Vacuum Tube (Valve) Radio and Audio Repair Guide: Faults on detector/AVC/AF amplifier stages

When a receiver uses a double-diode-triode with top cap grid connection in this stage a very simple and convenient test for the AF amplifier section is to apply a finger to the cap, whereupon a loud hum should come from the loudspeaker (assuming, of course, that the output stage has been passed as satisfactory). In the case of single-ended DDTs such as the 6SQ7 try touching the blade of a small screwdriver on the centre contact of the volume control, with this fully advanced. If no hum results measure the anode voltage on the valve, which normally will be between about 40V and 100 V. Very low or no voltage suggests that the anode load resistor has gone high or open, whilst a high reading points to the triode section of the tube (valve) itself, for which the easiest test is substitution.

A very occasional fault encountered by the writer has been the fitting by a previous repairer of the wrong value volume control such as a 5 K-ohm type being fitted in place of a 500 k-ohm. It is well worth measuring the resistance of the centre contact to chassis with the control at maximum to make sure all is well here.

When the AF amplifier is working well but no signals are heard the detector must be investigated. The diode sections of DDTs are usually extremely reliable and seldom give trouble. A remote possibility is hum due to heater/cathode leakage which may be cured by swapping over the detector and AVC diodes. In cases where the diodes are strapped try disconnecting each in turn. If the detector is working correctly, touching the meter probe (on a suitable voltage range) onto the anode of the preceding IF tubes (valves) should cause crackling from the loudspeaker, but compare the remarks regarding oscillation in the section dealing with IF amplifiers. With this in mind measure the voltages on the detector diode anode, which should be zero with respect to chassis. As mentioned elsewhere in the text, if there is negative voltage on the anode the result as far as the listener is concerned will be ‘ploppy’ reception with the odd few stations appearing suddenly along the dial with wide dead spaces between them. In extreme cases the detector may be cut off altogether, with no stations at all receivable.

First of all, check to see if you are working on one of those sets fitted with ‘squelch’ or some other means of obtaining ‘silent tuning’. It ought to be possible to switch off the feature, but if operating the switch appears to make no difference the first thing to do is to check the switch itself with an ohmmeter. Usually the switch closes to override the squelch and if it fails to ‘make’ a dose of switch- cleaning fluid may put things right. If not it may be possible to fit a replacement unless the switch is a special type made to work with an unusual operating system. Given that there is very little call for ‘silent tuning’ nowadays it probably wouldn’t hurt simply to short out the switch permanently.

When, as in the majority of sets, ‘silent tuning’ is not featured, proceed as follows.

If you have a circuit diagram for the set being repaired, check it to see if the cathode of the detector diode (usually that of a double-diode triode) returns to chassis via a bias resistor. In certain cases no cathode bias is used on the DDT, which will affect the test procedure. In the absence of a circuit a quick check is to measure the resistance between the cathode and chassis; if a bias resistor is in fact used you may expect it to have a value of between about 500 ohm and 5 k-ohm.

Assuming that there is a bias resistor, the anode of the detector diode should return to it via the diode load, which may be anything up to 2 M-ohm If there is no path the diode load is either open circuit or someone has rewired it incorrectly.

Now disconnect the lower end of the bias resistor from chassis and test for any resistance between the diode anode and chassis. Theoretically it should be infinity, any reading at all showing that a leakage is taking place, most likely via a coupling or decoupling capacitor. Try disconnecting one end of each of these in turn until the resistance returns to infinity.

In cases where a bias resistor is not fitted and the cathode is taken directly to chassis, the resistance of the diode load should appear between the anode and chassis. When the bias resistor is omitted ‘grid current’ biasing is used for the triode section of the DDT, effected by returning the grid to chassis via a very high tubes (valves) resistor, typically 10 m-ohm. It is essential that the negative voltage which is developed at the grid should not reach the detector diode anode, so a DC blocking capacitor is inserted between the two. To test this, check to see if there is any resistance between the grid of the DDT and the detector diode anode — if there is it should not be less than 5 M-ohm and a more usual figure would be well in excess of 10 M-ohm.

As with any circuits where high value resistors are employed in conjunction with coupling and decoupling capacitors, the latter must be completely free from ‘leaks’. This is particularly important with DC blocking capacitors used to prevent the detector diode anode from returning to chassis and thus receiving negative bias. It is here necessary to repeat the warning about keeping an eye out for previous work done on a set, in case someone has done some incorrect wiring around the volume control, especially if this has been replaced.

General AVC problems

Whichever type of AVC is employed check for leaky decoupling capacitors along the bias line. Note that in some cases negative delay bias is applied to the AVC diode from the negative HT line, this voltage also being used as minimum operating bias for the IF amplifier and frequency changer tubes (valves) (and RF amplifier, if used).

When a ‘clamp’ diode is used to delay the AVC check the resistor used to provide the positive clamp bias. This is likely to be of anything up to 22 M-ohm and experience shows that high value resistors are all too likely to go higher and higher until they become the next thing to an open circuit. When this happens and the clamp bias disappears the performance of the set may well be restricted quite badly.

This latter remark applies equally to the failure of any means used to delay AVC. As always, look out for previous and faulty repair work.

Amplified delayed AVC, being complex in design, naturally stands a greater risk of going wrong than conventional systems. This will usually result in the AVC action being restricted and causing severe overloading on strong stations.

The chief thing to remember is that no AVC bias can be developed without the negative voltage source to which the cathode of the DDT is returned. Measure this point to see that it is at least 50V negative with respect to chassis. If this appears to be in order, check the resistors used to connect it to the DDT cathode.

If much lower than normal, or even no negative voltage at all is found the fault is almost certain to lie in the HT smoothing circuitry. Check the resistance between the centre tap on the mains transformer HT winding and chassis to see if it corresponds to that given in the service data for the set. If not, check the resistance of the smoothing choke or field coil in case it is suffering from shorted turns, then the values of any series resistances used. As always, be on the look out for incorrect replacements fitted by an earlier repairer.

Because the AVC action depends so much on the amplification factor of the triode section of the DDT, any drop in emission in the latter will have a marked effect on the efficiency of the system. The best bet here is to check the tube (valve) by substitution.

Detection and AVC in ‘short’ superheats

A few early sets of this type employed a separate double-diode valve, but this arrangement was soon superseded by the double-diode-output pentodes such as the Pen4DD and EBL1/31. In nearly all cases to make up for the lack of an AF amplifier between detector and output stages the AVC was heavily delayed by the use of higher-than-normal cathode bias resistors. It was common to employ two in series with the AVC diode returned to chassis to receive the maximum bias, the pentode grid returned to the junction to receive its normal bias and the detector diode returned to the cathode to receive zero bias. Obviously there is a lot of scope here for things to go wrong, especially if a previous repairer has been careless. As for DDTs, check the various bias voltages and look for leaky capacitors or high/open resistors.

Indirectly heated double-diodes in battery receivers

Points to watch: first, check that the voltages on the anodes of the diodes are within limits. Manufacturers were shy about giving figures, but the following tests should be quite adequate. Connect the meter negative lead to chassis and measure the cathode voltage, which should be approximately 6V. Then check the signal diode to ensure that it has nearly the same potential. If it is much less the diode would be unable to respond to anything but very powerful signals and it would suggest that the resistor returning it to the cathode had gone high value or open circuit. The AGC diode, however, should be at zero or even slightly negative for the delay to be effective. A low reading on the cathode in a receiver such as the Cossor where a voltage divider across the HT line was used to provide voltage, suggests that one or other of the resistors has altered in value. In the type of circuit used by Decca, in which the cathode is connected directly to a bias battery, it theoretically would appear to be impossible for the cathode to be below par but wires have been known to break before now.

Secondly, note that the AF appearing at the detector finds its way to the volume control, in most designs via a capacitor of around 0.1 uF. Alternatively the latter would be in series with the slider of the control and the following valve’s control grid. The purpose of this was to obviate a DC path which would result in either the diode or triode receiving some unwelcome bias of the wrong polarity. There are some high value resistors around this part of the circuit, and as always capacitors must be absolutely free of leakage to ensure correct working conditions.

All makes of indirectly heated 2 V double-diodes ceased to be produced a number of years earlier than the majority of 2 V types and replacements will be harder to come by as a result. In most cases, with ordinary directly heated battery tubes (valves), so long as the filament and bulb are intact there should be some kind of performance available but it is possible for a cathode to lose its emissive properties and a simple ohmmeter check cannot guarantee that the 2D2, 220DD, etc., will work. It is worth asking the specialist tubes (valves) dealers if they have a replacement available, but otherwise Mullard published an official modification to enable a TDD2A double diode-triode to be used in its place. The tube (valve) holder does not need to be changed and only three connections have to be altered:

Pin 1 on 2D2 — disconnect and take to pin 5 on TDD2A.

Pin 2 on 2D2 — leave as found

Pin 3 on 2D2 — leave as found

Pin 4 on 2D2 — leave as found

Pin 5 on 2D2 — Disconnect and insulate.

In addition, connect the lower end of the detector diode load resistor to LT+, making sure that in doing so you don’t accidentally short out the GB supply.

This should enable the set to work as before but the volume will be reduced to a certain extent due to the lack of AVC delay.

Grid leak and anode bend detectors

These are found mostly in TRF receivers although a few early superhets did use them. Prior to about 1935 triodes were used almost exclusively, but after that date ‘straight’ RF pentodes became popular because of their greater gain. Little can go wrong with a triode grid leak detector other than the tube (valve) itself or the grid capacitor and resistor or the anode resistor. Experience shows that the most likely candidate for failure is the grid leak, because as with all high value resistors it can go very high indeed or even o/c. When this happens the negative voltage built up on the grid in the process of detection cannot escape back to chassis, and it continues to rise until the tube (valve) is ‘cut off’ altogether. Note that in battery receivers the grid leak should return to the positive side of the filament.

Because the grid leak detector generates its own grid bias the triode cathode normally is returned directly to chassis unless a gramophone pick-up facility is incorporated in the set which makes use of the triode as its first AF amplifier. In this case the act of switching the set to the ‘gram’ position, or in some cases by merely inserting the connecting lead from the pick-up, a cathode resistor is brought into circuit. Always check that the cathode does indeed return directly to chassis when the pick-up is not in use.

An opposite arrangement may be used in the case of an anode bend detector. Its grid has to receive a considerable amount of negative bias for the detection process, usually obtained by means of a cathode resistor of up to 20 k-ohm Check that there is a resistance of this order between cathode and chassis if the detector does not seem to work properly. On the other hand, for gramophone pick-up use the triode needs far less bias and there is usually some means of shorting out all or part of the cathode resistor. Check that whatever switching method is used that the cathode returns via the correct amount of resistance when either ‘radio’ or ‘gram’ is selected.

When an RF pentode is used for detection the principles are the same as for triodes in both grid leak and anode bend modes, the only difference being that the screen grid has to be supplied with voltage — but not very much. When a feed resistor from the HT line is used it is likely to be at least 500 k and maybe up to 2.2 Mf When readings were taken with an AVO ‘Seven’ the service manuals reported the screen voltage as ‘very low’ or ‘not measurable’. An AVO ‘Eight’ will give a reasonable idea of what voltage is present but to be on the safe side check the resistor on the ohmmeter. Clearly any leakage on the associated decoupling capacitor will have a considerable effect on the voltage (cf diode pentodes used in ‘all-dry’ receivers).

In some receivers, especially of USA origin, the screen-grid voltage is obtained from the cathode of the output valve, this being quite sufficient for a pentode used as a detector. The anode load resistor also will be high and it too should be checked on the ohmmeter.

‘Westector’ diodes

These were small copper-oxide rectifiers made by the Westinghouse Brake and Saxby Signal Company from about 1932. Several types were produced in two main groups, the ‘W’ series and the ‘WX’ series. They were suitable for use as detectors, the first up to frequencies of about 200kHz, the second up to l500kHz. Thus ‘W’ types could be used in the early superhets with low IFs whilst the second could function in LW! MW TRF receivers, although being stretched to the limit at the higher end of the MW band. They could also be used to provide AVC and in ‘battery economy’ circuits.

Although cheaper than equivalent thermionic diodes Westectors never really achieved great popularity. They appeared in a few of the commercial superhets of the early 193 Os, before double- diode-triodes had been introduced, but seldom afterwards. Apart from the fact that the DDT could do three jobs at once, at a time when the larger the stance number of tubes (valves) in a set the better, as a selling point a tiny component, hidden away under a chassis, simply wasn’t so attractive as a tube (valve) in full view. In fact, the only real instance of Westectors being used in a mass produced set was in the Wartime Civilian Receiver of 1944, when tubes (valves) were in short supply.

Experience shows that Westectors give very little trouble. The writer does not recall ever having to replace one in fifty years of radio servicing but there can be a first time for almost anything. The method of testing would be to disconnect one end of the device and compare its ‘forward’ and ‘backward’ resistances on the ohm meter. The forward resistance (the direction in which conduction takes place) should be a fraction of the backward resistance. If the two figures should happen to be much the same the Westector almost certainly would be faulty. If an exact replacement cannot be obtained it may be possible to press a small germanium diode into service.