Marconi TF1370A Wide Range R-C Oscillator

[Phil G4SPZ]

Some Forum members enjoy restoring and collecting signal generators, so I thought I would write up my most recent tussle with another big valved Marconi unit, particularly as less than 1 in 40 Success Stories involves vintage test gear. The TF1370A is not a common model, and it also had a few interesting faults that took some tracking down. It is an AF/RF generator dating from 1965 covering 10Hz to 10MHz in six switched ranges, providing sine and square wave outputs from 1mV up to 30V RMS and peak respectively.

My approach to vintage test equipment is to overhaul and repair, rather than restore. Test gear would have been routinely repaired by the technicians and engineers using it, and past repairs are therefore part of its history. I therefore make no attempt to disguise new components by re-stuffing or other techniques, and simply make sure that my soldering and workmanship is as professional-looking as possible. Old test gear therefore appeals to my lazy nature, by offering maximum fault-location challenges with minimum cosmetic or cabinet-restoration overheads! I also try to recalibrate the equipment so that it can be relied on to perform reasonably accurately in subsequent use in my workshop.

A friend from the local Amateur Radio Club donated the TF1370A to me, and on arrival it was dirty, but undamaged and otherwise complete. On internal inspection it was quickly obvious that Ham-Fisted Harry had been in there before me, aided and abetted by his apprentice Fumble-Fingered Fred... dotted throughout the chassis were numerous examples of ‘quick fix’ repairs using components which must have been to hand at the time. Amongst many other bodges were under-rated resistors, a small electrolytic capacitor wired back-to-front, several components hanging in mid-air on long bare wires, and capacitors made up of two in series, presumably because the repairer didn’t have the right value at the time. I debated whether to re-make all these repairs properly, but decided against doing so purely for cosmetic reasons; if an earlier repair was correct and safe, and not affecting operation, I felt it was part of the unit’s past history and could be left in place. It has to be said that the sheer number of previous repairs suggests that this particular TF1370A was not one of Marconi’s most reliable models, but I have no way of knowing how hard a previous life it had led.

After cleaning the internals of dust and checking for shorts, I applied power carefully through my variac and series-connected AC ammeter. The unit worked, after a fashion, although the output was far from a sine wave, and both amplitude and frequency were heavily dependent on the applied mains voltage. Before tackling further repairs I obtained a service manual from an on-line source, studied the schematic and ordered-in a number of components that I anticipated would need replacement, mainly the high-voltage electrolytic capacitors that I didn’t have in stock. Once these arrived I started work.

In common with many high-spec signal generators of this era, the TF1370A incorporates stabilised HT and LT power supplies, and I always start here as some stabilizer faults can cause excessive HT and consequent stress elsewhere in the circuit. The LT supply comprises a transistorised series regulator providing 6.3 volts DC for the heaters of the four valves forming the Wien bridge oscillator, and fortunately this proved to be in order. The +285 volt stabilised HT supply, on the other hand, would not give more than 180 volts and no regulating action was evident. Marconi used some complex feedback circuitry in their regulators which repays careful study, but the 6080 double triode series regulator tested good and so I replaced all the high voltage electrolytics and a few suspect decouplers, then checked the values of some likely resistors and found that several had risen in value. Replacing these restored proper control over the HT, and this was set to 285 volts, at which point the oscillator started to behave itself... for a short while only...

The higher regulated HT now began to cause components to fail elsewhere. Two resistors going open-circuit one after another in the oscillator stage stopped everything for a while, but once I’d found and replaced these I spent some time working steadily through, making voltage checks and identifying and replacing 'high' resistors and leaky capacitors, from the Wien bridge oscillator to the subsequent 20dB amplifier and Schmitt trigger stages in sequence. At this point I had a nice, stable sine wave from the master oscillator and a good square wave from the Schmitt trigger, but nothing more than short positive-going spikes from the output stage. Voltage checks revealed that V8, the cathode follower’s shunt compensation 6BQ5/EL84 pentode, was cut off despite its control grid being at +30 volts relative to the cathode! I replaced the offending leaky coupling capacitor and then found that the valve itself was faulty, presumed killed by positive grid voltage. I fitted a new valve, all voltages returned to normal and a pure sine wave burst from the output stage. The repair concluded with a few further adjustments, a soak test, and a check on the accuracy of the frequency and output calibration. Lastly, the instrument’s front panel and case responded to a good clean with warm soapy water.

The TF1370A joins its Marconi stablemates, a TF144H/4S and a TF2008, in my workshop and together they make a trio of beautifully-engineered test equipment which would have cost the original buyers many thousands of pounds. The attached photos show the finished unit, together with screen shots of the outputs. I paid virtually nothing for this gear, other than a handful of replacement parts and a few enjoyable hours of stimulating fault-finding. All I have to do now is find permanent space to house the TF1370A, which in typical 1960s Marconi style measures 20” x 14” x 11” and weighs in at 38lbs!

Phil

[glowinganode]

Well done Phil, good clean looking waveforms.
I had one of these many years ago, a super bit of kit but rather bench hungry.
I seem to recall it uses an obscure valve in the oscillator, I remember driving down to Croydon (Langrex?) to get a replacement set of valves.
Rob.

[Phil G4SPZ]

Hi Rob,

Thanks for your kind comments. Yes, that's the trouble with test gear from this era, it's all so big and heavy, but that tends to make it easier to work on. There's a lot of space inside the TF1370A!

The oscillator uses two EF184s and two E810Fs, the latter are indeed a bit obscure. Fortunately these valves are OK because, whilst I have many EF184s I don't have any E810Fs in stock. The National Valve Museum says: "Mullard describe the E810F as a special quality high slope pentode designed for use in industrial equipment where stability of characteristics and long life are required. It is a wideband RF device, but no frequency range is given. From the information available it appears that 40 - 100 MHz is a probable upper value. Oscilloscope amplifiers and other measuring instruments are likely candidates for this valve."

Kind regards,

Phil

[Phil G4SPZ]

A quick rider to this story. When I left my workshop last Tuesday evening, I inadvertently left the signal generator running at 700 c/s, and didn't realise my mistake till Thursday evening. It coped with this unintentional 48-hour soak test perfectly well, and kept the workshop warm into the bargain! During this period I reckon it had generated around 121 million cycles, and used mains electricity worth about 50p...