Marconi TF2002B problem.

[pwdrive]

Hello Alan yes it must have been, the freq. synth. in the HRO 500 uses the blocking osc technique so far as I recall.

Lawrence.

[Mark Hennessy]

So, do you think that the SCR-connected transistors are oscillators then? I wondered that - as they resemble UJTs, but wasn't immediately sure. My last experiments with transistors connected as UJTs were about 20 years ago...

Cheers,

Mark

[Skywave]

Gentlemen: thank you for your replies; appreciated.
Attached is an extract from the manual that explains the ccts. operation.

   

I might find the time today to have a further look and think about this. It is quite clear that he effective value of C14 / C15 is critical to correct cct. operation, but I'm sure I have either checked their values on a bridge or have replaced them. However another thought has just crossed my mind. I know that the 10 kHz divider isn't working. Perhaps the area where the 100 kHz and the 10 kHz dividers 'join' is where the culprit lies: there is no buffering / isolation between them . . . . .
And for sake of completeness (if nothing else) I will try to take a photo or two of the board in question - a picture says a thousand words, etc. . . . .

Al.

[Mark Hennessy]

Thanks Al,

No-one responded to my description of the circuit (#10), or my subsequent question in post #13. However, having read the description, it looks like my initial supposition (about the transistors being an SCR rather than an oscillator) was right, and that the values of those capacitors are chosen to respond to 10 pulses. That's reassuring, I guess...

It feels that we're pretty close to finding the culprit. I'd love to see the pictures so that I can understand just how difficult the mechanical issues are.

All the best,

Mark

[Skywave]

I might find the time today to have a further look and think about this.
Al.

And found a little time I did and looked a little harder I did too - and what did I find? Please read on - an 'interesting' tale . . .

Well for a start, C15 ('Suflex' type) was short-circuit! (Really weird: its capacitance value is spot-on, but put 1.5v. from an AVO on resistance range X1 across it - and phut: it's s/c! ) But the biggest mystery was TR4. The cct. lists this as a 2N3702, and indeed, that was what was fitted. When I replaced in 'on spec.', I didn't have one of those, so a consultation of the Towers Transistor book revealed that a BC212L was an equivalent: a few unused ones I have in stock. As usual, I checked the pin-outs of that and the 2N3702 prior to fitting the BC212L: both the same: the base connection is not the centre leg. However, now, today, a close exam. of the pcb revealed that the tracks and artwork are actually laid out for the centre leg to be the base. So, out comes the BC212L, in goes a 2N2906 (an old favourite of mine). As a pro-tem measure, the s/c C15 has been replaced by a trimmer. Using the 'scope and adjusting that trimmer produced a nice 100 kHz narrow pulse train at Test Point 3. And a subsequent check (using the 'scope and a freq. counter) on the main R.F. O/P, with the calibrator set to 100 kHz, gave 100 kHz beats exactly where they are supposed to be. Hooray!

So there we are - well, on that bit of the repair to the calibrator, anyway. But the point is this: when, in effect, you are given misleading / incorrect tech. data during a repair, it does not help!

Anyway, next steps: tidy up that bit and have a look at the 10 kHz divider. Then I'll report again (or ask for help . . . if necessary ).

Al. / Aug. 5th. //

[pwdrive]

Well done that man, I have had problems with them thar sufflexes in the past on other stuff, thats a fault for the record I think.

Lawrence.

[AlanBeckett]

You pair have frightened me as I've always had faith in Suflexes

I once had an obscure fault like that at work. The PCB had been out in the field for a couple of years and the Customer reported that very occasionally it stopped on an internal fault. All sorts of Production tests failed to find the problem, but yes, it would indeed stop once in a flood. Eventually, after much testing and head sctratching, it turned out to be a Transistor in backwards - C/E reversed. It just had enough gain to work most of the time.
You live and learn.

Alan

[pwdrive]

Alan, those sufflexes also used to fail in the old BRC/Thorn tellies especially round the IF/Video stage, mind you those sets got really hot inside back then. Maybe the one Al had was always "on the edge" from new as his generator is low voltage and probably runs relativly cool.

Lawrence.

[Skywave]

Things are looking on the up with this item - at last.

The 10 kHz section seems to be working O.K. Which makes sense, now, of course, since if the 100 kHz section isn't doing its stuff, the 10 kHz section can't do its job either!

So as far as the calibrator functionality is concerned, I reckon I'm 99% there. Why only 99%? Because when the 100 kHz cal. is selected, there are beats where there shouldn't be, i.e. between 100 kHz points. Most of these are substantially weaker than the genuine 100 kHz ones, and there is a disproportionately strong beat at 500 kHz intervals. Looking at the O/P waveform from the 100 kHz stage, although it is a pulse, it is by no means 'clean' as regards rise and fall times - so some sub-harmonics are in there.

The questions thus arising now are:
(a) Was it always like this? - i.e. when it was new and left the factory? To which I suspect that there is a reasonable chance the answer is "yes". And I say that since to my eyes and understanding, the mixer circuit design is a rather crude unbalanced affair. It seems strangely out of context when one considers the exorbitant and extensive lengths Marconi have gone to in many other areas of the design of this generator. Or am I being unkind / unreasonable / mistaken here?
(b1) Do I simply live with it? After all, it isn't a major feature of the generator and if I ever do need to know the O/P freq. accurately, it's easy to connect a freq. counter to the O/P: this generator even has a dedicated BNC O/P on the front panel for that very purpose.
(b2) And thirdly, as far as a 100 kHz cal. 'pip' is concerned, this is mainly only going to be useful of on the lower freq. ranges anyway: this generator tunes from 10 kHz to 88 MHz: above about a few MHz, the 100 kHz 'pips' will be very close.
And the same argument applies to the 10 kHz cal. 'pips' too ~ perhaps more so.
So I intend to spend a little time investigating purity of waveform - mainly for peace of mind before I 'blow the whistle' on this calibrator repair job.

Having said all that, and on the topic of repairs, when I first inspected this generator, I did discover no R.F. O/P on the lowest but one freq. range - but very soon after that, that condition / fault mysteriously disappeared. But now that fault it has come back. The 'good news' is that because this particular fault is restricted to one tuning range only and the unit is designed with separate osc. modules for each band, the 'search area' for the fault is thereby substantially narrowed. (Least, that's the theory & hope - at present. )

So enough for today: bed-time beckons. Today I have achieved more than I anticipated: I found I had some unexpected time available: it was put to good use.
More to come later: don't g'way now!

Al.

[pwdrive]

You are nearly there Al it's quit a complicated design, as for the weeker harmonics/birdies then as you say some further waveform and voltage analasys should confirm wether it's supposed to be there or not, I guess that the upshot is that if all the main markers are the predominant ones then the machine is fully usable, I think even in the RA 17/117 book of words it mentions the odd unavoidable harmonic or sub harmonic birdie?

Lawrence.