HP 8569B Spectrum Analyer: sweep issues on larger spans, A18 full multiband assy issue

A quick repair story from a kind contributor (Martin, you can find him at www.mjbrf.com).

(1) I have a unit that generally works OK but has stopped displaying spans above 2MHz/div.

(2) I still get a display on the higher frequency settings e.g. span/div frequency etc etc are displayed at top. However, no trace is displayed.

(3) If I manually sweep using the 100 MHz test signal / comb generator or external frequency source I can still identify the peaks of the signals in the higher span/div settings.

Note that there is a “NARROW” signal used in these machines, which will switch state when span is set above 2 MHz/div. So, best start with looking at all the circuits that are affected by the NARROW signal.


The problem was with the A18 (Full multiband assembly board).
When I scoped out the “over sweep blanking signal” (TP1) the signal was high as soon as the unit was set to >= 5MHz/div.
The signal into the board was OK though.

Anyway, I checked U3B opamp and the signal on the output was stuck high on all the higher ranges.

After inspection of the schematics and downstream circuitry ……
* Removing the multiband board allowed viewing of the higher span/div settings, without the full / multi band and over sweep blanking functionality.
* Isolating the over sweep blanking connector pin from the A18 board to the backplane allowed all ranges to be viewed, but without the over sweep blanking function. Note: The display still looked OK without the over sweep blanking.

Anyway, I changed U3 and U7, but still no luck with the operation. Then I re-checked all diodes and feedback resistors around the opamps.
On inspection, I found that the feedback resistor R29 for U3B (see schematic below) had failed open circuit and was effectively putting U3 into an open loop configuration and thus saturating the output.

The resistor has been replaced and it all seems OK at the moment.

8569b a18 full multi-band assy

This is a quite uncommon fault, on a low-power resistor – maybe a singular fault, or a lot of defective resistors, who can tell for sure.

HP 8663A Signal Generator: switchmode power supply repair

A heavy guest on my bench, a 8663A signal generator. These generators are exceptionally clean, perfect for close-in noise measurement and receiver checks. Still today, hard to find a cleaner source, especially not, if you are on a less than USD 30k budget.

Symptom – easy to describe. Unit turns on, but only briefly, then switches off; over voltage protection light activated at times; a lot of noise on the DC rails when shorting out the safety circuits.

After some probing, the culprit could be located: one of the input capacitors. While this is a common failure mode of other equipment, these caps don’t fail too often for such HP equipment, because only best-in-class components were used, and these are typically run cool, for long life. Still, one of these 32DR6593 SPRAGUE Compulytic caps failed (resistance about 100 kOhm, virtually no capacitance, rapid self-discharge when charged to 50 V for test, framed red in the schematic below).

8663a pwr supply schematic

These were replaced by 81D series Nippon Chemi-Con (Vishay) caps. To call this a successful repair, you might wish to check the ESR specifications – the SPRAGUE had about 0.25 Ohms, the Nippon 81D (680 µF, 250 V) has about 0.198 Ohm, good enough. Note that the 600 µF screw-type terminal caps might still be available, but they are pretty expensive, so I opted to for Nippon Chemi-Con, USD 2.50 per piece, surplus, rather than USD 50 for a pair of screw-type caps. I still think it is a good compromise, because this is not about restoring old equipment, but to make this unit working again, quickly, and at lowest cost.

8663a 32dr6593 data

8663a cap vishay 81d

8663a new and old caps

Some repair is also needed on the A7A3 board – there are 22 µF caps that provide a low impedance DC input to the switching transistor, these are essential for operation (framed green in the schematic). They still work, but were hot, and stressed, possibly overstressed, by the dead main cap. Their can be replaced by any good cap, I use Shiangchen GSA T axial caps, 105 deg C rated. Note that the schematic calls for 15 µF, but 22 µF (measuring about 28 µF) were present in the circuit.

8663a a7a3 pwr supply board

With the power supply disassembled, always a good idea to take out the boards for cleaning, and for re-tightening of the screws holding the various TO-3 regulators in place. Some of these were pretty loose (no wonder, with 30+ years of thermal stress on the boards).

After the repair, add thermal compound to the heatsink/cover – this power supply has a rather critical thermal design. Then, make sure to check the insulation resistance and electrical soundness/earth leakage, which is always a good idea after repair of switchmode power supplies.

HP8568B Repair Story: 249, 275 unlock

Quick story from a contributor, no pictures:

8568B Spectrum Analyzer. When switched on – following error codes were displays:
249 unlock
275 unlock

Findings (in German; in short, there was an issue with thw 275 MHz oscillator, C10 had been replaced before, along with Q5; replacing C10 with a high-quality adjustable cap and increasing capacity a bit, this increasing feedback, the issue could be resolved; root cause: oscillator did not start oscillation reliably when powdered up, because of a lossy C10 cap).
Der Übeltäter war auf der Platine A18
275 MHz Phase lock Oscillator..
Nach diversen Abgleichen fehlte plötzlich
das Ausgangssignal. Als ich das Modul
herausnahm, sah ich,
dass jemand vorher schon herumgelötet hat.
Das C10 war durch einen billigen Keramik-C
ersetzt, ausserdem war ein neuer Transistor
Q5 eingesetzt.
Am Werktisch liess sich zeigen, dass der
Oszillator nicht immer anschwang. Mit
Betriebsspannung ab und wieder anschalten
scwang der Oszillator dann wieder an.
Ich habe das C10 durch einen hochwertigen
Trimmer ersetzt, die Kapazität etwas höher
eingestellt (mehr Rückkopplung)
und der Fehler war beseitigt..

Es ist auch bemerkenswert, wie relativ stabil
der Osz. freischwingend läuft, trotz der
doch einfachen Schaltung.
Thank You Bodo for sharing this.

A13 30 MHz Reference Oscillator: a reasonably quiet oscillator, and a noise cable

A nice little oscillator assembly came my way, supposed to generate about 17 dBm at 30 MHz. Nothing special at first glance, but after checking out its internals, it appeared to be worth a more careful look.

a13 ref osc

A hand-made box, and even more labor intensive assembly work inside. All build by point-to-point wiring, using only the best components available, glass trimmer caps, filters, mica caps – most of these parts are still available today – about 100 USD bom, at least.

a13 upper side

a13 lower side

After a bit of reverse engineering, here the schematic, a modified Colpitts oscillator. Note: base resistor of 2N5109 is 150 Ohms.

a13 schematic

To measure phase noise, connected it to my HP 3585A spectrum analyzer (this is really a great piece of equipment, a bit heavy, but still best of class noise performance and holding this title for the last 35 years….). Connected the oscillator via a 6 dB attenuator, to provide a clean load to the output, rather than dealing with the imperfections of cables, adapters, and the analyzer input.

30 mhz ref osc floor0

Quite shocking, all this noise. The green trace shows the analyzer noise floor. Check, and re-check, still a lot of noise. Too much to be true. After 3 hours of tests, found the issue: a defective BNC cable. Center connector was fine, but both shields were non-connected.

a13 bnc plug

A bit more examination of these cable shows their lousy construction. Not bad for 2 dollars a piece, but you get what you pay for…. the shield is not even reaching to the plug – there is a 5 mm gap from the screen end, to the actual plug. So even if all would have been connected fine, the would still be a lot of leaking, from inside out, and outside in.

a13 rg-58u cable

Notice the BNC plugs – these have a somewhat uncommon construction, the dielectric is covered at the front… not quite according to BNC standard.

a13 bnc cable assy

Clearly visible, the cold solder joint…. Turns out, both ends were open-circuit at the shield.

a13 bnc cold solde

Finally, using a good quality BNC cable (also, using LMR-195 double-screened cable). Looking much better. Noise is down -115 dBc at 10 kHz from carrier. It’s good, but not great. I think one could do better, especially, considering all the pricy parts, and high-quality construction. A good target for a Colpitts osciallator would be better than -130 dBc, at 10 kHz separation.

30 mhz ref osc recheck1

Note the pink trace – this is the bad cable, terminated with a 50 Ohm resistor (with the shield broken at both sides, it is actually a 1 meter wire antenna, with an open-circuit 50 ohm resistor at the end).

Dell Optiplex FX160 – 2 GB – Spinpoint M9T 2 TB: Ubuntu server economizer

Most of my data, manuals, pictures, web sites, databases etc are hosted on a Dell PowerEdge SC1425 (Dual 2.6 GHz Xeon, 2x 3 TB RAID1), which is a great and affordable machine, but has some shortcomings – massive noise, and considerable power consumption; see earlier post PowerEdge SC1425.

Also, it is just one machine, and in case of some severe failure, all data could be lost in case of some unforseeable event like a big thunderstorm hitting the power line… not a frequent issue, but bad enough, if you have to go back to 6-12 months old partial backups.

Therefore, I decided to try a different approach – still use the PowerEdge as the main system, but run a more energy efficient and silent machine for the actual business. After review of various options, a Dell Optiplex FX160 appeared to be a good option. This runs on an Intel Atom 230 @1.6 Ghz CPU, and included 2 GB RAM. The price is right, USD 40 for the unit, including a 2 GB SATA flash drive, and completely fan-less, and noise-less.

fx160 board

For the storage, 2 GB ain’t enough. First, considering some flash drives, but these still come at a hefty price, for 2 TB storage…. A Samsung/Seagate M9T series will be good enough. These are small, silent, efficient 9.5 mm form factor drives.

fx160 spinpoint m9t

A quick test of the read/write speeds:

fx160 read write speed

Read/Write speeds are quite OK, >100 MB/s, plenty to saturate my internet connection, and mirroring the main Poweredge server won’t be an issue anyway – this is all done with the system life, and only by incremental update – not so many files will change from update to update.

Network performance during mirroring from the PowerEdge Server (aka arctur).

fx160  server nload

Quite a lot of power – in such a nice case; “acrux” refers to a star, and is the host name for this unit.
All is run on Ubuntu Server 14.04.3 LTS, including SAMBA (to work as a NAS for various Microsoft Windows based systems, SSH, Apache/2.4.7, and eventually, a printer server).

fx160 acrux