R308 VHF Receiver

Above, anxious to discover whether I'd wasted my £32, I've pulled the chassis partly from the case. It had only a single securing screw.

Much to my surprise the interior of the receiver is in excellent condition. I'd been concerned that, as I hadn't seen anything other than external crusty views before I'd acquired it, the set may have been a real problem to restore. As per usual the (12 volt) vibrator unit is missing from the power supply but that isn't a worry as I'll use it on a mains supply. The valves as you can see from that red one look brand new as does that screening can.... rather than being over 80 years old.

Those yellow and black cables carry power and the odd signal so that either the RF or IF/AF module can be detached from the chassis for any maintenance requiring them to be powered up. You can see the reason (inaccessibility) if you look at the underside views (particularly the RF unit).

Another view, showing the RF front end chassis. This uses three VR136 valves (= EF54) plus a pair of AW2 voltage stabilizers similar to the type 7475 fitted in the R206 power supply. In this view you can see that the front panel has received significant stress in its life which has distorted the top edge of the metalwork. A few years back I devised a simple way of fixing this sort of thing ** and if I can remember what it was I'll tackle that as soon as possible.. There's a crack in the plastic dial which is probably associated with the panel damage but as this has a flat profile and unlettered a new piece can be simply made.

** I placed a large furniture cramp (edge on) along the front, resting this on the bushes for the front lid (not supplied) securing screws then tightened it to remain in position. I then placed a small machine vise at the bend and tightened this against the cramp until the panel was straight.

When I removed that securing screw so I could sort out the bent panel I found a suggestion of "bike-shed" design. As you can see, the screw is fitted with a bush about an inch and a quarter long. This is so the screw can't be tightened too much resulting in a bent panel. The bush presses on a tapped bracket on the inside of the case and to prevent losing the bush it's secured by a small key punched through a hole in the screw. This rather extravagant piece of engineering may have been a penalty for using the R107 case?

Bike-shed design.

HDRS also had a similar design feature in its sledgehammer.

This addition of an S-Meter by the first radio ham owner (a really good idea) is basically the only modification that's apparent but strangely all its wiring has been removed as if the last owner had decided to return it to its original condition but stopped perhaps awaiting a replacement panel etc. The original area was used for headphone storage and had a metal compartment behind a hinged lid (note the space behind the front panel.. was this space left deliberately for what might have been another "bike-shed" idea by a committee?

SRDE would have had MoD representatives (or their 1942 equivalent) present at handover meetings to KB. SRDE would have concentrated on really complicated technical aspects but MoD would have commented on basically irrelevant points and being chair of any meetings would have minuted stuff like the design of securing screws and making space for the storage of headphones. In the event the screws were great but the receiver was a bit iffy.

Why was that octal socket fitted on that added panel? Maybe the receiver was pressed into service as an IF for an external home-brew front end because, having FM, it would make a decent 2m or 70cm receiver? At least in the 1950s to the 1960s.

A puzzle is the countersunk hole drilled in the identification plate at the top right of front panel.. maybe this was used to mount a single-screw-mount BNC connector?

I thought a perusal of old Practical Wireless magazines might provide a clue to suggested modifications to the R308? But no.. I didn't really find anything other than a few "for sale" adverts by dealers.

It's also not prolific on the Net, where I can see only a few examples... Serial Numbers "7", "68" and "185". Mine is "158" so all are in the in the KB tranch of 200 (see below).

The follow-up to the R308, the R216 (19-157MHz)

Now, a little on the background to the R308 receiver.

It was designed as a range extension companion for the much more common R107 and uses many identical parts, to no-doubt save money but more importantly to save development time, for example the outer case plus various miscellaneous components, such as knobs, switches, headphone and power sockets etc.

The way it was born is interesting. Design was almost completed by SRDE in 1942 when a contract was made with Kolster Brandes for its rapid development leading to the manufacture of 200 examples. The end result wasn't up to much so my guess is that first batch of 200 was the total quantity built before a replacement (to be called the R508) was decided on. The R508 may have been KB's baby but the replacement ended up as the "R216" manufactured by Ekco. More about this set can be seen by clicking the picture on the left.

Although KB had been making radios since the 1920s, SRDE must have decided they weren't sufficiently well versed in VHF equipment so switched to Ekco. One important aspect in the replacement was the use of a turret tuner rather than an outdated yaxley switch for range selection. Turret tuners were used in the WW2 receivers such as the R206 and the DST100 and offered much better repeatability of performance and repair & maintenance of the tuned circuits. For example, one of the bad features of the R308 (and the R1132) was in the operation of the VHF local oscillator when it would cease to oscillate due to absorption of the signal due to resonance in neighbouring coils and wiring.

The key tuning parameters for the R308 are as follows..

Ist IF = 9.72MHz; 2nd IF = 2.1MHz

Range 1 = 20 to 30MHz, Range 2 = 30 to 48MHz

Range 3 = 48 to 75MHz, Range 4 = 75 to 112MHz

Range 5 = 112 to 145MHz

I'm expecting to find a few shortcomings with this old receiver and, if I do, should I try and improve the design or just accept it for what it is?

The front panel has been poorly re-painted (fortunately the legends are engraved) so if I can remove all the control knobs (a common "major" problem) I can repaint it.

Below is a provisional circuit diagram stitched together from six pages. Click to see full size.

Components table

C1
A
B
25uF
C12
  A
-
2nF
C23
A
C
50pF
R1
A
C
7R
R12
A
-
700R
R23
A
-
22K
R34
  A
C
100K

C2
  A
E
8uF
C13
  A
-
2nF
C24
  A
D
50pF
R2
  A
-
22R
R13
  A
-
820R
R24
  A
-
27K
R35
  A
C
220K

C3
  A
B
2uF
C14
  A
B
500pF
C25
  A
C
5-47pF
R3
  A
-
22R
R14
  A
B
1K
R25
  A
-
27K
R36
  A
-
220K

C4
  A
-
2uF
C15
  A
D
400pF
C26
  A
E
3-30pF
R4
  A
-
42R
R15
  A
-
2.5K
R26
  A
-
27K
R37
  A
B
330K

C5
  A
F
500nF
C16 A
C
300pF
C27
  A
B
12.5pF
R5
  A
-
120R
R16
  A
-
3K
R27
  -
-
R38
  A
B
470K

C6
  A
E
100nF
C17
  A
-
280pF
C28
  A
B
10pF
R6
  A
-
150R
R17
  A
-
3K
R28
  A
C
27K
R39
A
-
470K

C7
  A
-
50nF
C18
  A
-
230pF
C29
  A
-
3-10pF
R7
  A
B
220R
R18
  A
D
4.7K
R29
  A
B
47K
VR1
A
B
5K

C8
  A
AL
10nF
C19
  A
-
200pF
C30
  A
-
6-8pF
R8
  A
-
220R
R19
  A
-
4.7K
R30
  A
C
68K
VR2
A
-
2K

C9
  A
-
10nF
C20   A D 200pF C31   A E 2-8pF
R9
  A
-
270R
R20
  A
-
10K
R31
  A
-
68K
VR3
A
-
1M

C10
  A
-
5nF
C21
  A
H
150pF
C32
  A
-
2-4pF
R10
  A
-
470R
R21
  A
-
10K
R32
  A
F
100K
-
-
-

C11
  A
B
4nF
C22
  A
G
100pF
-
  -
-
R11
-
-
-
R22
  A
B
10K
R33
  A
B
100K
-
-
-

Looking at the table above one can see the influence of two designers. The resistor values are linked to their numbers with higher numbers being higher value resistors. Different resistors having the same value and wattage are given codes A upwards. Condensers though are back to front with lower numbers having higher capacitances. Type C8 is exceedingly common being the sets standard decoupler.

To help identify the resistors, hence measuring the voltages, I've made the tables below which also indicate the main associated condenser.

Also shown is an underside view showing the positions of the valves (listing follows).

From a working receiver I'll add the voltages or notes later.

DESIGNATION	TYPE	COMMERCIAL	FUNCTION
V1A	VR136	EF54	RF AMPLIFIER
V1B	VR136	EF54	1ST OSCILLATOR
V1C	VR136	EF54	1ST IF AMPLIFIER, 9.72MHz
V1D	VR136	EF54	2ND IF AMPLIFIER, 2.1MHz
V2A	VR92	EA52	1ST MIXER, 9.72MHz
V3A	AW2	7475	VOLTAGE STABILIZER
V3B	AW2	7475	VOLTAGE STABILIZER
V4A	ARTH2	ECH35	2ND OSCILLATOR + 2ND MIXER
V5A	ARP34	EF39	3RD IF AMPLIFIER, 2.1MHz
V5B	ARP34	EF39	DISCRIMINATOR, 2.1MHz
V6A	ARDD5	EB34	AM+FM DETECTORS
V7A	6Q7G	DH63	AUDIO AMPLIFIER + AVC
V8A	VT52	EL32	AUDIO OUTPUT

R1	A	7R	V1D	G1	C8M
R1	B	7R	V5A	G1	C8R
R1	C	7R	V5B	G1	C21M
R2	A	22R	V5B	A	C24D
R3	A	22R	V8A	A	C10A
R4	A	42R	HEATERS	-	C8AA
R5	A	120R	V1A	K	C14B C17A
R6	A	150R	V1C	K	C6A
R7	A	220R	V4A	K	C8K
R7	B	220R	V1D	K	C8P C6C
R8	A	220R	V1A	A	C8C
R9	A	270R	V5A	K	C8S
R10	A	470R	V4A	AT	C22C
R11	A	-	JACKS	-	-
R12	A	700R	V8A	K	C1B
R13	A	820R	JACKS	-	-
R14	A	1K	V1C	A	C8F
R14	B	1K	V1D	A	C8O
R15	A	2.5K	JACKS	-	-
R16	A	3K	V7A	K	C1A
R17	A	3K	V8A	G2	C3B
R18	A	4.7K	V1A	A	C4A
R18	B	4.7K	V2A	A	C19A C19B C15C
R18	C	4.7K	V4A	A	C8L
R18	D	4.7K	V3A	A	C8U
R19	A	4.7K	V4A	AT	C8M
R20	A	10K	V3A V3B	STAB	C2A
R21	A	10K	V2A	K	C22A
R22	A	10K	V1D	G2	C8N
R22	B	10K	V7A	K	C4A
R23	A	22K	V1B	G2	C12A
R24	A	27K	V6A	A1 A2	C20C C20D
R25	A	27K	V7A	A	C3A
R26	A	27K	V1C	G2	C8E
R27	A	27K	V1A	G2	C14A

R28	A	27K	V4A	G4	C8J
R28	B	27K	V4A	G4	C8J
R28	C	27K	V4A	AT	C22C
R29	A	47K	V1B	G1	C30A
R29	B	47K	V4B	AT	C8X
R30	A	68K	V4A	GT	C8K
R30	B	68K	V4B	GT	C22E
R30	C	68K	V6A	K1	C23C
R31	A	68K	V5B	A	C4D C8Y
R32	A	100K	V1A	G1	C8A
R32	B	100K	V1C	G1	C8A
R32	C	100K	V4A	G1	C8G
R32	D	100K	V1D	G1	C8M
R32	E	100K	V5A	G1	C8R
R32	F	100K	V7A	G1	C22F C22G
R33	A	100K	V5A	G2	C8T
R33	B	100K	V4B	AT	C8X
R34	A	100K	V3A V3B	-	-
R34	B	100K	V4B	G4	C8V
R34	C	100K	V5B	G3	C8W
R35	A	220K	V5B	G1	C27B
R35	B	220K	V6A	K2	C23C
R35	C	220K	V7A	A	C3A
R36	A	220K	V5B	G2	C4D C8Y
R37	A	330K	V7A	D1 D2	C6E
R37	B	330K	V7A	D1 D2	C6E
R38	A	470K	V6A	K1	C23C
R38	B	470K	V6A	K2	C23C
R39	A	470K	V8A	G1	C13A ***
VR1	A	5K	JACKS	-	-
VR1	B	5K	V5B	K	C8W
VR2	A	2K	V5A	K	C8S
VR3	A	1M	V7A	G1	C4A

The front panel paintwork is in very poor condition, being very flaky and can be readily brushed off.

Also, the whole front was covered in what seems to be dried mud. I was cleaning some of this away when I noticed this Headphone/Line jack socket escutcheon was fitted upside down.

This suggests the front panel has at some time been detached and painted (with an unsuitable paint type or without properly cleaning the metalwork). At first I thought the escutcheon was fitted accidently upside down,but, it's possible the owner wanted to re-position the smaller hole to suit an alternative jack plug without a thick shoulder?

Below a view of the underside of the receiver and the three individual modules.

Click any picture to see full size.

Here are three separate views of the underside... click any to see full size

Here's a view of the RF (aerial) coils through the aperture at the bottom rear of the RF module.

Because of the layout of the coils and their close proximity I can see a potential problem in absorption of RF by adjacent coils unless a shorting ring is present on the wavechange switch. This might result in flat spots in one or more tuning ranges and, if the same design is used in the oscillator section, either a similar flat spot due to reduced oscillator amplitude or even the oscillator dropping out..

Above the coils in the previous picture you can just make out the mixer diode. Here's a better view of the VR92 (=EA50). There's some tarnishing on its pins so it may be necessary to remove the valve and clean them.

Is that elastoplast valve remover original?

Reminds me of an acceptance meeting regarding equipment to be fitted in a satellite ground station...

"Vot is dat... string?"

"It's NATO codified string!"

Where to start on tackling this receiver? As with nearly all my collection I'd like to see if it can be made to work. The biggest drawback of course are the aged components, particularly the old paper condensers which will leak. In this case, instead of just removing each in turn and stuffing with modern components (my preferred choice is to use high voltage chips), I intend to connect a high voltage from a bench power supply and gradually increase the voltage from zero to 250 (or 320 volts for the main rail). This done I can measure any leakage and at least replace any condensers that are getting warm. To this end I'll use a thermal camera.

Here's the first results. I didn't have my thermal camera to hand so basically relied on smoke (wisps appeared when the drain was 82mA).

I located a convenient point on the 250 volt rail which happened to be the top of a large orange resistor (R17A).. the feed to the audio output valve V8A screen grid. Cranking the variable HT bench supply to 250 volts resulted in a current drain of 55mA. This gradually increased and at 82mA I switched off the supply. R17A was pretty hot as was the metal-cased condenser shown here C3B (decoupling the screen grid to ground). I clipped its ground lead and repeated the test. Oddly the result was exactly the same with a draw of 82mA before I switched off. Again C3B was hot, so clearly its metal case (clipped to the chassis) was connected to the negative terminal. This is marked with the wavy black ring (right). The condenser is marked as you can see here, being made in April 1944.

Checking with a meter proved the outer case was indeed shorted to the negative terminal but the DC resistance to the positive terminal was a few megohms. The latter condition is typical of old condensers and doesn't really hint at breakdown when HT is applied. The way that current gradually increases suggests a slight short which reduces in resistance as the condenser gets warm and the short is compressed in-line with internal expansion of the innards.

Clipping the positive lead and repeating the test resulted in a constant current of 24mA. Waiting for a few minutes proved this was unvarying other than a couple of mA up and down. One of the voltage stabiliser valves V3B (AW2 = 7475) was illuminated so a little of the drain will be due to this. If you look at the schematic you'll also note several resistors configured as potentiometers which will also draw current. The remainder might be leakage at the metal rectifiers and leaky decoupling condensers.

I reckon it'll be safe to now apply mains power and see what happens.

Applying power wasn't straightforward. The on/off switch was lazy and continuity between the pair of mains pins to the transformer primary was intermittent and high resistance. I checked the fuses and found these were is a fairly poor corroded condition. I decided to bypass the switch, fuses and 2-pin mains connector and applied power directly to the mains transformer primary, using the +10 and 230 volt pins (these were already wired). The pair of dial lamps lit and nothing was heard from the loudspeaker after a couple of minutes so I plugged in a pair of headphones. The centre socket wasn't very loud so I plugged into the RH socket and heard a very loud 100Hz hum. Trying the various knobs and switches (significantly the wavechange switch) showed the receiver was probably in a generally working state, but the loud hum meant that the HT was likely to be unsmoothed, or maybe the metal rectifiers were bad, so I turned off the set.

These excerpts from the schematic show the power supply circuitry. C2 x 4 are 8uF and C8 x 2 10nF.

Sure enough all four 8uF condensers measured "open circuit" on my capacitor tester so I checked with my universal tester and initially it was confused. C2E was 75nF then a varying 459Kohm, C2D was just a few nF then give or take 44Kohm, C2C was 3nF and C2B only 14pF.

The HT rectifiers are Selenium types. W1A seems to be a single full-wave bridge. This needs to supply only a small current for the voltage stabilisers and the first local oscillator V1B. W2A comprises a pair of much larger rectifiers as it has to supply all the valves bar V1B.

On the right is the area causing a problem with Slydlock fuses shown as F1A, F1B and F1C. The mains switch is quite a large robust affair and a squirt of switch cleaner will bring it back to normal.

The dotted area looks a bit confusing but represents a short cable terminated in an octal plug which can be inserted into one of two positions.. either mains or battery operation.

I cut the positive connections to the four bad condensers and loosely soldered temporary replacements, soldered a pair of mains connections directly to the 240 volts terminals on the mains transformer, plugged in a pair of headphones, switched on my signal generator to 80MHz AM at 10dBm and turned on power. The dial lamps came on and, after only 30 seconds, I heard a healthy hissing sound and a weak audio signal. By chance I happened to be switched to Range 3 and tuning the knob a few degrees the 80Mhz signal came in very loud and clear. Whatever its weaknesses the R308 seems to be extremely reliable.

Following are some experimental pictures of a couple of resistors preceded by a picture of the area under review.

This picture is interesting because it shows R25A which is part of the anode feed to V7A, a 6Q7 whose anode current should be no more than 1mA. the 27K resistor should be dissipating about 27mW but from its temperature it's running at a lot more than this.

The schematic reveals the answer. There's a 2uF decoupling condenser which must be very leaky.

The proof is the tiny adjacent resistor which is 220K (R35C also part of the 6Q7 anode feed) and this would be incrediby hot if it were drawing the same current as R25A.

This is R7A which is the 3K screen grid resistor for the audio output valve a VT52 (=EL32). G2 of this valve should be drawing around 4 or 5mA. This equates to about 75mW disipation in 3Kohms so its temperature looks a bit high.

I already disconnected its 2uF decoupling condenser which was short circuit. My guess is that either the resistor has changed its value or the EL32 is not performing correctly.

When I looked at the picture of the hot resistors I realised that their temperatures and physical sizes didn't really mean much to me, for example... what temperature does a half watt resistor run at when the voltage across it draws enough current to sink half a watt? The temperature of course will be dependent on ambient, cooling airflow and heat sinking etc but take a resistor in air with no airflow.. what happens?

I set up a simple experiment, choosing a 200 ohm carbon resistor made sometime between 1938 and 1946.

This measured about 12mm or 1/2 inch long and 4.5mm or 3/16 inch diameter which I assume is rated at half a watt.

I waited about one minute before measuring each temperature. Figures are approximate.

From these results the 27K is definitely too hot and the 3K probably about right.

VOLTAGE	CURRENT	WATTAGE	TEMPERATURE	NOTES
10V	50mA	0.5W	49C	-
12V	62mA	0.75W	58C	-
15V	68mA	1W	74C	-
17V	89mA	1.5W	91C	slight smell
20V	100mA	2W	116V	hot smell
24V	125mA	3W	165C	very hot smell
28V	143mA	4W	179C	burning smell, discolours

Now a picture of the EA50 mixer diode followed by the 250 volt HT rectifiers. The diode has a piece of elastoplast stuck over the body.

Whilst looking over the R308, having discovered its reservoir and smoothing condensers were all bad, I wondered if the same rules applied to selenium rectifiers (two of its three pictured above) as valve rectifiers. In particular the maximum value of the reservoir condenser. Digging out a 1961 valve data manual I was surprised to read the term, not "maximum", but "minimum" value for the reservoir capacitance, unlike vacuum tubes which is the reverse. The R308 designers chose 8uF for both reservoir and smoothing, but perhaps this was a practical decision based on physical size or even the price of larger ones?

Bearing in mind selenium rectifiers have a relatively high resistance, then any capacitor surge would be quite low. I'd decided to look for new capacitors and restricted my search to 450VAC motor start types, being easy to mount (with their fixing stud), but I'm now thinking I can use normal electrolytic types as I have plenty salvaged from written-off lift equipments.

Those two rectifiers shown above have an inked code "280/LU530C" written on their ends. Each has three connections so will each be half a bridge (two diodes in series). Maybe 280V @ 530mA ?

I removed the suspicious 2uF condenser decoupling that 27Kohm resistor that's getting very warm and tested it. It's marked 2uF at 350/425 volts but both my capacitor testers said it measured 5uF with less than 0.5 ohms ESR, but my general component tester weirdly said it was an 11 volt zener diode one way round and an 11pF capacitor the other. Maybe the case is shorting to the positive terminal, but no... neither terminal was shorting. Still puzzling I connected it to my bench HT supply and found it broke down at 100 volts and went dead short. Presumably this was only temporary with the 27Kohm current limiter as it had previously recovered.

Next job is to fit new Slydlock fuses.. due here today (11 Sept 2025), then new reservoir and smoothing capacitors. The fuses arrived and I discovered that in order to remove the old ones I had to remove the power supply. Surprisingly I found that this aspect of the receiver design was really good. The power supply could be detached after removing just four 0BA screws and its cabling was long enough to balance the chassis out of the way.

The three fuses are located bottom left of the front panel adjacent to the 4-pin plug for 12V battery power.

These fuses are... top to bottom 12VDC, live mains, neutral mains.

The mains plug fits on those two large pins (typical in 1944)

This is the view revealed once the power supply chassis has been lifted away (the lowest fuse has already been replaced).

Each Slydlock fuses is held in place by a 4BA screws and nut.

A design weakness can be seen in that the connections to the fuses are made with large nuts and washers which are extremely close the the chassis.

The fuses turned out to be different to the ones I'd bought. Those very long threaded posts are an integral part of the old fuses and cannot fitted to my new ones.

This meant I had to modify the replacements so they'd fit.

Connections to the existing heavily tarnished fuses are made by long threaded posts which screw into the innards of the fuses. The new fuses are different. I had to remove the two brass fittings from them and drill and tap these to accept those posts. In fact I dispensed with the old posts and used shorter M4 bolts suitably modified to fit. Are the originals 4BA?

I spent ages working out the best approach but in the end each fuse was modified within 10 minutes. I drilled through the existing tapped hole (=M4) right through the fitting and threaded the rear hole by tapping through the top as shown. This is to ensure the top and bottom holes line up and are not angled. The rear of the fuse has indentations which are drilled out to accommodate the threaded posts.

For the fuse to work mechanically the modification must be very accurate otherwise the slide and lock doesn't work smoothly.

This picture shows a modified fuse. The plastic posts were cut from the old fuse and serve two purposes. One is a safety feature as the threaded posts go through the front panel, but essentially that nut screwed tightly prevents the post from rotating when connections are made to them.

This exercise was really annoying but as the original fuses were in poor shape and in fact open circuit there wasn't an option if the receiver is to be kept original.

Below, all three Slydlock fuses installed.

The next task was to fit reservoir and smoothing capacitors.

I selected a pair of 220uF x 450V reservoir capacitors and a pair of 470uF x 450V smoothing capacitors but I'd not noted which yellow sleeved wire was which except to mark them with stripes 1 to 4.

One choke is fitted at the end of the chassis and the other inside the power supply chassis but which was which? L11A serves the 320V supply and L12A the 250V supply. I worked out that the latter is physically larger and underneath the chassis but I turned on the mains power and measured the voltages at the yellow wires. This enabled me to identify which was which. The voltages on the 4 wires emerging from under the chassis read 257V,183,195 and 20V. The highest (257V) is the 320V supply at the rectifier output whilst the 20V is a leaky figure from the regulators from the 250V rail. 195V is the input to L12A and 183V the 250V rail.

Having established the identity of the wires I connected the four capacitors and checked the sets voltages. The 320 volt supply settled at 313V and the main supply dropped to 238V after the valve heaters had warmed up.

The PSU (opposite) will be tidied up soon. Although the chassis can easily be detached from the main receiver chassis it's not too easy to access.

Now that the receiver is powered correctly, and with steady HT voltages, I can investigate its operation. The first job was to identify all the control knobs.

Now that the power supply is producing correct voltages I began a bench test (using wires soldered directly to the mains transformer), initially with that signal generator peeking out above the receiver and headphones, but after about 30 seconds the Wavetek told me it wasn't happy having failed its self test. Next was the Rohde & Schwarz (on the right of the receiver) which turned on, passed its self test and prepared to give me a test signal.

A really annoying feature of this signal generator is its AM settings as the default is a stupid value like 60KHz. Anyway I set it to 25MHz, minus 10dBm and 1KHz AM. The R308 responded but with the dial showing a slightly high frequency. Resetting the signal generator I found all the R308s five ranges were working OK. One of our strongest FM stations is Classic FM at around 100MHz so I set the R & S to 100MHz FM and tuned this in. I then removed the signal generator and connected an aerial. Tuning a little to one side I heard Classic FM at good strength but somewhat distorted. I tried various knob twiddling and the results were pretty much the same but with the limiter knob working in FM. Audio volume was perfectly adequate (in fact some controls noisy enough to be quite painful on my ears) but when I checked the loudspeaker it seemed completely dead. The speaker is marked 140ohm and measures around 200ohms but is completely silent. That's the next thing to sort out. I recall the speaker on my last R107 (physically it looks exactly the same) was also duff.

This is the audio output stage with a 6Q7 driver and EL32 output stage. As you can see it uses a common transformer for phones and speaker and a line output also (derived from the phones output) as shown below.

Of note (left) are three components viz. C3A (currently removed but probably not important with the new 470uF smoothing capacitor), C1B which is likely to be bad, and the commonly bad coupling condenser C13A. I'll wap that first, then renew C1B.

Checking the voltages across R3A and R17A will define the results.

Why the designers went to the expense of providing L15A and L15B (and C10A, C11A and C11B) is very odd. It reminds me of an exercise during the design of an air defence system at Plessey back in the early 1980s. A small group of three engineers were recruited and given the job of overseeing the design of the interface between the air traffic controllers, fighter interceptors and missile people. They spent an inordinate amount of time and money coming up with special VOGAD amplifiers and hugely expensive transformers and filters.

The day of the big demonstration of their efforts using specially purchased test equipment that used artificial speech. A queue quickly formed.

"This will show you the difference between a basic amplifier and our super duper filters". Sure enough one of the two tests produced beautifully clear speech and the other was muffled. Unfortunately all the neutral listeners chose the basic circuit. The reason was that the headphones and microphones used in the system were already designed for ideal audio characteristics so adding another layer of 300-3000Hz audio filtering didn't work too well.. it sounded muffled.

I was surprised to dscover that C13A had very little leakage. When C13A was disconnected the EL32 anode rose from 152 to 157 volts. The joint anode screen current therefore dropped from about 19.6mA to 18.6mA.

Notwithstanding, I'll remove C13A and test it but fit a modern equivalent.

The dissipation of the 3K resistor feeding the anode and screen is about 1.2W and hardly changed. A check across the cathode resistor R12A supported the figures above (12.72V over 700 ohms = 18mA). During testing I noticed reception of Classic FM sounded almost identical in AM, CW or FM so some work is needed to resolve this.