The "Grand" Hallicrafters SX-71
Written by Bryce Ringwood   
The Hallicrafters SX-71 receiver was manufactured from 1949 to 1954. It cost from about 190 to 254 USD, so was quite an expensive radio for the time. It is a single/double conversion superhet covering 0.56 to 54MHz in 5 bands. On the two lower frequency ranges it is a single conversion superhet with an intermediate frequency (i.f.) of 455 kHz. On the upper three ranges it is a double superhet with a first i.f. of 2025 kHz.
The bands covered are:
560 - 1600 kHz
1650-4700 kHz
4.7 - 13.4 MHz
12.8 - 34 MHz
46 - 56 MHz
Measured sensitivity was better than 1uV at 30MHz (Selectivity - Normal)
Reception modes are AM, CW and FM (Narrow Band). It is possible to receive SSB transmissions by backing off the sensitivity control.(As you can see in the video).
The set uses an eclectic mixture of valves:
6BA6 - RF Amplifier
6AU6 - Mixer
6C4 - Local oscillator
6BE6 - 2nd Conversion oscillator/mixer
6SK7(2) - 1st and 2nd Intermediate Frequency amplifier
6SH7 - 3rd IF (Limiter?)
6H6 - Detector/Noise Limiter
6AL5 - FM detector
6SC7 - BFO / 1st Audio
6K6  - Audio (Someone has put a 6F6 in this set. TODO - get the right valve.)
VR150/30 - Voltage stabiliser
5Y3    - Rectifier (In this set, there was a solid-state replacement because the transformer had no 5 volt winding.)
New Front Panel and Cabinet


B.F.O Pitch, Reception Mode, Crystal Phasing, Tone, Volume, Sensitivity, Main Tuning, Bandspread, BFO ON/OFF, ANL ON/OFF, Receive/Standby, Phones. The set has an 'S' Meter calibrated 0-9 and db.
This particular set was given to me by a radio amateur, who did not have the time to restore it. If the truth be known, neither did I, but it seemed a pity to throw it away, even though, frankly, it was a rusted mess. 
SX-71 Rust and all
The Front Panel was rusty - like everything else!
I began by ensuring the audio stage worked and the power supply was OK and all the basic checks. Someone had replaced the original mains transformer with a monster that was providing too much HT.  I began by calming things down with a 1K 20Watt series resistor on the HT. I replaced the 6BE6 2nd mixer, and the set began to work intermittently. 
The next step was to remove the tuning capacitors from the chassis so that the thick layer of rust and grime could be removed using a Dremel with a wire brush. The final rust was removed with deoxidene - a phosphoric acid based rust remover. In other countries, you might want to use "Naval Jelly" or Jenolite. The front panel was removed and stripped of all paint using paint stripper. The exposed rust was also removed with deoxidene and the panel and parts were re-sprayed with a silky black finish - not quite gloss, not quite matte.
My wife is used to be a draughtsman, in the good old days before CAD was invented, so she did the front panel lettering using normal draughting pens, stencils and acrylic white ink. I asked several stationers for Letraset, but was greeted with blank stares. Clearly Letraset has been disinvented. As it turned out, the stencils were easier to use and the acrylic ink doesn't wipe off.
The set had a ton of bypass capacitors from every era - the 50's to the noughties and looked a real mess. I broke my rule and changed most of them, for the sake of aesthetics. 
The remaining problem was an intermittent I.F. transformer (secondary - a real *** to track down). This was thought to be caused by an aging mica capacitor used to tune the coils to resonance.IF Transformer - SX-71 They were replaced with modern ones. Of course, mica capacitors are "not supposed to age and in any case, there aren't any high voltages across them". I'm sorry to say, it just ain't so. Spikes of interference can put huge voltages across those capacitors, eventually making them intermittent. The modern equivalents have somewhat higher "Q" than the originals, with the result that the IF stage(s) can go into oscillation. The cure is to use anode or grid-stoppers. I slightly "stagger tuned" the IF and reduced the value of one of the cathode by-pass capacitors and hoped for the best. So far, so good. 
The bandspread dial had been incorrectly strung, so this was another minor detail to sort out. 
Finally, I checked the alignment, and anally tuned every band until it was correct to the nearest Hz. The only remaining problem is an intermittent fault with the AVC (I think) which causes the sound, but not the signal to almost disappear.

SO - What's it Like?

If I had had the money to purchase this in 1950, I would have thought it was a terrific radio. Any 7-year old (my age at the time) would. The maker claims it to be a sensitive radio - and so it is, with a better than 1uV sensitivity, it would have received just about anything. I'm not totally sure, but I think it would have received UK TV sound at that time. It would certainly have received 6m Amateur transmissions from the USA, and probably police cars too. I used to listen to them on an ex-government R208 receiver. Nowadays, you will be able to listen to security guards, as I soon discovered while trying to align the set on the 6 metre band.
The set has electrical bandspread on the amateur 80, 40, 20 , 10 and 6 metre bands, but not the 15 metre band, and the recent additions. Later models of the set had the bandspread dial calibrated for 15m. (RUN 4 versions).To use the set for broadcast, it is best to set the bandspread dial in a little bit from its set position, then use the main tuning roughly to the station you want to listen to Next, fine tune with the bandspread. The tuning drive uses dial cord, rather than gears, but seems reasonably (but not entirely) free from backlash. It has quite a nice feel. The tuning scale is large and easy to read - very important for for the class of user this set appeals to! 
At the time of writing, I have not made any attempt to align the FM discriminator, or the IF other than the bad transformer. The crystal filter seems to work quite nicely, with the "phasing" able to remove annoying heterodynes from interfering stations - but at the expense of audio quality. The winners in this notch filter category are the National NC-109, the Drake R8 and the iCOM PCR-1000 .
On a random 3m antenna strung around various objects in the workshop, the set received all the usual broadcast stations. On the amateur bands, SSB reception is possible by backing off the sensitivity control on strong stations. On 20m, stations from San Francisco and Nevada were heard in conversation with French stations. (The AVC is off when the BFO is on). The video (To be posted) will give you some idea of the set's shortcomings when resolving SSB. 
Shortly after I got this set working, an SX-130 came in for a little attention. This is a set from the hey-day of the valve era. Performance was quite similar (subjectively), but the SX-130 has a product detector for SSB. Instead of double conversion, the SX-130 uses very high "Q" I.F. transformers operating at 1650 kHz to provide the selectivity and image rejection. This leaves a "reception gap" from 1600 to 1700kHz. The SX-130 does not receive 6m or FM mode. The overall gain of the SX-130 fell off markedly as the reception frequency increased.
Of course, I prefer the SX-71, but that's probably because I spent/wasted so much time on it. The other reason may be that the double conversion arrangement provides additional gain on the higher frequency bands.
Both of these sets would benefit from a crystal calibrator. (See the circuit for the Morse Code Beacon under projects, but maybe use a more common toob, such as an EF91/6AM6). 

Some questions

Why did they use a different valve for the AM detector and FM discriminator ? I can't get my head around the use of a 6H6 and 6AL5 in the same set. I thought the 6AL5 was merely the miniature version of the 6H6. 
This is a double conversion receiver - why not use a tunable 2nd oscillator to provide equal bandspread at every tuning point ? I think the Eddystone 830 was one of the few sets to do this. There weren't many other radios that used this technique, to the best of my  knowledge. There was one military miniature set that did this, using detents at every MHz to click the first oscillator into position. For the life of me, I can't remember its make (Something like "Military Equipment Ltd") or designation. You would probably have to use a separate RF preselector like the Barlow-Wadley uses, or use a very small tuning range +- 50kHz, say, which would also be satisfactory.
I feel a project coming on...
Another concern is the position of the crystal filter. In sets , such as the Eddystone 940, R-390A, NC-109 and many others, the sharp I.F. filters are all placed immediately after the mixer stage, or 2nd Mixer in the case of double superhets. The idea of this is to prevent blocking and cross modulation caused by a nearby strong station on a weaker station that you are trying to receive. The idea being to keep the gain as low as possible prior to the selective stages of the set. If high gain mixer stages are used before the filter, then its very easy to get a strong signal causing several volts to appear at the mixer - and if there is no filter to cut it out, it can produce (theoretically) several hundred volts by the time it gets to the detector, wiping out anything that you really want to hear. (The triple superhet Barlow-Wadley fights a losing battle to reduce gain by using diode mixers. The design is inevitably gainy, so it is best just to use its built-in whip to keep signals low.  The Barlow Wadley designers still placed the selective filters as early as they could in the circuit.)
All this should be considered against the design concerns of the time. The designers wanted the set to be sensitive and to have good image rejection. They also wanted it to be easily able to tune AM and CW signals, but were not designing for SSB, which was to become popular later on. Things like noise factor (reducing the internal noise from the RF stage) and cross-modulation were not a big concern, since the bands were not yet crowded with high-power stations, as they became to be in the late '60s and '70s. As always, they would be worried by cost.
(I am still in the process of restoring this set, in between using it a lot).
Joomla template by a4joomla