Path: ...!eternal-september.org!feeder3.eternal-september.org!news.eternal-september.org!.POSTED!not-for-mail
From: Cursitor Doom <cd999666@notformail.com>
Newsgroups: sci.electronics.design
Subject: Re: Distorted Sine Wave
Date: Sat, 1 Jun 2024 21:46:10 -0000 (UTC)
Organization: A noiseless patient Spider
Lines: 169
Message-ID: <v3g4r2$2u0a6$2@dont-email.me>
References: <48bd78e1-7da8-3bba-2879-d22962203fa3@electrooptical.net>
	<9olh5j9al34fhrebr4grqq8h6c8javjpp1@4ax.com>
	<seth5jlmgu2gv6lr61m31jk2q94073rvtk@4ax.com>
	<1n0i5jh257hiinlj2dhaatlo11s33m5n0e@4ax.com>
	<9k2i5jpfhu3ncfpm28ukusrok4hugal80s@4ax.com>
	<c6rj5j1l1gfoskul3nnvudf3nc57017k84@4ax.com> <v3d1eb$2anif$1@dont-email.me>
	<go6k5jd7754174pb4k7ko1suh89acqervm@4ax.com> <v3dia1$2anif$4@dont-email.me>
	<rcom5jpoi28rc5hucm2p2dcsiu6es8mt6m@4ax.com> <v3fsbp$2u0a6$1@dont-email.me>
	<0v3n5jh1mrclpfh9jru7k4n54v9ovs03q5@4ax.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Injection-Date: Sat, 01 Jun 2024 23:46:11 +0200 (CEST)
Injection-Info: dont-email.me; posting-host="722a1db93b34a23f770c1bceecb5bb8e";
	logging-data="3080518"; mail-complaints-to="abuse@eternal-september.org";	posting-account="U2FsdGVkX18nvsdcgxILMxHHa9cxuIQxE26ufrouCa0="
User-Agent: Pan/0.149 (Bellevue; 4c157ba)
Cancel-Lock: sha1:nuyPbc48ejuejAAeuBKXczJfybs=
Bytes: 8731

On Sat, 01 Jun 2024 17:18:03 -0400, Joe Gwinn wrote:

> On Sat, 1 Jun 2024 19:21:29 -0000 (UTC), Cursitor Doom
> <cd999666@notformail.com> wrote:
> 
>>On Sat, 01 Jun 2024 14:28:35 -0400, Joe Gwinn wrote:
>>
>>> On Fri, 31 May 2024 22:17:37 -0000 (UTC), Cursitor Doom
>>> <cd999666@notformail.com> wrote:
>>> 
>>>>On Fri, 31 May 2024 14:48:28 -0400, Joe Gwinn wrote:
>>>>
>>>>> On Fri, 31 May 2024 17:29:47 -0000 (UTC), Cursitor Doom
>>>>> <cd999666@notformail.com> wrote:
>>>>> 
>>>>>>On Fri, 31 May 2024 11:34:46 -0400, Joe Gwinn wrote:
>>>>>>
>>>>>>> On Thu, 30 May 2024 16:24:25 -0700, john larkin <jl@650pot.com>
>>>>>>> wrote:
>>>>>>> 
>>>>>>>>On Thu, 30 May 2024 18:53:29 -0400, Joe Gwinn
>>>>>>>><joegwinn@comcast.net>
>>>>>>>>wrote:
>>>>>>>>
>>>>>>>>>On Thu, 30 May 2024 14:56:30 -0700, john larkin <jl@650pot.com>
>>>>>>>>>wrote:
>>>>>>>>>
>>>>>>>>>>On Thu, 30 May 2024 16:06:47 -0400, Joe Gwinn
>>>>>>>>>><joegwinn@comcast.net>
>>>>>>>>>>wrote:
>>>>>>>>>>
>>>>>>>>>>>On Thu, 30 May 2024 15:02:44 -0400, Phil Hobbs
>>>>>>>>>>><pcdhSpamMeSenseless@electrooptical.net> wrote:
>>>>>>>>>>>
>>>>>>>>>>>>On 2024-05-29 18:59, Joe Gwinn wrote:
>>>>>>>>>>>>> On Wed, 29 May 2024 22:11:47 -0000 (UTC), Phil Hobbs
>>>>>>>>>>>>> <pcdhSpamMeSenseless@electrooptical.net> wrote:
>>>>>>>>>>>>> 
>>>>>>>>>>>>>> Cursitor Doom <cd999666@notformail.com> wrote:
>>>>>>>>>>>>>>> On Wed, 29 May 2024 13:42:13 -0700, john larkin wrote:
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>> On Wed, 29 May 2024 21:43:54 +0200, Arie de Muijnck
>>>>>>>>>>>>>>>> <noreply@ademu.com>
>>>>>>>>>>>>>>>> wrote:
>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>> On 2024-05-29 19:07, Cursitor Doom wrote:
>>>>>>>>>>>>>>>>>> Gentlemen,
>>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>> Whilst fault-finding on my HP 8566B spectrum analyzer,
>>>>>>>>>>>>>>>>>> I've found the 10Mhz reference oscillator is generating
>>>>>>>>>>>>>>>>>> an 'unsatisfactory waveform'
>>>>>>>>>>>>>>>>>> which may be causing the device to be unable to lock
>>>>>>>>>>>>>>>>>> it's main PLL. I've come across this waveshape before,
>>>>>>>>>>>>>>>>>> but mostly with oscillators I was building and in the
>>>>>>>>>>>>>>>>>> process of trying to iron out the wrinkles of and
>>>>>>>>>>>>>>>>>> certainly NOT a critical reference oscillator from a
>>>>>>>>>>>>>>>>>> respected manufacturer. Can anyone tell what's most
>>>>>>>>>>>>>>>>>> likely going on here?
>>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>> https://disk.yandex.com/i/z6fYbeVfPRK7aA
>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>> Looks like reflections in the cable. Try the 50 Ohm
>>>>>>>>>>>>>>>>> termination.
>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>> Arie
>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>> If the drive is a sine wave, a cable can't generate that
>>>>>>>>>>>>>>>> 2nd harmonic.
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> I don't understand how a reflection can account for it
>>>>>>>>>>>>>>> either.
>>>>>>>>>>>>>>> THe cable's only 4' long! However, with the 50 ohm input
>>>>>>>>>>>>>>> enabled, the 2nd harmonic disappears. It's just one of
>>>>>>>>>>>>>>> those inexplicable mysteries that no one knows the answer
>>>>>>>>>>>>>>> to. :)
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> That?s pretty diagnostic. There must be an LC filter on the
>>>>>>>>>>>>>> output?mis-terminating it will cause all sorts of
>>>>>>>>>>>>>> frequency-response whoopdedoos.
>>>>>>>>>>>>> 
>>>>>>>>>>>>> It also occurs to me that if there is a diode in series with
>>>>>>>>>>>>> a resistor somewhere, the impedance presented to the feed
>>>>>>>>>>>>> coax may be 50 ohms for positive input voltage, and say 10
>>>>>>>>>>>>> Kohm for negative.  At the very least one could get an
>>>>>>>>>>>>> inverted reflection on negative.
>>>>>>>>>>>>> 
>>>>>>>>>>>>> Joe Gwinn
>>>>>>>>>>>>> 
>>>>>>>>>>>>Yeah, or an emitter follower.  Good point.
>>>>>>>>>>>
>>>>>>>>>>>Plus some LC filter wiggles, to distort and smooth things.
>>>>>>>>>>>
>>>>>>>>>>>I dug around and found a copy of the HP 8566B spectrum analyzer
>>>>>>>>>>>service manual.  The 10 MHz ref input is an amplifier driving a
>>>>>>>>>>>mixer,
>>>>>>>>>>>with not hint of for instance a TTL input.  So, the problem
>>>>>>>>>>>must be elsewhere.  Or, it's just busted.
>>>>>>>>>>>
>>>>>>>>>>>.<chrome-extension://efaidnbmnnnibpcajpcglclefindmkaj/https://
>>>>>>xdevs.com/doc/HP_Agilent_Keysight/
>>>>>>HP%208566B%20Troubleshooting%20&%20Repair%20Vol.%201.pdf>
>>>>>>>>>>>
>>>>>>>>>>>Joe Gwinn
>>>>>>>>>>
>>>>>>>>>>334 pages! Where is the issue?
>>>>>>>>>
>>>>>>>>>The pagination was unclear, so I didn't see a way to say.  How I
>>>>>>>>>found it was to look at the schematics.  Look for module A22.
>>>>>>>>>Don't think search works on such images, but visual search didn't
>>>>>>>>>take that long.
>>>>>>>>>
>>>>>>>>>Joe Gwinn
>>>>>>>>
>>>>>>>>Too much work for free consulting.
>>>>>>> 
>>>>>>> I took another look.  There is a page reference on the right edge,
>>>>>>> near the bottom, which is not marked as a page X of Y, but is.
>>>>>>> 
>>>>>>> Anyway, look at pages 59 and 77.  At 77, look for A22 in the lower
>>>>>>> left region.  Just above, look for INT and EXT jacks.  The EXT
>>>>>>> jack is where an external reference enters.
>>>>>>> 
>>>>>>> Joe Gwinn
>>>>>>
>>>>>>I don't believe the 10Mhz ref osc is the problem, John. The 2nd
>>>>>>harmonic distortion goes away when the scope input impedance is set
>>>>>>to 50 ohms. There is some slight distortion on it, but not enough to
>>>>>>cause an out-of- lock error. Furthermore, since that osc is the
>>>>>>pace-setter for every other module in every other loop in this
>>>>>>analyzer, its failure would give rise to way more error messages
>>>>>>than a mere "YTO unlock" as it stands at present. The manual
>>>>>>suggests the most likely areas where the fault is located are in one
>>>>>>of the boards A19, A20, A21 or A11.  If A22 were the culprit, there
>>>>>>would be over a dozen error messages.
>>>>> 
>>>>> What is the 10 MHz signal power level at the EXT input in these two
>>>>> castes, 1 Mohm and 50 Ohm?  The expected range is 0 to +10 dBm.
>>>>> 
>>>>> Joe Gwinn
>>>>
>>>>7.68dBm on 50 Ohms 2.68V P-P on 1 Meg
>>>>
>>>>No issues there AFAIC.
>>> 
>>> Not so fast there.  It's the scope that is being set to 1 Meg or 50
>>> ohm,
>>> and at 1 Meg we are observing the drive into the 50 ohm input
>>> impedance of the EXT ref input of the spectrum analyzer.
>>
>>Unless I misunderstand what you're saying, that's not the case, though.
>>What we are observing is the output of the analyzer's 10Mhz reference
>>oscillator taken from a BNC socket on the rear of it which HP have
>>thoughtfully provided and fed directly into a scope switchable between
>>1M and 50 Ohms.
>>
>>> That 2.7 Vpp over 50 ohms is about +13 dBm, which exceeds +10 dBm.  So
>>> the EXT input may be over-driven. Install a 5-dB inline attenuator and
>>> see what happens.
>>
>>It's not an external input. They've put it there in order to make it
>>easy to check the frequency of the reference oscillator and adjust it to
>>precisely 10.000000Mhz if necessary (after a *minimum* 72 hour warm-up!)
> 
> Maybe.  But it's simpler to just try the attenuator.
> 
> Joe Gwinn

If I can find a reverse attenuator, I'd be happy to try it.