Hallicrafters S72R

[Michael Watterson]

I've been looking for one of these interesting full coverage AM (except LW) battery/mains valve sets at an affordable price for years. There is a LW version, but rarer and more expensive.
It's rather unusual and quite severe on the battery pack, twice consumption due to having EIGHT valves!

See http://www.radiomuseum.org/r/hallicraft_s_72_s72.html

Some photos soon.

I replaced about six leaky waxies. It uses quite a few 5nF ceramic caps as couplers and has some sort of combo ceramic capacitor part on the audio out 3V4 (DL94).
I replaced one leaked, corroded LT 100uF smoothing cap into original card tube. The others reformed OK.
1T4 = DF91
1R5 = DK91
1U4 is a little like a DF92, I think, so only serious AGC on the 1T4 RF stage.
1U5 is an improved 1S5 (reduced microphonics, 1S5 is DAF91) with different connections.
Arrrgh!

It destroyed the DL94 when I connected it to the 110V AC shaver outlet. (The 3V4 in it when it came was no emission).

I'd noticed that the BFO was intermittent. It's got EIGHT battery valves in two paralleled series chains.
It was an Intermittent 0V connection on one of the two serial filament chains in parallel. That's why BFO was intermittent!
So the the two chains ended on separate tags on the main HT smoothing electrolytic and one was corroded. I've replaced wires on both "chains" to the 0V point and put them on same tag, using a single wire stripped in the middle and folded, so that can't happen again!

BFO now fine on the battery pack.
Working now on 110V Shaver adaptor.

Now to finish restoring case. Some torn bits on cloth and hinges were badly bent.
Brass front very corroded and speaker cloth filthy!

It doesn't use the 1R5 (DK91) as a mixer/osc. It uses separate valves for mixer and osc, so it's reliable to 30MHz even at end of LT life. It's got 2 x 1U4 as IF amp and a tuned RF amp (1T4) and separate valve as 455MHz local oscillator for BFO (CW or SSB).

3V4 (DL94) as single audio out, I checked grid volts as it gets warm, but was OK.

Chassis working fantastic on SW on bench with a 2m wire. Will try loop aerial on MW after dark.

All service info on web seems to be earlier version(s). I think mine is 1958.

[crackle]

A rather complicated battery portable for use on the beach.
Is V8 the BFO?
Its great its all working again now.
Mike

[Michael Watterson]

On Sams "Photofacts"
V8 = 3V4 (DL94) is output
V7 is BFO, 455kHz oscillator, 1U5, an improved 1S5 (DAF91) but incompatible. I don't know exactly how it works, except g2 is tied to anode to "triodise" it. I don't understand how it's mixed with IF.
The diode part and a double pole switch on C17 and C13 are probably the ANL missing on my S72R, only seems to be on early 1949 S72 models. I can't imagine it worked very well.
V6 (1U5) is the normal diode detector feeding volume control and pentode audio preamp (like a DAF91, except not) feeding V4 the output pentode.

Actually that's only on the Sams "Photo facts" folder dated 1950. The "operating and service instructions" are far more detailed and omit the ANL capacitors, diode connection and double pole switch.

The Hallicrafters official circuit does have
V8 = 1U5 is BFO, 455kHz oscillator
V6 = 3V4 (DL94) Audio out
V5 = 1U5 diode / pentode Detector / audio preamp.
V7 = 1R5 (DK91) is oscillator only, with V2 (1U4, approximately a DK92) as Mixer, RF and Osc both to g1!

Switching on the BFO disables AGC, pot R8 (500K) is between +90V and 0V and drives g2 of the 1T4 RF amp and 1U4 first IF amp, for RF/IF gain. The normal AGC line from detector diode (V5 1U5) drives the 1T4 g1 in normal way, so there is never AGC on the pair of 1U4 via their g1, which makes sense as they are not variable mu. The trick of using g2 voltage for gain control is common on sharp cut off tubes.

I'd appreciate an explanation how the 455KHz oscillator is mixed into IF!

The early version looks like a perforated metal grill is on speaker, mine and any photos I've seen are all S72R and cloth at speaker.

I need a stronger 110V transformer. The shaver transformers are designed to power limit at under 20W to meet safety specs. The unit supposedly takes 25W, though that would be 217mA, which sound far to high. I can't see why it's not 13W to 15W.

I replaced the selenium rectifier with a 1N4007, leaving it in place and using the cathode tag, so I put heat shrink sleeve on the anode of 1N4007 to the dropper resistor on the AC/DC in side.
I think if I had a decent rating 115V transformer, the R33 (22 Ohms 2W) would need an additional series resistance.

[Michael Watterson]

Some other links
http://www.ohio.edu/people/postr/bapix/s72.htm

he says

A manual and schematic are available from BAMA in djvu format. Pages from the Sams photofact are also on BAMA. I found an obvious mistake on the photofact. Sams indicates a current draw of 0.37 amps at 117 VAC. In later testing, I found that the set drew 0.22 amps at 117 which is in keeping with the 25 watts as specified in the Hallicrafters manual. I determined that my set was likely Run 4 which conformed to the manual on BAMA. An earlier version of the S-72 had a separate noise reduction switch on the front panel and located the bandspread capacitor on the underside of the chassis. My version places the bandspread cap on top of the chassis.

Why does it take more than say 120mA? Nominal 50 mA types, two sets of series in parallel. Must be mostly due to R29A, 270 ohms to 0V

[Michael Watterson]

I'll upload photos tomorrow.
Now safely in the restored cabinet. I added 47 Ohms to replace drop of selenium rectifier and wire snapped of at body on the existing 22 ohms. So I replaced it too.

My local radio guru suggests the BFO couples into IF just by leakage/stray pickup!
Not too hard to tune SSB on 7.1MHz or CW on 7.0MHz, good reception 5pm today on the built in telescopic aerial.
No BFO on mains, only on battery, I need a better 230 V to 115V transformer. I'll check though that the filament dropper hasn't gone high and what the AC is dropping to.

[crackle]

You mean there is more than one radio guru in Limerick.
Mike

[turretslug]

My local radio guru suggests the BFO couples into IF just by leakage/stray pickup!

May well be - the AR88 does this with its BFO, the manual talks about "electrostatic coupling", in practice this amounts to wiring proximity between BFO and final IF amp! Good enough for pre-SSB use, where HF gain is adjusted manually for a satisfactory beat note strength on CW but some add a small coupling capacitor for less tricky SSB demod.

[Cathovisor]

My local radio guru suggests the BFO couples into IF just by leakage/stray pickup!

May well be - the AR88 does this with its BFO, the manual talks about "electrostatic coupling", in practice this amounts to wiring proximity between BFO and final IF amp! Good enough for pre-SSB use, where HF gain is adjusted manually for a satisfactory beat note strength on CW but some add a small coupling capacitor for less tricky SSB demod.

As does the FM tuner head of the Pye FenMan II: the local oscillator of the FM tuner couples into the mixer solely by stray capacity.

[Michael Watterson]

[, in practice this amounts to wiring proximity between BFO and final IF amp! Good enough for pre-SSB use, where HF gain is adjusted manually for a satisfactory beat note strength on CW but some add a small coupling capacitor for less tricky SSB demod.

Curiously the BFO coil and circuit isn't that near the IF!
I tested it with SSB on 3.6MHz and 7.1MHz, as well as Shannon Air Radio (about 5.5MHz), which must have a slight carrier as it works slightly on some AM sets. It was actually OK on SSB, with the "bandspread" tuning it. CW was easier to tune, but that's true even on my modern HF radios.

Before any restoration, as received.

After restoration

Insides with a pair of phones I have in the manual's mounting position:

You can see my replica RCA pack (with alkaline D cells for LT and 60 off alkaline AA cells for HT, pinched from my RCA Globetrotter. My Pye "World radio" has same pack, but in Ever Ready package.

[Michael Watterson]

Front of chassis

Underside, caps replaced, but before I added extra dropper.

Centre top, beside wave change.
The mysterious pair of caps with wire around them replace a single waxy. The coil is a series RFI trap. This goes between 0V line and the chassis. As is common with USA AC/DC direct plus battery portables in USA, the chassis isn't at DC 0V, but only an RF ground.

The orange carton had a corroded 100uF 50V cap on the LT supply. It has a new cap inside it.

[crackle]

That looks a very well made and featured radio, in pretty good condition too by the looks of it.
Well Done
Mike

[Michael Watterson]

Rear view

It was filthy and corroded on outside, speaker cloth badly faded but had moved. I touched up cloth with a metallic "sharpie" to match un-exposed part!

There are some torn bits I need to repair, I put matching brown acrylic paint on the bare wood for now.

it's quite an amazing performance battery set. I think maybe Sony ICF5900 was the first better consumer transistor set with BFO. Most cheaper "World" radios have no BFO.

Edit:
Also rear door hinges were badly bent.

[turretslug]

That is an unusual and intriguing set, and must have been recognised as a very good performer in its time, having separate LO and 2x IF amps put it above the crowd. That LO/mix valve choice and arrangement is particularly noteworthy, no doubt there was credible technical basis to it at the time. Definitely a set that uses your battery wrangling know-how!

I'm sure I recall something about Shannon airmet (perhaps others) using reduced carrier SSB, maybe that helped older, simpler aircraft receivers use some kind of pullable lock product detector set-up. It does make it discernable using an ordinary AM set, albeit with muffled, "puffy" audio.

[Michael Watterson]

That is an unusual and intriguing set, and must have been recognised as a very good performer in its time, having separate LO and 2x IF amps put it above the crowd. That LO/mix valve choice and arrangement is particularly noteworthy, no doubt there was credible technical basis to it at the time.

Yes, also separately tuned RF preamp. I think only Amateurs/Hams would have realised in 1950s to 1970s how much better it was.
Almost all the Military and Commercial radios use a separate oscillator for stability and separate mixer. The USA often used Octodes & Heptodes as mixer / Osc on mains sets, but from mid 1930s recognised that even on domestic mains sets the separate Mixer/Osc, if even one envelope was better, the ECH and later UCH series. But Military and Commercial sets up to transistor era tended to use a separate triode oscillator, often using gas voltage stabiliser and either Pentode, Hexode, 7660 beam switching tube, or ring diode mixers.

The Octode/Hexode, or self oscillating mixer triode* (or Pentode) has several problems:
1) The upper frequency limit is lower specifically on Octodes and Heptodes (1R5/DK91 and DK96, though the DK40 and 1L6 and DK92 are better, as mixer osc unreliable above 15MHz, fine as oscillator only to over 30MHz)
2) A lot of Oscillator leaks out of RF in pin, hence on VHF the EF80 preamp is more to block LO leakage than actually add gain! RF pre-amp even on USA MW only sets to reduce re-radiation! VHF battery valve sets need very careful alignment to avoid being transmitters. All Battery sets with no RF amp and Octode or Heptode re-radiate LO badly. Nasty on a long wire!
3) A strong signal will Frequency modulate the oscillator! Causes distortion.
4) An adjacent strong signal will pull oscillator as well as causing FM. This causes distortion and reduces IF level.

(*Pentodes and Hexodes have been used as combined mixer/osc in early superhets, but EF80 and DF97 are actually wired as triode mixer/oscillators. Early transistor sets sometimes used separate mixer/osc, but by 1960s all consumer AM sets used triode like combined mixer/osc single transistor. Inferior to separate mixer /osc. The Sony ICF5500 Captain 55 uses separate mixer and osc on VHF, which is unusual for a consumer set. It's more "up market", but the ICF5900 is a bigger step as it has BFO and dual conversion on HF, it reuses part of FM IF as 1st IF on HF.