Racal RA63D

[wd40addict]

I've owned a Racal RA17 for over 20 years and have always been very pleased with it, the stability and fine tuning were a revelation after the typical valve shortwave receiver. The RA17 was launched around 1955 and pre-dates the widespread adoption of SSB, although it does have a BFO so these signals can be resolved with a bit of fiddling. Racal obviously realised this was a weakness and produced a series of add-on adaptors which gave various SSB and ISB capabilities. They work by taking a feed from the RA17 100kHz IF output and processing as required, containing their own audio stages. Recovered audio is not routed back through the main receiver.

The various add-on units are much less common than the RA17s themselves, very rarely coming up for sale. I have kept an eye out for 20 years and only seen two.

The RA63D unit under discussion here is a member of the RA63 family, these are the most basic units offering a choice of upper or lower sidebands only. A small 2.5" built-in speaker allows monitoring of the audio, there is also a high impedance headphone socket on the front panel and further connections on the rear panel.

Photos below show RA63D with RA17 and general views of the internals.

The members of the RA63 family are as follows:

RA63A: For receivers with 100kHz IF output, uses LC oscillators.
RA63B: As A, but with American plugs and sockets.
RA63C: For receivers with 465kHz IF output.
RA63D: For receivers with 100kHz IF output, uses Xtal oscillators.
RA63E: For receivers with 455kHz IF output.
RA63G: Ruggedised RA63A
RA63H: Ruggedised RA63B
RA63K: Ruggedised RA63D
RA63L: Ruggedised RA63E

What happened to F, I & J is anyone's guess! Presumably the non 100kHz versions were aimed at other manufacturer's receivers.

In wanting to service these units another problem arises: The units themselves are rare and the data is rarer still. There does not seem to be a surviving RA63D service manual so work has to be based on extrapolating from the RA63A manual.

How does it work?
The incoming 100kHz IF is mixed with 118kHz from the first Xtal oscillator to produce an 18kHz IF. It is then bandpass filtered by what appears to be a 3rd order Butterworth with a response of 14.5-21.5kHz. This filtered signal then passes through either a low pass or high pass 18kHz filter depending on whether upper or lower sideband is required. The wanted sideband is then mixed with 18kHz from the second Xtal oscillator in the product detector to recover the original audio. This signal is then filtered by a 3 pole low pass filter and a bridged T to remove any 18kHz carrier. The audio then passes through the volume control and into Racal's typically strange use of an RF pentode used as an output valve. Presumably this was done to minimise the different types of valves required to be held as spares.

Differences with the RA63A: Obviously two Xtal oscillators in place of LC types and a first LO of 118kHz rather than 82kHz. On the RA63A the first LO has a front panel fine tuning knob. The RA63D is fixed frequency. The high and low pass filters are materially different in both appearance and electrically to those shown in the RA63A manual.

My unit had no obvious damage although some screws were not original. Opening it up everything was there except the monitor loudspeaker... This was a bit ominous, had someone started to strip the unit down for spares due to some terminal fault? Gently powering up on the variac and connecting a 100kHz oscillator then swinging either side produced the expected outputs albeit at much lower sensitivity than detailed in the manual (30 or 100mV in, they can't seem to make their minds up!) for 100mW into 3R.

I then tried it with the RA17, results were very disappointing - low sensitivity and distorted audio. Needless to say the waxies were all very leaky upsetting the bias on the output stage. Now to tackle the low sensitivity: In the RA63A the output from the first LO is given as 5.6V, I was probably getting 1/10th of that. Feeding in an external 118kHz oscillator @ 5.6V gave the expected sensitivity. Not having the 'D' service manual I traced out the circuit of the first LO, mine was not the same as diagrams floating round on the net nor some other Racal equipment using a 118kHz Xtal LO. I tried tweaking the circuit to match those diagrams - I got increased output, but the Xtal wasn't entirely happy starting and would sometimes also fade away. I think 60 years of Xtal ageing meant the original circuit was no longer the optimum match, by rearranging the original circuit as shown with the trimmer as a series element I could adjust to a perfect 118kHz and starting and running were fine and I had approx 5.6V output.

The missing speaker was replaced with a Visaton FRS7 4R version which was a perfect fit. Now connecting to the RA17 the Lower sideband sounded pretty good, but upper was a bit 'pinched'. Frequency response sweeps of both filters showed the upper was a bit peaky, further investigation showed that one of the 5 inductors was open circuit :-( Carefully opening the pot core I found that the break was just as the wire left the coil and was able to repair it - phew! Now I don't have the correct alignment instructions for this filter, but in this modern age frequency sweeps are easy to achieve - connect PC soundcard to AM input of test oscillator. The attached plot shows the difference in response between the upper and lower filters. The lower is very flat, upper less so. I've adjusted the upper as best I can but some of the pot core trimmers are jammed solid so this is about as good as it gets. All other filters I have not attempted to adjust.

How well does it work? Very well, clear crisp audio. I can see that sometimes the fine tune on the A would be useful, but I'm not complaining. The incorrect screws were replaced with chrome versions from Ebay, although one of the originals was unbelievably jammed it eventually gave way. The headphone socket I rewired to be low impedance so an external speaker can be used if desired.

[wd40addict]

Here are 3 versions of the 118kHz oscillator. First the version used in the RA121D, then the original circuit in my unit and finally my modified version which fixes the low sensitivity and allows correct tuning.

[wd40addict]

Here is the RA63A circuit, plus the 18kHz Xtal oscillator from the RA98. I'm not sure how they got a crystal to work at this low frequency, but it's something to do with being dual element...

[wd40addict]

Finally here's the comparison of the final frequency response between the upper and lower filters.

[Craig Sawyers]

I'm working on several units simultaneously.

1. My original RA17II
2. A somewhat rare RA17W, an ex-GCHQ special
3. A NOS RA121 (the sideband converter with the funky CRT)
4. RA37 LF converter
5. RA63 sideband converter
6. RA98 sideband converter

That is a lot of sideband converters. But when they come up, why not?

The only thing I'd like to get my paws on is an RA66 panadaptor. They are particularly rare. There was one on eBay a couple of months ago, but the price ended up too hot for my pocket.

Craig

[wd40addict]

My receiver is also a RA17ii. Which RA63 version do you have?

[Craig Sawyers]

RA63A

Craig

[1954Fender]

I'm keeping my eyes open for a '63 but have an RA121D which has the pretty CRT to watch if the program material gets a bit on the dull side. I also have an RA98 which has an interesting motor driven, magnetically coupled variable capacitor tuned oscillator tracking the carrier. As Craig said above about converters, why not?