Repairable Spectrum Analysers?

[mag_therm]

I am thinking of purchasing an old spectrum analyser to add to the vintage test gear already on my bench.
I realize that the only reason my old gear ( mostly '70's vintage)  is still working,
 is that I have service manuals and the components are mostly leaded or dip etc.

So I am seeking suggestions from those who repair stuff, for a  basic vintage spectrum analyser that has available service manuals.
Performance specs are not high, I play with shortwave radio, and inverters.

I looked at Avcom HP and Tek. It seems Avcom service manuals are not publicly available .

Thanks!

[Johnny10]

HP 8590 Series still have service manuals and were made for quite a long time.
I see Parts units for under $200.

[Stray Electron]

  The absolute King of repairable Spectrum Analyzers has got to be the HP 141 series. All of the service manuals, parts lists, schematics,etc are on line and they sold about a million of them so parts, accessories and a wide variety of plug in units are readily available. Also because there are so many of them in private hands, there is a lot of help available over the internet.  Add to that that the HP stuff seems to hold up much better with age than Tektronix or most other brands. HP probably also had the widest variety of accessories such as Preselectors than any other brand. Except for size and weight, IMO the HP 141s are your best choice by far.

[mag_therm]

Hi Johnny and Stray,
Thanks very much for the infos!
8590 look good, lots available, but a bit above budget.
A 141T/8552B/8553B is on the way and have downloaded service manuals from Bama.

I might need to add some bracing to the electronics bench!

[vk6zgo]

If you have a Tek 7000 series mainframe, a 7L12 or similar is a possibility.

I know the TE "gurus" treat them with disdain, but I found the 7L12 quite a capable unit. .
That was at work----- I wish I had one of my own to fit the 7613.

At a previous job, we used HP 180 mainframes with a SA plugin (can't remember the model).
They were much more convenient for roadtrips or air travel than the big old 141!

[0culus]

I would recommend HP all the way for this. By far will have the broadest community of folks online who are willing to help when things go wrong or you need a part, besides having service documentation available.

If you have the space, consider an 8568B. 100Hz-1.5GHz, it tends to not be as desirable as the 8566B and thus can be had for cheaper (which is  a shame, because the 68B has it beat in some specs, like phase noise). It will allow you to easily see all your harmonic content for the frequencies you're interested in. Very nice box, and it has two inputs, one DC coupled and one AC-coupled input that's rated to 50VDC (I would not recommend approaching this rating, but it could save you from an expensive mistake).

[David Hess]

If you have a Tek 7000 series mainframe, a 7L12 or similar is a possibility.

I know the TE "gurus" treat them with disdain, but I found the 7L12 quite a capable unit. .
That was at work----- I wish I had one of my own to fit the 7613.

I get the feeling that the Tektronix 7000 series spectrum analyzers are much more difficult to work on because of their form factor.

[Wallace Gasiewicz]

Good choice with the HP 141. Also lots of options.
I repaired one once and the Groups.io bunch was a big help
I also fixed a Marconi 2380 2382...still use it
But it is even BIGGER, imagine that!
Much harder unit to repair but good backup with the Marconi Groups IO
I think the Marconi is a better unit within its limited freq range,  but HP 141 is THE classic and is very good, more parts easily available.
Perfect choice for the Vintage Equipment Buff. Also very usable.

[coppercone2]

unless you already have alot of RF equipment, get a SA with a manual and keep the frequency down

the higher the frequency, and the smaller it is, the harder it is to repair. Do not go into unknown waters unless you are pretty much gambling, have no real need for the equipment and have at least another SA and function generator ready that work on the same frequencies on the SA.

Also, do not attempt this if you don't have a real empty work table to test it on that can be occupied for extended periods of time. Usually its not as power supply type job, its a head scratcher. I have seen like  5+ faults on a single unit before that were all over the place.

Also beware of the hardships of the tin smithing or exorbitant prices involved with repairing the complex electromechanical switches in the front panel of older equipment.

I would say its perilous work to repair SA.  I find it to be an unpleasant job.

[mag_therm]

I fully agree, Coppercone.

I have restored (at least to working) a Tek 466 storage scope and  Wavetek 3000
  500 MHz synth sig gen.

The 466 is a mechanical nightmare toward the front panel. However the circuits are straightforward to diagnose.
I had to use non original concentric pots off another model, and turn some adaptors and parts on the lathe.

The Wavetek is mechanically smart, with with about 8 plug-in copper can modules containing most circuits.
I added a fan to blow over the modules as some were hot inside, with plastic ttl and tant electros.

But Wavetek 3000 has about 4 PLL looping through various modules and is very difficult to find a fault.
Inside the RF modules, the RF paths are of metal can transistors, feedthru caps, and coils cast into  pools of solder.
I don't know how they ever built them.
If ever a failure occurs in there it will be all over.

Today I am making brackets for the work bench.
 I will make a place so the "new" HP141 can lay disassembled for a long time if necessary.

[mag_therm]

The Hewlett Packard 141 spectrum analyser with 8553B/8552B 0 ~ 11MHz / 0 ~110MHZ has arrived.
Initial testing ( Wavetek 3000 gen) indicates that it is fully working over freq ranges  and all controls functional except for one attenuator step.

Some of the knob dials have faded graduations and one die cast handle is cracked.
The RF section will release but will not slide right out.

So overall I am pleased, as it was a low price "For Parts" unit.
Thanks to all for the suggestions.

[0culus]

Enjoy! Cracked knobs and stuff are a way of life with old gear. sphere research is a good source for this sort of thing.

[Wallace Gasiewicz]

Sounds like you got a great deal.
These things are eminently usable.
A trick I use to make numbers on old knobs visible is to clean the knob then put on some "white out" let it dry and you can scrape the excess off with your fingernail.
This works with knobs that have indented numbers/letters/gradations, do not remember if the HP 141 has indented numbers on dials.
The white out yellows with dirt over time and looks "original" also easily undone.
You might want to consult the Groups.IO bunch https://groups.io/g/HP-Agilent-Keysight-equipment for hints to get your RF unit out.

The low frequency units seem to be the most expensive, possibly because of audiophoole use.
Wally

[George Edmonds]

Hi

The modules in a 141 can be a nightmare to remove, the problem is that tend tend to get dropped and mechanically distorted.  You may find that once removed they cannot be replaced, check with a square for trapezoid distortion.

The last one I repaired took well over a day to first get the modules out and then to straiten them up to get them back in the mainframe, the way that the interconnect plugs try to reengage is a good clue as to what is happening.

Standard problem with the 141 is the precision voltage reference failing to strike as a result the crowbar operates and blows the PSU fuses, this can be a very intermittent effect.

George G6HIG Dover UK

[coppercone2]

Hi

The modules in a 141 can be a nightmare to remove, the problem is that tend tend to get dropped and mechanically distorted.  You may find that once removed they cannot be replaced, check with a square for trapezoid distortion.

The last one I repaired took well over a day to first get the modules out and then to straiten them up to get them back in the mainframe, the way that the interconnect plugs try to reengage is a good clue as to what is happening.

Standard problem with the 141 is the precision voltage reference failing to strike as a result the crowbar operates and blows the PSU fuses, this can be a very intermittent effect.

George G6HIG Dover UK

one of my modules in that old HP microwave analyzer without the preselector came out like a banana, thankfully it went back in like a banana and it works. And I bought it direct and did direct pickup from a company everyone here knows, I can't imagine what you can get with the postal services.........

fixing it would be like restoring a 4th dimensional van with panel beating.

Btw you will be disapointed by your bending skills in large equipment, even to do the easy things you need to spend a pretty big sum of money on various steel or aluminum parts to act as weird anvils.. almost feel like you need to weld some custom stuff up if you want to get it really nice without extensive metal bending experience. Its pretty repressive if you just want to experiment with RF without being an equipment nut.

[mag_therm]

Thanks for the useful comments.
For the old projects I have previously repaired, aim was to make it work, even using parts not original or remade etc.

The knobs are not engraved, just smooth aluminum skirts of conical section.
So we had a family competition to see who could do the neatest with a fine point permanent marker.
....Unfortunately I was the winner!

Before trying to get the modules out, today I am welding  frame and lathe turning axles for a trolley so to move it around on the work-bench without dragging it by the handles.

The 8556A (low freq) modules for sale are scarce. I saw one or 2 bundled with modules already in this 141.

[George Edmonds]

Hi

Having read my own previous posting I realized that I may have given the wrong impression.

The problem with the plugin modules is not caused by damage to the main frame, in shipping or elsewhere.

I is caused when the two plugin modules are separated from each other.  To do so the RF module has to be removed vertically from the IF module after operating a latching mechanism.  At this point the RF module appears to have a great desire to jump off the bench and say hello to the floor, forgetting that the floor is frequently a hard place to visit.

Hence the damage to the RF modules.

In pre Covid days, which are now but a dim and distant memory, when Amateur Radio events still took place in the UK, 141's with multiple RF modules sold for around the $250 US mark in the UK.

Great Idea to make up a mobile mount for it.

Although old and large it is still a very nice SA and good bet to use as it has NO digital manipulation of the results.

George G6HIG Dover UK

[mag_therm]

Hi George,
The trolley is working well. It reduces the flexing and scraping of the feet. I will put up some photos maybe in the sticky when it is finished and painted.
Actually I might make another trolley for the sw receivers and smaller scopes etc here.

The 141  modules pull forward just enough to open the connectors then the is a click that sounds like something spring loaded, or maybe a rf grounding strip hanging up.
From some tags inside this was a Navy unit, and one tag says that the RF/IF can only be used with this particular main frame serial number.
I wonder if it is locked in?

Today I will search the manuals for some diagrams. I have to take the right side diecast frame off anyway to repair the broken handle.
I think I will use a splint with epoxy and small screws into tapped holes.

Thanks for the notes.

[George Edmonds]

Hi

The tag is probably to do with maintaining the calibration between the IF and the mainframe.

Cannot recall a grounding strap, but it has been about two years since I last worked on one

From what I remember the module latching system is not of the best design

George G6HIG Dover UK

[mag_therm]

Modules are out now.
Cause was a missing screw on back wall, allowing it to angle over, binding on the big pins.
All neat and clean inside.

[Stray Electron]

Mag,

   Just out of curiosity, how old is your HP 141?  The first two digits of the serial number added to 1960 should tell you the year that it was made. The 3rd and 4th digits tell you what week. The next character should be a letter and that will tell you where it was made. I suspect that will be an 'A' for America.

[0culus]

Small correction: the first two digits tell you the year of the engineering revision. It may be close to the actual year for many instruments because there were constant revisions being made.

[mag_therm]

Serial Numbers
141T  :1337A 08619
8553B: 1215A06638
8552B: 1505A08169

Navy Metcal Program calibration tags:24 Mar 1986 & 24 Mar 1988

Now I am milling a flat across the fracture in the diecast of the lower front handle.
Then  to add a splint of 6061 epoxied and screwed.

[mag_therm]

HP 141T : Here is the splint I added across the broken handle connection.
Next is to flatten the bottom cover, straigthen a cross bar along lines mentioned by coppercone.
Then put it all back together.
It won't be handsome but hopefully will be handy.