Monday, December 10, 2018

New Toy On The Way **Updated**

My trusty old Systron-Donner Model 1702 signal generator took a hit to the front panel on the move out here. 100% my fault for not securing a 75-lb item properly, allowing it to slide around on the floor of the trailer. This one doesn't have rack handles, so there's about 18" of front panel that's pretty exposed, and the knobs slammed into something on the trip out here.



The level adjust pot for an external modulation source sheared off, and the level adjust pot for the internal modulation source is bent, and spins freely. These pots had 1/8" shafts on them, so I'm pretty sure they use a sub-miniature "instrumentation" sized pot, which used to be readily available. If Digi-Key/Mouser don't carry them, then Google is my friend, but these were standard parts for decades.



I'm just really, really, happy, happy that the thumbwheel switches used to set the output frequency survived unscathed. These things were unusual in their day, and even the surplus dealers are probably out of them by now.



So, *if* I can get the parts, is it "worth" repairing? These generators typically command $750~$1k on the used market, a bit less than I paid for this one almost 20 years ago. And they're ruggedly built, rock solid, low phase-noise (quiet!) generators that seem to never wear out or fail, unless some dolt doesn't pack it properly for shipping.....

So, yes, to me it's worth fixing. Even if the pots cost $25 each, I'll have a good signal generator for fifty bucks!

BUT.......what to do in the meantime.......?

Why, get another generator, what else? Besides, there are some very valid (and interesting) RF measurements that require two generators.

I wound up with this one, an HP 8657B, OPT1, OPT3. Option 1 is the world famous HP OCXO ovenized crystal oscillator, called the "High Stability Timebase", and Option 3 is improved noise performance/switching time in certain kinds of pulsed modulation, something I doubt I'll use.



There were several available in my price range (< $1k), but this one had several advantages.

First, the installed "High Stability Timebase" is about the best crystal oscillator HP ever made, and they were used in many, many critical applications where an "External Reference" was not available. In some factories I've worked in, it was the highest precision 10MHz source on-site, and wound up being distributed throughout the plant as the "Master Reference".

It's also the "heart" of the various GPS Disciplined Oscillator designs out there that not only give you the superb short-term stability and spectral purity of a crystal oscillator, but also the long-term stability of an "Atomic Clock". That's a whole 'nother post that I'll be doing since I want to get my GPSDO running again to use for my own 10MHz Master Reference. This way my frequency counter, signal generator, and spectrum analyzer will get the enhanced accuracy and stability of a high performance (or used to be) Master Reference.

Secondly, it had the standard front panel output jack. Many of these instruments were used in Automatic Test Equipment racks where everything was under PC control on the HP-IB or IEEE-488 bus, and the operator was along for the ride, and to load/unload modules and hit "Enter" when prompted. All the normal front panel RF out put jacks were "optioned" out to the back panel because the "Back of the Rack" was where everything happened in ATE, signal routing wise.

Thirdly, the place is an ISO-everything certified Calibration Lab, and the instrument has been tested for 100% functionality, and comes with a fresh calibration.

Oh, and it came with FREE shipping, a $75~$150 expense otherwise.

And I (finally) finished the towel rack project in the upstairs "wife's" bathroom.


One down, countless to go.....

Update on the generator..... 

The generator has an output corresponding to where the thumb switches are set, and the output attenuator (think 'Volume Control')  works correctly.

The control with the bent shaft also appears to work, at least according to the display that goes active when you select that function.

However......when I tried to look at the generator output on my Spectrum Analyzer, I see.....nothing. I tried using a different signal source, which I can read on my frequency counter, but I can't see that signal, either.

And finally......I can't even see the built-in 100MHz "CAL" signal, which also drives the frequency counter.

If you can't see the signal from the built-in CAL source, you're SOL. Looks like I'll be buying another Spectrum Analyzer sooner than I thought...... 

14 comments:

  1. Well, I understand the towel rack. These rest?

    ReplyDelete
  2. I actually got both, but am most impressed that you had the foresight to mount the towel rack on wood.

    ReplyDelete
    Replies
    1. One of my previous posts had the reason for the wood strip.

      There had been 4 or more towel racks installed there, with 4 sets of drywall anchors all drilled and pushed into the wall.

      4 sets of screw anchors in that small an area means it was all plastic bits and Spackle, and it fell apart, and back into the wall, when I took the rack down to see why it was always loose on the wall.

      Yep....fist-sized holes in the wall, so the board covers up the two holes.

      Delete
  3. I just had to say, I recently relieved myself of ownership of a 60's vintage Siemens single channel scope I bought in 1972, a Heathkit dual channel scope and a kit built signal generator, all of which had sentimental value but no practical use for me and I needed the shelf space.

    ReplyDelete
  4. I have a good scope, but I just found out last night my spectrum analyzer is dead.

    ReplyDelete
  5. If it's fixable, good. If not that HP unit is a find!

    ReplyDelete
    Replies
    1. Well, I can read the output on my counter, and it agrees with what the thumbswitches are set to.

      Sucks that my Spec A is dead because I wanted to see if the modulation pot still works. When that pot with the bent shaft is active, turning the knob shows the modulation level is changing. Guess I'll have to fire up a receiver to see if I can hear the modulation level change.

      Delete
  6. I have the next sig gen down, the 8656B. It has been good enough for what I need. Really good find on the one with their best OCXO, though!

    ReplyDelete
    Replies
    1. I really wanted to get an 8662A, but they're several hundred to several thousand bucks more expensive, depending on options.

      The 8662A is considered one of the best HP signal generators ever made in terms of spectral purity and other noise. The cheapest one of those I found was $1200, going up to about $4k for one that looks like it just came out-of-the-box.

      If I'm gonna cough up $4k for a piece of test gear, it'll be for an HP Spec A!

      Delete
    2. Yeah, I've got hours upon hours in front of an 8662/8663.

      I'd like to get my hands on one of those Siglent or Rigol analyzers to see if they're good, but if the budget really was $4k, I'd look seriously at the Tektronix Real Time Spectrum analyzers. I test flew one of the RSA306B analyzers, which is the lowest BW one they have, 6.2 Gigs, and priced around $3900. They need a pretty kick ass laptop because it does a lot of work. OTOH, the one I had was every bit as good as the ~$100K RTSA we had in the lab, just not as wide BW.

      Mine is just an 8558B. That needs work.

      Delete
    3. I pulled the trigger on an HP 8594E the other night. Got real lucky, and found a super clean one for "only" $988, shipping included. It "only" goes to 2.9GHz, but it's an HP, which makes me smile.

      I was able to get the Tek 494 Spec A up and running. The "problems" I thought I had were all Operator Induced, and it actually works and shows the internal 100MHz, -20dBm CAL signal.

      Maybe it's just me, but I always thought Tek analyzer control functions worked "backwards" from HP, and I've made the same mistakes many times going back and forth between HP and Tek Spectrum Analyzers.

      Delete
    4. Oooh! Congratulations!

      Spectrum analyzer lust. I could really go for an 8594E. Or something newer and better than what I have.

      Delete
    5. After using HP analyzers for DECADES, I just can't get used to the bass-ackwards controls on the Tek.

      I had a Tek 494, which at least had a keypad selectable frequency input, similar to an HP, but like a dummy I sold it, and kept the 494 because.....26GHZ!!!

      And I have a nice Kron-Hite function generator with a blown power supply (hope it didn't take anything else out), and a nice HP 100MHz RF sweeper that needs to be gone through.

      And ALL of this stuff needs the covers removed, and dirt blown out. My HP microwave frequency counter might need a new fan too.

      Delete

Keep it civil, please....