Hallicrafter S38B restoration: the saga

[worsthorse]

I am not much of a boat anchor guy but I really like the pre-jetsons space agey look of some of the broadcast and communications receivers from the mid-fifties. Combine that with too much free time for looking at ebay and this happens:



This is Hallicrafter's take on the All American Five, a four band AM/CW receiver with a band spread control and requiring an external antenna called the S38B (I think there were five versions of this radio, all based on the same design). It was manufactured between 1947 and 1953. I am going to learn how to restore old radios by restoring an old radio, I guess.

Here's what it looks like inside:



A lot of dirt and grease up top. A little less underneath. I did a brief power up test and, of course, it didn't work. Probably one or more bad tubes, the filaments of which are wired in series.   After doing a little research, here are the first things I am going to do:

Inspect for damage and mods made by previous owners,
Replace the existing wax paper and electrolytic caps,
Clean the chassis, top and bottom,
Install an isolation transformer and fuse (more on this later).

After doing the inspection, I found minor damage to the speaker cone (which I can repair) and no obvious mods or rework other than the lamp wiring replacement, so I set about cleaning the chassis. Here's what it looks like after four or five hours of degreasing and removing corrosion, etc.  The tools of choice: simple green, a mild degreasing solvent, brass-x, toothbrushes and a pile of cloth rags.



The smear on the chassis is from varnish left behind when someone scraped a label off the chassis.  It was removed after this photo. Next up, recapping...



The bottom after the recapping:



The S38B, like all the AA5 radios, was heavily cost-engineered to the point of being dangerous (hence the isolation transformer...) with the result that the physical placement of components and wiring markedly effects the radio's performance.  When replacing components ignore matching lead length and component placement at your peril. I did not.

While I wait for replacement vacuum tubes to arrive, I am going to install the isolation transformer and see if figure out how the circuit actually works.  I haven't tried to decipher a tube-based circuit in, I dunno, forty years?

[bob91343]

I appreciate and commend your efforts.  Having said that, I am amazed at the great popularity of these cheap radios.  The S-38 series had to be the worst possible short wave radio.  The original S-38 had six tubes, including the BFO that they eliminated later in favor of making the IF amplifier oscillate.  They made all user accessible surfaces of nonconducting material to justify the transformerless design, and had to struggle a bit to connect an antenna and headphones but they found a way.

I recall, even in my youth, reading the Hallicrafters motto 'the radio that amazes the experts' and thinking yeah sure, amazed that anyone would try to sell something with such poor performance.

Yet they are still in use and those who use them like them.  I once owned one but was quick to sell it.  The later versions changed design to a slide rule dial style.  Lafayette imitated the appearance with a somewhat better radio.  And National imitated it with the SW-54.

I never felt that an uncalibrated bandspread dial was of any value.  In those days, hams hated dials with strings and preferred gears.

Because I am so old, I remember many of these units and have played with them.  I formulated opinions which have stood the test of time.

[worsthorse]

I appreciate and commend your efforts.  Having said that, I am amazed at the great popularity of these cheap radios.  The S-38 series had to be the worst possible short wave radio.  The original S-38 had six tubes, including the BFO that they eliminated later in favor of making the IF amplifier oscillate.  They made all user accessible surfaces of nonconducting material to justify the transformerless design, and had to struggle a bit to connect an antenna and headphones but they found a way.

I recall, even in my youth, reading the Hallicrafters motto 'the radio that amazes the experts' and thinking yeah sure, amazed that anyone would try to sell something with such poor performance.

Yet they are still in use and those who use them like them.  I once owned one but was quick to sell it.  The later versions changed design to a slide rule dial style.  Lafayette imitated the appearance with a somewhat better radio.  And National imitated it with the SW-54.

I never felt that an uncalibrated bandspread dial was of any value.  In those days, hams hated dials with strings and preferred gears.

Because I am so old, I remember many of these units and have played with them.  I formulated opinions which have stood the test of time.

*laugh*  No argument with anything you said. I am fairly certain that the S38B is not going to replace my FT1000D or even any of the single band CW only rigs I own, like the QCX, in my shack. That said, it is kind of cool looking and works okay for casual SWL which is all I expected.  And given the circuit simplicity and construction, it is a good introductory restoration project. If I get it working I will set it up in my office next to the equally anachronistic Numechron Tymetor desk clock.

[worsthorse]

Installing the isolation transformer and a fuse.  First why...



In the cost-savings derby, power supplies and transformers just aren't going to fly. The S38B used a series string of filaments, powered by the AC line instead. That means it has an unpolarized plug, which allows the chassis to be hot even if the switch is in the off position.  The fix was to isolate the chassis from the case, which is better than nothing and still not very good, from a safety perspective. Here's the schematic:



I broke the connection at the power input and inserted a 35VA isolation transformer and fuse:



This is the only place the transformer will fit. Caution required, because the speaker is right above it and mis-placement means the case won't fit...



Isolation transformers are a topic of much animated discussion. My only concern is safety. My only worry is if adding one will cause problems because there's no earth ground connection... 

[retrolefty]

We seem to have followed similar paths. My first SW/Ham receiver was a Hallicrafter SX-110 in the 60s, looked really great but performed poorly. Spent many years trying out others receivers and stopped collecting with a Drake R-4A. Now that is a nice looking receiver, great tuning feel. and nice performance. Now they buy a little SDR module dongle and your receiver is a video game on your PC. Blah. However the waterfall display sure looks useful. 

[bob91343]

We seem to be in accord here.  My current transceiver is a TS-940S with TS-440S backup.  Those old tube radios couldn't hold a candle to these 'modern' units (mine are from the 1980s).

There isn't much SWL fodder these days.  Foreign BC seems to have died for the most part.

The isolation transformer is a good idea and, since you have gone that far, you can defeat the grounding isolation in the radio and connect antenna and then ground the chassis.  You may experience hum problems but those were endemic to the radio.  They saved on a speaker by making it small and got the side benefit of it not responding well to hum frequencies.

As I recall, the MSRP was $39.95 for a long time and then rose to nearly double that.

Hallicrafters made the prettiest radios.  While the competition was busy making nice radios, they couldn't sell as many as Bill Halligan could because of the snazzy looks.

I owned an SX-25 for many happy years.  My transmitter was Hallicrafters also, the HT-18, supplemented by a James Millen 90811 (I think).  But I am giving away my age.  Look on my QRZ page to see an old pic from about 1950.

Bob K6DDX

[worsthorse]

 

hum... lots and lots of hum.  more hum than is warranted. modulation hum and standby mode hum.  so... there's leakage somewhere or a short somewhere, or a bad rivet, or some wire in the audio section is picking up hum from some wire in the power circuit.  first quick check, swapping tubes, did not change anything.  second quick check, all the recapped capacitors are in the right places and the isolation transformer is wired correctly.

note to self: yeah, i know it didn't work when you got it but it would have been better to recap, replace bad tubes, and check performance before adding the isolation transformer. because now you have to take it out to do exactly that.

so i will unwire the isolation transformer and see what the hum is like. and try not to electrocute myself in the process. 

[bob91343]

I suspect much of the hum is from power supply ground loops.  To do it right they would have had to spend a nickel more, thus decimating the profit.

I recall helping a friend with an audio mixer that had hum.  The factory said no problem we have an update so he sent it in.  Came back, same hum.  I looked it over and made a few changes and the hum was inaudible.  All because some of the filter capacitor ripple current was allowed to flow through a path common to the signal.

In the S-38 the first thing is to see if the 12SQ7 has heater to cathode leakage.  If not, it's a loop.  Or possibly a high ESR filter capacitor.  Is the hum clean or raspy?  Clean comes from power line; raspy comes from filtering.

[rcbuck]

My first receiver as a novice licensee in 1962 was the S-38E. The entire 80 meter was covered in less than an inch of dial space. The band spread control helped but you get used to hearing 5 or 6 CW signals at a time. Your ears are the best filters. I still prefer a bandwidth of about 2 kHz when working CW. I made a lot of contacts with the S-38E and my Ameco AC-1 transmitter.

[worsthorse]

Haven't been able to work on the S38 for almost three weeks. Finally had a chance to get back to it today.  The radio is working but has two sorts of hum problems. First, I can hear hum when the radio is in standby mode. Second, the hum is louder when the radio is turned on but I have not tuned a station.  Third, there is modulation (aka "tunable") hum: the hum level is correlated with the strength of the received station.

Since I wasn't able to test the receiver before installing the isolation transformer, I couldn't rule out that the transformer as the problem, so took it out of the circuit and connected the mains directly to the AC circuit again.  Powered it up and no change.  Powered it up with the two prong plug reversed. No change.  Powered it up through an external isolation transformer. No change.

Good news because it is now unlikely that adding the transformer caused the hum problem.

So… time to check the capacitors I changed out, along with looking for questionable solder joints and bad rivets (the grounding points for many parts of the radio connect to a ground lug strip that is riveted to the chassis).

Here's what I found:

I made a wiring mistake with one section of the multi-section capacitor that filters the AC. The rest of the capacitors were wired correctly and the soldering was good.  So, fixed my mistake.

I found several old solder joints that were, at best, questionable.  Fixed those.

I found two questionable rivets. Drilled them out and replaced with machines screws and lock washers.

Power the radio back up.  Much better!  

The hum on standby is very very low now. Not gone but I wouldn't expect it to be given the nature of the design. The hum "between stations" is the same level as the standby hum.  I then ran through the three tests: AC direct, AC direct with the plug reversed, AC through the external isolation transformer. Actually improves a little in the last case. 

The modulation hum is still there, though not quite as bad. Did some research and here's the hypothesis:  Very strong BCB signals are picked up by the house wiring, which is a very long wire antenna, get modulated by the 60Hz power line, and injected into the radio through the power cord. The closer a station is tuned in, the more of the hum gets through the filters around the IF stage.

I need to think more about that as a possibility. I could test it by putting an appropriate capacitor across the line. I could also do some tube swapping and see if there's a leakage issue in one of the tubes. But first…

put the isolation transformer back into the circuit and check to be sure that the hum problem doesn't re-appear and give the chassis one more going over, looking for poor solder connections and bad rivets.

[bob91343]

I am guessing that you have been spending more time on this poorly designed radio than the original engineers did.

Well, maybe not poorly designed as much as designed to a poor $pecification.

[worsthorse]

I am guessing that you have been spending more time on this poorly designed radio than the original engineers did.

Well, maybe not poorly designed as much as designed to a poor $pecification.

*laugh* perhaps. figuring out how stuff works and, when i can, making stuff that doesn't work so well work better is a lot of fun.

[worsthorse]

I was able to spend a few hours at the bench, working on the S38B.  After testing the radio with and without the isolation transformer, it was clear that the isolation transformer wasn't making the hum problem worse.  After looking at radios of this era that had a proper multi-tap transformer, which looked like this (from Elements of Radio Servicing):



I duplicated this with my isolation transformer, although with just one winding since the filament string uses line voltage.  On the input side, I added a in-line fuse. The on/off switch is on the primary side, too. There is a 0.01mfd X2 capacitor across the primary side.  This is in place to reduce modulation hum and it works better, or at least as well, as a 0.01mfd from one side of the AC line to ground.  With these changes, I could keep the original two prong plug while making the radio safer to use.

After all that, the modulation hum is gone but there is still a low level hum that can be heard even in standby mode. There are, I think, two reasons for this. First, the radio design, which I can't really do much about. Second, I have power line noise here that shows up to a greater or lesser extent on every radio (no matter how expensive) around here. That being the case I am going to declare victory on this one.

On the outside, before re-assembling it, I cleaned the case, cleaned up the paint splatters, repaired a couple of dings in the paint, and cleaned the plastic window (carefully, it scratches ver easily).  Now it looks like this, not quite brand new but not bad:

Here's the front



The back



And the inside, from the back



I spent the evening tuning around and listening to AM SW stations and, yes, CW on the amateur bands.  As expected the amateur CW sub-band covers a very very small portion of the dial but the band spread actually works pretty well.

The weirdest thing was listening to FT8 signals on a radio built in 1953ish. I was briefly tempted to see if I could decode FT8 and decided not to torture myself with it.

Anyway, this one is done. At this point, reducing the hum further would require time and energy I want to invest in other projects.

[bob91343]

Good job!  It looks great in the pics.

One thing gnaws at me, that you left the filament string on the primary side.  At least that's what I think you did.  The secondary side needs some sort of control so that its voltage doesn't drift to a dangerous level.  The heater to cathode ratings of the tube need to be respected.

The hum is either due to a ground loop, a leaky heater to cathode, or insufficient filtering.  Is it smooth 60Hz, or raspy open grid, or 60 Hz with harmonics from insufficient filtering?

You are wise to avoid attempting to decode FT8.

[worsthorse]

The filament string is on the secondary of the isolation transformer as expected. Sorry for the confusion.

It might be a ground loop somewhere that I missed. I have swapped the most likely suspect tubes with no change.  I've also recapped it and added a filter cap on the primary side of the isolation transformer. It is a smooth hum and it is apparent when the radio is in standby mode.  I could start looking around with an oscilloscope, which would be the next step, but I am going to let it go for now, though I may come back to it after I get a couple of other more pressing projects in hand.

It does look pretty good though, doesn't it?   

[CaptDon]

I owned an S38B bought used from Jim Strauss K3JFL at
J.R.S. Distributors in York Pa.  I wanted to listen to Icelandic
broadcasts on 15.770mhz and my Kenwood twins R599/T599
did not have those SWL bands included I did extensive mods
to the S38 including huge filter caps (1000uf@200vdc), replaced
all coupling caps and cathode caps and did a 'peak' alignment
which may have made it a bit more narrow banded but slightly
more sensitive. The weather, moon phase and time of year
hugely effected the signal strength. Signal hops at that frequency
tend to have 1200 mile intervals and I was at about 2800 miles
a bit beyond the second hop but I loved to listen on that radio.
It had a much better speaker put in and had a great AM SWL
sound to it. My girl at the time was from Iceland and spoke also
Danish and German and she loved that little radio listening to
Danish, Norwegian and German broadcasts. The KW Twins were
Ham Only bands so she liked to play with the S38.

[bob91343]

A smooth low pitched hum is likely pure 60 Hz.  It may be magnetic coupling from the transformer to the voice coil or output transformer.  When in standby, what's still energized (besides the heaters)?

[Noddy]

You could have used the space to install a proper and safer power supply instead of an 'isolation' transformer'
You could then have earthed the chassis as opposed to never really knowing which mains lead is active and which is neutral.  Isolation is better than the spine tingling and horrible 'mains' filament arrangement to chassis but not good. A 'proper' tranformer power supply and using 6 or 12 volt filament tubes in parallel fed bya filament/heater supply at that extra low voltage.

I own one of these sets with 'proper'power supply and then found B+_on the S-meter...!! I don't think they are much of a set but were built to a "$' market. Use th components to build a better receiver on a circuit from say ARRL/RSGB/Practical Wireless....Include an rf stage then it will be a good little bandspread receiver.