Marconi TF2015 sig gen - excess FM on CW

[ronbryan]

My Marconi TF2015 began to intermittently lose lock with its associated TF2171 synchroniser and so I separated the two instruments and started to look at the TF2015 signal generator 'counter out' output, set to CW, on a Philips PM6671 timer counter.

The frequency at 10.7MHz seemed to be varying randomly a few kHz around the set point, but the triggering on the counter seemed to be very fussy. The TF2015 'counter out' output was found to be very low, well below the >80mV specified in the manual. This signal was traced through and found to be caused by an unsoldered centre pin in the TNC bulkhead connector on the rear panel. It seemed that the silver plating on cable and pin had given good contact without solder for many years. It was not an easy connector to access, requiring removal of board A16 the RF output pcb.

The triggering on the counter was improved, but the excess random FM was unchanged. I suspected the power supply, but the 22V and 19V rails seemed to be clean. A further test, running the generator from the 'battery' input from 28V from a bench supply, gave no improvement.

The tuning voltage for the selected varactor oscillator is available on the 'sync / sweep in' BNC socket and viewing this voltage via a screened cable to an AC coupled scope showed a randomly varying voltage of at least 10mV pk-pk. The tuning voltage summing junctions have inputs from several sources including' fine' and 'extra fine' tuning, plus several FM modulator circuits, which made analysis difficult, but scope checks indicated that TR1, a BFW11 on fine tuning board A21 seemed to be at the centre of the problem. This was confirmed by removing the 19V supply from board A21. This source follower circuit is needed to provide a low impedance drive for the external modulation, according to the manual.

I changed the two 1M CR25 TR1 gate bias resistors R10 and R11, which seemed to improve the situation a little, but the problem was still present. TR1 position was found to be filled by a BC184C transistor , rather than a four pin BFW11 fet, so I must have an early production circuit board for A21. Anyway I changed it for a BC184 that I had, but it made no difference, so I refitted the BC184C. Finally I removed C3, a Mullard box style type 344 0.1uF capacitor and tested it - it seemed intermittent and one leg was not properly attached.

I replaced it with a STC PMA polycarbonate capacitor, because it was good quality component and had the correct lead pitch. I was pleased to find that it solved the problem and the random variations of the tuning voltage were reduced to around 1mV pk-pk.

Ron

[Nymrod121]

Thanks for posting this Ron; your explanation along with the photos and diagram made for an interesting and informative read. I applaud your tenacity!
Best wishes
Guy

[mhennessy]

Hear hear

I must admit that those capacitors have been 100% reliable in my experience, and I suspect that might have coloured by judgement, meaning that I might never have suspected it! Still, "when you have eliminated the impossible, whatever remains, however improbable, must be the truth"

I have a 2015 on the tuit pile. Unfortunately, one of the plastic cams that operate the various attenuator sections has broken in two. Fixing that will require some serious effort - I suspect that gluing it will be hopeless and some sort of replacement will have to be manufactured.

Compared to other RF generators, I really like the small form factor. It's interesting that they chose to use a whole load of separate oscillators rather than the complicated frequency shifters and multipliers that were used in beasts like the TF2008. I'm looking forward to seeing how well it performs, as space is always at a premium. Shame it's not a TF2016 though - that one is more useful for domestic radio work, as it covers 10kHz to 120MHz.

[ronbryan]

Here is a picture of the underside of the TF2015 sig gen, showing the cam operated attenuator, which as Mark has noted, is prone to failure due to fracturing cams.

Part of the strain on the plastic cams must be caused by the tightness of the grub screw that locates the cam on to the double 'D' operating shaft. I think it would be worth a try to remove the grub screw, glue the (correctly orientated) broken cam together in situ, using a 'all plastics' superglue and associated primer, leave the grub screw out and use a couple of spring rings (external circlips) slipped on to the shaft to give lateral location to the cam.

This suggestion, if deemed viable, permits repair of the broken item without having to dismantling the shaft to slide on a re-manufactured cam.

What have others done when faced with this repair issue?

Ron

[cmjones01]

I've done repairs to very similar cam-operated attenuators in the Tektronix 7A29 1GHz scope plugin. In this case they're not held by grub screws, but are a push fit (possibly moulded on, actually) on to splined sections of the shaft. The plastic shrinks and cracks, and then the cams slip round so the attenuator doesn't work properly. I've got them running by adding a fillet of epoxy resin (Loctite 3430 because it's what I happen to have in the workshop) against the shoulder of each cam. I don't know how long they'll last but they should be good for a few more years.

Your circlip idea sounds like a good one if there's space.

Chris