2) nope. But I was told that it uses some kind of transformer, so the block diagram in the service manual is deceptive. I furthermore heard RUMORS of the one for the 4195A working on the 4953A if you slinked around with the cables. Sadly, this is all I know - so far, all managed to elude me.

Yes, I believe the general shape of the circuit is known or can be deduced from available literature.? The best summary I have found is here:
? ?(about the 4291A RF impedance analyzer---similar vintage from the same design group in Japan)
But it would be nice to see the innards.? The difficult aspects will be those which cannot be represented in a schematic diagram.

This is my first post to the group, so hello everyone.? You've been a great resource when I've needed to search for information on prospective or actual TE purchases; thanks :)

I recently acquired an Agilent 4396B spectrum/network/impedance analyzer with all of the options (1D5 oven, 1D6 time-gated measurements, 010 impedance measurement) in non-working condition.? The main problem was some stressed SMD electrolytics on the A50 DC-DC converter board, as discussed in service bulletin 11.? There is also something odd in the source assembly, but it seems to be after the ALC power detector; current theory is the semi-rigid coax link from the output attenuator to the bulkhead N connector, or perhaps the N connector itself.? I'm documenting the repair here:



Looking ahead I have a few questions (and will undoubtedly have more as I dig into the analyzer's capabilities):

1) The calibration constants are stored in an EEPROM and updated over GPIB, but as far as I can tell the actual commands and format of the data are not documented.? The only approved way to update the calibration is to use an HP-UX program called "ADJ4396B" running on an ancient HP controller.? Is this likely to be a problem in the long run?? Should I consider sending the instrument out for calibration (assuming the cal lab even has access to this magic software)?? This is for home / hobby use only, so it likely doesn't matter, but I wonder if a window of opportunity is closing as these units get older and the last support is dropped.

2) Has anyone actually laid eyes (or camera) on the insides of the 43961A impedance test fixture?? I have no need for it except curiosity, but it would be fun to leverage the 010 software in my unit.? It could probably be a worthwhile learning experience to build something equivalent at lower frequencies.

3) Not a 4396B question per se, but what is the currently recommended GPIB interface solution in 2019?? I would prefer not to use any NI stuff or the VISA library, but if that's the price for being able to run useful third-party software (like KE5FX's phase-noise tool), I will.? Although in that case I would still prefer options which *can* work with the VISA library but can also work without them.? Cross-platform compatibility with Linux and Windows would be a plus.

Regards,
Mark

On Mar 21, 2019, at 5:48 PM, Dave_G0WBX via Groups.Io <g8kbvdave@...> wrote:

Be careful "bench testing" Yig Tuned Ocillators" (YIG = Yttrium Iron
Garnet by the way *.)

Older designs that use + and - polarity supplies can be destroyed if you
get the power up sequence wrong, the little FET that "is" the oscillator
dies. Similar can happen (or at best some damage that makes them
produce less signal power) if they are run without a RF load of some sort.

*

Plus Wikipedia has some good references too.

From past experience working professionally servicing Spectrum Analysers
to component level that used YIG Oscillators, check the power supply
rails that feed the coil drivers, also any degaussing scheme, if such is
fitted.

A sudden and permanent step change in the current/frequency
relationship, and also a change in the apparent dynamics of the coil
driver (stability issues) can be an indication of shorted turns within
the tuning coil. If, there are any documented figures for the tuning
coil inductance, that can of course be checked with a LCR bridge, at low
power.

Take care..

Dave B (G0WBX)

--
Created on and sent from a Unix like PC running and using free and open source software:

Yeah, but you have factor in the suits.


On Thu, Mar 21, 2019 at 4:46 PM Dave McGuire <mcguire@...
<mailto:mcguire@...>> wrote:


? Pretty sure the 8753D is newer than that, but even if it isn't, JPG
was *never* suitable for that sort of imaging.? A good choice (perhaps
the only "right" choice) in 1995 would have been GIF.? Heck, even the
awful BMP would've been better than JPG!

? ? ? ? ? ? ?-Dave

On 3/21/19 4:44 PM, W2HX wrote:
> I don't know the age of these VNAs but PNG was only invented in
1995. Probably needed a few more years before they became popular
such that an engineer would consider using it. Maybe that explains it?
> ________________________________________
> From: [email protected]
<mailto:[email protected]>
<[email protected]
<mailto:[email protected]>> on behalf of Dave
McGuire <mcguire@... <mailto:mcguire@...>>
> Sent: Thursday, March 21, 2019 4:36 PM
> To: [email protected]
<mailto:[email protected]>
> Subject: Re: [HP-Agilent-Keysight-equipment] I'm officially lost
now with this 8753ES
>
> On 3/18/19 3:32 PM, pianovt via Groups.Io wrote:
>> Yes, PNG is great for screen shots. The problem with the 8753D/ES
>> graphics is that the firmware guys didn't do the best job. They
added a
>> graphics dump, but they only implemented jpg. To make it worse, you
>> can't download it via GPIB, it has to be stored to a floppy. At
least, I
>> don't see a command to download the JPG via GPIB. If anyone here
knows
>> how, please let me know.
>
>? ?Wow, I can't imagine that any engineer who was able to get hired
by HP
> wouldn't understand why JPG isn't appropriate for screen shots or
other
> line art. (or anything else that isn't a continuous-tone natural
scene,
> which is what JPG was designed for, and the only thing for which it
> works well)
>
>? ? ? ? ? ? ? -Dave
>
> --
> Dave McGuire, AK4HZ
> New Kensington, PA
>
>
>
>
>
>


--
Dave McGuire, AK4HZ
New Kensington, PA