R216 Receiver

 

 The R216 is a portable "Intercept Receiver" used in the 1950s for example for eavesdropping on VHF radio signals emanating from East Germany. Several types of receiver were used for this general purpose including various Eddystone models but, comparing the two makes, the R216 (made by Ekco in the early 1950s) I think is superior to the Eddystone types (and which anyway do not offer the portability of the R216)and looks more professionally made... no doubt to an unrestricted "cost plus" budget. Frequency coverage is pretty wide and the receiver can demodulate AM, MCW and FM or CW with selectivity at 30 or 120KHz. (The R216 looks similar to the R210, a short-wave receiver). During 1965 the R216 receiver was fitted in a new Canberra variant, the B Mk6 which was effectively a sort of GCHQ outstation (ref RAF Strike Command 1968-2007: Aircraft, Men and Action).

 Band

 From

To

 1

19MHz

30.5MHz

2

29.5MHz

47MHz

3

 45MHz

 69MHz

4

67MHz

103MHz

5

100MHz

157MHz
 

 

 As you can see in the above picture the set uses a 4-section turret tuner, seen below from a couple of views.

 

 

The use of a turret tuner allows the coils for the different wavebands to offered up to the RF valves without the need for extraneous wiring (like that used in the R208). 

 

Above is a view showing the chassis of the IF/AF unit (note V12 is missing ... perhaps used to replace another in a more important role). V12 is used to provide an IF output (at 4.86MHz) for connection to alternative demodulation equipment.

Below right is the RF chassis which includes the front-end plus the crystal calibrator (top) and left the IF/AF chassis. The curved aluminium panel marked "Serial 593" encloses the rear of the dial. Should the serial number originally have matched that of the receiver which is 597? Perhaps an inadvertent switch was made when mods 1 & 2 were being incorporated (see mod record panel below)?

 

 

 

 
 The valve line up is slightly odd as it uses a mixture of both battery and mains valves possibly because VHF battery valves were relatively uncommon, although one might query the use of the EF91 as crystal calibrator and 1st IF amplifier (the simple explanation may have been to just keep the valve heater types common to each of the two chassis, but then... why use the EB91.. was there no suitable 1.4 volt equivalent?). Note that the valve listing doesn't include an RF mixer. For some odd reason the designer chose MR1, referred to as a "Westinghouse silicon diode rated at 19 volts and 80uA" which carries the code "KMSA" which I'm assuming was the manufacturer's part number:-

 Valve Designation

 Valve Type

 Equivalent

Function 

 V1

 CV138

EF91

1st RF Amplifier

 V2

CV138

EF91

2nd RF Amplifier

 V3

CV138

EF91

Local Oscillator

V4

CV287

QS150/15

Voltage Stabilser

V5

CV138

EF91

Crystal Calibrator

V6

CV131

EF92

1st IF Amplifier

V7

CV785

DF91

2nd IF Amplifier

V8

CV785

DF91

3rd IF Amplifier

V9

CV1758

DF92

BFO

V10

CV1758

DF92

4th IF Amplifier

V11

CV1758

DF92

Limiter/AM Detector

V12

CV1758

DF92

IF Output buffer

V13

CV140

EB91

FM Discriminator

V14

CV785

DF91

AM AF Amplifier

V15

CV140

EB91

Detector & Noise Limiter 

V16

CV1758

DF92

AF Output

 MR1

 CV291

 KMSA

 Microwave silicon diode
 
 

Above and below... RF and IF/AF circuits of the R216. Click to see a larger image complete with component values.
 

Of course you'll have noticed the R216 uses an external power supply: Two versions were used, AC and DC connecting via the multi-way "Plessey" connector PLE on the front panel (and SKE on the PSU). Below is an enlarged view of the plug with its connections.

Note that the external DC power supply includes a second battery switch made by a link in the receiver across pins M and L which is seen once the lead from the DC power supply is plugged into the receiver. M is fed by the fuse following the power supply main on/off switch connected to battery positive terminal and L is the input to the battery power supply main circuit. Maybe a method of saving battery life when the DC power supply is left switched on when a receiver isn't plugged in?

 

 

 You'll note that the wiring takes account of sharing currents to minimise cable losses and the resultant voltage drops.

 

 PIN

 VOLTAGE

 CURRENT

 NOTES

 A

 0V

400mA DC 900mA AC
Ground return for V6, V7, V8, V9, V10, V11, V12, V13, V14, V15, V16

 B

 1.4VDC

 400mA DC
 Filament supply for V7, V8, V9, V10, V11, V12, V14, V16

 C

 6.3VAC

900mA AC max
 Heater for V3 + V5 + dial lamp return

 D

 250VDC

43mA DC
 HT supply for valves V1, V2, V3, V5, V6

 E

 100VDC

26mA DC
 HT supply for valves V7, V8, V9, V10, V11, V12, V14, V16

 F

 -19VDC 

15uA DC
 Grid bias supply for valves V6, V7, V8, V14, V16

 G

 6.3VAC

900mA AC
 Heater for V1, V2, V6, V13, V15

 H

 0V

1.2A AC
 Ground return for V1, V2, V3, V5

 J

 19VAC

300mA AC max
 Dial lamp supply voltage (return via Pin C)

 K

 6.3VAC

 -
 Not shown in circuit diagram, not connected

 L

-

 -
 DC power supply battery positive linked to Pin M

 M

 -

 -
 DC power supply battery feed linked to Pin L

 
   

 This is a close up of the R216 power connector. You'll note that there are two polarising techniques used. The plastic insert has a couple of protruding pips and the outer aluminium shell has five slots cut in the inner section. Within the assembly there are also six slots allowing the inner plastic part to be located correctly. The five slots give six available orientations.

These connectors were made originally by Plessey and their code number for the "Free Socket" mating connector to this is 508-1-40009-320. The corresponding NATO code is 5821-99-638-1648.

Weald Electronics now make these connectors under their LMF/LMG range. The code for the mating part for the R216 plug is LMF/1/40009/320 clearly lined up with the original Plessey code to avoid confusion.

The relationship between those five shell slots and orientations is given by the last digit in the final 3-digit part of the code number. So you can have 320/321/322/323/324/325. The picture opposite has the polarisation pip between Pin A and Pin B symmetrical between the two pairs of shell slots. This is given the 320 orientation number. As the pip rotates through 60/120/180/240/300 degrees the number increases by one 321/322/323/324/325.

For those unfamiliar with battery valves I'll explain the need for a grid bias supply. Most valves, battery or those with indirectly heated cathodes ("mains" valves) draw anode current when an HT supply is connected. The value of this current is determined by the valve characteristics which call for a voltage between control grid and cathode. In the case of "mains" valves the addition of a cathode resistor will result in a negative control grid voltage (bias). The higher the resistor, the greater the bias and the less the anode current. This method of biasing is usually termed automatic bias and roughly maintains a constant average anode current. With battery valves the use of a cathode resistor isn't practical so each valve may need a separate grid bias voltage to set its anode current. The designer will work out the optimum value for bias voltages. One could just ground the bias lines but this would result in increased anode currents, and although may not affect operation much (it might produce distortion), it will reduce the life of the HT battery. Another point to mention is the decoupling capacitor used in auto-bias. One can tailor the effect of auto-bias on the frequency of the signal being amplified by altering the cathode resistor decoupling capacitor, making it anything from nothing to several tens of microfarads. Although the anode current of the valve is roughly governed by the cathode resistor (say 600 ohms) working at DC, a 25uF capacitor will have an impedance of about 6 ohms at 1KHz thus shunting the 600 ohm cathode resistor and much lessening the instantaneous biasing of the control grid and producing a 1KHz current at the valve anode in excess of the DC value.

A bias voltage can not only modify the valve anode current, but with some specially designed valves called variable mu valves, increasing the bias will reduce the gain of the valve. This technique is very useful in receivers which might have to deal with signals ranging from 1 microvolt to hundreds of millivolts. To handle this range of signals an AGC or AVC circuit is usually employed. The amplitude of the received signal is converted to a negative control voltage and applied as feedback to the amplifying circuits. Their gain is reduced, the amplitude drops and the amount of feedback is reduced. Very rapidly the receiver will establish a more or less constant output no matter what the received signal strength. This protects an operators ears but there should ideally be an S-Meter as all signals will sound much the same. It depends on exactly what the receiver was designed for as to whether an S-Meter was included or not. For some reason the R216 does not use AGC, instead it has an RF gain control. This sets the overall bias level for those valves arranged to be controllable.

In the R216 you'll see CV138, EF91 valves which are variable mu types, as are the CV785, DF91. Type CV1758, DF92 is not variable mu and if you examine the circuit diagram you'll see that these are operated under zero-bias condition (control grids DC-wise tied to chassis).

The first step in getting the receiver working is to build a power supply. As I already have a general purpose power unit (see below) for running an R1155 separately from the T1154, I decided to modify this so it would be suitable for the R216 whilst still enabling it to be used with the R1155 (this means isolating the HT negative from ground because the R1155 uses this to develop an extensive network of bias voltages.. perhaps via a front panel switch which can be used to ground or open HT negative from common ground).

  

The R216 requires several voltages which I'll summarise...

1.4 volts DC @ 400mA. This supply is extremely critical as it's used to feed the filaments of the battery valves. I chose to use a linear voltage regulator type LMS1585ACT-1.5 whose output is a nominal 1.5 volts (1.47 to 1.53 volts) at up to 5 Amps and which requires an input voltage of 2.9 volts to 14.5 volts. The device requires only input and output smoothing capacitors plus a "minimum drain" resistor of about 150 ohms across the output. The circuit is fitted to the left of the right hand mains transformer above and is driven from the 6.3 volt winding of the transformer on the left and comprises a full wave bridge rectifier, the LMS1585ACT-1.5/NOPB, two 2200uF smoothing capacitors and a small 47uF output capacitor (needed for stability). A 150 ohm load resistor is fitted across the choc block terminals carrying the output voltage. Because of the low voltage it's important to minimise the output wiring length to avoid voltage loss when current is drawn. Using four 3.3 ohm resistors wired in parallel this supply drew 1.8 Amps with the output voltage maintaining over 1.47 volts.

100 volts DC @ 26mA. This is the HT voltage for the battery valves and is supplied from the basic HT line of 300 volts. In the centre of the baseboard you'll see a tagstrip carrying two plastic power transistors type BUT11. The left transistor is a series pass regulator with its base connected to a pair of 51 volt zener diodes connected in series and fed with a 22Kohm feed resistor from the 300 volt HT line. The emitter of the BUT11 is maintained at 100 volts. I tested this supply using a 1.1Kohm load resistor drawing 88mA with the output dropping from 100 to 99.8 volts.

250 volts DC @ 43mA. Using the same technique as the 100 volt stabiliser, the base of the second BUT11 is fed with a chain of five 51 volt zener diodes and a 8.2Kohm resistor to provide an output of 250 volts at its emitter.

6.3 volts AC @ 3A. This is supplied directly by the right hand transformer, above.

-19 volts DC @ 15uA. This voltage needs to be perfectly hum-free. I fitted a third mains transformer to the baseboard for the bias supply (and also the dial lamp supply). The total power drain for this circuit is pretty low so the transformer needs a rating of no more than say 10VA. To ensure compatibility with the dial lamp AC supply (as in the official R216 power supply) the DC output will be provided by a half-wave rectifier followed by a twin pi-filter using two 22 Kohm resistors and three 2200uF capacitors.

19 volts AC @ 300mA max. In the official R216 power supply this voltage is supplied by the same winding as the bias supply and I used a small standard transformer with twin windings to match the requirement.

  Metering of the various voltages is carried out via the rotary switch in the centre of the front panel connected to a voltmeter and to monitor HT current I used a milliammeter. This indicates the drain of the main 300 volt HT supply, but because the HT supply also provides feeds to the 100V and 250V stabilisers I arranged the wiring to take the feed for the zener diodes from the input of the milliammeter rather than the output. This ensures that only the HT consumption of external equipment will be monitored. The latest part of construction is the addition of stabilized voltages minus 19 volts, 100 volts and 250 volts (left to right above). I used an 18 volt zener for the 19 volt supply (a PNP series pass transistor) and 51 volt zener diodes using two for the 100 volt supply and five for the 250 volt supply (both with NPN series pass transistors).

A permanent R1155 supply cable is fitted with a Jones plug and the supply to the R216 is fitted with a 12-way Plessey plug.

The right hand transformer (top picture) carries twin 6.3 volt 3 Amp output windings tapped at 4 volts and 5 volts. The twin outputs will be wired so that they can supply either 6.3 volts 6 Amp or 12.6 volts 3 Amps. The main HT rail is wired for the R1155 to supply outputs from before and after the HT choke. Here's a circuit diagram (liable to change).
 

 Note the need to have dual switches for the voltmeter because of the need to monitor one negative supply as well as the four positive supplies. The meter has a 0 to 3 scale which is read as 300V for the HT supplies, -30V for the grid bias voltage and 3V for the filament voltage. The movement required 300Kohm for each HT range, 30Kohm for the grid bias and 3Kohm for the filament voltage.

Because the R1155 has an isolated HT requirement (+250V including -30 volts bias) as well as needing a little extra voltage (not necessary for the R216) I used a 2-pole switch and a relay, the relay is activated by the switch when R1155 operation is required and disconnects the 6.3V heater winding from HT negative (that connection is required for the R216) with the second pole of the switch adding a small reservoir capacitor to the swinging choke circuit used for the R216. I used the swinging choke circuit to keep the HT voltage down as the transformer secondary winding was a little high for the task, probably capable of a maximum, off load, of around 370 volts with a typical reservoir capacitor. The position of the switch in the circuit is important as there is a risk of damaging the HT regulator circuits if their HT return is removed, so I placed the switch in the 6.3 volt grounding connection (Black) as shown above. The R1155 heater circuit wiring connects one side of the 6.3 volt supply to the R1155 chassis which is some tens of volts higher than HT negative so disconnecting the PSU heater ground connection in this power unit eliminates the risk of the R1155 bias circuits shorting out (note: when a T1154 is used in conjunction with the R1155 the heater supply has to be DC rather than AC because that same circuit operates the T1154 send-receive relay which has a 6 volt DC coil). Unless I add this 2-pole switch the cabling carrying power to the R216 might need to be disconnected from the PSU terminals before plugging in an R1155 because R216 Plug Pin A carries HT ground and a grounded heater connection as well as the filament return and the dial lamp supply. The relay (which may not have been strictly necessary but gives some flexibility for later modifications) is driven from the 30 volt supply used for the -19V R216 grid bias. I added a 220 ohm resistor to reduce the coil voltage to exactly 24 volts.
 
 The power supply is built on a piece of chipboard and the front metal panel is isolated, so circuit ground hasn't any real meaning, other than the interconnection circuitry between the various output supply voltages. If the PSU used a metal chassis and this was connected to HT negative it would have a potential difference of some tens of volts compared with the R1155. A significant advantage of using wood is that parts can be moved around easily as design changes are made. Some more work is required to fit the R1155/R216 switch and the addition of the voltmeter wiring for monitoring the bias supply. Once everything is working I can tidy up the rats nest of wiring.

 Here's a picture of the R216 plug wired up. I found a drum of cable with 4-cores plus screen, so two lengths of cable will do the job as the R216 needs 10 connections with two being duplicates of earth return for HT and LT. I used the screens of the two cables connected together giving me enough wires for carrying the various power connections. The ends of the cable were soldered to tails secured in the choc bloc.

Once completed I decided to test the receiver.....

Note that the plugs and sockets use a fine thread not the very coarse type used in later equipments.

 

 After checking over as much as I could I plugged in the receiver to the new power supply. There was a slight crackling sound and smoke rose from somewhere inside one of the two chassis. Clearly something was amiss but I couldn't work out what component was burning so I disconnected the various tails and measured the current for each. The 1.4 volt supply was drawing 360mA and as one valve is missing (it's only the IF output buffer so not critical) this current is correct. Next I connected back the 6.3 volt supply and this stayed at the same level so I judged it to be OK. Next the bias supply which measured 53uA and the voltage remained constant. The 19 volt AC supply drew 0.9A which is too high but it didn't change when the illumination control was turned. I suspect the phase is wrong. Looking at the proper power supply circuit, the 19 volt supply is derived from the same winding as the 6.3 volt supply, but the dial lamps are supplied by the voltage between the 19 volt and 6.3 volt windings (ie. 12.7 volts). I should therefore modify my PSU which presumably has the opposite phases for the two voltages resulting in something odd happening. As there's a spare 12 volt winding on the bias transformer, I'll use this.

Next, I checked the HT currents. 100 volts drew 12mA and 250 volts drew 9mA which might be OK as these measurement were taken with the valves cold. On the other hand, only the regulator should draw HT current, and that from the 250 volt rail so what explains the 100 volt drain? I can see there's two resistors across the 100 volt rail at V11. These add up to 55K so should draw 1.8mA leaving 10.2mA unaccounted for. The RF chassis doesn't use the 100 volt rail so the drain must be in the IF chassis.

As all the measurements, bar that small 100 volt drain, appeared to be fine I plugged in the receiver once more... no smoke and the combined HT current measured around half scale on the milliammeter. As the total HT current is 69mA I reckoned this wasn't too critical. Plugging in a pair of headphones gave me a loud hiss on the FM settings, dropping to about half the volume on the narrower bandwidth setting. AM and CW gave me a loud hum which dropped off a lot when I reduced the AF gain. Connecting a length of wire to the aerial socket did crackle slightly but absolutely no signals were present even from the crystal calibrator. Apparently the IF module is working but not the front end.... Maybe the problem is associated with the smoke? A short-circuit condenser may have open-circuited a resistor? That would explain why the smoke didn't reappear a second time. What would smell hot... well a drain through a 10K resistor from 250 volts would be 25mA and dissipate 6 watts and if the resistor turned into a charred mass of say 50K this would dissipate around a watt or so and get pretty hot. The 5mA HT current wouldn't really make much difference in the sets consumption.

 

A second test... this time I fitted a spare 1T4 into the unoccupied DF92 V12 socket in case I decided to monitor the IF output. Turning on the power supply gave me a quarter scale current reading which slowly rose to half scale (something like 35mA). I guess this is because the battery valves have a very short warm-up time and the mains valves need up to 20 seconds or so. After a couple of minutes there was a hot smell but I couldn't easily figure out from where it was coming. I also turned on my spectrum analyser to see if any RF signals were present but drew a blank. This implies the RF chassis has a problem, perhaps a short-circuit condenser decoupling the HT line? If this had burned out it's associated resistor (the smoke when first turning on the receiver) it would all make sense. R15 and R21 are prime candidates especially the latter. A simple check is of course to simply remove the can from V4 and see if the voltage stabilizer is lit (but it's in an awkward position.. and I noticed it lighting up when power is applied to the receiver).

Another test could be to remove all the mains valves and see what current if any is drawn from the 250 volt supply. Unless I've missed something, only the QS150/15 would draw any HT current... I unplugged the interconnecting cable from the RF chassis and disconnected the filament supply. I measured 9mA from the 250 volt supply which disappeared when V6 was unplugged which seems fine, but I noticed that the HT rail wasn't too high, being only 230 volts dropping to 205 volts with the RF chassis plugged in, when the HT draw is 33mA (which equates to an additional 24mA from the RF chassis). Clearly the power supply isn't producing enough voltage so I'll have to add extra reservoir capacity, or remove the 250 volt stabilizer because it's not necessary. I'll do this once I've fixed the fault.

I had noticed something slightly odd when I carried out the initial checks. Rotating the RF gain control dramatically increased the HT current. This is a combination of the 100volt and 250 volt rails. Some of the change is current drawn from the 100 volt rail because two IF valves are controlled by the pot although only V6 uses the 250 volt supply. Of course I did hear a lot of audio noise and this would drive the audio stages to draw more current.

I then did another test with the filament supply disconnected. This would of course prevent the battery valves from drawing current. In fact I measured the 100 volt draw and it measured 1.8mA (this is the current drawn by the resistors at V11 as calculated above) and remained at this no matter what setting was the RF gain control which is correct. With the RF gain control set at full the 250 volt draw was 33mA, but if I set the RF gain at zero the 250 volt HT current dropped to 10mA. V6 drawing 9mA at full gain will drop to maybe 2mA at minimum gain. But that leaves 8mA from the RF valves rising to 24mA at max gain. This is because the RF valves are fed from a bias supply which is combination of the RF gain pot setting plus a potential divider formed by R43, R44 and R46 so their HT current will change with the setting of the RF gain control. All this seems about right.

I suppose I can unplug the RF unit and test the IF chassis by injecting varieties of 4.86MHz signals. To make things easier I should first find a suitable 2-pin headphone plug and a suitable RF plug for the IF output. This I did, fitting two old fashioned wander plugs to a standard quarter inch jack socket and plugging in a loudspeaker. I turned on my signal generator and tried listening for signals. I did notice that using FM I could hear a bleep when the generator was tuned to 4.86MHz, the IF. Cranking up the FM deviation did however produce a good response using an input of 50mV to the aerial plug. Switching the receiver to AM I could hear the FM signal so both AM and FM are working at least at the IF of 4.86MHz. Nothing could be heard when using the various wavebands with the generator tuned to corresponding frequencies such as 28MHz or !00MHz so the local oscillator or possibly the mixer is not working. The RF gain smoothly changed the audio volume so my guess is the RF amplifier may be working. The stabilizer is lit but there's still that burnt smell coming that seems to be coming from the RF chassis.

Swapping V2 and V3, type EF91 should prove the oscillator valve, unless both are too tired... , but no that didn't work and it was a real problem replacing the oscillator valve. It's held in place by a metal clamp and difficult to see the holes in the socket for lining up the valve pins. Of course the mixer might be u/s as it's a CV291 a special, extremely sensitive device (see a sheet carrying data on this). Also here's some more data including a picture. I have a few spare ones in the workshop if mine has been damaged. I prodded the signal generator on the thin end of the diode and the IF amplifier responded but nothing heard when prodding the fat end. Duff mixer diode or then again that burning smell might be relevant?

 

 

 Frequency of operation

 6000MHz

 Min back to forward resistance ratio

 8:1

 Min forward resistance

 265 ohms
 Max amount by which performance is below "standard best"

 3.0db

 Max noise temperature ratio

 2.6

 Nominal IF impedance

 350 ohms


I measured some resistors and voltages. Two 15K each measured 18K so not important. V2 anode went from 158 to 79 volts, drawing 7.6mA through 18K to 216 volts (15K reading high) and V1 anode was 116 volts fed from 230 volts and 18K (15K reading high) drawing 6.3mA at max gain. V3 (triode-connected) anode was 88 volts fed from 141 volts so is drawing 5.8mA through 9.2K (the 8.2K was high). All these measurements seem to me to be OK. I wondered about the three coax cables which mix the calibrator output to the aerial but placing a signal on the output cable proved they were probably blameless as there was no improvement. I probed the oscillator with an oscilloscope probe but no RF could be seen.

Just in case the oscillator stops working when prodded I'll tune my general coverage Japanese receiver to listen for the R216 oscillator in case there's something odd about it. Absolutely nothing heard. Next I fitted a loop to the end of a lead plugged into my signal generator which I tuned to 4.86MHz with 200KHz FM and heard a really loud signal. I then tuned the generator to 105.16MHz then to 95.44MHz connected an aerial and tried to tune in Classic FM on 100.3MHz but failed. Switching on the R216 calibrator to 5MHz I listened on my Japanese receiver to 95MHz and found that I could hear the 19th harmonic of the crystal tuning nicely when the R216 passed through 95MHz. This proves the RF amplifier is at least partially operating so I'm left with the oscillator and the crystal mixer. Could the latter be faulty and as a consequence killing the oscillator? Looking at the circuit diagram I can't see that this is likely, so could there be a failed oscillator component? Something did smoke when I first turned on the set so was this one of the critical parts? For example there's a path to ground from C36 connected to V3 anode.


 
   I found the reason for the local oscillator not oscillating. I'd noticed the lamp for a particular scale didn't always come on so I wondered if the turret switch was bad at the oscillator contacts. There are three contacts for the oscillator K,L and M and by using a strong light I could see the last contact M, wasn't touching the turret coil contact because the spring strip was straighter than its neighbours. By carefully levering the metal strip I was able to make it look the same as the others, but testing proved the oscillator still wasn't working. The three contacts are: a cathode connection at K where you see 1.2Kohms to ground. That was OK. Next a small condenser connected to the anode, L, and thirdly a ground connection M. It's clear then that the centre of the three contacts should be grounded when turret contacts are good. In fact this wasn't happening and wiggling the centre strip showed it was fractionally clear of the turret contact but grounded when it was bent slightly. I then applied some pressure to the centre strip (with the turret disengaged) until it looked level with the other two strips. Whilst doing this I noticed a barely perceptible movement at the material where the three strips were soldered. As the oscillator was now working I didn't follow this up until I turned the chassis through 90 degrees so I could read the dial. The oscillator stopped and I found that the RF chassis had a tiny amount of play between one side and the other. By pressing the front towards the back I noticed the two screws A and B were nowhere near tight (in fact you can see about 1mm gap). Someone had perhaps been investigating the oscillator problem and slackened the screws, failed to notice the contact problem, then just left them. I applied pressure to line up the two sections of the chassis, tightened the two screws and now all is well.
 Now that the local oscillator is running I can tackle alignment. Tuning the dial to 100.1 and with wide FM selected I could hear Classic FM which is supposed to be 100.3MHz. The sound is badly distorted and using narrow FM, AM or CW the overall gain is too low to hear the signal. As I still haven't discovered the source of the smoke when I first turned on the receiver, no doubt that will account for at least one problem.
 

 

 One reason for lack of audio output and distortion was very poor IF adjustment. The IF is 4.86MHz and the coils are predominantly anode loads with condenser coupling to the following stage.

The tuning of each coil is via a dust core at the top.

Very carefully I managed to roughly tune each IF stage to the correct frequency and receiver sensitivity increased dramatically, however a couple of cores were damaged because of the force needed to adjust them.

 

 
 On the left are two examples of IF transformers and above is the BFO coil. Each core has been secured by masses of wax dripped into the cans so I decided to pull off each can (just two 8BA screws in the top) and remove the wax.

 

Each core has a slot at either end so I removed each one and turned it over. To remove the cores I first removed each aluminium can (2 screws and just pull the can vertically upwards) scraped away as much wax as possible, then applied heat from a soldering iron carefully pressed onto the top of the core until slight smoke rose from the heated wax, then unscrewed the core using finger pressure. I then removed most of the white locking compound from the core. The threads are the same as those square dexion nuts and I checked each core freely accepted a nut and each coil former freely accepted a dexion bolt before re-inserting the core with the better slot uppermost. It's a long job as there are ten coils needing attention, however all went well and all cores are now adjustable.

The next stage is to align the IF strip to 4.86MHz and centre the BFO. I don't have any tuning instructions so I'll assume that IF alignment is carried out in the narrower bandwidth setting in AM mode (TR1, L31, L32, L33, L34, TR2, TR3) then switching to FM and continuing with the limiter stage L37 and the Foster-Seeley FM discriminator tuning coil L39.

Below are the results of IF alignment with a signal injected at V8 Pin 6 with V8 removed with bandwidth set to 30KHz then 120KHz. The vertical scale is set to 7dB steps and the horizontal scale 50KHz steps.

 

 

Just after I'd checked the IF response there was a sudden increase in HT current and some smoke from the HT regulator in my PSU. I traced the latter to cable insulation touching a hot component, but the surge in HT current I suspect was a recurrence of an earlier fault I've yet to isolate. Reducing the HT voltage resulted in the excess current not being drawn although there seemed to be instability within the front end of the receiver, perhaps a leaky condenser?

 

 

 Most of the decoupling condensers are enclosed in metal tubes soldered to the chassis with their earth connection soldered at the exposed end so it can easily be unsoldered. The plan to identify the culprit is to supply an HT voltage and check for something running warm.

You can see some examples on the left. There's a power supply connector between the two chassis which can be unplugged to make the job easier.

 

Front-end alignment looks tricky. I know it needs sorting out because dial accuracy was a few hundred KHz out in the FM broadcast band and signal strength for Classic FM improving greatly when the main tuning condenser trimmers were adjusted. Each waveband is adjusted within the turret and the various trimmers look awkward to access.

 A little more investigation continued. I modified the HT voltage regulators in the home brew PSU to reduce zener current because once I'd added an extra reservoir condenser the HT had risen and the zeners were running quite hot. Connecting the receiver showed it was consuming 60mA once the mains valves had warmed up. Overall gain was now resulting in a hiss from the loudspeaker, but with some roughness and hum. I turned to the FM broadcast band and I was able to hear the crystal calibrator set to 5MHz at 100MHz or so on the dial, but Classic FM was weak. After a little while I was aware of smoke and this was coming from a tiny 10 ohm wirewound resistor which was running white hot. I'd noticed the lamp for this band wasn't coming on, although the others were OK when their bands were selected. Presumably the lamp circuit is short-circuit.
 

 

 The R216 low voltage wiring is sightly unusual.

The dial lamps are switched by contacts linked to the turret and are arranged to let the operator know to which band he's listening. The lamps LP1-5 are festoon types rated at 12 volts 2W (about 167mA) and their brightness can be varied with RV3. The various resistors are designed to limit the lamp voltage, perhaps to improve their longevity?

The lamp supply is connected between the 6.3 volt (Pin C) and the dial lamp supply of 18 volts (Pin J). Pins A and H are ground connections.

There are a several wires in the power cable feeding the LT circuits so that resistive losses are kept low.

You'll note the presence of RF chokes in some of the filament and heater wiring, used in conjunction with decoupling condensers, to reduce interaction between stages and RF leakage.

 There seems to be three faults remaining (plus the odd one or two I haven't yet discovered)... firstly the short-circuit dial lamp, secondly the low front-end gain. I'd noticed that putting a long wire on the aerial pin didn't produce a crackle and thirdly there's still that sudden surge in HT current I haven't tracked down. One possibility is a decoupling condenser has failed in the RF stages so that RF gain is reduced? For example those decoupling R2 and R10 resulting in a low screen grid voltage. Each time I get set to tackle one fault, another seems to pop up. Again I tested the set, but this time I removed the valves from the IF amplifier chassis and monitored the HT. The HT registered a stable level at less than 10mA but adding valves suddenly caused the current to shoot up and the voltage to drop. Next, I disconnected the 100 volt and 250 volt supplies and fed these from my two Solartron HT PSUs. With the voltages set at exactly 100 and 250 volts both currents remained stable at 25mA for the former and 35mA for the latter. I was able to check the FM broadcast band without any problems arising... so the fault must lie in the home brew PSU. Time to do some calculations on dissipation.

In fact, when I carried out load testing on my power supply I discovered the problem was breakdown of the series pass transistors. These were plastic versions of the BUT11 (the BUT11AF) designed for pulsed operation in switch mode power supplies. I substituted two BUT11A mounted on a pair of heatsinks. During testing I changed the zener diode resistors as I'd calculated their values without allowing enough base current for the transistors. Further testing showed the 100 volt regulator transistor was running fairly warm so I added a 3.3Kohm ballast resistor in its collector feed. This moves the dissipation from the transistor to the resistor (for example... running 25mA at 100 volts from an HT rail of 300 volts results in wastage of 200 volts which equates to 5 watts of heat and 10 watts at 50mA. The ballast resistor absorbs 82 volts at 25mA and 164 volts at 50mA soaking up about 2 watts and 4 watts respectively).

A genuine R216 power supply, complete with a proper cable, is due to arrive here in a couple of days (see below).

When checking the receiver further it seems that most if not on all the other wavebands the local oscillator is not working. Maybe it's misalignment of the turret tuner with respect to the springs located on the RF amplifier chassis? If the turret and the chassis are not precisely lined up then the studs will not be in contact with the springs at one end or the other. I think this is a design weakness because if things can go wrong they usually will go wrong. Well, after scrutinising the mechanical assembly I decided the turret could be removed by slackening and/or removing screws in the end plate. This I did and all was revealed... for starters I'm not the first to detach the turret because I spotted black pen lines where something needed to be marked before removal. It was clear that the springs which mate with the turret have caused problems before. I also noted that several 6BA screws were missing and something that may explain the smoke and the overheating resistor. See below.
 
 

This green resistor is fitted at the rear of the turret and is provided for dropping the HT feed to the RF amplifier valves. The smaller 470K resistor is associated with V4, the voltage regulator valve.

This small tag panel was broken away from its mountings and this could quite easily short to the chassis.

I think I can replace it.
 
 

 The repaired tagstrip. I had difficulty getting solder to adhere to the tags because the thing was ancient and oxidised but it serves its purpose and will stop any short circuits that I attribute to the damaged one. I can't figure out how the original tagstrip had been damaged. It's possible the turret adjusting lever, which is visible above, had been wedged behind it.

The lever can be moved to accurately position the rotational position of the turret so the coil locating pips are precisely lined up.

Below is a view showing the pips that engage with the turret visible once the turret has been detached. The set of three on the left are the contacts for the oscillator coil. One or more of these fails to make contact on all but one waveband.

 
 
   Here you can see that the three springs have lost their set. They're supposed to be bulging slightly outwards so that pressure is maintained on the turret pips. To fix the problem I think I'll need to unsolder the right hand ends, bend the spring slightly then resolder the springs. I suspect these have been worked on previously because the end spring is a different colour and the pip looks very slightly different to the others?
   The task was a lttle easier than I imagined. Before unsoldering anything I placed a dense plastic pad under each spring to give it less flexibility but then I discovered that there's enough play in the hole marked 628 to lift the spring to the same level as others. I did this to the three oscillator spprings plus one other so that they all now lie in the same plane.

 What about the faulty dial lamp? I fitted a new 12 volt lamp in place of the duff one (open filament) and not seen a recurrence of the overheating of R73. Maybe there's something other than a short-circuit to explain this (but see the broken tag strip above). It does remind me of bias shorts in the R1155... Could the burning resistor perhaps be due to a bad grounding connection in the home brew PSU?

Whilst the turret is out I checked the various resistors which would normally be inaccessible. All read high in value so it will be prudent to replace them.

 Without a detailed description of removing and refitting the turret I had to work this out by trial and error. Removing the endplate securing screws allowed the turret to just pop out. That part was easy. Refitting involved struggling with spring tension so I unscrewed the nut on the spring end. This allowed the turret locating lever to drop away. The turret could then be placed in postion. I turned the wavechange knob half way between Range 4 and 5, then fitted the turret so that it lined up midway between Ranges 4 and 5. I then turned the knob to check the setting lined up when the coil pips were located correctly. Then I screwed back the endplate ensuring the front and rear bearings were in place. The turret then turned relatively freely. Next the locating spring needed to be refitted. I screwed the nut back so it was located against the bar (as above) and checked the lever was located correctly in the slot at the end of the turret. Next the spring needed to be stretched and its hooked end positioned over the peg within the metal slot. To do this I threaded some string through the slot and around the hook and then tied the ends to form a loop. By gently pulling the looped end the spring stretched and located on the peg. At this exact point the string snapped but the hook dropped onto the peg. The final task is to confirm the correct coilset is engaged with respect to the knob setting then adjust the lever at the end of the bar adjacent to the resistor panel (which you can see by scrolling up to the picture of the damaged tagstrip).

You'll note that the yaxley switch seen above the spring is correctly selecting the lamp circuit.

Also note more missing screws... this time at the panel covering the filmstrip.

 Having got EMER 384 which carries lots of alignment instructions, I can hopefully now more easily track down the reason for receiver deafness. After having carried out more testing, which confirmed the turret is working correctly, I can use the test data to confirm or otherwise the performance of the IF amplifier. A test signal of AM at 30% depth should give me 5mW at the 150 ohm output sockets. What sort of level should be input to the IF strip? The data doesn't tell me this in so many words, but it does specify that a 10uV 4.86MHz CW signal applied to the IF side of the mixer diode should give me a decent tone using the BFO. If this works out then the deafness must be attributable to the RF front end (including the mixer diode), otherwise the problem is in the IF strip. I've already noticed that all the IF coils tune nicely which is a good sign. Currently I can hear local FM broadcasts but no other broadcasts on any waveband. My signal generator can be heard but that needs to be at best 50uV and sometimes more than 10mW across the different bands.

Part of the receive problem is fixed. The IF strip is prone to instability when the maximum gain setting is used and retuning the stages upset the centre frequency. I found this out when I injected 4.86MHz directly into the IF input socket on the chassis instead of forcing it through the front end. I set the spectrum analyser to 4.86MHz and proved the calibration by adding a signal from my signal generator and found the IF amplifier was several KHz away from 4.86MHz. It was quite awkward to shift it back and I soon discovered that I needed to back off the RF gain control to get the amplifier to centre on 4.86MHz. Once the shape of the response curve was correct and dead centre on the correct frequency I gradually turned up the RF gain. Instead of the shape of the resonse just getting a higher peak as the gain increased, it flat-topped and got wider and wider as the control was turned up. Backing off the input signal and increasing the gain showed that the amplifier began to get unstable. This seemed to be due to interaction between input and output leads used for testing. Finally, after checking the amplifier was properly centred on 4.86MHz I disconnected the monitor lead from the grid of V11 and re-plugged the lead from the mixer. I was then able to complete the IF tuning by adjusting V11's input at TR3 and V6's input transformer TR1.

That flat-topping I'd noticed at V11 grid.... I wonder if it's the so-called "noise limiter" working its magic?

The FM adjustment method in the EMER sounded pedantic so I adopted a much easier method. First L37 is tuned for maximum noise then T39 tuned to almost minimum noise, then L37 retuned slightly for max noise and finally T39 tuned for minimum noise. This lets me listen to broadcast stations without distortion if wide bandwidth is selected and careful tuning employed.

Next, I'll need to align the turret tuner RF circuits. As there are 4 trimmable coils for each waveband the difference between best and nominal adjustment will be very significant, but before tackling this I decided to check the mixer diode. Testing these types of microwave diode is not easy because a normal test meter would probably destroy it so I looked in my spare parts box and spotted a couple of Schottky diodes marked "HP5082-2835". I bought these diodes from a local Tandy shop that was closing down, probably about 1985 and checking their datasheet they look ideal for a new R216 mixer. The new diode was easy to fit by just soldering across the terminals at the original diode socket (so if ever a new CV291 becomes available it can easily be fitted). I turned on the R216 and immediately heard the familar racket of a shortwave band. That was because of the last test I'd done... I'd switched to the lowest range with the BFO on and a long wire connected and heard precisely nothing... apart from broadcast FM stations on range 4 the receiver had refused to ackowledge anything on any band (apart from millivolts from my signal generator). So there's the answer to receiver deafness. Mixing was presumably taking place at the first IF amplifier but only if sufficiently strong signals were present... the duff diode was acting purely as a tiny capacitor and coupling the RF and local oscillator signals to the grid of the EF91 where slight non-linearity within the valve produced mixer products. Essentially the IF amplifier was a valve down and the substitute mixer only working on strong signals.

I attempted to measure the old diode now it's clearly u/s and I could see absolutely no trace of anything except possibly a minute capacitance. It seems the capacitance of the diode socket and local wiring must have been responsible for coupling the RF from the front end to the IF amplifier. No wonder the receiver was deaf... the previous owner probably tested the mixer diode with a multimeter.

 
 

 Left, the defunct CV291, early 1950s microwave diode, and on the right a new HP5082-2835 from my junkbox. This can be used as a mixer diode said to be suitable for use up to UHF so should perform adequately.

The anode goes to the turret contact (left) and the diode to the IF input transformer (right).

 
 The R216 power supply designed for the receiver (shown below) uses a much simpler means of supplying the correct voltages because the mains transformer has been wound to supply exactly the right voltages and power.

 

I've printed two circuit diagrams here. The first came from the "Illustrated Parts List" and strangely is wrong. No doubt the technical author misread a document and no-one bothered to check his work. I've indicated the correct markings in red... This now lines up with the condenser numbering ie. C2/C6 are 350V working and C1/C5 are 150V working. 

 

 Now the correct circuit diagram. I wonder how much confusion resulted from the errors?

 

 I opened up the new power supply but closed it again because of the very strong smell coming from inside the case. Both the R216 and its PSU are sealed and it took a couple of weeks to disperse the smell of government surplus equipment from our conservatory where I first opened the receiver case. I opened the PSU in the garden and here are pictures of its interior...

The first view shows the mains transformer and two HT chokes ZA44291, plus the LT filament choke ZA44292, on the right. The two HT (250v + 100v) rectifier valves are type CV493 or 6X4. This has two anodes which are wired together as a single-phase rectifier in the circuit shown above.

 

Here's something very odd... in the view above in the lower left corner you'll see the mains input socket. If you look closely you can see the connector is mounted on a metal plate and this is stuck to the front panel with araldite. I discovered this when fitting a mains cable. On my lead the 3-pin plug orientation pips were turned one position away from those on the socket and as the mains lead came with a mating socket I decided to swap that for the one fitted to the PSU. Just unsolder three wires, unscrew the securing ring and fit the new one. I noticed that the straight edge designed to mate with the flat on the connector was poor and allowed the connector to turn so I tightened the locking ring a little more and the result was that the new connector dropped out carrying the plate with it. Up till then I hadn't noticed the plate was just glued to the front panel... I weighed up the options and decided to fit an IEC connector, then decided against this, found a tube of superglue and used the contents to re-secure the plate. Whoever fitted the plate took care to touch up the front panel so that the repair wasn't obvious. It seems the repair was not done by the Army, but by Andy, G8JAC who in 2007 removed an amateur-fitted plastic replacement for the original socket.

An alternative solution could have been to drill shallow holes in the rear of the front panel and self-tap the aluminium plate in position. I rejected that idea as it risked penetrating the front so I used superglue. It's currently setting and if its secure it'll have to do....

There are several electrolytics. Identifying these led me to discover the errors in the circuit diagram shown first. The largest is C6, a 32uF x 350V for 250V smoothing, C2 is an 8uF x 350V for the 250V reservoir, C5 is 16uF 150V for 100V smoothing and C1 16uF 150V for 100V reservoir.

C4, 2uF 150V; C8, 2uF 150V; and C9, 2uF 150V are for smoothing the bias supply and C3 and C7 both 1000uF 6V are used for smoothing the filament supply.

 

 The large black rectifier is used for the filament supply (also in picture below).

 

 The black rectifier is a Sentercel full-wave selenium device used for the filament supply.

 

 Above you can see a set of high stability resistors for the metering circuits together with a tiny mains rectifier MR4 for monitoring the PSU input voltage.

 

 On the right you can see a pair of Yaxley switches incorporating an interlock so that mains voltage can't be changed when the receiver is turned on.

Also, another view of G8JAC's Araldited plate, top right. Once the plate had fallen away you could see there was just a large hole with no metal left for mounting the socket. The superglue repair worked fine and the mains cable fits.

 

 On the left is the filament rectifier MR1/MR2 and top centre (end on) the bias supply rectifier MR3.

Most people advise reforming capacitors before running them at full voltage, but I know that the PSU worked OK in 2007 so I didn't expect a problem. As this PSU uses valve rectifers it's fine to connect an external variable voltage HT power supply to the HT line and very gradually crank up the voltage with an eye on the test HT current which should be held back to say no more than 10mA or so (bearing in mind there are bleeder resistors in place). Once you've reached say 30 to 50% of either 100 or 250 volts in this example and the feed current has dropped away nicely the capacitors should be serviceable the full voltages can then (less) slowly be applied, whilst still montoring the current. The 100 volt and 250 volt rails have 22K and 620K resistors consuming about 4.5mA and 0.5mA respectively. If you have a bad capacitor one of two things will be noticed. First the feed current will be next to nothing which implies an open circuit or high resistance capacitor, or the feed current may be erratic and prone to increase rather than decay.

Large capacitors in old military equipments can be paper types rather than electrolytics and these are usually constructed in exactly the same way as those dreaded wax tubular types. In my experience these large paper types can either be perfect or leaky and if really bad will run hot, then very hot then expire.

With 250 volts applied to the PSU the meter read 7.5 and with 100 volts applied the meter read about the same. Next, I'll check the low voltages are present and at the correct pins. Powering the PSU proved it was working. Off load of course, the voltages were miles away from the levels specified for the R216. Being used to modern circuitry where most voltages are regulated to a couple of percent it's a bit alien to see a 4.13 volt supply destined for 1.4 volt valves, however when the leads were all connected the voltages settled down to their correct levels. I noticed the 1.4 volt output was a little low but gradually asserted itself to 1.3 volts... no doubt the low voltage electrolytics needed some reforming and will slowly improve.

I plugged in my VHF aerial and turned on the set with the proper PSU in place and was pleased to hear the FM broadcast band full of stations. The highest band is now receiving a fair number of aircraft transmissions but I did notice that the local oscillator seems to cut out before the lower end of the highest band is reached. I'm guessing this is possibly due to a low emission EF91 because I had a similar problem wuith a Racal RA14 oscillator. Also noted is an annoying effect of the audio on AM reducing in amplitude as a signal is tuned... quite possibly the action of the noise limiter which is no doubt engaging to prevent too much headphone output.
 
 

 Return to Reception