Collins 18S-4

[Opcom]

This 18S-4 transceiver came home with me. I had given it to a friend many years ago. He had wanted to try it out for its proper use CW & AM, but the manual never surfaced and he didn't want to chance it in light of the fact that it's channelized. He liked to tune around looking for CW contacts, being expert in that mode and with a very good ear, he enjoyed picking weak signals out.  Unfortunately, this came from his estate, but I'm glad it was not pitched out otherwise gotten rid of.

In the meantime a dynamotor showed up, a manual also showed up in a building I was cleaning out, and this topic was posted a few years ago.
https://amfone.net/Amforum/index.php?topic=33807.msg261431;topicseen#msg261431

Anyway, I got it back. A fitting it should be on the shelf out in the Bunker, and the manual is before me now. It's amazingly clean on the inside and maybe I can revisit getting it working. One major issue is the rear connectors, but it should be possible to use Molex female contacts to fit those pins. I'm told the 350C-5 shockmount, which is the shelf that secures the radio and includes the connectors is rare and unavailable.

[Opcom]

for some rerason only one attachment can be posted. The (more attachments) doesn't seem to work on this PC. Here is anothe rimage.

[Opcom]

next

[Opcom]

nameplate

[Opcom]

The manual for the 18S-4-C has dates of 1953 to 1960.
It could be considered a channelized version of the ART-13, and the father of the 618S-.

The set has 10 settable channel frequencies covering 2 to 18.5 Mhz. 10 more channels can be used if each one is within 1% of the frequency of a set channel, according to the manual. The 314S-4 control panel has a 20-position channel switch choose the correct crystal. Frequency selection is mechanical with no tuning or resonating discriminators/servos

Channel selection in the unit is by a position-seeking motor and switch setup, with a long shaft from front to back carrying the many channel selection wafers for the receiver and transmitter oscillators and tuned circuits. The motor is held in a resilient frame mount so that in case of excessive torque, the motor frame moves and a limit switch is engaged to stop power to the motor. This protects the gears in the motor unit. The reason given is that the torque required to turn all of the switches can become excessive. This all should be inspected and cleaned/lubricated like any moving parts in this old an equipment. the reprint manual should be available through usual channels.

Channels are tuned up by individual adjustment of slug-tuned coils for each RX and TX stage arranged in the box in the center of the unit, and the output tuning is done by tapped coils and fixed capacitors in the front section. The caps can be individually connected or disconnected if needed. Loading is not adjustable but is set for an expected 50 Ohms. This '50-Ohms' scheme is popular with some other aircraft HF radios like the 618S-1.

Transmit is at the crystal frequency or is doubled.
TX CRYSTALS: XTAL fundamental = F < F=6MC > XTAL 2nd harmonic = F
12AU6 OSC
6AG7 BUFFER
1625 BUFFER/DOUBLER
813 CLASS C RF OUTPUT.
Power is 100W up to 16MC, and 80W to 18.5MC

The modulator starts right out with a 6V6 driving a pair of 811As. A T-17 carbon mike is needed to get enough voltage through the input transformer to drive the 6V6.

The 18S-4C has a diode clipper, the 18S-4A does not. There are a few other differences between the models, mostly to improve audio and comply with more FCC regulations.

The receiver is a single conversion type with a 455KC I.F.
RX CRYSTALS: XTAL fundamental = 455KC higher than F < F=7.5MC > XTAL 2nd harmonic = 455KC lower than F

6BA6 RF AMP
28D7 OSCILLATOR
12BE6 MIXER
Z101 COLLINS 455KC MECHANICAL FILTER
6BA6 1ST I.F. AMP
6BA6 2ND I.F. AMP
6BA6 3RD I.F. AMP
1N52 CRYSTAL DETECTOR
12AL5 NOISE LIMITER + AVC
12AU7 AUDIO DRIVER + SIDETONE OSC
12AU7 BFO
28D7 AUDIO OUTPUT

Reciever audio output 28D7 twin beam power tube is set up in push-pull with no bias. Output is "minimum of 50mW" into 600 Ohms for headphones. The receiver uses another 28D7 for its oscillator. The 28D7 is interesting as a power amplifier which is designed to operate from 28VDC anode voltage and can make 600mW as push-pull class A2 amp.

A meter on the front shows battery volts, grid current, and plate current. 'normal' is in the center area marked in red. There is no other metering, though the companion manual or automatic antenna tuners have meters for those purposes.

Power for the set is 28VDC input, 38A maximum. No 400Hz required for the transciever.
B+ and relays run on 28VDC, everything except the 811As and 813 run from a 400V sypply, and the high voltage is 1150V.
A dynamotor with separate 400V 0.75A and 750V 0.35A sections in series supplies the 400V and 1150V.

[KA3EKH]

K4CHE in Delaware has had one of the set on air for a while now. Think a dozen or so years ago found him the appropriate antenna tuner at Dayton. he built up the rack and connectors. He has some pictures on his web page and also may be a good source of information on doing things like building up the tuning elements for the radio. His web page on that set is:

http://k4che.com/Collins%2018S-4/Page%202%20Station/Page%202%20Station.htm

[w7fox]

Patrick,
Thanks for posting this interesting piece of equipment.  I bought one in 1977 at the local surplus store, and have wondered what it was ever since.  What a treasure chest of parts for $25.  The store had a pile of them in the corner that nobody wanted.  My homebrew equipment uses tubes, sockets, modulation transformer, coils, switches and even the case that houses my linear power supply.  Thanks for letting me know it was a 18S-4.  I wish I had bought two and kept one.
Best regards,
Fox

[Opcom]

Reading the manual way past the point of TL;DR.

It looks like all 10 channels could be devoted to complete coverage of one or two bands within the unit's range. Not fired it up yet - no time so far and have paid work on the bench.

(Note that Going outside the 2-18.5MHz range would require modification to some of the coils. The 813 is only good to 30Mhz at full ratings.)

There are 5 coils in series for the 813 plate circuit. Each coil is tapped in 1-turn increments. The 10 position mechanical channel switch selects, for each position, a tap on this series coil arrangement and a capacitance to make up the tuned circuit with a 50 Ohm output. (the switch has multiple wafers and also selects the other coils in the TX chain as well as the RX chain)

The operator control box switch has 20 positions, so that one of two crystal frequencies can be sent through the circuits set up by each of the ten mechanical positions of the channel switch's wafers. This gives the 20 channels advertised.

On any one of the ten mechanical positions, the two crystal frequencies can be up to 1% apart. Putting those 1% ranges end to end gives a cool 10% bandwidth on any given center frequency.

Therefore, as a single band transceiver for CW and AM phone the 18S-4 looks really versatile. It's also possible that a VFO could be used instead of individual crystals.

Presumably, the 10% bandwidth gives enough coverage to use every frequency on most bands, and even a couple of whole bands at the upper frequencies of it's designed range - thinking 40+20+15M.
The 1% spec is for stated performance. it may be OK to bend that a little. I wonder if the highest Q is in the RF amp output circuit.

BAND: COVERAGE (WIDTH, presumed # of switch positions needed)

160M: 180KHz (band is 200KHz wide)

80M: 350Khz. (band is 500KHz wide)

40M: 700Khz. (band is 300KHz wide, 5)

20M: 1430Khz. (band is 350KHz wide, 3)

15M: 2100Khz. (band is 450KHz wide, 2.14? or 3)

10M: 2800Khz. (band is 1700KHz wide, 6)

I could be wrong about what actually works, but it casts the unit in a much better light for those with favorite bands.