RCA AR88D Receiver
Written by Bryce Ringwood   

The RCA AR88 was produced during World War 2 as a receiver for point-to-point communications. (i.e. between individual transmitters and receivers, as opposed to broadcast.) Two main versions were made, namely, the AR88LF and the AR88D. The LF version covered long-wave as well as medium and short-wave (From 75kHz to 30.5MHz in 6 Bands). The AR88D covered 535kHz to 32MHz in 6 Bands as follows: Band 1: 535kHz to 1.6 MHz, Band 2: 1.57 to 4.55 MHz, Band 3: 4.45 to 12.15MHz, Band 5: 16.1 to 22.7 MHz and Band 6:22.0 to 32.0 MHz. Notice that for bands 1 to 4, the ratio of start frequency to end frequency is approximately 2.5:1. One might have expected the AR88 to have just 5 bands, with Band 5 covering approximately 13 to 32 MHz (as is done with the Eddystone 940, for example.) Instead, the designers chose to split the last band into two, giving the AR88 additional bandspread on the highest ranges.




The set has 14 valves with 2 6SG7 RF Amplifiers and 3 6SG7 Intermediate frequency amplifiers. The local oscillator is a 6J5 and the mixer a pentagrid 6SA7. The audio amplifier is a 6K6, and another 6J5 is used as a BFO for CW reception. 6H6 valves are used as detector and noise limiter, and a 6SJ7 is used as first audio stage. All this is powered by a 5Y3 rectifier and VR150/30 voltage stabiliser for the local oscillator stage.


Noise Limiter Clipping Level
Antenna Trim
BFO Pitch
ON/OFF, Stand By, CW
Band change
RF Gain
Mode - AGC, Manual, Noise Limiter

Some sets have an 'S' meter.

In order to provide excellent selectivity, the set designers chose to provide double-tuned IF transformers for each IF stage. The 3 sharpest selectivity positions used a single crystal filter. IF frequency for the AR88D is 455kHz. For the LF it is 750kHz (To be confirmed)

Many AR88 receivers were made without an 'S' meter because of a shortage of movements, but the set was designed to have a meter calibrated in decibels. This meter works the “wrong way round” compared to norS Metermal meters. In other words, the needle moves to the left with increasing current. The set to be described had a standard meter with dial calibrated in 'S' units. (Tracing its ancestry, the set turned out to have belonged to a colleague of mine – he had certainly done a neat job on the meter.)

After the war, a large number of AR88s became available on the surplus market and were used by radio amateurs as the receiver of choice until single sideband became popular from the mid 1960s onwards. From this period on, it became popular to modify earlier AM radios, that lacked sensitivity, or the stability required to resolve SSB stations. Strange “mushroom pentodes” and other valves would replace the 6SG7 RF stages improving the sensitivity but degrading other areas of performance. Actually, the AR88 was quite capable of performing perfectly adequately, provided it had been correctly maintained , however, using it as an SSB receiver was not as comfortable as using a receiver designed for amateur band use, so many of these radios ended up being discarded, or like the one to be described, stored in a chicken coop. For anyone interested in broadcast listening on short wave, the AR88(D or LF) will provide a fine introduction. If you live in a city, the noise level from PLT, ADSL lines,Compact Fluorescent Lamps and so on will be the limiting factor to your reception ability – NOT the fact you are using a noisy pentode valve as the first RF amplifier.

Mushroom Pentode
Mushroom Valve

The set in this study was given to me by a rather disillusioned gentleman. He had intended to restore the AR88D to its former glory and had sent the front panel to be resprayed. The person to whom he had entrusted the work went out of business, and he lost his pristine front panel and was left with a somewhat battered panel for an AR88LF. He had two chassis and two cabinets – one for the LF and the other for the D. Alas, the LF was beyond redemption.

I began by stripping all the paint off the front panel, then straightened it and filled it. (I almost decided to do a new front panel from aluminium sheet.). At first I tried to use wrinkle finish paint, but it proved impossible to fill in the lettering. Maybe someone can tell me whether the AR88 front panel uses wrinkle finish, gloss or matt black.lettering In the end, I opted for matt black. I resprayed the cabinet with gloss black over the wrinkle finish to hide the scratches, assembled everything and had a reasonably decent looking radio. (The other cabinet was sprayed grey and used for an R390.)

I removed the mushroom valve in the scecond RF stage and replaced with the correct type. The first RF stage has been replaced by an ECC88, which for the time being I have left in place. I switched on and – nothing. Then something, then nothing again. The fault was traced to a bad solder joint inside the penultimate IF transformer. This must have been a fault since the set was made and probably gave trouble throughout the radio's life. Medium wave worked, but was pretty lifeless. This was caused by an open circuit medium wave coil, which had to be rewound. Somehow, the set still had all its original trimming tools and Allen key, and I realigned the front-end without any trouble. The set now worked, but the power supply capacitors had decided to take on the role of oil-heater and were taken out of circuit, the modern electrolytics to replace it fitted under the chassis with ease. (Don't be tempted to replace power supply caps with larger values, just because you can – the life of the rectifier valve will be seriously reduced.) I also replaced one of the bathtub style bypass capacitors by emptying the contents and putting modern capacitors inside.

The audio output transformer had been replaced with the one from the AR88 LF. On these sets, excessive headphone listening causes the output transformer to fail (presumably because it has no load – which is bad.) Many AR88s have the 6K6 replaced with a much easier to get 6V6. Biasing has to be altered to suit the replacement valve – and 6V6s are still made today.

Finally, I attempted to re-align the IF stages. Alignment of the crystal filter was straightforward and went well. I have to confess that alignment of the two “broad” selectivity positions was not as successful. The result is OK – but not as perfect as I would like. (Equipment used was a sweep generator and 'scope to display the selectivity curve.)


In use the set is the most stable of all the conventional superhets and the ability to resolve weak signals is also just as good as a modern radio – but that is probably because of the ECC88 modification (and is a bit of a disincentive to put things back the way they should be. ) The "native" sensitivity of an AR88 is 7 microvolts, compared to less than 1 for a modern receiver. But that is probably RCA playing it safe.

The tuning dial is difficult to read once you have passed theage of 40.

If you want to purchase an AR88 – look out for the broken audio transformer (This can be repaired/replaced) and also beware of the trimmer capacitors – the metal lining can flake off, making the set impossible to align. They can be replaced with Philips trimmers, or modern types, but its something better avoided. Inspect the set carefully to make sure it has not been modified. - these radios were used by radio amateurs keen to extract the last drop of performance no matter what.

These radios were designed for the reception of AM and CW signals.  An unmodified AR88 will not resolve single sideband (SSB) very easily - you will have to back off the RF Gain and position the BFO correctly. The selectivity is probably best set at position 2 - sharp, but no crystal filter.

For CW reception, there are three sharp positions. There is no crystal phasing control to nothch out an unwanted transmission. If you align the set so that the notch is present, you will have to use the main tuning control. Some skill and perseverance is needed to use this set well, but in order to do this, the tuning drive has to be in really good condition - any backlash is going to make operation somewhat challenging.

Also challenging is the alignment of this set to specification. Stagger tuning the IF transformers to give the correct selectivity requires a certain degree of patience. Assuming I have got it somewhere near correct, the set has good bandwidth for normal AM reception in position 1. The selectivity narrows to about 8kHz in position 2 giving some rejection from adjacent signals. The crystal positions severely distort the AM sound.

Tuning accuracy is dependent upon the scale. The AR88 has no crystal calibrator - but no matter, you can buy a 1MHz crystal oscillator module from almost anywhere to do the job. You can't say "I'm going to listen to the BBC on 17 640kHz", set the frequency and hey presto, like you can with a more modern receiver. You are going to have to work with the set until you know exactly where every station happens to be.

The AR88, Eddystone 940, Murphy B40 and the original R390 were all designed with 2 RF amplifier stages. All those tuned circuits at the front of the receiver helped to prevent image interference. When you were tuned to 3255kHz - you knew that was what you were receiving and not 4165 kHz. The reason for 2 RF stages in the "R390 non A" is obscure to me. The problem is that there is simply too much gain. Strong signals can begin to behave like another "local oscillator" and modulate nearby weaker signals causing a litany of problems. The Eddystone 960 - a solid state version of the 940 must have been very bad in this respect. A reaction set in, so that later designs of the 70s had no RF amplifiers and used valves like the 7360 as a mixer fed straight from the antenna. It was rumoured that a certain military radio used ultra fast power diodes in the front end as a balanced mixer. I don't think that can be true.

My own view is that the cross-modulation thing was over rated. Yes it happens. Less so in the southern hemisphere and less so today. I think the AR88D or LF is still a fine set for listening to shortwave AM stations.




Joomla template by a4joomla