Repair : HP 8753A Network Analyzer

[hendorog]

Now I just need to fix the supercap/battery backup on day.

It would be a good idea to double check that this is actually faulty before replacing it (which is not what I did!) - some of the normal behaviour of the 8753 makes it appear like the supercap is faulty. It's a bit of a pain to replace it and then find that nothing has changed...!

For example saving the instrument state doesn't save the calibration. I think the display settings are also lost and need to be saved and recalled using the option in the display menu.

[Howardlong]

Now I just need to fix the supercap/battery backup on day.

It would be a good idea to double check that this is actually faulty before replacing it (which is not what I did!) - some of the normal behaviour of the 8753 makes it appear like the supercap is faulty. It's a bit of a pain to replace it and then find that nothing has changed...!

For example saving the instrument state doesn't save the calibration. I think the display settings are also lost and need to be saved and recalled using the option in the display menu.

Hmm, maybe it's simple operator error then. Those are the exact symptoms. So you need to recal and reset display settings each time you switch on... as I've been doing for the past 15 years, assuming that the battery was dead?

[hendorog]

Now I just need to fix the supercap/battery backup on day.

It would be a good idea to double check that this is actually faulty before replacing it (which is not what I did!) - some of the normal behaviour of the 8753 makes it appear like the supercap is faulty. It's a bit of a pain to replace it and then find that nothing has changed...!

For example saving the instrument state doesn't save the calibration. I think the display settings are also lost and need to be saved and recalled using the option in the display menu.

Hmm, maybe it's simple operator error then. Those are the exact symptoms. So you need to recal and reset display settings each time you switch on... as I've been doing for the past 15 years, assuming that the battery was dead?

Thats correct, you probably already know about this, but you can get around the recal annoyance with a GPIB adapter and KE5FX's VNA util.
http://www.ke5fx.com/gpib/readme.htm#VNA

[Howardlong]

Now I just need to fix the supercap/battery backup on day.

It would be a good idea to double check that this is actually faulty before replacing it (which is not what I did!) - some of the normal behaviour of the 8753 makes it appear like the supercap is faulty. It's a bit of a pain to replace it and then find that nothing has changed...!

For example saving the instrument state doesn't save the calibration. I think the display settings are also lost and need to be saved and recalled using the option in the display menu.

Hmm, maybe it's simple operator error then. Those are the exact symptoms. So you need to recal and reset display settings each time you switch on... as I've been doing for the past 15 years, assuming that the battery was dead?

Thats correct, you probably already know about this, but you can get around the recal annoyance with a GPIB adapter and KE5FX's VNA util.
http://www.ke5fx.com/gpib/readme.htm#VNA

Pretty much all I use that GPIB toolkit for is screen shots, I should try harder!

Thanks for the tip.

[TiN]

Got the reinforcements today:



Color LCD kit NewScope-9 from Simmconn Labs.

Money shot on controller board:



Good old Spartan-6 FPGA, die of memory, Winbond SPI Flash ROM with bitstream, few 384 buffers and 3.3 Vreg. It takes digital data from 8753A display bus and translates it into LVDS stream for TFT panel.
Of course there is more to it than just interface converter, as even original CRT module in HP 8753's has own VRAM, processor, chargen and vector to raster engine.

Kit comes with nice industrial grade panel and crappy CCFL HV inverter. That dodgy Jicon capacitor there has to go.



Now need to find time to install it into unit and see how it goes. Bit not easy, as whole VNA is teared apart and have it's guts spread all over the apartment  .

[TheSteve]

Ditch the entire CCFL and go to LED's.

[TiN]

Yey, it's alive.



Well, sorta...

Unknown-condition 8753A with rusty boards.. Check..
LCD kit working with color... Check...
Replacement HP N-SMA connectors 4pcs.... nope..200$USD?
Hardware and knowledge to test functionality of VNA....nope....200-1000$ for attenuators, couplers, loads?
S-parameter test kit?....nope....1500$
Calibration kit for Type N up to 3G...nope....500$?

Uh-oh...
I'd rather get another rusty 3458A then for time being! 

[hendorog]

Yey, it's alive.

Awesome - and its better than new cos its in colour!

Right, here we go, VNA on a budget is my new speciality

Replacement HP N-SMA connectors 4pcs.... nope..200$USD?

These might be another option - i.e. dump N and use SMA instead. I would have saved a lot of money in adapters if everything was SMA!

http://www.ebay.com/itm/HP-Agilent-Bulkhead-Connector-3-5mm-26-5-GHz-/381775153701?hash=item58e3928a25:g:u8gAAOSw37tWBFtm

S-parameter test kit?....nope....1500$

T/R test set is a much cheaper option here, the only downside is you have to reverse the DUT yourself. There are a couple of upsides too - more flexibility and very slightly better specs due to a simpler signal path. They go for around $200-$300 normally. You want the 50 ohm 85044A model not the 75 ohm 85044B.
Here is a really cheap one which is a bit beaten up ($75):

http://www.ebay.com/itm/HEWLETT-PACKARD-85044A-TRANSMISSION-REFLECTION-SET-300KHZ-3-0GHZ-/201693688646?hash=item2ef5e17346:g:GxUAAOSwMHdXRfqb

You need a precision adapter for the test port, which are not that cheap. If you use a cheap adapter instead on the test port then the adapter will become the limiting factor in the (uncorrected) directivity of the system. The adapters and cables for the other ports are not as critical.

Building a T/R test set is another option and is pretty straightforward. The uncorrected directivity isn't as good - but an OSL calibration corrects for directivity. I've got a couple of PCBs I can send you that you just need to plop connectors and some parts on.

Hardware and knowledge to test functionality of VNA....nope....200-1000$ for attenuators, couplers, loads?

Splitters are easy to build, couplers are harder to build but are available on ebay pretty cheap. I have a couple that have excellent performance considering they were never designed to be used in a VNA!

Get some SMA terminators, attenuators and test cables from eBay. Reasonable quality stuff is cheap enough.

Calibration kit for Type N up to 3G...nope....500$?

I've got one of these ($71):
http://www.sdr-kits.net/Webshop/products.php?50&cPath=5

[Howardlong]

Definitely agree with standardising on SMA rather than N.

I have the 75 ohm T/R test set rather than the S parameter test set, and use it with a minimum loss pad to convert it to 50 ohms with some loss of dynamic range.

Don't get too uptight about super expensive cables, the unit will cal out any reasonable cable losses and minor mismatches.

The key is to have it calibrated at the DUT, and that often means making your own cal kit and test jigs if you're unit testing a subsection of a PCB. Having a proper cal kit though is pretty much essential to ensure that the VNA is working correctly as well as testing those devices which happen to terminate in your cal kit's connectors.

[TiN]

HP-Agilent-Bulkhead-Connector-3-5mm

Ordered 4pcs.
It's not an SMA however, it's 3.5mm
So I think I'll still need to get 3.5mm-N cables in future to connect regular N-type test set.

[hendorog]

HP-Agilent-Bulkhead-Connector-3-5mm

Ordered 4pcs.
It's not an SMA however, it's 3.5mm
So I think I'll still need to get 3.5mm-N cables in future to connect regular N-type test set.

They are compatible with SMA. I was just too lazy to type 3.5mm (now look what you made me do
http://na.support.keysight.com/pna/connectorcare/What_mates_with_what.htm

The other option is to use SMA-N adapters when you get the test set. At least you will only need 3 (or 4) now

I have the 75 ohm T/R test set rather than the S parameter test set, and use it with a minimum loss pad to convert it to 50 ohms with some loss of dynamic range.

Do you need 75 ohm N connectors for these with the larger centre pin? Also, do they top out at the 2GHz spec or can you use them a bit  beyond that? I've observed that the A model doesn't go much beyond its 3GHz spec.

[Howardlong]

I have the 75 ohm T/R test set rather than the S parameter test set, and use it with a minimum loss pad to convert it to 50 ohms with some loss of dynamic range.

Do you need 75 ohm N connectors for these with the larger centre pin? Also, do they top out at the 2GHz spec or can you use them a bit  beyond that? I've observed that the A model doesn't go much beyond its 3GHz spec.

No need for 75 ohm N connectors other than the one that's part of the 75 to 50 ohm minimum loss pad connected to the test set. Everything else is 50 ohms. I leave the MLP permanently connected to avoid self-inflicted operator error. You do lose 5 or 6dB of dynamic range as a result of using the MLP. There is nothing practically to stop you running the 75 ohm TR test set to 3GHz, and I've never noticed any particular problem in doing exactly that.

The 75 ohm test set came with my 8753A when I bought it around 2000, and I never bothered to change it.

[orin]

Yey, it's alive.



Well, sorta...

Unknown-condition 8753A with rusty boards.. Check..
LCD kit working with color... Check...
Replacement HP N-SMA connectors 4pcs.... nope..200$USD?
Hardware and knowledge to test functionality of VNA....nope....200-1000$ for attenuators, couplers, loads?
S-parameter test kit?....nope....1500$
Calibration kit for Type N up to 3G...nope....500$?

Uh-oh...
I'd rather get another rusty 3458A then for time being! 

I got quite a long way with an 8753C with these and a few adapters.

The couplers came from ebay, the splitter and attenuator from Mini Circuits directly, but I'd be surprised if you couldn't find them on ebay for less.  The couplers are only rated to 2GHz - I don't recall how well they did to 3GHz.  If I get a chance, I'll do an S11 calibration and look at the reported directivity.

To do the adjustments, you'll need the proper 11667A splitter, 8482A power sensor and 436A or 438A power meter... it starts getting expensive!  436As are cheap enough, but good luck finding an 8482A that works.

Having said that, as long as you didn't mix up the samplers, the adjustments saved in the EEPROM are probably OK as long as you do a user calibration with a good cal kit.

I was thinking of building up my own test set, but then an 85046A came up on ebay that I got for $350.  Port 1 fails the directivity test, but I can replace the bridge with one from an 85044A that I got for $120.

[TiN]

Okay I won, that HP 85044A and it's shipped. Shipping is 92$, while box itself 75$  .

Also 3.5mm bulkhead connectors arrived. Two are decent shape, I'd say 5- and two other are okay 4). Should be good enough for 3GHz VNA, as these rated 26.5Gig originally .
Will post macro photos later.

However I don't have suitable cables yet.

[hendorog]

Okay I won, that HP 85044A and it's shipped. Shipping is 92$, while box itself 75$  .

Also 3.5mm bulkhead connectors arrived. Two are decent shape, I'd say 5- and two other are okay 4). Should be good enough for 3GHz VNA, as these rated 26.5Gig originally .
Will post macro photos later.

However I don't have suitable cables yet.

Heh, welcome to the shipping company shareholders club.

If you can cope with having the test set directly in front of the VNA then these jumpers look pretty good:
http://www.ebay.com/itm/MARCONI-MICROWAVE-12cm-SMA-MALE-to-SMA-MALE-COAXIAL-LEAD-/351862994501?hash=item51ecab3645:g:jo8AAOSw8gVX8lpj

[TiN]

Well, the HP 85044A test set still has N-type ports. And that pesky APC7...

Oh well...

Promised photos of connectors. I'd appreciate if someone knows their original HP/Agilent part-number.



I bought 4pcs. N-SMA on the right is original from 8753A.



Two connectors are decent shape, two other bit worse. But nothing too bad.

Close-up of inner collet:



"Air" isolator, clearly visible by backlit photo on the right. Of course it's not air in full meaning of the word, but most of gap space is empty, hold just by the transparent plastic spacer.

To actually install them into 8753A's frame, some metal work would be required, as new connectors have only one notch on threaded mount, while old N-SMA had two notches on both sides.



I'll just mill excessive metal on the frame with Proxxon

And everything would look somehow like this:



Why all this trouble? Here are macro of original N-type ports:



Probably not so bad as most VNA-nuts would say to me, but I don't like that crud and rust on collets and housings. And you can see dents on the front outer shell, likely due to physical drop impact. All GHz's will run away 
Hopefully this would be enough to get something going, once I receive that 85044A. I'll get some cheapo N-SMA cables, and maybe APC7 to SMA or N adapter.

[vtp]

Probably not so bad as most VNA-nuts would say to me, but I don't like that crud and rust on collets and housings. And you can see dents on the front outer shell, likely due to physical drop impact. All GHz's will run away

OK, I will be then the first to say that those do not look too bad. You do know that the connectors can be disassembled completely, even the inner collet visible on your picture can be removed from the dielectric support? Would be a lot easier to clean those disassembled.

EDIT: If you do not want to clean those N-SMA connector blocks they are readily available for example on epray. The same bulkhead adapter has been used in a huge number of HP/Agilent instruments.

As for standardizing for SMA I'd rather use N connectors on instruments and then convert to SMA or similar with a good quality adapter. VNA needs to be calibrated anyway at the end of the connecting cable or adapter for measurements.

N lasts pretty much forever, SMA not so well at all. And APC3.5 is really easy to destroy with sub-standard SMA connector - all it needs is to have the center pin on male SMA connector to protrude too much (think of cheap chinese cables and adapters) and the APC3.5 center collet is a tiny flower after one insertion/removal cycle.

As for calibration kit, I would suggest brand new from any manufacturer that comes with actual coefficient data. I bought one "good" used 85033D from a cal lab and it was a complete disaster on what comes to female standards that get used the most. For the sake of peace of mind, I have a brand new 85033E from Keysight now.

[hendorog]

There are also these SMA-N if you choose to stick to the original:
http://www.ebay.com/itm/272427568928?_trksid=p2055119.m1438.l2649&ssPageName=STRK%3AMEBIDX%3AIT

I'm just using reasonable SMA-SMA cables and a lot of cheap SMA-N adapters on mine. The adapters are a risk but I couldn't find decent matching N-N cables or N-SMA cables at a good price.

You will want a precision APC-7 adapter though, otherwise the nice directivity of the 85044A is degraded by the reflections from the adapter. The one I have is this one (same seller):
http://www.ebay.com/itm/7mm-to-SMA-f-Adaptor-OMNI-SPECTRA-OSM-2082-2700-00-/252575012486?hash=item3acea4d686:g:T20AAOSwbsBXiy9z

The first one I received had been dropped and had a 'flat spot' on the SMA connector which affected the return loss. The seller sent me another one FOC which is good.

[TheSteve]

Came across a really cheap 8753A today, so I am joining the club!

[Theboel]

if any one has 8753 family with whatever ugly looking condition and willing to send it to overseas  please let me know

[TiN]

Be careful, I'm this close to accept your wishes . Just looking at RF cables, calibration kits and APC7 or test set prices make me sick already...

[Theboel]

TiN I understand it very well and I watching my messages box very impatient   
 

[TheSteve]

Mine may be for sale at some point. Assuming it works or I can get it working I will be on the lookout for an affordable test set and the 4 jumper cables. Overseas shipping would be expensive but I do have access to a pretty heavily discounted UPS service.
I'd like the LCD upgrade for it but that is already twice what I paid for the 8753A including shipping. I already have an 8714C to compare it to so it should be a fun project either way.

[Nemo1956]

I have a 8753C where the tube is going flat now and very dark. So where did you get this LCD display?
Is it a simple up grade where can I find out about this please.
Thanks.

Paul

[TiN]

Welcome, Nemo.

All info was linked in my article , LCD kit was bought directly from Simmconn labs, model NewScope-9. It was quick and easy deal.