Repairing a 1961 Jukebox Amplifier

You may remember that in the last issue I made a comment about the variety of items that I'm asked to repair. Not long after I wrote that a friend rang asking me if I'd look at the amplifier from a jukebox! I said bring it over, and I'd have a look.

The complaint was that the sound was weak and, my friend added, the valve heaters don't seem to be very bright. I soon had the amplifier chassis in front of me on the bench. It was from an AMI Continental 1, made in 1961.

The auxilliary chassis

As well as the main amplifier chassis, there was a smaller auxilliary chassis containing a power transformer and switches, power sockets, fuses etc. This auxilliary chassis had mounted upon it several USA type two pin sockets into one of which the amplifier was plugged. The other sockets were for the turntable, changing mechanism and various display lights.

At first I thought that the transformer on this auxilliary chassis was a 240 to 120 volt type for powering the rest of the jukebox. But I realised that it was too small to do this. I found that the USA-type mains sockets were all 240 volt. So the amplifier was 240 volt as well, despite the type of power plug used.

It turned out that the auxilliary chassis contained a 28 volt (no load) DC power supply for some purpose unknown to me.

Close inspection revealed an interesting thing: both the amplifier and the auxiliary chassis were made in Australia! The amplifier chassis looked the same as the USA ones that I have found on the net - but all the original components were Australian made: Ducon capacitors, IRC resistors etc, typical of the era. This poses the question: did AMI have a factory in Australia?

The amplifier

The amplifier itself has nine valves, with a pair of 6973 beam tetrode valves in push pull in the output stage. The rest of the amplifying valves were dual triodes: 12AX7's and 12AU7's with a 7025 in the first stage. The power rectifier is a 5U4GB. The 7025 is like the 12AX7, but is a low hum, low microphony version.

My first task was to test the valves. I rather hoped that the 6973's were good, as I didn't have any replacements to hand. Two of the 12AU7's were weak, no problem; I had plenty of those but the 6973's measured only 77% and 42% Gm (mutual conductance).

It turns out that 6973's are in high demand, and fetch a high price. There is a currently made version, the 6973EH, which is readily available but a bit of probing on the internet revealed that they are not well regarded. I started to wonder if there was another valve that I could substitute.

Then I remembered a conversation I had with Darryl Kasch, must be 15 years ago at the time of writing this, about a valve that he often swapped for service data from the UK. This valve was the 6CZ5 and Darryl happened to mention that they used them in jukeboxes!

A bit of research revealed that yes indeed, the electrical characteristics of both types were identical. The 6CZ5 was used here in Australia in televisions as the frame output amplifier, mainly by AWA. I did have three 6CZ5's, in uncertain condition. I also recalled reading in a Radio, TV & Hobbies magazine, in relation to their 1961 23 inch television project I think that the 6CZ5 had been replaced by an improved type, the 6EM5, which could be used in place of it. I had plenty of 6EM5's so it would be very convenient if they would serve in place of the 6973's.

It turns out that the 6973 and the 6CZ5 have very similar characteristics to the good old 6V6 and the 6AQ5. I couldn't find audio amplifier data for the 6EM5 but I knew it must be close to the 6CZ5 to be compatible in vertical amplifier circuits. The 6EM5 has a higher heater current than the others but I though this wouldn't be a problem as the power transformer seemed of generous size. Another difference between the 6EM5/6CZ5 and the 6973 was that the control grid in the latter was brought out to two pins. I had to change over only one connection under the chassis to make it suitable for both the 6973 and the 6EM5/6CZ5 pin connections.

I did also consider the 6BW6 as this also has the same characteristics but the pin connections were very different. The way I had modified the connections a future owner could use 6973's if desired with no further changes.

Repairs

Back to the repairs, I found that many components under the chassis had already been replaced, including all the electrolytic and paper capacitors, and many of the resistors. I found a couple more resistors gone high and replaced them.

I could not fault the two Ducon chassis mount multi-section electrolytic capacitors so I left them in circuit.

Power up

It was time to try the amplifier. I selected two 6EM5's that tested close to 100% and plugged them in. I attached a speaker and switched on. All seemed well but then I noticed a blue glow in the right 6EM5. Before I could turn off the glow brightened, the anode glowed and fireworks flashed inside the envelope. The valve expired with a rude blurt from the speaker as I switched off. I checked the bias supply, this amplifier uses fixed bias, but I found no fault.

I concluded that the valve was gassy to start with so I tried another but the anode in the new valve started to glow red. As this happened the bias voltage dropped away. Grid emission I concluded. I fetched a fourth 6EM5 and all seemed well.

I should note that the output valves are driven pretty hard in this amplifier, with 350 volts HT. Also the grid resistors are 470kΩ, much higher than usual for fixed bias. The 6973's are specifically designed for this sort of service so I did expect to have to go through a few 6EM5's before I found some that stood up.

Soak test

I hooked the amplifier up on my test bench with some good music and a good speaker for a long soak test. I also monitored the bias voltage. Over the next week or so the amplifier played for about 20 hours without a hitch. It didn't sound too bad either.

I was puzzled about the owner's original comment that the valve heaters didn't seem very bright. Did he think that the amplifier was 120 volt, seeing it had a USA type plug? I asked him and he said that no, he had tried it on 230 volts. The apparent brightness of valve heaters depends greatly on the ambient light level, so I had to conclude that he had observed it under bright conditions. Also, the 5U4GB being directly heated, doesn't produce much light.

He was very pleased with the result though - and impressed with the sound.