughly in the zone (1.5 + 3.8 GHz) but not exactly right.  Also, level is low (approx -20 dBm) and has lots (several MHz) of frequency jitter/drift.    It's looking like the YTO is the baddie, but I will build an extension cord for the 10 pin cable that connects the YTO to the rest of the source to probe the signals right on the YTO pins, make sure not some bad voltage.

Any hints/suggestions?


Arrrrgh...

听起来很糟糕



     以下为原文

  sounds bad.

是的，我总结说YTO失败了。
我在ebay上买了一个完整的替代品8753C，价格低于修理YTO的报价。
由于一个晶体管发生故障，废弃整个仪器似乎是浪费和羞耻。
我希望有人能以合理的价格进入源维修业务。



     以下为原文

  Yes, I concluded the YTO had failed.   I bought a complete replacement 8753C on ebay for less than the quotes to repair the YTO.  It seems such a waste and a shame to scrap an entire instrument because one transistor has failed.   I wish someone would go into the source repair business at a price that makes sense.