Rohde & Schwarz SMT03 Repair
RF Signal Generator
Part 1
– Specifications, Features and Overview
Specifications:
Features:
Introduction:
5kHz is audio, right? 1KHz certainly is and the Rohde & Schwarz SMT03 is specified down to 5kHz at full output and down to 1kHz with reduced output levels and at the other end its bandwidth extends out to 3GHz, a ratio of nearly 1 million to 1!
For almost all audio work our Audio Precision test sets 200kHz bandwidth is totally adequate, but these days with the plethora of electronic devices in the environment equipment interoperability or Electro-Magnetic Compatibility (EMC) is a key a requirement and the ability to test at RF frequencies is now an important part of the tool kit
In almost all countries radiated and conducted emissions testing is requiremed to be allowed to sell a piece of electronic equipment and in many countries further restrictions are placed on a piece of equipment’s ability to function correctly under the influence of external radiated and conducted electromagnetic fields. Known as susceptibility testing, an RF signal source is an essential part of this testing
The RF source is often modulated to simulate the effects of modern communications systems and testing is sometimes required into the GHz range
The standards published in many jurisdictions are minimum requirements and many manufacturers of electronic equipment will go further to ensure a interoperability in electromagnetically noisy environments as no one wants cell phone modulation intruding into their audio mix
While the path to compliance with EMC standards is via EMC lab a lot of the preliminary work can be performed on the work bench using the appropriate tools. One of those tools is an RF signal generator that be used to help design filters to block electromagnetic interference and test a designs robustness
Overview:
The Rohde & Schwarz SMT03 is a 5kHz to 3GHz RF signal generator featuring a variety of modulation modes including AM, FM, φM and an optional low frequency oscillator can be used as a pulse modulation source or for frequency sweeps. It has a mechanical attenuator able to output signal amplitudes from -144dBm to +13dBm over all frequencies and up to +16dBm over a more limited range
A large graphic LCD display features an easy to use, menu based system for parameter selection and setup
Other options include a choice of front or rear mounted connectors, rack mounting kit, stereo FM generation
Importantly for EMC immunity testing the low frequency generator can sweep the carrier over a specified range in percentage steps
This particular unit has the low frequency generator & front panel main output connector along with the rack mount kit although this has been removed. It dates from 1995
A full set of specifications is too detailed to be included here but the specification sheet can be found at R&S SMT03 Spec Brief for those who are interested
Construction:
The SMT03 is a fully modular design, a cast aluminium alloy main frame chassis contains a small motherboard that supplies power, digital and low frequency analog signals to the installed RF modules. The RF modules are each enclosed in metal shields and are inserted like cassettes from the top of the unit on plastic slides In addition to their connection to the motherboard the RF signals are routed via a series of hard-line and flexible SMA or SMB cable assemblies both between modules and the external connectors
Modules fitted to the SMT03 based configuration include, from front to back, the RF synthesizer, 1.5GHz and 3GHz output modules along with the optional low frequency generator (not visible)
SMT03 Top side showing RF cassette modules
SMT03 Underside showing motherboard and interconnects
Apart from the RF cassettes, the mechanical attenuator is fitted to the chassis immediately behind the output connector along with a small control PCBA and the modular power supply is attached to the rear of the signal generator. It connects to the motherboard via a 50 pin D-sub connector. Finally the front panel module containing the microprocessor, display and controls is inserted from the front of the chassis connecting to the motherboard via a set of ribbon cables
The modular construction allows R&S to use the same chassis for several different RF products simply by swapping in different modules. The power supply is also common across several device models and appears to sourced from an external supplier
The general grade of construction is excellent, within the RF cassettes, extensive use of shielding between sections, shielding gasket help to provide isolation between different circuits. PCBA’s almost entirely utilise double side mounted surface mounted components with the top layer reflowed and the bottom layer waved. Several LSI chips appear to be of R&S proprietary design The LCD display is fitted with an EMC shielding glass to prevent stray RF exiting the unit
Circuit & Design:
The SMT03 synthesizer cassette generates signals from 67.5MHz to 1.5GHz that can be modulated by FM or φM via a single loop synthesiser operating from 750MHz to 1.5GHz with other frequencies generated by a PLL with a fractional divider
The output of the frequency synthesizer is sent to the 1.5GHz output module that provides AM modulation and a set of low-pass filters to produce the lower frequencies down to 5kHz through down conversion by mixing with a 600MHz signal. The signal is amplified and levelled with the values of a look up table held in RAM in the front panel microprocessor module
The 3GHz module doubles the frequency from the 1.5GHz cassette, passes the signal through a series of band-pass filters and amplifies and levels the RF amplitude as above
The Low frequency option cassettes uses direct digital sythesis to produce low frequency signals of different wave shapes up to 500kHz for frequency sweeps and pulse modulation
An extensive on-board monitoring system is utilised front panel monitoring of many test points within the generator
Automation is via GPIB or RS232. External triggering and oscillator connectors are available on the rear panel
1.5HGz Output Module
Part 2
– Repair & Restoration
Strange power-up behaviour:
The signal generator had been performing well for a years before developing and fault where, after switching the unit on, after exactly 3 minutes the beeper sounds very faintly for 10 seconds and the screen becomes dark. Once the buzzer stops, the screen then has a solid image of many, mostly horizontal, lines. The screen then remains like this and the beeper repeatedly sounds at regular intervals
The fault was traced to a failed jelly bean 74ACT573 transparent latch gate with an output stuck low and with the IC replaced everything worked perfectly again!
Front Panel PCBA with Intel i960 system microprocessor hidden under the 50-way ribbon cable
Power supply module at rear of chassis
Capacitor trouble:
A few more years passed without incident before a second fault appeared, the unit simply wouldn’t power up. After probing the power supply rails on the motherboard and finding no voltage, this fault looked like it might be easier to solve being a power supply problem
While power supply problems are often relatively straight forward, on the SMT03 the power supply is a complex unit and built into a very small form factor for its output capability using a network of of PCBA’s interconnected in a variety of devilish ways
There are seven different power rails and none of them were working properly with the exception of the standby supply that avoids the need to have the mains supply switched from the front panel and allows the instrument to be controlled remotely
After getting the power supply module apart a visual inspection showed little that was unusual apart from one of the electrolytic capacitors had a can that had buldged slightly at the top
Not immediately expecting a capacitor failure to be the cause of a complete power supply failure some time was invested in trying to fault find the problem but without success
Further inspection showed that it wasn’t just a single capacitor that had bulged but many…
Non-isolated (mostly) primary supply side
A couple of the suspect capacitors were removed from the PCBA and underneath some of the liquid capacitor electrolyte had escaped, in one case causing some corrosion of the underlying copper traces
There was only thing for it, remove all of the capacitors!
Once all of the capacitors were removed from the circuit boards the true nature of the power supply failure became apparent, in several areas of the PCB the resist covering the copper traces or the copper traces themselves were partially corroded and near the 20-way ribbon connector a narrow trace controlling the state of the power supply had been completely severed
It was noted that all of the failed capacitors were Nippon Chemicon LXF series, long life types rated for 105C
Unfortunately, the SMT03 power supply contains mostly LXF capacitors, 34 of them!
All of the LXF caps were tested for leakage and ESR and even though not all of them showed any defects it was prudent to change them all for new capacitors. Wanting to replace the capacitors with similar parts, the newer equivalent Nippon Chemicon LXY series was selected for the job
The PCB was carefully cleaned of all traces of the electrolyte and any areas that showed signs of resist corrosion were re-coated with a protective film. Any copper traces that showed signs of corrosion were also repaired
Isolated (mostly) secondary supply side
During capacitor removal some adjacent parts received a considerable dose of heat from the rework and looked the worse for it and so these were also replaced. Additionally a handful of small, low impedance SXE type capacitors were tested and showed higher than normal ESR and also replaced. The remaining very large, bulk smoothing KME type capacitors tested OK and were returned to service
It is thought that part of the reason for the capacitor failures is due to the very compact design of the power supply that relies on forced air to keep parts cool. The small capacitors are tucked into tight corners surrounded by power devices radiating heat where airflow would not readily penetrate and this probably contributed to the failures
On the other hand, the much larger, bulk smoothing capacitors are all placed around the periphery of the boards and would receive much better cooling in additional to only partly being in the radiant field of hot, power components as well as themselves having a much larger surface area to dissipate their own heat, so these parts seem to be far less stressed and still serviceable
It appears that there was a known problem with several manufacturers capacitors called the Quaternary Ammonium Salt issue that affected, amongst others, Nippon Chemicon LXF series capacitors. In response to the problem, and after it had been addressed, Nippon Chemicon issued a statement that their newer series capacitors do not contain quaternary ammonium salts: NCC QAS Statement
Parts replaced during power supply repair!
Part 3
– Calibration and setup
Calibration:
Using the internal calibration functions of the SMT03, basic calibration was successfully completed
However, without the official Rohde & Schwarz test program only one of the advanced calibration functions can easily be performed, calibration of the internal 10MHz frequency reference
This is performed by using an known external frequency reference to a frequency counter and comparing the SMT03 10MHz. By tuning the the SMT03 internal oscialltor using the internal reference oscillator calibration utility a frequency exactly equal to the external reference can be acheived
In this case an ovenised, 10MHz reference oscillator is used after allowing a couple of hours warm up to allow it reach temperature equilibrium with the surround ambient temperature. Equilibrium can be discerned when the frequency stops drifting and the oven current draw stabilises. The oscillator is powered from an low noise, linear power supply regulated to exactly 12VDC
In the case of this particular instrument, the SMT03 internal reference oscillator offset needed to be set to ‘2236’ at a nominal ambient temperature of 20C
Following checks on the other basic specifications, the SMT03 signal generator is once again fully functional and should give many more years service!
SMT03 during internal reference oscillator tuning
Further Information:
For more information about the repair and restoration of Rohde & Schwarz test equipment please check out: http:/www.bymm.de
In German, but with loads of detailed information for anyone interested in maintaining these great pieces of test gear
Big thank you to Marc for his help, humour and inspiration with getting the SMT03 back on the road!