Another deaf Racal RA 17

[der Sammler]

Hello!

Against clear commands of the better half, I decided unilaterally to expand my collection of tube receivers by one RA 17L.
Either Faith favours your spouse's judgement, or a dark curse rests upon those who worship glowing items.

So I am looking now at a deaf Racal RA 17L (Serial number N6180) receiver.
Powers up nicely and the BFO functions correctly, yet you only get static.

Have a noise generator, only "comes through" on the 0 Mc/s position.

Verified the voltage on all tubes, results are OK.
So I have changed all tubes exept V4, V20 and V24.

As indicated on the Manual I have removed V9 and turned on the Megacycles dial: no drive at TP3
Reeplaced V9 and turned the dial: I get voltages variations between -2.9 and -3.0 depending on whether I am on an Mc/s point or not.

The first VFO has been extracted and connected to the side (at first I assumed this section to be defect...) and the only unusual feature I can notice is high DC Voltage on the screw of C24 (the 37.5 Mc Band Pass Filter)

Having reached the limits of my (most limited) knoledge, I must appeal to the collective wisdom of the web:
Any suggestions on the nature of my Racal's deafness?

[Radio Wrangler]

Do you have any knowledge of the history of your RA17? Possibly the worst thing would be if someone had tried to fix it by "adjusting" all the filter stages in the Wadley frequency converter.

Your next step is to inject some signal into the 2 to 3 MHz IF stage and see if you can tune into it, and whether the receiver is reasonably sensitive here.

If the VHF filters in the Wadley system have been disturbed, you will need the help of someone experienced, and a sweep generator to correct them.

David

[vosperd]

So many things can cause the RA17 to be deaf.
The high voltage on the screw of C24 is normal.
Does the calibrator work on cal or bfo?
I have had faulty V3 and bad contact on the 1mHz crystal causing problems.
You should be able to see if the filter caps have been moved from their original positions.
The silver mica caps in the filters can be a problem. They can be changed without retuning the trimmers.
Can you get anything at all. On wide band and widest IF filter can you find any broadcast stations?
Don m5aky

[Herald1360]

Welcome, oh collector!

As you may have realised, the RA17 falls in the same class as the girl with a curl- "when she was good, she was very very good, but when she was bad, she was horrid".

When they were new, they outclassed pretty much anything else, at least on drift performance, but like any thoroughbred, they need to be kept in perfect tune.

Being ex Racal I'm slightly biassed of course

Unfortunately I never had anything to do with RA17, it was obsolescent by the mid '70s and in any case I was a Transmitter bloke. I do recall it wasn't always easy to set up even in the works. Sometimes all that could be done was to undo all the trimmers and just start over

[der Sammler]

A friend has lend me a signal generator that will do 2 MHz, attached to the IF stage it will tune in.
If I take the noise generator and tap into the yellow lead out of the first VFO it is possible to tune into this also. Weird because it is a noise generator and on all other receivers it tunes into it regardless of how they are set.
Anyhow on this "noise generator aknoledged" position over the second VFO you can tune into a "noise bump" on places you would expect a potent AM station (such as 800 MHz for Germany). Will not work on other well known stations around 6 MHz...

This set comes to me out of the Polish border via e-bay, not much to say on the backround...
The original antenna connector has been reeplaced by a metric one, many of the valves I reeplaced where soviet made, I would have to guess that the Set has been on the eastern block for quite some time.
The screws on the filters have no markings of any kind, either that has not been tempered with, or the guy who did the tempering had the correct screwdriver.
Can´t tell if the filter caps have been moved (do not know the original positions) but it is cleat that some have been reeplaced!

Calibrator makes blunt noises on (more or less) the right spots, took the 1 MHz signal out of the back of the receiver and fed it into my R&S EK 07D (nasty German Boatanchor) over a capacitor, tuned into 1 MHz the signal generator is OK.

Attempted to inject signal directly into the filters from the noise generator and the signal generator, odd creature this Racal, it would not tune. yet it tunes 2 MHz out of the IF stage...?

What am I missing... I wonder what the guys at the Racal works did to get these beasts to work, sure hope it did not involve a pentagram an a dead chicken...

[Radio Wrangler]

Pentagrams and dead chickens are easily obtainable

In the factory, and a major military service centres, they had lots of custom equipment dedicated to setting these receivers up.

Without the hardware, an RA17 can still be set up properly, but you will need an oscilloscope and a sweep generator capable of going up to about 50MHz and performing narrow sweeps.

As it's had a lot of 'intervention' I think you should work slowly, one section at a time. If you can find someone experienced, with appropriate test equipment, it would be good.

David

[Skywave]

der Sammler: I am somewhat confused by what you have written in post 5 (various reasons).

The first thing you must do is establish if the 2 -3 MHz section of the receiver is working properly. The way to do that is by using a signal generator (not a noise generator) that will inject signals in the range 2 - 3 MHz into the input of the 2 -3 MHz receiver part. If you do not get the corresponding 100 kHz signals out, then that implies that the 37 MHz (something like that frequency - check the manual) is missing. In which case, proceed to investigate the cause of that signal loss. That investigation will take you back to the mixer where the 1st. VFO and O/P from the harmonic generator are mixed to produce that '37' MHz signal.

Al.

[vosperd]

It looks like you have the crystal oscillator and vfo2 working.
Whilst you are carrying out the checks that others have suggested it would be useful to check the output of VFO1, 40.5 - 60.5 mc/s. An oscilloscope would be handy.
It is possible that the mc/s dial is not accuratly aligned 6mc/s could be either side of the cursor. I was not able to get mine to track accurately across the range.
It is always difficult with a set that you have not had working as you can be dealing with more than one problem.
Good luck, as a previous posting says - they are great sets when they are working.
Don m5aky

[der Sammler]

OK!

Have removed PL4 from the socket (my best guess on where the 2-3 MHz Stage starts) and injected into PL4 2 MHz coming out of a Philips PM5132 signal generator set to sine wave.
It thumps regardless on how I tune the first or second VFO, If I set the generator below 1 MHz the signal dissapears.

Where (and how) do I ook for this 100 kHz signal?

Is this 37.5 MHz signal supposed to go into pin 7 of V10 out of that odd looking L shaped filter?

The voice of the thread has spoken: I must find an oscilloscope!!

[Sean Williams]

I think a couple of evenings spent reading the manual might be instructive.

Must just be me, but when faultfinding a RX I start at the speaker and go backwards - Audio, Detector, IF, Mixer, LO, RF and Antenna.

Admittedly the Wadley loop complicates things a little, but the same methodical approach will work, and in my experience there will be a multitude of faults lurking.

I would be a little dubious of any Eastern Bloc tubes - the RA17 is a fussy beast when it comes to valves.

HTH
Sean

[Skywave]

This is really a follow-on and an expansion of what Sean has said above.

When working on a complicated beast like the RA-17, you really do need to have a block diagram drawing in your head all the time. If you can add to that an outline of what frequency ranges lie where, how they are derived and have a clear picture in your mind as to how the whole she-bang works, that really helps. You can derive all that knowledge from studying (note: not 'reading') the manual - and then studying it again, slowly and in depth - several times, if necessary. Not until you really feel that you understand how each bit works - and why - can you then sensibly begin a comprehensive fault diagnosis.

As for "which end do I start at", I always choose to work from aerial to 'speaker: which method you use is largely a matter of personal choice. What isn't a matter of choice is the need to take a systematic, step-by-step approach in testing and analysis . . . with that block diagram in your head. The random "I'll try this, then I'll try that" scatter-gun attitude at best is a waste of time: at worst, by ignorant mal-adjustment, you may introduce a fault that wasn't there in the first place.

Oh yes: having the cct. diag. in from of you helps too. And a 'scope + freq. counter (both that accurately display up to 100 MHz) are not luxury items here, either.

Another case of the old motto: patience, persistence and perseverance usually wins the day.

Al.

[Radio Wrangler]

Fortunately most of the valves in an RA17 haven't had their prices jacked up by the audio fools.

I would certainly start by working my way along the structure, replacing valves with the correct types. Starting at the speaker (they sometimes die too) and working in reverse-signal-flow direction is best. If you work from the antenna end you need a signal generator and a spectrum analyser (or a very comprehensive receiver). Working from the speaker, you just need the sig gen and a pair of ears.

You'll need a source of low level audio, then a source of amplitude modulated 100kHz, then AM tuneable from 2-3 MHz

The problem is, you don't know just what has been done to this radio.

If you can post photos, people here can compare them to their radios and help to spot modifications.

Cheers
David

[Skywave]

Yes David, there is a good deal of logic in working from the 'speaker towards the 'front end' on a totally dead RA-17, despite my earlier remark. That remark was based on my experiences with my one RA-17 which over the years has developed the occasional fault of not being a totally dead set, so the symptoms (plus machine knowledge) have usually indicated where to start first - and that's always been in the non-audio, non-100 kHz sections.

Having said that, although I agree that the use of a spectrum analyzer (or a very comprehensive receiver) will make repair easier under some circumstances, it has not been my experience to need such kit, but then I have not had the need to check or re-tune the 40 MHz band-pass filter on mine (and hope I never have to, either!)

Al.

[Radio Wrangler]

You don't ven thing of touching those filters without some sort of sweeper and display (with one eentsy weentsy little exception)

I use a synthesised spectrum analyser with a tracking generator. It makes life much simpler. The spectrum analyser is amazingly sensitive if I want it to be, so I can use very lossy, very non-invasive probing techniques to split the thing up.

That exception..... coupled resonator bandpass designs have each resonator targeted at the exact GEOMETRIC centre frequency, and the pair of poles it creates are pulled away by the coupling to the adjacent resonators (upstream and downstream) It is therefore possible to kill the adjacent resonators by shorting them, and then peaking the resonator in between them to the square root of the product of the upper and lower 3dB frequencies wanted. Do each resonator in turn and the filter should work as designed... the difficulty is in probing and exciting for resonance without affecting the tuning.

There's more on this technique in Zverev for those prepared to dig. The RA17 pre-dates Zverev's seminal tome.

David

[amornummi]

Try putting some numbers in for the 40 Mc/s filter.
The 3 Db points are not specified in my RA117 book)
Say 39.350 / 40.650 gives 39.995 Mc/s (required flat passband)
39.3 / 40.7 gives 39.994
39.2 / 40.8 gives 39.992
39.1 / 40.9 gives 39.990
39.0 / 41.0 gives 39.988
An excitation loop will have to be very loosely coupled indeed not to shift the freq, and to get the required accuracy of a few kc/s will need some kind of phase detection.
The lid has to be on , with short wire soldered across unused trimmers. Not easy!
I will have a play next time I have a set on the bench.
The 2k2 filter term resistor usually needs replacing.
I find the mixer gain plus filter loss is about two. Measured at first mixer grid (tp2) to second mixer grid (tp3) with hi z probe and HP141.
C108 has serious effect on the 40Mc/s filter response ...

[Radio Wrangler]

When I do the one-resonator-at-a-time method I use a spectrum analyser with tracking generator to drive and receive from a hybrid coupler (return loss bridge) with a small coupling loop on it.

The HP8553 spectrum analyser (0-110MHz) plug-in is the one you want, but the matching counter/tracker box is rather rare and very desirable. Alternatively and added sig gen with a counter will make an accurate marker.

I can just get a discernible dip in the trace from these filters with my analyser on 0.5dB/division before I see closer coupling pull the dip.

Cheers
David