Advice for new HP8510c owner

[EternalStudent]

I was laid off for a few months and spent my time on eBay doing research on parts and managed to put together a HP/Agilent 8510C vector network analyzer system.


I got the HP8510C/8514B/8511A/83621B GPIB and connection cables and managed to get a Gore FB0HA0HC048.0 VNA Test Port Cable but I have been unable to find a calibration kit except for a 7mm one in the 85051A.

I am just waiting on the last pieces of this system to put it all together and see if it all works as described.

Any advice anyone can give me on putting this together and what to avoid would be very helpful.


With very careful research I managed to get everything under 5k total and I don't want to break things by putting it together wrong.

I understand the 8510C has 3.5mm connections which can be connected to 2.92mm and sma connectors which I assume does not break the 8510C.

Does anyone have any advice on how to calibrate this for 3.5mm/2.92mm short of spending 1k on a calibration kit?  I am hoping to get to at least 18 Ghz of usable bandwidth as my first start up point.

I also picked up a 8569B spectrum analyzer with HP 11971A and HP 11971K mixers as part of the total cost so I can cover the range of the HP8510 with a spectrum analyzer as well.

For testing I picked up a Lindgren Model T/T RF Copper Brass Enclosure 12"Cube w EC-2005-1C EMI/RFI Filter for testing very small PCBs or assemblies up to roughly 10 GHz.

I previously have a log antenna and lower frequency antennas for EMI/EMC testing.

I have never done RF design except a radio project in school, I took prop and rad and have a basic understanding of the issues but have not manufactured anything yet.  I am looking for any project ideas that I could perhaps use the new goodies to make something I could sell even if the profit is below the cost of my time.  My new job has every other Friday off and I want to spend the time diligently doing projects to expand my skills both in RF and outside RF.  So I want to work on worthwhile projects that can be cash flow positive even if they don't pay my time at market rates.

Any words of advice are welcome =)

EDIT:  In looking over the totals maybe it was closer to 6k or 7k spent.

[Jay_Diddy_B]

Hi,

I would start off by making your own calibration kit. This can be done very cheaply. I won't be good enough for any serious work, but it will get you familiar with the network analyzer.

Here is a link that goes into a lot of detail:

http://www.qsl.net/in3otd/electronics/VNA_calkit/SMA_female.html

You will need to build something like this:



Picture taken from here:

http://hforsten.com/cheap-homemade-30-mhz-6-ghz-vector-network-analyzer.html

These will work well to 1 GHz

Regards,

Jay_Diddy_B

[Neganur]

Reading all those different coax sizes in the same sentence makes me urge you to read about RF connectors (there's a great The Signal Path episode for example)Especially since you mentioned you're new to RF.


PS. Cal kits won't come cheap for the higher frequencies, and it's easy to buy something broken without knowing it. I'll see if I can find some of the old threads from when I asked the same questions.

Edit: on 3.5mm and 2.92mm (better not try with normal SMA, they just don't obey the mechanical tolerances) Gore. Takeaway is, you can do it with high quality connectors, but imo most connectors are not high quality if we're talking "eBay" and "affordable".

[EternalStudent]

Thanks for the information.

I am piecing together a calibration kit online buying used good parts.

Seems I need  3.5mm male - short, load, open and 3.5 mm female - short, load and open.

Then for the feedthrough I can just connect my 3.5 f to the 3.5 m right?

At $100-$200 each it adds up =(

[bjcuizon]

Reading all those different coax sizes in the same sentence makes me urge you to read about RF connectors (there's a great The Signal Path episode for example)Especially since you mentioned you're new to RF.

Here it is...

[rastro]

Eternal,

Congratulations on your 8510 VNA - this ought to keep you busy for a while.  You must have got a great deal on theses units if your budget is under $5K; I know the HP 83621B series usually list/sell over several thousand dollars on eBay.

I have a system with an 8515A (26.5 GHz) using a 8340B (26.5GHz) RF source.

It's really frustrating when you realize to cost of calibration standards, connectors, and cables.  All precision mechanical - the electronics seems less expensive pound for pound.
I am not an expert but here is some guidance from my experience putting together an 8510 system.

Even with out a cal-kit you should be able to run the system and get reasonable results.  That is you should be able to functionally check your system without running the calibration routine.  At the very least you should get a decent M/F 50-ohm load for reflection reading and some trust worthy adapters.  There is also a short document showing all the cabling and configurations for a typical system.  I'll include at the bottom a list of documents you can find on line in pdf format.

Some tips on putting your system together:
1.  You can use a  $15.00 straight male to male cable (37-pin??? connector) for the cable between the IF and Display rather than the $100.00 HP cable.
2.  You can use high quality used HP/Agilent bulkhead connectors as adapters.  These are typically taken off scraped equipment.
-- I have used a couple of HP 3.5(m) to 3.5(m) on the 2 ports of my 8515A (basically a barrel connector).  HP PN 08513-20017;  These are rated to 26.5GHz.  I typically sees these listed for $30-50.00 but have purchased mine for $15.00 to $20.00.  Look for a US sellers who has a history of selling HP test equipment and parts.  Make sure they are listed as "used" and not "repair".
-- I've also bought some HP/Agilent  N to SMA bulk head.  These can be found for around $10.00 to 15.00 as there are more out there.  I can't remember the P/N.
3. I'm still working on a more complete cal-kit solution.  However you can find some intermediate solutions for the lower ranges. 
-- sdr-kits.net in  the UK sell several kits to support the VNWA.  I bought their 'Female SMA SOL 4 pcs Calibration Kit of Rosenberger' kit for my 8753C.  You can also duplicate their Amphenol base kit ordering from mouser.  The Rosenberger pieces are harder to order and the kit was the easy way to go and they really where not marked up significantly over the piece parts.
-- Even if you don't have a HP level cal-kit it is still a good idea to have a precision 50-ohm load like HP 909X series.  These would also be included in a complete HP cal-kit.  You can also find these at ebay but they are not cheap - expect to pay in the $100.00 range.  However this will provide some confidence regarding your system - compare against the internal loads in the 8510 systems test set.
--- I have a set of 909 loads.  I also use mini-circuits ANNE-50-8 loads these are surprisingly good price performers.  Not as good as the HP 909 but certainly runner up with the Rosenberger load in my SDR-kit. 
-- If you join the yahoo WVNA group there is member based in Germany who has done a lot of work characterizing these SDR-kits and has some printed reports posted on line in English these include important coefficients. 
-- Ultimately to get the most accuracy out of the system  especially at the upper RF ranges you will have to acquire bonafide calibration standards (HP/Agilent).

Document To Find Online:
8510C...Service Quick Reference Guide (08510-90317.pdf) pages = 52
PAGE            DESCRIPTION
-----------------------------------------------------------------------------------------
9      Basic system cabling diagram showing back of system "preflight checkout"
51      8510C System Level Troubleshooting Block Diagram

Other Manuals:
8510C...Operating and Programming Manual (08510-90281.pdf) pages = 391
8510C...On-Site Service Manual (08510-90282.pdf) pages = 439
85395...Using Your Upgraded Receiver (85395-90007.pdf) pages = 10
8510...Key Word Dictionary (08510-90280_8510KeyWords.pdf) pages = 826
HP 115755 Performance Upgrade Package (hp_115755_8510C_install.pdf) pages = 26
HP 85108L Pulsed-RF Network Analyzer System (85108-90036.pdf) pages = 122
HP 8510 Quick Reference Guide(08510-90292.pdf) pages - 115

-rastro

[hendorog]

Thanks for the information.

I am piecing together a calibration kit online buying used good parts.

Seems I need  3.5mm male - short, load, open and 3.5 mm female - short, load and open.

Then for the feedthrough I can just connect my 3.5 f to the 3.5 m right?

At $100-$200 each it adds up =(

Once you have the parts, then you need the coefficients which characterize those parts. Those coefficients come with a proper cal kit, and are the thing you pay $$$ for.

If you can get your parts measured on a calibrated VNA - i.e. by someone who has an official cal kit - then you should be able to determine the coefficients using the script at the link already mentioned by Jay_Diddy_B. That should get you results pretty close to a bonafide kit.

I haven't done this myself, so YMMV but intend to if I ever stumble across someone with a proper VNA cal kit

[rastro]

Another alternative.

Kirkby Microwave Ltd also sells a kit that is tested on Agilent VNA against an HP/Agilent calibration kit and the coefficients and other parameters are provided in formats for 8510 or 8753.  Just search '7 GHz Low Cost SMA Connector Calibration & Verification Kit 'for VNAs HP Agilent' on eBay.  Kit is around $400.00 plus shipping.  I am not affiliate with this vendor nor do I have the kit.  I do know he posts to the HP yahoo group and is very knowledgeable with HP VNA's.  I think his background is a PHD in electronics.

[Neganur]

In medical physics but don't let that stop you from talking to him.

However, I think his kits may not be suitable for all the way to 18GHz (which is what the OP asked for). He still might have a suggestion. Perhaps some of the ecals on eBay work with the 8510?

[hendorog]

In medical physics but don't let that stop you from talking to him.

However, I think his kits may not be suitable for all the way to 18GHz (which is what the OP asked for). He still might have a suggestion. Perhaps some of the ecals on eBay work with the 8510?

Also he doesn't list any 3.5mm kits, only SMA. I know they are compatible but I expect the return loss of the load will be degraded.

[EternalStudent]

Maury Microwave individual parts were available for about $130 each.

I don't have the parameters yet but I am pretty sure these will exist somewhere.

I have my shorts m/f, opens m/f (I think), and a female load and am bidding on a male HP load.

I still need to get a precision wrench.


All the parts are at least 18 GHz with a few 40 Ghz so I will be able to expand over time.


When I say eBay I mean eBay not cheap Chinese parts on eBay =)

If the used parts are all working I should have a pretty good calibration set for ~$800-900 but realistically if its under $1200 I will be happy.

[rastro]

You may find this page helpful for your HP/Agilent/Keysight calibration standards:

http://na.support.keysight.com/pna/caldefs/stddefs.html

Most of the individual Muary Microwave calibration standards I've seen do not have part numbers or revision markings making it difficult to know exactly what part you have to match with VNA settings.

-rastro

[darrell]

The Maury standard definitions are near the end of their calibration kit manuals (available on the web) and have not changed in decades. One thing to be aware of is the Maury loads have a bunch of tuning screws and if dropped, can be detuned. I've got a newish 8031B4 with only 20 dB return loss at 6 GHz. It still has ~ 50 dB RL at < 1 GHz. I've seen more bad ones than good ones. A DC check won't show anything wrong. The only way to find out would be a known good kit, a sliding load or an airline. I did a TRL cal and then measured all of mine.

The newer HP/Agilent/Keysight cal kits are made to tighter tolerances than Maury, but you have to watch out for damaged precision slotless connectors. Older HP was made by Maury.

Assuming 1 male and 1 female test port, through is just a direct connection. With same gender test ports, through is a challenge as I don't think 8510C supports SOLR. Swap equal adapter was the commonly used method, but is obsolete with SOLR. For that you need a M-F and F-F or M-M adapter with equal S Parameters.

If you are doing opposite gender test ports, TRM calibration is an option. You don't need opens for that. Just shorts (same delay required for M and F) and loads. It can be more accurate than SOLT as the uncertainty of the opens isn't included.

I have 4 pairs of 3.5 mm opens and might be willing to sell a pair of older Maury ones in decent condition. PM me if interested. I might be able to come up with a pair of Maury loads, but they would likely be somewhat out of spec and older.

I'd recommend a set of connector gages when you can find affordable ones. There are lots of SMAs with protruding pins out there that can cause damage. I gage everything that comes into the lab and have chopped a lot of bad connectors off. This is particularly prevalent on cables that might have been assembled by someone without gages.

I second the suggestion of the used HP bulkhead 3.5 connectors. They are great.

[rastro]

Does anybody know if it's possible to replace the pins on the front ports of the s-parameter test set?

[Jay_Diddy_B]

Does anybody know if it's possible to replace the pins on the front ports of the s-parameter test set?

If you are asking about APC-7, you might want to read this:

http://literature.cdn.keysight.com/litweb/pdf/11591-90003.pdf?id=1391701

Regards,

Jay_Diddy_B

[rastro]

Does anybody know if it's possible to replace the pins on the front ports of the s-parameter test set?

If you are asking about APC-7, you might want to read this:

http://literature.cdn.keysight.com/litweb/pdf/11591-90003.pdf?id=1391701

Regards,

Jay_Diddy_B

Thanks; this is good info but I should have been more specific.   I was looking for something like the HP 8515A front ports.  I'm not sure what this connector/type is called/Part-Number??.  It's supports 26.5GHz and is sized to accept 3.5mm.  I know these are typically use test cables specifically mount to these connectors (expensive!).  I cheat and use my 3.5mm barrel/bulkhead to a couple of short 3.5mm gore cables. 

I remember seeing some procedure on a similar connector that the pins could be swapped out - but can't remember where.  My current unit is fine because I use 3.5mm barrel connectors as connector savers.  However, I may be getting a spare that is missing/damage port pin.

-rastro

[Kire Pûdsje]

For the lower frequencies (< 2 GHz), a kit as shown in one of the first posts can be made reasonably accurate. Especially if you have access to a well calibrated setup, so you can characterize the standards with C0, C1 .... etc.
For frequencies above that, use TRL calibration. This method is essentially self calibrating. The only definitions needed are the line impedance and the approximate electrical length (to identify the proper half-plane).
Please also read up on insertable/non-insertable calibrations (adaptor removal).

[KE5FX]

Does anybody know if it's possible to replace the pins on the front ports of the s-parameter test set?

If you are asking about APC-7, you might want to read this:

http://literature.cdn.keysight.com/litweb/pdf/11591-90003.pdf?id=1391701

Regards,

Jay_Diddy_B

Thanks; this is good info but I should have been more specific.   I was looking for something like the HP 8515A front ports.  I'm not sure what this connector/type is called/Part-Number??.  It's supports 26.5GHz and is sized to accept 3.5mm.  I know these are typically use test cables specifically mount to these connectors (expensive!).  I cheat and use my 3.5mm barrel/bulkhead to a couple of short 3.5mm gore cables. 

I remember seeing some procedure on a similar connector that the pins could be swapped out - but can't remember where.  My current unit is fine because I use 3.5mm barrel connectors as connector savers.  However, I may be getting a spare that is missing/damage port pin.

-rastro

You're probably referring to "NMD" connectors, an acronym for HP's original Network Measurements Division.

3.5mm is interchangeable with SMA, but you need to be careful to use only good-quality SMA hardware with it.  The 3.5mm connectors use a relatively-fragile suspended center conductor that won't tolerate a lot of lateral force from a poorly-centered SMA male center pin, or one that flexes with the cable it's attached to. 

They are somewhat tougher than the literature suggests but should still be treated with more than the usual respect.

[darrell]

I remember seeing some procedure on a similar connector that the pins could be swapped out - but can't remember where.  My current unit is fine because I use 3.5mm barrel connectors as connector savers.  However, I may be getting a spare that is missing/damage port pin.

-rastro

There's a procedure to replace that part in the service manual. Even if you can get it from Keysight, the cost is probably more than another Ebay 8515A. I've had that apart on an 8517B and it's fairly simple. You'll have to make or buy a thin wrench to get it off.

[EternalStudent]

I just discovered the HP8530A exists and that it is a slight difference from the HP8510C

Anyone able to describe the difference?


Also anyone familiar with the HP83650 and the HP83640 are they interchangeable if you are willing to give up 10 ghz (20ghz?)  or do the high bandwidth test sets have software that only allows the 83650 if I manage to find the 110 Ghz set for example?  Or even the 8517B test set at 50 ghz?

How important is having a pulsed test set like 85110?  I need two 8360L type sources for a pulsed setup correct?  That is not cheap.

I just got my VNA cable and its 3.5mm m to f so I have 3.5mm (m) to 3.5mm(m) and will need to use a (f) to (f) adaptor.  I purchased a HP bulkhead for $70.

With the Maury Microwave calibration odds and ends and the 26.5ghz bulkhead I think I am about ready to power up the system and start my guesses regarding calibration parameters.

I have been delaying to get the precision wrench and a autotransformer (I meant to say ferro-resonant power conditioning transformer) and UPS to try and reduce things that can destroy the system.

I picked up Labview Home for $50

http://store.digilentinc.com/labview-home-bundle/




Its a legal copy for home use only but since the HP8510 has GPIB I figured there may be someone else out there interested in Labview for cheap.

Thanks for that advice you saved me $100 =)

[rastro]

Eternal,
I have seen the HP 8530A's for sale but couldn't find a whole lot of information.  They probably have different firmware at the very least.

I'm not sure why you are using an autotransformer.   Are you referring to something like a variac?  I use a Staco L1010VA which has a volt and amp meter when I do an initial check of individual instruments.  Usually when I get a new instrument I check that the mains fuse and AC setting is correct then gradually power it up with a my L1010VA.  If I don't see high amps issues I plug it into a regular outlet.

You probably want to do a minimum sanity check on the system before attempting the port calibrations.  The system should get reasonable readings using a quality 50-ohm load and cable even without a port calibration.
1. Run the system self test.  This only verifies the Display and IF unit.
2. Look at the reflection S11/S22 for P1 and P2 with a quality 50-ohm load hooked up.  This should verify the coupler for return loss.
3. Hook up a cable between P1 and P2 and again look at reflection S11 and S22.  This should verify the internal 50-ohm terminations available on each port.
4. Verify transmission (cable between P1 and P2) of P1/P2 S21 and S12.  Verify transmission couplers.

It's best practice to use proper torque wrenches; however if you hand tighten snug with your fingers you are probably not going to over tighten the connector (8.0Lb/in for a 3.5mm).  For example if I use a proper torque wrench even on an SMA (5.0Lb/in) I cannot unscrew it using just my fingers.  The 3.5mm are on even tighter.

-rastro

[rastro]

I thought I'd provide a sort of sanity check baseline for an HP 8510C with an HP 8515A S-Parameter Test Set. 
The measurements where done on a system that did not use any calibration corrections.  So any measurements includes basic errors in the system.  Even without any calibration corrections you should get results that are 'qualitatively' reasonable.    Also all RF connections where made finger tight snug and still got pretty decent results.  The normal/correct torque ~8lb/in for 3.5mm connectors.

The first measurements are to check Port 1 & 2 reflection into a 50-ohm load.
This is using a good quality termination HP 909D (3.5mm/26.5GHz)
Pictures show S11 & S22 and respective smith plots.

[rastro]

HP 8510C Baseline Sanity Check Continued:

The next set of measurements use a quality cable between Port 1 & 2:
- Reflection S11 and S22 and respective smith chart are run.  This uses the HP 8515's internal 50-ohm terminations.  It is not as good as the the prior S11/S22 measurements in part because it is going through a cable.  However it can still be used to determine any meaningful issues with the internal terminators of the test set.
- Transmission S21 and  S12 on Ports 1 & 2 show that gradual linear drop off with higher frequency as would be expected. 

rastro

[Zeren]

If you are OK with planar measurements, you can make microstrip or GCWG multiline NIST TRL cal kit. As connectors you can use southwest endlaunch, which you can reuse many times. At minimum you need only 2.

[EternalStudent]

I picked up another 8510B in the form of a 8530 with the option to switch back and forth to the 8510B.

That and the monitor were $300 total.

So now I am trying to figure out if I just purchased a spare 8510 or if the microwave receiver part of the 8530 is uniquely valuable.

How large are antenna testing ranges typically?

I live an hour or so away from some EXTREMELY cheap land...

 

[rastro]

I picked up another 8510B in the form of a 8530 with the option to switch back and forth to the 8510B.

That and the monitor were $300 total.

So now I am trying to figure out if I just purchased a spare 8510 or if the microwave receiver part of the 8530 is uniquely valuable.

How large are antenna testing ranges typically?

I live an hour or so away from some EXTREMELY cheap land...

 

Any update on your first system?  Have you got it up and running?  Did you manage to piece together a calibration kit solution?

-rastro

[EternalStudent]

I am in the process of pricing a 45 U 19" rack and shelves (full ones are $86 each!!).  Which is currently looking like a $500-1000 expense because I need a rack that can be disassembled and carried.

I still need the power conditioning for input AC before I am going to power it up and a SMA physical gauge to make sure the precision connectors don't get damaged by cheaper connectors.

I just started a new job and between getting ready, driving there, working, driving back from work I am spending 15 hours a day on work.

Its not that I haven't had time on the weekends but when your work week is ONLY work its much easier to do non technical things on the weekend.

I have just switched to Orcad and PSPICE for work and may be buying the home version of MatLab to learn work software and ramping up on that is ahead of this on the order of weekend activities.


So I have moved from plenty of time but no money to no time but sufficient money =)

Once the rack arrives I will assemble the system and power it on to see if its DOA but I may not have play time until work settles down a bit.

I think I have all the components of a calibration kit in the form of individual pieces of Maury Microwave load, open , short.  They are located near me so I can ask them questions and may get a response.


Part of the purpose of buying this kit was to get proficient in another field of EE to give me better job flexibility to find something closer or to work for myself.

I apologize for being so excited but proceeding so slowly but this fits my present circumstances.

[rastro]

Going back to work is one of those good problems have.  Glad to hear you are back in the job market.

Getting the racks is a good idea.  I have a make-shift wooded rack for my system.  It's not beautiful but it does the job and I probably don't need to worry about moving it anytime soon. 

Since the majority of my equipment purchase is eBay I like to check out my purchases early due to the eBay resolution time line.

Anyway good luck going forward.
-rastro

[hendorog]

I had the Schaffer power line adapter blow up on my HP 8515A, you might want to replace those modules before you get it all racked up.

Thanks for posting those return loss pics rastro, good to know.

[EternalStudent]

The wheels of progress grind slowly.


I have just spent $800-900 buying a 42u 19" rack and shelves and a power conditioning transformer to protect the hardware and allow me to reclaim my work bench.  I previously said auto transformer but what I meant was ferroresonant power conditioning transformer.  Once that arrives I will start the incredibly basic operation of putting the setup together and making sure its not DOA. 

But the interesting thing is that I work 9/80 meaning I get every other Friday off and am applying to work at Maury Microwave and at an antenna range with that every other Friday.

I probably will be unsuccessful getting hired at either but I basically offered to work for cheap/free if they give me access to an RF engineer to ask questions =)

I also picked up a mechanical positioner (FLIR/DIRECTED PERCEPTION PTU-D300 PRECISE PAN TILT) and two antennas (WIRELESS EDGE ANTENNAS MT-485025) for less than $300 total as a package deal on eBay.  I already own 2 log-periodic PCB type antennas and a 10 khz to 30 MHz dipole antenna.  So I still need to get a wideband reference antenna solution and find a directional between 30 MHz and 800 MHz.

My concept has been to put together my own antenna range to make use of the HP8530 and I found some land for $5000 - $10,000.
I am in the USA and the government gives away or sells super cheap all manner of things if you can transport it to your own land.
I have seen all kinds of things I would like to own but I have no place to put.
I have seen free 5 meter satellite antennas on craigslist for anyone who can come pick it up for example.


But as I was headed in that direction (and am still interested in owning my own antenna range) I realized that all the money is in EMI/EMC compliance testing.

So the plan right now is to purchase 1 or more site suitable for an antenna test range and compliance testing and then to wait for a steel structure building to come up either free or super cheap in government auction.  Then I will pay for a foundation and build a building suitable for EMI/EMC compliance testing and for the permanent part of a antenna test range. 

I apologize for moving in such slow motion but the wheels are turning.

[rastro]

It sounds like you're moving on a grand scale.

[EternalStudent]

Just a little update.


1)  Got the ferro-resonant transformer and started to build my cheap 19" rack    and the angle iron is the wrong angle  so the easiest part is causing me troubles lol.

2)  Purchased a 85110 ($800) and 8517b ($1025) kits to allow me to do pulsed S-Parameters to 18 ghz (I purchased 11720A from ebay $60) and to do continuous to 50 ghz (provided I get a suitable source which I don't presently have).

But the interesting part is I discovered OML see omlinc.com who seem to do after market modifications to HP hardware to extend the usability and allow use of non HP/Keysight hardware with the 8510c or late firmware model 8510b.  They claim they can modify 8510c to 325 ghz if the right 85104 or 85105 is owned and the modifications done.

So if any of you have a 8510c and 85104 or 85105 and are looking to get to 325 ghz....

[EternalStudent]

I got some time to finish the rack and load the equipment but still haven't finished the ferro resonant transformer hook up or powered on the hardware.

It was a nice feeling to give myself the presents I had purchased but not had time to assemble =)

[DL8RI]

Hello everybody,

allow me to revive this Zombie here
Since a few weeks, I'm also a happy "new HP 8510C owner".
My Setup is 8510C, 8514B, 83620A. Also got an 3650 Cal Kit from Anritsu with the data-disk.

I got the Unit as defective (it was) and have it now working again. Of course, also with my test-set, the Schaffner-filter did go off. I believe another one did bite the dust longer ago, whenever I switch the VNA on and all fans start, it smells like burned filter in the first seconds. Need to check and replace this one too.
What I was a little bit hesitant to do, is to change the clock-battery on the CPU Board (I have the feeling it is weak). If I change the battery and by accident the 3V are lost during the process, am I losing more than just the time?
HP's service-manual just says, the battery is not field-replaceable and the whole board needs to be switched. Some other Units from HP have cal-data or something, that may get lost. I would like to avoid "bricking" my Unit.

After the repair, I observed that my VNA has a significant lower dynamic-range on port #2 (please see picture), is that normal?

I also consider getting a HP 8516A-Testset to go up to 40 GHz, still searching for a good offer (but am not in a hurry). If anybody has one and is willing to sell....

Thanks a lot

Edit: Since people like teardowns, attached a picture of the 20 GHz Test-Set open.

[jjoonathan]

Possibly relevant:

In  the  “High  Forward  Dynamic  Range  Configuration”  (option 
003) the  Port  2  coupler  is  reversed  to optimize  dynamic  range  in  the  forward  measurement  direction.  Because  the  b2  sampler  is  connected to  the  coupler  through  path  instead  of  the  coupled  arm,  there  is  less  isolation  between  the  b2  sampler and  Port  2,  and  the  b2  power  level  will  be  higher  than  bl.  This  configuration  is  better  for  wide  dynamic range  or  reciprocal  devices  or  components  like  filters,  cables,  or  antennas,  where   S12 is  generally  not measured.  However,  as  a  result  of  the  lower  isolation  between  the  test  port  and  the  sampler,  “sampler bounce,” appearing as an occasional spurious response related to the  VT0 frequency, can be noticeable  in  this  “asymmetrical”  test  set.

Keep an eye on the raw directivity figures for the test sets. 8515A (the 26.5GHz one) is a standout in this regard. If you go up in GHz you go down in directivity; make sure the tradeoff computes for your use case.