Roberts R77 testing and restoration

[getters]

Yesterday, I became the proud owner of a Roberts R77 in very good condition requiring minimal restoration effort.

First job was to remove the chassis and pull out all of the valves, visually inspect all components and soldered joints, dust, and apply switch cleaner to the on off switch, circuit board, sockets etc. and allow to drain. During this period, all resistors were measured (this can be done in situ), noting that all but R12 had no tolerance band indicating 20%. R12 was gold (5%). All were within tolerance.

I usually replace all capacitors in a set (excepting mica, polystyrene and ceramic) but this example was untouched and it would be a shame to spoil its originality if unnecessary. So, the next step was to switch on at the front panel switch and slowly ramp up the HT voltage from an external power supply in 10V increments, holding for around half an hour at each setting, to a maximum of 90V (valves absent). During this period, the leather case was cleaned using a pH balanced detergent (baby shampoo) applied using very lightly dampened cotton buds and cloths, followed by a rinse with clean water similarly applied. Over wetting may cause staining. After allowing to dry, Dubbin was applied.

On completion of the voltage step ramp, the radio was switched off at the front panel switch and the time taken for capacitors to discharge noted, monitoring the voltage at pin 3 of the empty DL96 socket. This took perhaps a quarter hour to fall below 30V indicating that the electrolytic and some decoupling capacitors were reasonably good.

90V was re-applied, instantly this time, and voltages measured at empty socket pins where HT would be found. Results worthy of note were around the penultimate stage DAF96. A low 50V on C20 due to the high value (10 Meg) g2 dropper resistor R9, and a low 75V on C22 due to the high value (2.2 Meg) RL resistor R10.

It was known that a leaky coupling capacitor to grid 1 of the output stage commonly causes the valve to overheat so this was an area of concern. However, the voltage measured on grid 1 (pin 6) of the output stage, DL96, was <5mV even though the grid leak resistor R11 was also 2.2 Meg, and the voltage drop measured across the resistors R9 and R10 were zero as expected. Obviously the test meter was loading the circuits it was measuring with its 10 Meg input impedance. Calculating the output voltages of a voltage divider with lower resistance value of 10 Meg confirmed that all measured results were as expected.

For confirmation, I needed to know how much current DL96 was drawing. This would cause a voltage drop across the 512 ohm output transformer primary DC resistance, and the service data sheet provides the anode voltage. I inserted the valves and switched the set on. The radio instantly burst into life, LW as well as MW, with such selectivity and volume which made me even reconsider realignment. All specified voltage measurements were satisfactory.

The AGC voltage, measured at the volume control R7, varied between -0.35V (no station) to -1.2V (strong station), the latter condition giving a reading of about -0.9V when measured at the waxed capacitor C5, this being after the 2.2 Meg resistor R6 comprising the low pass RC filter. Again this was attributed to loading by the multimeter input impedance. Operation was deemed to be satisfactory and the capacitor pardoned.

I then focused on freeing the volume control. It was evident that the indicator dial was fretting against the back plate but the knurled knob just wouldn't pull off despite switch cleaner lubrication. Eventually, I discovered that by rotating it clockwise to its stop and pinching the dial against the metal back plate to prevent rotation, the knurled knob could be gently rocked back and forth whilst pulling it off. Progressively, it disassembled. The knurled knob is a standard D style push fit with metal spring clip and is separate from the dial which, using a taper, pinches a brass tapered split ring onto the shaft.

A further generous coat of Dubbin was applied this morning followed by re-assembly and the restoration is complete. Batteries next. I was amazed there were no significant faults or components which needed replacement. I imagine it will provide many years of enjoyment in service before re-visiting the circuit.

[Steve_P]

Nice machines these. I'd be inclined to replace 'that' cap though for future safety and reliability.

Cheers,

Steve P.

[getters]

You're right, Steve. For reliability and safety (generally), it's best practice to always replace them, and to use Y1 grade capacitors on aerial inputs, tape outputs, across mains etc. where applicable. I would normally replace unreliable electrolytic and waxed capacitors and I was tempted to replace C21 and C22. Please forgive my thorough account which follows but you raised an interesting point which, I believe, is worthy of detailed examination. This amount of detail is probably OTT but it crystallised my own thoughts on risk assessment, and what better place than this site to record it? So, just how forgiving is this particular circuit design?

In a battery set like this, a shorted electrolytic would cause batteries to get hot accompanied by no sound - not a good scenario if left playing unattended! I've just measured 1A current for 1 second through a shorted new ZnCl2 PP3. This equates to 90W initially, reducing sharply as discharge continues, perhaps for 15 minutes or so. A Google search found the following useful PDF;

http://friedrichengineering.com/web_...%20Battery.pdf

I don't recall seeing a fuse inside the R77 and maybe designers thought that this eventuality was unlikely or inconsequential. I'll fit a 50mA fuse in the battery pack to prevent any potential heat damage, allowing me to slumber in peace. This would be a good idea even with a new capacitor.

If the electrolytic goes open circuit, it will cause audible distortion or total malfunction, and if it develops high leakage it will reduce battery life, both without damage to the set or myself. With a battery in-line fuse, C23 can safely stay.

Regarding the waxed AGC capacitor, open circuit, short circuit or leakage would all result in some audible effect without damage. The worst case scenario is C5 going short which will bias DF96 at 0V. Page 9 of the Philips data sheet for this tube shows that Ia would be around 1.7mA under these operating conditions (Va = Vb = 85V and Rg2 = 39K) which wouldn't exceed the 0.25W limiting value for Wa. It would also slightly reduce Rk of DAF96 making Vg1 go more positive but this valve is running under starvation conditions and there will be no damaging effect. The waxed AGC capacitor (C5) can safely stay.

I believe that the red capacitors are ceramic and rarely fail or go leaky. Where used as HT decoupling capacitors through dropper resistors, their failure would be largely inconsequential from a safety viewpoint. A very leaky or shorted capacitor on g2 (or G4) would safely lower Ia and result in an audible fault condition. An open circuit wouldn't alter the DC grid voltage, and therefore wouldn't overload the valve, but would place some AC there resulting in feedback through the power supply perhaps causing motorboating or reception problems. Decoupling capacitors C6, C12, C13 and C20 can safely stay.

Where used as coupling capacitors, much depends on the circuit. C19 open circuit would result in no sound output, that's all. Shorted, it would effectively make DAF96's Vg1(DC) more positive. Worst case, it would approach Vk DC causing 0V bias. The set would malfunction but, under starvation conditions with a maximum possible Ia of 90/2200 mA, there would be no damaging effect. C19 can safely stay.

The coupling capacitor, C22, and the low pass AF filter capacitor, C21, (I'm unsure to which you refer) are a different matter, and I was tempted to replace them. However, I believe these capacitor types are generally very reliable, and the precautionary DC measurements I made confirmed this.

C22 open circuit would result in no sound output, that's all. A short would raise Vg1 of DL96 but this would be self limiting to some extent as DL96 runs into positive grid current creating serious distortion. DL96 would be destroyed in minutes or tens of seconds. Worst case scenario (and a short is unlikely), if left unattended, I lose the output valve. High leakage is more likely (but not measured) and will cause a very much slower demise, albeit without audible warning. The DL96 in this circuit has a reasonable safety margin, operating at -10.4 x 0.56 = -5.8V bias where, according to the Philips data sheet, -4.1V bias represents the safe limit.

C21 open circuit would cause excess treble. Short circuit would bias DL96 at 0V, and the Philips data sheet shows that Ia would rise to around 13.5mA, causing DL96 to overheat with 1.2W dissipation, approximately twice its limiting value. It may take several minutes to destroy the valve, and there would be no sound output as a warning. Worst case scenario (and a short is unlikely), if left unattended, I lose the output valve. The effect of leakage is as described for C22 above, again without audible warning.

For completeness, C14 open circuit wouldn't alter DAF96's DC bias and hence Ia. C14 shorted would cause 0V bias on DAF96, causing no damage under starvation conditions as previously deduced. C14 can safely stay. C24 open circuit would cause a harsh sound, and shorted would give no sound, both without damage. C24 can safely stay.

The rest are tuning capacitors and faults would immediately be noticed through loss of functionality without component damage. Whilst reception is good, these can safely stay.

In summary, I'll keep a watchful eye on C21 and C22. Worst case, I loose the output valve, either quickly or at a very much more leisurely pace. I don't think it likely and I'm more concerned about putting a 50mA fuse in the battery pack.

Thanks again for your comments, Steve. If I've made a mistake somewhere in my deductions or overlooked something, please advise.

Regards,

Bob

[Silicon]

I used to use Dubbin (on leather) until someone told me that it supported fungal growth which would lead to rotting of the material it was applied to.

Can anyone confirm this theory?

[getters]

I also read these internet warnings on microbial attack before applying the Dubbin I had at hand. Neats-foot oil was one of the specified alternatives. A little research showed that both contain natural products, i.e. microbial nutrients, and both would ultimately be susceptible. Non nutrient synthetic leather preservatives were also available and may prove to be a better choice, but it occurred to me that unless they actually contain fungicides, problems can feasibly still occur during general handling and use where surfaces become contaminated with organic nutrients such as dust (skin), body oils and spills.

It is commonly understood that microbial attack is accelerated by hot and humid conditions, though it can occur under dry conditions, and is often discovered after periods of long term (usually damp) storage without regular inspection. Since it is rarely hot and humid (or overly dry) in my home and I intend to regularly use the radio, I thought it was much more important that the leather was supple to prevent the splitting and tearing often seen at buckle folds and stitches on vintage items.

I also read remarks warning against application of Dubbin to stitching, again due to accelerated microbial attack. Rightly, wrongly, or stubbornly, I applied it to the stitching to prevent any tearing there and (sparingly) to the zip fabric and zip for lubrication.

I argued that microbial attack would be slow and readily visible in my own climate and the Dubbin could always be removed at some later date using a suitable solvent, then replaced with an alternative product if problems arise.

Apart from its highly inflammable nature, isopropanol is generally safe to use on most plastics, fabrics and rubbers although it can remove some ink markings. It's generally a little milder in this regard than methylated spirit (ethanol). It would certainly remove most of the applied Dubbin and, today, I found it specified as a diluent for a surface consolidant in the last of the links below, suggesting that it may be safe on leather (as always, test it on an inconspicuous area before proceeding).

Interestingly, before cleaning the case, there was some surface fungal growth, not on the leather but on the synthetic material inside the battery zip compartment where the name plate is attached. This simply washed off with a cloth moistened in the pH neutral soap solution I used.

The part used tin of Dubbin I used was at least 20 years old and showed no signs of fungal or bacterial attack, off odours, drying out, oxidation etc.

Today, I looked a little further into microbial attack and it's difficult to find a definitive answer on this. There are commercial interests, conflicting statements, and experts remain cautious in their opinion on leather dressing, presumably because there are so many leather processing variables and so many other degradation mechanisms besides microbial attack which may be more important or need to be considered. Further, some items may have significant value - if this applies, then it's probably best to leave it to the experts.

Neats-foot oil is often cited in the Military and Federal specifications and standards, e.g. TO 00-25-245 - free download from www.everyspec.com .

The leather itself may already contain a fungicide. See MIL-DTL-32067.

Here's a very good link which expresses caution in the over application of dressings;

http://cool.conservation-us.org/waac.../wn19-209.html

Regards,

Bob

[Station X]

One post moved to a new thread here:-

https://www.vintage-radio.net/forum/...d.php?t=120055

[Heatercathodeshort]

I thought Roberts radio's were all covered in Rexine or vinyl. Dubbin is for leather ie horse saddles and is intended to penetrate and waterproof the leather. It serves no purpose on non penetrating finishes.

For decades, after cleaning, I have simply used a spray on polish such as 'Pledge' and buffed them up with a soft cloth.

Just my method and it works fine. J.