Help Needed on Atwater Kent 185A

[Paul Mathew]

I have restored an Atwater Kent 185A model, by replacing most of the usual suspects such as electrolytic, and wax paper caps, and replacing some perished wiring. All went fine and the radio jumped back to life. It's an export model, so is designed for 220V mains. The measured mains voltage hovers around 240VAC from my household socket.
The first disturbing issue I discovered was a reading of 220VAC to my house ground (independent of the radio chassis) from the antenna wire. I also have 220VAC to my house ground from the top cap of the 2A7 oscillator/1st detector tube. Both these points are connected to T1 and T2, so I'm wondering where this mains voltage is coming from, and also why 220V and not 240V?
I haven't removed the chassis from the cabinet yet to check, but I'm thinking that possibly C15 leaking/shorting to the chassis may be a problem, and the mains supply voltage is getting into the system via the chassis?
The second issue is that the volume dwindles down to a hardly audible level after a short while of playing. Not sure if there is any connection?
Looking forward to any wise counsel!

[TrevorG3VLF]

I do not know the mains system in SA but in the UK all metal work in the house should be connected together to give no opportunity for a shock.
This 'earth' can be very different from true earth outside the house. I would suggest you put a peg in the ground a few metres away from the house and measure the voltage difference. There has been considerable discussion on running amateur radio aerials into a common earth house. My brother had trouble shearing his sheep. It turned out that the steel shed was at 80V relative to true earth due to a faulty supply transformer.

[Amraduk]

Originally Posted by TrevorG3VLF

...in the UK all metal work in the house should be connected together to give no opportunity for a shock.

Yes, but more importantly, the metalwork should be connected to EARTH, otherwise, if a fault occurred, it could potentially become live and present a shock hazard!

Regards,

Dave.

[1100 man]

Originally Posted by TrevorG3VLF

My brother had trouble shearing his sheep. It turned out that the steel shed was at 80V relative to true earth due to a faulty supply transformer.

Which illustrates exactly why metal or substantially metal buildings must not be fed from PME supplies!! (Wiring regs) As the earth & neutral share a common conductor back to the supply transformer, a neutral fault will render all the 'earthed' metalwork live!

Paul,
Strangely enough, last night I was investigating exactly the condition you report on the radio I was working on. I could measure 140V between it's chassis and earth. More details can be found here

https://www.vintage-radio.net/forum/...26#post1068826

Nick

[D Cassidy]

Hello.
Certainly the chassis should not be at mains potential. You yourself have correctly diagnosed that C15 could be leaking or have a short circuit. If the Radio was mine I'd disconnect it and re run the test. Some leakage is often inevitable but to have such a high voltage on the chassis to me sounds like a fault. If removing C15 cures the problem I would fit a modern 0.1uf X type capacitor across the mains input leads to the transformer and not return it to chassis, if you do keep it original then use a Y rated capacitor. Using an earth to the earth connection will add an extra safety measure.

[Herald1360]

C15 may not need to be leaky for you to read a high voltage from chassis to earth- if you're using a high resistance meter such as a quality dvm.

Try earthing the chassis via a 100k resistor. If the measured voltage drops significantly C15 is probably OK. It wouldn't do any harm to replace it by a suitable Y-rated part, though.

The fadeout.... check dc conditions around the output valve carefully- something may be causing runaway there.

[D Cassidy]

Hello.
The practice of fitting a capacitor from one side of the mains to chassis on an AC set is far from ideal. It can carry mains borne interference often exacerbated if no earth is used. A wax capacitor in this position too can leak, so it really should be an automatic replacement.
On a course I did a few years ago the lecturer always advised using a filter cap across the mains (X type) and not to return a capacitor to chassis on an AC radio from the mains.
If keeping the cap from mains to chassis don't be tempted to increase its value, the cap in this position acts as a capacitive dropper so you could get a nasty shock with a larger value if you accidentally touch chassis and earth.
Regarding the 20 odd volt discrepancy due to your mains and the 220v input of the radio it may be an idea to calculate the required series resistance required on the mains feed to bring it down to spec, a resistor as close as possible to the calculation could be fitted in the set, I was thinking about the bolt in gold coloured ceramic types, I'm not sure of their make though. It would save over running your radio.

[Paul Mathew]

Thanks for the replies.
I disconnected the replacement cap, C15, (I used a .0047mF class X2 275VAC) which was the closest I had. This has resulted in the chassis to house ground dropping to 130VAC! More questions than answers!!
The saga continues.....

[D Cassidy]

Hello.
That could well be correct. As Herald says try a 100k between chassis and ground and if the voltage goes well down or even zero then its just normal leakage. One other item to check is the power switch if its chassis mounted is a possible leak in the insulating material. Also try reversing the mains lead or at least confirm you are switching the live side of the mains.

[ms660]

This might or might not help to explain what your meter is reading....If say the meter is 10meg impedance and your mains is 240 volts and the mains to chassis capacitor is 0.0047uF then that in effect is a potential divider of 677k and 10meg, the voltage at the junction will be approx. 224 volts, when you disconnect the capacitor the voltage will drop substantially because the only capacitance now is the capacitance due to the transformer and any other stray between the mains and chassis which means that Xc increases somewhat compared to Xc of a 0.0047uF, that's assuming the capacitor was connected to live side.

Lawrence.

[Paul Mathew]

Thanks gentlemen, that makes a whole heap of sense. Will make the tests suggested and revert.

[Paul Mathew]

Following useful advice, I started with the 100k resistor from chassis to house earth and the voltage dropped to 4.68V, so I fit a 3-core mains lead and earthed the chassis. This sorted out the high potential on the chassis.
I left C15 completely out of the picture.
As you can see in the attached pic, it was quite a challenge getting to all the caps, but I managed to replace C12 & C14 which I'd previously deemed too difficult to extract. Quite a bit of the wiring was perished beyond redemption. Some replaced and others I dressed with heat shrink.
Tackling the low level audio, I followed further advice to check the voltages around the 2A5 output tube. The voltage on the anode, pin2 was 260V and screen grid 2,was 274V. The specs here, http://www.r-type.org/pdfs/2a5.pdf,
indicates the voltages should be equal. Not sure how critical these values are?
I started poking around with my voltmeter and bridged R11, which resulted in a surprising improvement in the volume! I reckoned that the voltmeter, probably being around 10Meg, could be replaced with a fixed resistor of the same value. This proved successful, to a point, but another 10Meg resistor in parallel to R11 improved the audio further.
A bit of luck is always a good thing!
I will try and get R11 out of circuit to test for value, and see if it's out of spec, if I can.
Anyway, a good result, and thanks for all the input!

[Herald1360]

If R11 has gone high, it will seriously drop the anode supply volts to the first audio amp triode.

It's probably worth checking all the old composition resistors for value. Most can probably be tested in circuit if there's no parallel dc path between their ends.