Juke amp frolics

[Alex R]

I have this issue with an old jukebox amp - no schem available. (It's from a v rare Chantal Meteor juke, amp built by Calne Electronics, a Type A 12/58). It uses a magnetic cartridge. Treble response is poor when listened to, and shows up when scoped as a steady downward line between 500Hz and 5KHz or so. By the time you get up to 5KHz the signal is about 20% of the bass amplitude. Same through the amp at each decoupling point.

I've been through the amp recapping etc. I should say this amp had had a lot of piecemeal repair down the years, and may not be quite correct circuit-wise. The input stage is an ef86, pretty much similar to the AC15 first stage. The odd thing is, my sig gen output shows a similar downward frequency curve when scanned up the audio range, as soon as I connect it to the amp. Disconnect it and the curve disappears, amplitude constant whatever the frequency as you'd expect. So it doesn't seem the amp has a problem - something to do with the input, maybe impedance? I've tried it with and without a cap on the input, and with various ground ref and grid stopper resistors, with no change. I wondered whether the input socket arrangement had some capacitance or something, but when I removed it from the circuit and connected direct to the ef86 base, same issue showed up.

I really feel like I'm missing something. Any idea what?

[Chuck H]

Back in the day most phonographs included circuits to accomplish the RIAA equalization standard. It sounds like this could be what you're seeing. Look for some kind of low pass filter in circuit behind the test point. Perhaps somewhere along it's various repairs a mistake was made regarding compensation for such a circuit.?

[Mick Bailey]

It does seem that you have an impedance problem. Does the signal generator response fall if the EF86 is removed?

[Mick Bailey]

Good point about the RIAA equalisation. I'm wondering though how this would affect the signal gen output.

[Chuck H]

Testing for loading by removing the input tube is a good call. If it's a circuit somehow connected to the input doing the damage then the signal from the genny will stil load. If not it may be a bad tube. I thought about the possibility someone removed a matching transformer from the input circuit but I would think the input would be a terrible noisy place to use one because of the low cartridge signal requiring a lot of amplification.

[Alex R]

Really helpful, thank you! It's off the bench now, hopefully will get it back up tomorrow and try the valve removal idea, just didn't think of that. If it's still there with the valve out then there are v few components could be causing it. Really should have thought have that!

Interesting idea re the RAAA. But it doesn't sound good (have been back and to to the customer's house trying the damn thing!). And the ramp-down happens even at the input. So that pulling the valve will track it down.

I shall report back in a day or three.

PS this forum remains the best resource on the net for guitar amp repair. Hopefully we can keep it quiet from the types who frequent the fb amp repair pages: "it's your caps', 'must be dry joints' etc etc

[Helmholtz]

What is the output impedance of your signal generator?
If it's high the effect could be explained by a loading effect due to a low and frequency dependent input impedance.
As any magnetic cartridge needs RIAA equalisation, the input amplifier might be wired as an active filter using f-dependent NFB connected to the input grid.
This would result in low and f-dependent input impedance.

Best try to draw up the schematic of the tube wiring.

[Mick Bailey]

If the equalisation was the only thing affecting the bench testing the amp should work correctly with the cartridge input but it doesn't, which makes me suspect that there's a fault rather than a design intention.

I've used a 2xEL84 jukebox amp with an EF86 preamp for guitar and it made a really good amp with the equalisation left in place and just with its tone and volume controls. It still had plenty enough top end with a Strat.

[Helmholtz]

But it doesn't sound good (have been back and to to the customer's house trying the damn thing!).

How did you test?
Using a vinyl record or a guitar as source?

Is this a restoration or a guitar amp conversion project?

[Chuck H]

How did you test?
Using a vinyl record or a guitar as source?

Is this a restoration or a guitar amp conversion project?

This^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

[Alex R]

Okeydoke, thanks to all for the input. When I pull the ef86 as suggested, the condition disappears and the sig gen output is linear at all audio frequencies.

Inputs are: in the jukebox, a magnetic cartridge, vinyl records. In the workshop, a Thurlby Thandar TG215 sig gen with 50 ohm or 600 ohm outputs, using the 600 ohm.

I am restoring a jukebox amp to use in the jukebox, nothing to do with guitar. The fault is weak treble, this is observable in the jukebox when we play it, also on the bench in the output and at every useable node in the signal path. It's also there on the input, which has a 0.1 cap in line, a 100k earth ref and a 68k grid stopper. I did draw out the schem, hopefully accurate, I'll attach it. There's no official schem available, and there has been a lot of work in here, not all of it very professional, so the schem is probably reflecting some haphazard fixes. The jukebox has been like this for a while. I have recapped it and all the valves are good.

It is possible I am now looking at artefacts of the test setup, but I don't think so as the scope shows something close to what we hear4 when it's in the box (been in and out three times now as I try to get the treble back. Advice very welcome.

[Helmholtz]

The schematic gives no clue as to excessive f-dependent loading, at least as long as the tube is not shorting.

1) Did you disconnect the cartridge when using the signal generator?
2) Are you sure it's a magnetic cartridge? What DCR does it read?
3) Did you test the cartridge by connecting to an other amp?

[Mick Bailey]

The nominal impedance for a magnetic cartridge is 47k ohm, though there is a fair bit of latitude and much variation in preamp specs. A 600 ohm signal gen should work fine with the circuit and not suffer from the loading you're seeing. The reason I mentioned using a jukebox amp for guitar is to illustrate that even with the RIAA EQ left in place there's enough treble and bench testing should give acceptable results when listening to a regular signal source or even putting a guitar through it.

Do you have a replacement EF86 to try? The fact that the signal gen loading disappears with the tube removed is puzzling (though it eliminates the input circuit) and I'd want to at least eliminate the tube as well.

[35L6]

I think that .01 .033 divider is your RIAA , pimative as it may be . Try lifting the .01 and see what happens . Clean the stylus ?

[Helmholtz]

I think that .01 .033 divider...

It's not a divider but a low pass filter..