Ekco AD65.

[60 oldjohn]

Thanks Graham, It is the Bakelite piece in the second picture I need the dimensions from. Especially the length it protrudes into the case.
I suspect the double ended threaded piece is an addition to take up the slack for the bent bracket. I guess it would originally have had a bolt and washer straight into the case. Any other AD65 owners with sets out there with the double ended spacers?
Graham, Thank you for the kind offer but I think I may use Milliput and I have some threaded bushes to make them.
John.

[Station X]

John.

I think the spacers are needed. without them the chassis would "bottom" on the lower/rear chassis fixings when the upper fasteners were tightened. This would probably start pulling the upper the inserts out of the Bakelite. There's also be the possibility of the chassis pivoting fore and aft on the lower/rear fixings.

I guess a pile of washers on a set screw would do the job, but would be awkward to fit. Washers on studding?

[60 oldjohn]

Graham
Thought things might have changed after you get the speaker frame straight, or maybe not.
John.

[Station X]

Today I removed the speaker mounting bracket and straightened the crossbar which is an integral part of it and supports the tuning scale pointer shaft bearing. I made sure that the bearing was perpendicular to the crossbar.

10mm spacers are needed between the upper front chassis fixings, which are moulded into the cabinet, and the chassis. This applies regardless as to whether the cross bar is dished or not.

With the chassis fitted into the cabinet, the pointer is clear of the speaker and clear of the tuning scale.

However the pulley driven by the tuning knob and the pulley attached to the pointer are not in line and there is a danger that the tuning drive cord linking them will slip off. This leads me to think that the cross bar should be dished to some extent.

I'll experiment further tomorrow.

[Radio Wrangler]

Oh, re a comment made earlier.

Epcos were the passives arm of Siemens. They have an excellent reputation for the quality of their components. Capacitors, pot cores, ceramics and SAW devices.

I didn't know of any link between Kemet and Rifa. Rifa started as an offshoot of L M Ericsson of telephones fame. I first came across their switch mode power supply modules... they started out as converting -48v phone exchange battery voltage to +5 and +/-12 for modern electronics. Back in the 1980s, their non-shorting paper capacitors were considered to add flexing capabilities to the legs of bees.... Little did we know.

I suspect that most other capacitors designed for the surges sometimes seen on mains networks may also have limits on longevity, but I admit I flinch when I see one of those golden blocks.

David

[Station X]

Originally Posted by Station X

I checked voltages and compared them with the service sheet. All were OK except for the screen grid of the VP1321. This should vary between 115V and 145V depending on the setting of the volume control. It actually measured 145V to 200V. I've put this down to the low emission of the VP1321 and will attempt to source a new one.

I managed to get hold of a used VP1321. Checked it on the CT160 and it tested good as new.

Put it in the set and reception of LW is much improved.

However the screen grid voltage still measures 145V to 200V. The fact the screen grid voltage approaches the HT voltage at one extreme of the volume control indicates that the screen grid current must be almost zero at this point. I've also noticed that only the upper third or so of the volume control rotation is actually functional. Below this the set is silent except for mains hum. The increase in volume is smooth over the working part of the control, varying from zero to full volume, which tends to indicate the wiper is OK. I should point out that the volume control circuit in this set isn't conventional. It controls the bias voltage on variable mu valves.

Started to check the VP1321 screen dropper resistors of which there are two in series, but this isn't easy because the Manufacturer's and Trader service sheets show them as having different values.

Manufacturer's sheet R19 (no value given) equates to Trader sheet R6 15K. 15K fitted.

Manufacturer's sheet R18 15K equates to Trader sheet R7 50K. 47K fitted.

There are similar problems with the resistor in series with the volume control.

Manufacturer's sheet R17 300R equates to Trader sheet R8 2K, 330R fitted.

The volume control looks original and measures 2K across the track. This is as per the Trader sheet. the Manufacturer's sheet give no value. I'll check that the wiper is making contact over the whole of its travel and whether the pot is lin or log.

I'll also have to consider whether the screen grid voltages I'm measuring are down to my using a DMM with a high resistance input rather than a 1000 ohms per volt meter on an unspecified range.

[Station X]

Manufacturer's component designations.

Checking pre-restoration photographs shows that:-

R18 is 50K.
Incorrectly shown as 15K on the manufacturer's sheet.
Correctly shown as 50K on the Trader sheet.
47K fitted during repair.

R17 is 300R.
Correctly shown as 300R on the manufacturer's sheet.
Incorrectly shown as 2K on the Trader sheet.
330R fitted during repair.

Unable to check R19 on photograph as it's hidden under a tag strip.

[Station X]

Checking my notes and forensic examination of the contents of the bin I threw the old resistors into, shows that R19 was 15K, which was the value I replaced it with.

As a result of help received in this thread:-

https://www.vintage-radio.net/forum/...ad.php?t=98891

I shunted my meter with a 200K resistor to simulate a "Standard Avometer" on the 1200VDC range. The VP1321 screen grid now reads 116VDC to 156VDC which is much closer to the 115VDC-145VDC specified.

So I've been looking for a non-existent fault.

I'm still puzzled as to why the volume control only raises the volume above zero in its last few degrees of rotation. Is this normal behaviour for these sets?

[Herald1360]

Well the volume control appears to work by changing the bias on the IF and FC valves. If they're not exactly as original or gone low emission or Rs around the circuit have drifted off value that might affect the operation.

[Station X]

Very true, but all valves are now well up to spec and test voltages correct.. All off spec resistors have been replaced, but with nearest modern value, so 56K replaces 60K for example.

The B2 Spy Set volume control operates in a similar manner by changing the voltage on the control grids of variable mu valves. The volume control on that is a bit all or nothing too.

I was hoping that someone with a working AD65 would be able to tell me what's normal. I don't want to go looking for a non-existent fault.

[PJL]

The reflex bit makes it difficult to understand the interaction between volume control and the AGC action of the reflex design.

The best you can do is check the cut-off of the FC13 & VP1321 occur in the right part of the volume control. These sets assume the valve characteristics remain the same over the production life of the valve. I have a Pye set which uses a similar circuit and requires component changes if you use a later production valve.

[Station X]

Yesterday I returned to working on this set after a long break.

There's been some discussion here as to whether the bar across the front of the speaker, which supports the tuning scale pointer pivot, should be dished or whether the dishing is a result of this set having been dropped. A bit of experimenting showed that the bar needs to be dished by about 1/4". This enables the pointer to move freely between the loudspeaker surround and the tuning scale, whilst also enabling the tuning cord pulleys to line up with each other. Initially I found that the pointer did not remain a constant distance from the speaker surround when rotated. This was cured by placing a piece of tubing over the pivot and bending it in the required direction.

Having got the pointer working correctly, I noticed that the "lashings" securing the flying leads feeding the dial lamp on the pointer were coming unravelled. Resisting the temptation to use miniature cable ties to secure the leads, I relashed them using waxed twine.

Working on these sets is difficult, as no matter how the chassis is positioned in order to gain access to the underside the dropper resistor and pointer are very vulnerable should the chassis suddenly tilt. I fashioned a wooden stand to support the chassis upside down with plenty of clearance underneath for the dropper. The stand is crude and over engineered, but it cost me nothing and when I've finished with it, it can be used as firewood. It certainly makes the set much easier to work on.

The wave change switch could be rotated through 360 degrees. Inspection showed that the protrusion on the rotating paxolin part of the switch was pushing the stop lugs aside. I removed the circlip from the shaft and pushed the shaft inwards so that the rotating part was clear of the fixed part. The stop lugs were then gently bent so as to better intercept the protrusion. This cured the problem. I would have liked to have fitted a shim washer to the spindle so as to put more tension on the switch contacts, but I didn't have one of a suitable size.

As the tuning capacitor approached full mesh the set would suddenly become very crackly, indicating that the plates were touching. The resistance across each gang of the capacitor was checked with a meter. The circuit is such that the front gang of the cap is not shunted when the set is set to MW. The meter showed infinite resistance in all positions of the cap. The middle gang is shunted by a high resistance, which remained constant as the cap was meshed. The rear gang showed 18 ohms until the point where the crackling started whereupon it reduced to a couple of ohms. This was obviously the faulty section. Viewing the plates through a magnifier showed nothing obviously wrong. Pushing thin paper between the plates didn't make the short disappear either. Then I noticed that the outermost rotating vane was of the type which has radial slots. The short manifested itself just as the last "sector" started to mesh. I bent this sector out very slightly, which resulted in a complete cure.

The working range of the volume control only extends over the last few degrees of rotation as the control is rotated towards max. The volume control circuit is unusual in that it works by varying the cathode voltages of the frequency changer and IF Amp/AF Amp valves. This in turn varies the voltage on the control grids of these valves which are of the variable mu type. I checked all the resistors in this part of the circuit ie R4, R5, R3, R1 and R8 (Trader Sheet) and could find nothing wrong. The voltage on the volume control varied from 0 to 18V as the control was rotated, this being what I expected. Sound was first heard at the 4V point, reaching a max at the zero volt point. I concluded that the fault, if indeed there is a fault, must be elsewhere, possibly in the AGC circuits.

[Station X]

Originally Posted by PJL

The reflex bit makes it difficult to understand the interaction between volume control and the AGC action of the reflex design.
The best you can do is check the cut-off of the FC13 & VP1321 occur in the right part of the volume control.

Checked all the resistors in the AGC circuit ie R13, R14, R15, R5 and R12 (AGC delay). All OK. All capacitors on the AGC lines have already been replaced as part of recapping.

Measured the voltage on the AGC diode using my Fluke DMM which has a 10 Mohm input impedance. Reading was about 400mV regardless of whether a station was tuned in or not.

Put the output valve on the CT160 and tested the diodes which were both OK.

Temporarily replaced C13, the 1000pF capacitor which feeds the IF signal to the AGC diode. No change. Original cap put back.

Disconnected C13 which resulted in a drop in AGC voltage and an increase in gain. Reconnected it.

Not sure about the constant AGC voltage. Perhaps the RF signal isn't strong enough to bring the AGC into operation? I'm using a six foot wire aerial indoors. I'll try a 150 ft outdoor wire and a signal earth tomorrow.

[PJL]

As it should be OK, how about a full alignment next.

[Station X]

I had another play with this today. The first thing I noticed was that the voltage measured at the AGC diode was positive wrt chassis, whereas an AGC voltage should be negative.

Disconnected R13 from its junction with R14 and connected a meter to the end of R13. I now saw a negative voltage whenever a station was tuned in. This voltage varied with the modulation of the carrier, but this is to be expected with the AGC line disconnected. It also varies with the setting of the volume control which in this set is actually an IF gain control. So it looks like I have a problem on the AGC line somewhere.

I attempted to adjust the IF transformers to make sure they were peaked, but was unsuccessful. The windings are tuned by what look like compression type trimmers. I don't know whether the dielectric is mica or air. A brass threaded rod passes through the plates and on the end of the rod is an insulating washer and a brass adjusting nut. It appears that the plates have taken on a "set" and it's only possible to increase the capacitance, not decrease it. I'm going to have to remove the cans from the transformers in order to access the trimmers. Hopefully I'll then be able to bend the plates so as to reduce the capacitance when they're uncompressed.

I can see that this set is going to provide many more hours of fun.

[Station X]

I said that this radio would provide hours of fun and it's certainly doing that.

I decided the lack of negative AGC voltage was down to the set using delayed AGC. About 6 volts of delayed AGC in fact, so the AGC voltage won't go negative until it exceeds that level. Therefore I reconnected the AGC line.

Next I turned my attention to the IF Transformers. Removing the covers showed that the trimming capacitors, which have mica insulation, had developed a "set" at close to maximum capacitance. I carefully bent the moving plates outward to allow for some adjustment. After replacing the covers I set all the trimmer nuts about one turn out to give an approximate alignment. Interestingly, the paper insulation inside the covers was date stamped 1 October 1934.

I then switched on the set to find it was absolutely dead. Took some voltage readings. HT was down a bit, but allowing for this, the voltages on the frequency changer and IF/AF amp valves were OK. Anode and screen voltage were low on the O/P valve Its cathode voltage was high, about 9V whereas I'd previously recorded a reading of 6V (it pays to keep notes). It took me a while to work out that the control grid was actually floating. Although the top cap connection was firmly in place its inside was tarnished and it was actually insulated from the grid. I cleaned the inside of the top cap and the set then worked again. To confirm my diagnosis I removed the top cap and watched the cathode voltage change from 6V to 9V accompanied by silence.

Out with the signal generator, not forgetting to disconnect mains earth from it, as this is a live chassis set. Connected SG to A and E terminals via a capacitor as suggested in the service sheet, set it to 110kHz with 99% modulation at 1000Hz. Nothing heard from the speaker. Turned SG output level up and swept the frequency a few kHz each side of 110kHz. Still nothing. Shifted SG to control grid of frequency changer as per service sheet with same negative result.

Tested SG by scoping the output, listening to it on an LF receiver and checking the frequency with a counter. Everything is perfectly OK.

The set works though, so I decide to try a different approach. Looking at the set's tuning scale, which isn't fitted yet, shows that with the pointer set vertically the set is tuned to 380 metres or 790kHz. Set pointer vertically, switch to MW, inject signal at 790kHz and adjust the oscillator trimming cap until the signal appears. Cap peaks nicely. It's the only LO adjustment there is on medium wave. Sweeping the SG and receiver shows that MW coverage is from 507kHz to 1442kHz which isn't far out.

Set the pointer to a few random MW frequencies and sweep the SG. There are two responses at each point, a strong one and a weak one, the weaker one being the image frequency. The two frequencies are consistently 218kHz apart indicating that the IF is tuned to 109kHz. Attempt to peak the IF transformers. Two caps peak nicely, one runs out of adjustment when screwed right in and one has no affect at all.

Test the AGC. The voltage starts to go negative at about 14mV of RF input. With the SG set to max output of 2V, the AGC voltage is -5V. This of course depends on the setting of the "volume" control.

Do a crude sensitivity test. With 1000Hz 99% modulation, the signal is just audible in the speaker with 130uV of RF at the aerial socket.

So it looks like I have an IF problem, but I'm not sure what. I've checked the resistance of the IF Transformer windings and they're all close to the 100R specified. I'll try injecting a 110kHz signal further from the frequency changer and nearer to the speaker.

[Alistair D]

Graham, disconnecting the earth of the sig gen is not safe. Do some experiments with the earth connected and say 100pF between the generator's earth and the radio chassis. You may have to go up or down in value to get good results. Ensure that the voltage spec of the cap can withstand mains across it. There are many disc ceramics that can withstand kV so it should not be a problem finding one.

Al

[Radio Wrangler]

I'd keep the sig gen to about 80% mod. I doubt whether many generators can really do 99% and it's more than many receivers can handle without clipping/bottoming.

David

['LIVEWIRE?']

Generally speaking, I've always understood that 40-50% mod. is adequate. Is there something specific about the AD65 which requires the mod. level to be 80%? (I've never had any dealings with an AD65, btw.)

[Station X]

Unless specific settings are specified in alignment instructions, I generally run mod frequency and depth at whatever they happen to be set at. In this case I turned everything up (and down) in hope of getting a signal through.

I have established that the set's IF circuits are tuned very close to the specified 110kHz. There is obviously something, probably something very simple, wrong with the set or my SG.

It must be remembered that this set was a basket case with many wires cut, components with the wrong values fitted and so on. It's quite possible there are errors in my rewiring. The need to remove the tag strip in order to replace components on the underside meant that the "replace one component at a time method" could not be used.

In cases like this I find it's best to leave things for a few days and go back to the problem later. That's what I'm doing with this set.