Noise on VCO voltage ramp

[Cursitor Doom]

Gentlemen,

I've got this weird problem with the network analyser I posted about
recently. The problem *seems* to be noise spikes on the DC voltage
control ramp which are causing spurious spikes in the frequency domain
on the RF output when viewed on an RF spectrum analyser. The peculiar
thing is these spikes on the RF output are not random; they come and
go according to the point the control voltage has reached as it ramps
up. Curiously, I cannot see anything amiss with the ramp signal when I
check it with a scope. I'd have expected to see at least *some* sign
of instability, but it appears rock steady. I'm just wondering if an
ordinary oscilloscope is really fast enough to pick out these
transients, though. Maybe they're there even though I can't see them.
Any suggestions as to what other instrument might be better suited to
this purpose? A DSO, maybe? Or if not, optimising the settings on the
analogue scope to have the best chance of spotting them?
thanks!
--

No deal? No problem! :-D

[Phil Hobbs]

Finding an asynchronous glitch is hard on an analog scope.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

[Cursitor Doom]

On Sun, 12 Jan 2020 15:21:39 -0500, Phil Hobbs
<pcdhSpamM...@electrooptical.net> wrote:

>Finding an asynchronous glitch is hard on an analog scope.

It is to a major extent reproduceable, though, Phil. I can manually
override the ramp and set a voltage which will show up a glitch in the
frequency domain, which is a useful plus. You reckon it would be
better to use a DSO for this? I do have one.

[Phil Hobbs]

Frequency domain glitch? You mean a discrete spur?

A glitch is a transient time-domain animal. A DSO can save a single
instance, or (with averaging) ignore the noise and asynchronous
background and just reproduce the glitch itself.

[Jeff Liebermann]

On Sun, 12 Jan 2020 14:36:27 +0000, Cursitor Doom
<cu...@notformail.com> wrote:

>I've got this weird problem with the network analyser I posted about
>recently. The problem *seems* to be noise spikes on the DC voltage
>control ramp which are causing spurious spikes in the frequency domain
>on the RF output when viewed on an RF spectrum analyser.

In other words, it's picking up garbage from something in your shop
(or cave). I suggest you start by moving or removing all the
switching power supplies, wall warts, light dimmers, desk lamps, or
gizmos within a few feet that might have a switching power supply
inside. Next, cover the network analyzer with some aluminum foil to
see if the glitches are being delivered via RF. My desktop wi-fi
access point does that sometimes. If that also fails, drag your
network analyzer and scope to another part of your shop (or cave),
that's away from all the gizmos, and see if the problem persists. In
other words, determine if the garbage is internally generated,
externally radiated, or conducted via the power lines, clip leads,
cables, etc.

If you really want to see the noise, glitch, trash, garbage, or
whatever signal, set you scope so that the horizontal trigger is off
and the horizontal sweep is free running. Then, slooooooowly vary the
horizontal sweep until you see fairly stable glitches. If the noise
is periodic, you should be able to see something.


--
Jeff Liebermann je...@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

[Cursitor Doom]

On Sun, 12 Jan 2020 17:17:04 -0500, Phil Hobbs
<pcdhSpamM...@electrooptical.net> wrote:

>Frequency domain glitch? You mean a discrete spur?
>A glitch is a transient time-domain animal.

I could have put it better I guess. What I mean is I'm seeing thin
spikes in the RF output spectrum of the VNA at other frequencies than
the desired one. I'm guessing these spikes are the result of unwanted
transient voltage spikes present on the DC VCO control voltage which
sweeps the RF output frequency from 4Mhz through to 1.3Ghz.

>A DSO can save a single
>instance, or (with averaging) ignore the noise and asynchronous
>background and just reproduce the glitch itself.

Well, Jeff has posted some advice for me on that aspect which I intend
to try out tomorrow using my Tek DSO which is probably much better
suited to the task than the analogue one I've hithertofore been
using..

[Cursitor Doom]

On Sun, 12 Jan 2020 14:22:35 -0800, Jeff Liebermann <je...@cruzio.com>
wrote:

>In other words, it's picking up garbage from something in your shop
>(or cave). I suggest you start by moving or removing all the
>switching power supplies, wall warts, light dimmers, desk lamps, or
>gizmos within a few feet that might have a switching power supply
>inside. Next, cover the network analyzer with some aluminum foil to
>see if the glitches are being delivered via RF.

Yeah, I did try all that and it didn't help, Jeff. Also there's no
glitches or noise when I pull the RF signal away, so at least it's not
something coming from the spectrum analyser.

>My desktop wi-fi
>access point does that sometimes. If that also fails, drag your
>network analyzer and scope to another part of your shop (or cave),
>that's away from all the gizmos, and see if the problem persists. In
>other words, determine if the garbage is internally generated,
>externally radiated, or conducted via the power lines, clip leads,
>cables, etc.

I *suspect* noise is somehow getting onto the DC VCO ramp voltage and
manifesting as spikes in the RF output of the VNA accordingly.
Fortunately, there is a jack in the back of the VNA for the
application of an external VCO control voltage, so I plan to avail
myself of that with a big old linear bench supply and see if the
problem goes away.

>If you really want to see the noise, glitch, trash, garbage, or
>whatever signal, set you scope so that the horizontal trigger is off
>and the horizontal sweep is free running. Then, slooooooowly vary the
>horizontal sweep until you see fairly stable glitches. If the noise
>is periodic, you should be able to see something.

Thanks, Jeff. Duly noted!

[Phil Hobbs]

On 2020-01-12 19:18, Cursitor Doom wrote:
> On Sun, 12 Jan 2020 17:17:04 -0500, Phil Hobbs
> <pcdhSpamM...@electrooptical.net> wrote:
>
>> Frequency domain glitch? You mean a discrete spur? A glitch is a
>> transient time-domain animal.
>
> I could have put it better I guess. What I mean is I'm seeing thin
> spikes in the RF output spectrum of the VNA at other frequencies
> than the desired one.

Which is a discrete spur. Gotcha.

> I'm guessing these spikes are the result of unwanted transient
> voltage spikes present on the DC VCO control voltage which sweeps the
> RF output frequency from 4Mhz through to 1.3Ghz.

If there's a forest of them, spaced at equal intervals, they could be
coming from a periodic time-domain source.

>
>> A DSO can save a single instance, or (with averaging) ignore the
>> noise and asynchronous background and just reproduce the glitch
>> itself.
>
> Well, Jeff has posted some advice for me on that aspect which I
> intend to try out tomorrow using my Tek DSO which is probably much
> better suited to the task than the analogue one I've hithertofore
> been using..

Yeah, I'm fonder of analog scopes than JL is, but my beautiful Tek 2467
hasn't been used in quite awhile, whereas I use the 1180x, TDS 694C and
TDS 784As most days that I'm in the lab. For manual driving, the 2467
has the best triggering of the lot, but it's very hard to beat a DSO for
general use.

[Cursitor Doom]

On Mon, 13 Jan 2020 00:24:38 +0000, Cursitor Doom
<cu...@notformail.com> wrote:

>>If you really want to see the noise, glitch, trash, garbage, or
>>whatever signal, set you scope so that the horizontal trigger is off
>>and the horizontal sweep is free running. Then, slooooooowly vary the
>>horizontal sweep until you see fairly stable glitches. If the noise
>>is periodic, you should be able to see something.
>
>Thanks, Jeff. Duly noted!

Oh, and btw, I know you like details such as this, I'll be using my
Tek 2232 DSO for this measurement. I don't really get on with digital
scopes, but seems like there's little alternative for this job. :(

[Cursitor Doom]

On Sun, 12 Jan 2020 19:40:23 -0500, Phil Hobbs
<pcdhSpamM...@electrooptical.net> wrote:

>> I'm guessing these spikes are the result of unwanted transient
>> voltage spikes present on the DC VCO control voltage which sweeps the
>> RF output frequency from 4Mhz through to 1.3Ghz.
>
>If there's a forest of them, spaced at equal intervals, they could be
>coming from a periodic time-domain source.

There *are* equally-spaced spikes at the lower end of the spectrum, up
to about 100Mhz, but thereafter they assume a far more random
appearance - although they are NOT random as they reappear at the same
place on the spectrum and at the same amplitude with each sweep. Plus
there *is*lower-level truly random noise around 500-600Mhz - so looks
like 3 different and unrelated faults to fix!! I do love my vintage
test gear, but I often spend more time fixing it than using it! :(

>Yeah, I'm fonder of analog scopes than JL is, but my beautiful Tek 2467
>hasn't been used in quite awhile, whereas I use the 1180x, TDS 694C and
>TDS 784As most days that I'm in the lab. For manual driving, the 2467
>has the best triggering of the lot, but it's very hard to beat a DSO for
>general use.

I only have one DSO out of 13 scopes in all: the Tek 2232. It's
nothing special; it was okay in its day I guess, but I'm sure John's
cheapo Chinese Rigol could piss all over it in a head-to-head
"test-off" lol.

[Phil Hobbs]

I'm the other way round: 2 analogue scopes out of 9 working ones: a 475A
and a 2467. The others are TDS 784As, a TDS 694C, 11802, 11801C, HP
54something, and an Owon something-or-other.

[Jeff Liebermann]

On Mon, 13 Jan 2020 01:46:08 +0000, Cursitor Doom

Bad idea. The Tek 2232 has both digital and analog storage:
<https://www.tek.com/datasheet/2232-digital-and-analog-oscillscope>
I suggest you use it as a conventional analog scope for the
untriggered horizontal sweep. If the frequency of the glitch source
is stable, you should be able to "tune" it in by adjusting the
horizontal sweep (with the trigger turned off). That works much
better in analog mode than in digital. Better yet, go back to using
your analog Tek 2465A for looking at the VCO control voltage.

Incidentally, I just noticed that you screwed up in your original
posting. You provided most of the details, a few of the numbers, but
failed to describe the "weird problem with the network analyzer" and
what "spurious spikes in the frequency domain on the RF output" might
be. Is that like a spurious RF signal appearing on the SA swept
display? If so, are you picking up some junk from nearby radio or TV
stations? If you can see it, what's the frequency on the SA? I can
also imagine that your test setup or cabling may be causing these
problems. A probe on the VCO line could easily pickup RF junk from
everywhere.

All this goes back to my three commandments for getting a sane answer
to questions on Usenet:
1. What problem are you trying to solve? Just one or two lines and
leave the details for later.
2. What do you have to work with? What's your level of expertise and
what tools and test equipment do you have available?
3. Where are you stuck and what did you do to get there?

[Jeff Liebermann]

On Mon, 13 Jan 2020 00:18:13 +0000, Cursitor Doom
<cu...@notformail.com> wrote:

>On Sun, 12 Jan 2020 17:17:04 -0500, Phil Hobbs
><pcdhSpamM...@electrooptical.net> wrote:
>
>>Frequency domain glitch? You mean a discrete spur?
>>A glitch is a transient time-domain animal.
>
>I could have put it better I guess. What I mean is I'm seeing thin
>spikes in the RF output spectrum of the VNA at other frequencies than
>the desired one. I'm guessing these spikes are the result of unwanted
>transient voltage spikes present on the DC VCO control voltage which
>sweeps the RF output frequency from 4Mhz through to 1.3Ghz.

I would guess(tm) that they are mixer spurs produced by the two
oscillators used to produce the network analyzer RF output signal. A
clue as to their relative amplitude compared to the RF output would be
helpful.

Look at the spurious stuff coming out of the network analyzer. Now,
vary the network analyzer frequency and see how they move. You can
tell the order (harmonic number) of the spurious components by how
much they move. For example, if you change the network analyzer 1 MHz
and the spur moves 3 MHz, than you are dealing with the 3rd harmonic
of one of the two input signals to the mixer in the network analyzer.

I also suspect that you might be picking up RF from nearby radio and
TV broadcast stations. Install a small antenna into the RF IN port of
your spectrum analyzer and see if there are signals at the same
frequencies that you are seeing them on the network analyzer RF
output. If so, you might check the shielding and grounding in the
network analyzer.

>>A DSO can save a single
>>instance, or (with averaging) ignore the noise and asynchronous
>>background and just reproduce the glitch itself.
>
>Well, Jeff has posted some advice for me on that aspect which I intend
>to try out tomorrow using my Tek DSO which is probably much better
>suited to the task than the analogue one I've hithertofore been
>using..

Yep. Failure to observe my advice will result in divine retribution.

[Cursitor Doom]

On Sun, 12 Jan 2020 20:21:53 -0800, Jeff Liebermann <je...@cruzio.com>
wrote:

>Yep. Failure to observe my advice will result in divine retribution.

I wouldn't dare, Jeff. Thanks again for the tips; they will be
followed up....

[Cursitor Doom]

On Sun, 12 Jan 2020 19:59:52 -0800, Jeff Liebermann <je...@cruzio.com>
wrote:

>Bad idea. The Tek 2232 has both digital and analog storage:
><https://www.tek.com/datasheet/2232-digital-and-analog-oscillscope>
>I suggest you use it as a conventional analog scope for the
>untriggered horizontal sweep. If the frequency of the glitch source
>is stable, you should be able to "tune" it in by adjusting the
>horizontal sweep (with the trigger turned off). That works much
>better in analog mode than in digital. Better yet, go back to using
>your analog Tek 2465A for looking at the VCO control voltage.

I'm afraid I don't have that option at the moment. The 2465A blew up a
couple of days ago and has now joined my pile of "awaiting attention"
junk. :( I didn't blow it up, by the way; it managed it all by itself.

>Incidentally, I just noticed that you screwed up in your original
>posting. You provided most of the details, a few of the numbers, but
>failed to describe the "weird problem with the network analyzer" and
>what "spurious spikes in the frequency domain on the RF output" might
>be. Is that like a spurious RF signal appearing on the SA swept
>display?

I think you may have missed the clarification I posted. Here it is
again:

"I could have put it better I guess. What I mean is I'm seeing thin
spikes in the RF output spectrum of the VNA at other frequencies than
the desired one. I'm guessing these spikes are the result of unwanted
transient voltage spikes present on the DC VCO control voltage which
sweeps the RF output frequency from 4Mhz through to 1.3Ghz."

> If so, are you picking up some junk from nearby radio or TV
>stations? If you can see it, what's the frequency on the SA? I can
>also imagine that your test setup or cabling may be causing these
>problems. A probe on the VCO line could easily pickup RF junk from
>everywhere.

I think that's a very valuable point there. One possible 'infection
vector' for noise is the method I'm using for sniffing the VNA. As you
may know, spectrum analyzers of this vintage were very intolerant of
DC on the input socket. There doesn't even seem to be a margin of
error! So any DC riding on the signal you're trying to obtain a
spectrum for and you blow up the front end. And I have a bit of a
complex when it comes to blowing up test equipment. I'm very, very
cautious these days, so I've just used a few turns of wire to couple
the signal from the VNA to the SA. This removes the DC risk, but does
mean there's no shielding around the sniffer coil. And right next to
these items there's my linear power supply powering up my Yaesu 857d.
Now the linear supply should be fine, being linear, but I'm just
wondering if the Yaesu, even in receive mode, might be generating some
stray spurious that the SA is picking up somehow. I'll switch those
items off when I try it again. My working practices are a bit slapdash
I must admit, but no worse than the average hobbyist I would imagine.

[Gerhard Hoffmann]

Am 13.01.20 um 13:09 schrieb Cursitor Doom:

> I'm afraid I don't have that option at the moment. The 2465A blew up a
> couple of days ago and has now joined my pile of "awaiting attention"
> junk. :( I didn't blow it up, by the way; it managed it all by itself.

Probably that hybrid that sends all those 24xx to their heaven.
Runs too hot. Don't expect to get a replacement hybrid.

Now back to the weird spectra that my Agilent 89441A presents under
remote control.

And while I'm at it, how does that Wehnelt cylinder stuff work in an
analog scope? I have a 200 MHz Iwatsu since > 35 years that I kinda love
and that has a trace sharpness / astigmatism problem.

cheers, Gerhard

[Cursitor Doom]

On Mon, 13 Jan 2020 13:28:46 +0100, Gerhard Hoffmann <dk...@arcor.de>
wrote:

>Probably that hybrid that sends all those 24xx to their heaven.
>Runs too hot. Don't expect to get a replacement hybrid.

No, it's not that chip that's gone. And those are perfectly happy to
run hot, the problem comes when people accidentally block the air
vents directly behind the chip, in which case yes, of course it will
blow.

[DecadentLinux...@decadence.org]

Cursitor Doom <cu...@notformail.com> wrote in
news:cfmo1ftqkv08v7jb4...@4ax.com:

> On Sun, 12 Jan 2020 20:21:53 -0800, Jeff Liebermann
> <je...@cruzio.com> wrote:
>
>>Yep. Failure to observe my advice will result in divine
>>retribution.
>
> I wouldn't dare, Jeff. Thanks again for the tips; they will be
> followed up....

Divine intervention is what happens when your design fails to impress,
and God decides to "let the smoke out" of your circuit.

[Fox's Mercantile]

On 1/13/20 6:09 AM, Cursitor Doom wrote:
> I'm afraid I don't have that option at the moment. The 2465A blew up a
> couple of days ago and has now joined my pile of "awaiting attention"
> junk.:( I didn't blow it up, by the way; it managed it all by itself.

My 2445 lost it's trace.
I replaced the HV module. $56 from a guy in Greece.
Works fine now.


--
"I am a river to my people."
Jeff-1.0
WA6FWi
http:foxsmercantile.com

[Jeff Liebermann]

Divine retribution and intervention are quite different. Divine
retribution is where a lightning bolt from out of the clear blue sky
strikes the transgressor down following a chronic failure to RTFM and
follow the instructions. Divine intervention is where the
participants in some technical or political discussion are unable to
reach a suitable conclusion or compromise, and both need to be
reminded who is the boss.

Also, none of my better designs have impressed anyone. In most cases,
they attracted far more criticism than praise. Whenever I provide
something to the design review committee for target practice, I
consider it a very good sign if they take the time to tear it apart
and offer their collective misjudgment. However, if the committee is
passive and offers little criticism, I'm certain that I've done
something fundamentally wrong.

[Jeff Liebermann]

On Mon, 13 Jan 2020 12:09:44 +0000, Cursitor Doom
<cu...@notformail.com> wrote:
>I'm afraid I don't have that option at the moment. The 2465A blew up a
>couple of days ago and has now joined my pile of "awaiting attention"
>junk. :( I didn't blow it up, by the way; it managed it all by itself.

See my comments on divine retribution elsewhere in this thread.

>I think you may have missed the clarification I posted.

Yep. I read your comments after I posted my comments. I often read
Usenet threads starting from the most recent, rather than in
chronological order. That saves me the embarrassment of answering
questions that were previously answered, but also causes me to miss
important added information.

>My working practices are a bit slapdash
>I must admit, but no worse than the average hobbyist I would imagine.

Yep.

[DecadentLinux...@decadence.org]

Jeff Liebermann <je...@cruzio.com> wrote in
news:i0fp1f1a14te9qm4o...@4ax.com:

> Divine retribution and intervention are quite different. Divine
> retribution is where a lightning bolt from out of the clear blue sky
> strikes the transgressor down following a chronic failure to RTFM and
> follow the instructions.

<https://en.wikipedia.org/wiki/Roy_Sullivan>

<https://funfactz.com/amazing-facts/walter-summerford-lightning/>

[Cursitor Doom]

I've now uploaded 3 photos of what I'm seeing. The most obvious noise
(apart from being obvious) is the 'comb' across the screen. I'm
getting the most spurii at the lowest 'fundamental' (around 5Mhz) but
as I sweep up in frequency the number of spurs declines markedly so by
200Mhz there are only 4 spurs remaining and they're much wider apart.

The other form of noise I mentioned in an earlier post is the more
attenuated and spread out one visible just to the right of the
centre-frequency line. This noise, unlike the 'comb' is constantly
flickering and random.



https://yandex.com/collections/card/5e1dfafd874e97d27c81ef46/

[Cursitor Doom]

On Sun, 12 Jan 2020 20:21:53 -0800, Jeff Liebermann <je...@cruzio.com>
wrote:

>Yep. Failure to observe my advice will result in divine retribution.

Not to mention ingominous failure!
Well, the traces I'm seeing (and have now posted images of to the
group) do bear a certain distinct similarity to what you described,
Jeff, so if that be the case I'm wondering if one of the oscillators
is being over-driven. I'm guessing that might give rise to 3rd
harmonic spurii, don't you think?

[Jeff Liebermann]

Sorry for the late reply, but I had to waste the last few days dealing
with the California Dept of Motor Vehicles to renew my drivers license
and obtain a Real-ID endorsement. I passed both but am still
recovering and might need a few more days to calm down.

Interesting spectrum analyzer display and description. The most
interesting part is how you managed to produce the photo, without
bothering to mention the:
1. Network analyzer test Frequency
2. SA (spectrum analyzer) horizontal MHz/div.
3. SA vertical full scale, dBm per division, and dBm full scale.
4. Whether you were misusing a scope probe (again) or feeding it
with 50 ohms. Also, any attenuators or couplers in the RF path.
5. Did you remember to turn of the 1, 10, 50Mhz markers on the
HP8754A?
6. Any other numbers which might be useful, especially those you
previously mention, but which I'm too lazy to dig through your old
postings to find.

Had you supplied some numbers, I would have dived into the HP8754A
network analyzer service manual, and determined the frequencies of the
two oscillators (3.6 - 4.3GHz minus 3.6 - 3.0GHz) used to produce your
unspecified output frequency. I would then compare the possible
intermodulation mixes of these two frequencies, with the mess on your
spectrum analyzer screen, and see if any of the spurs are the result
of oscillator feedthrough, intermod mixing, or overdriving the
spectrum analyzer and creating the spurs in the SA. However, since
your photos were not accompanied by any numbers, I can't do this for
you. Incidentally, you might want to ask in one of the HP test
equipment forums for what a working HP8754A RF output looks like on a
spectrum analyzer. I would not be surprised if the mess on your
screen is quite normal.

Incidentally, if you download and read the HP8754A service manual,
you'll find a large number of Performance Tests. Performed properly
in the manner specified, you should get a better idea if your network
analyzer is functioning as expected.

Also, this is your second warning about providing insufficient
information and numbers. Provide numbers with your problems or
prepare thyself for the wrath of the radio gods. The recent failure
of your Tek 2645A scope was a poorly aimed warning shot (I was aiming
for your Tek 2232A scope) and a hint of things to come.

[Cursitor Doom]

On Wed, 15 Jan 2020 21:05:43 -0800, Jeff Liebermann <je...@cruzio.com>
wrote:

>Sorry for the late reply, but I had to waste the last few days dealing
>with the California Dept of Motor Vehicles to renew my drivers license
>and obtain a Real-ID endorsement. I passed both but am still
>recovering and might need a few more days to calm down.

My sympathies. I've heard about the queues in those places and they
must be pretty bad indeed to make the international news! Is it true
they recruit retired-disabled ex-NYC taxi drivers for their counter
staff?

[snip missing numbers]

Yeah, I know. I'm sorry but you know about as much as I do! If I'd
zoomed out on those shots there's a bezel around the outside of the
screen on that SA which is *supposed* to show all that info.
Unfortunately, it's bust. This is a partly working unit I picked up at
a ham rally and am in the process of re-commissioning. Unfortunately,
my No.1 RF SA (the 8566B) is currently out of action with a recent
"YIG oscillator unlocked" error message, so there's nothing to fall
back on. It's kind of tough to do stuff with all this busted
equipment. Even the equipment I picked up in prime working condition
is letting me down lately.
All I *can* say about those photos is the SA is set on the 0.01-1.8Ghz
display range, I'm sniffing the VNA output with a few turns of wire
rather than a direct connection, the 40dB attenuation is set (and
seems to work!) and the VNA output frequencies in each picture are the
filenames for each of those pictures.
So the situation is "not ideal" but you have to work with what you've
got, I guess.

[Cursitor Doom]

On Wed, 15 Jan 2020 21:05:43 -0800, Jeff Liebermann <je...@cruzio.com>
wrote:


I just noticed some points you mentioned I forgot to cover in my
earlier post...

>1. Network analyzer test Frequency

Shown in picture title for example 10Mhz.jpg means the VNA was set to
10Mhz.

>2. SA (spectrum analyzer) horizontal MHz/div.

Full band mode so ~180Mhz/div.

>3. SA vertical full scale, dBm per division, and dBm full scale.

No reliable info.

>4. Whether you were misusing a scope probe (again) or feeding it
> with 50 ohms.

Neither - sniffing with few turns of wire.

> Also, any attenuators or couplers in the RF path.

Nope. Just the 40dB attenuation setting of the SA.

>5. Did you remember to turn of the 1, 10, 50Mhz markers on the
>HP8754A?

Yep, no markers.

>Incidentally, you might want to ask in one of the HP test
>equipment forums for what a working HP8754A RF output looks like on a
>spectrum analyzer. I would not be surprised if the mess on your
>screen is quite normal.

Good suggestion.

>Incidentally, if you download and read the HP8754A service manual,
>you'll find a large number of Performance Tests. Performed properly
>in the manner specified, you should get a better idea if your network
>analyzer is functioning as expected.

I do actually have the original physical hard-copy of the service
manual and was working my way through those performance tests when I
got side-tracked by something else whilst waiting for the required
phase-matched power splitter and some new terminations to arrive. I
really MUST get back to that as they have since turned up in the post.

I also need to try out some other poster's suggestion re sweeping the
VCO ramp with DC from batteries, too. So I have plenty to get on with
whilst you digest my replies, Jeff. :-/

[Cursitor Doom]

The VNA was putting out 10dBm in all those photos, btw. I'm reasonably
confident that much at least is 'pretty accurate.'