Juke amp frolics

[The Dude]

Agree, but still possibly the cause of the issue.

[Helmholtz]

Yes, lifting the 0.01 cap will increase treble response.

[Chuck H]

Yes, lifting the 0.01 cap will increase treble response.

Not quoting as an argument to your point, but rather supporting it. If the unit was originally designed as per parameters in the day I might question what the speaker compliment is in the system. That would have been considered to balance frequency response when it was built but may not be original. And...

Agree with Mick that the EF86 is still suspect even though it's reported to be good. Like Enzo said, a tube tester can tell you if a tube is bad but not always that it's good.

[Mick Bailey]

Another possibility for treble loss is the capacitor that's in parallel with the EF86 plate resistor. Double-check that the capacitance is correct.

[eschertron]

I'm wondering why there's a 220p cap from plate to HT (I've seen that before, not so much of a puzzle) AND the .01u from plate to ground. Kinda makes the plate resistor bypass cap insignificant, dunnit? Can the .01u cap be verified as true to the original design?

[Alex R]

Thanks for the input. Here's an update. But first, your questions answered.

No I haven't tested the cartridge. Next time I go over there I'll take a little amp to try it. And a meter to check DCR. But the rolloff happens with the sig gen input just the same.

I'm pretty sure it is magnetic as there is so much gain in this amp. There's not much info on the Chantal Meteor but it is aming high quality wise and so a magnetic cartridge wouldn't be unexpected. But yeah the fault seems independent of the cartridge.

No there is no way of being totally sure of what the original circuit was, as quite a few new components have been fitted over time, but the basic config is pretty clear. I've tried it with and without the input cap, and with and without the plate resistor bypass cap, and with different values for the screen bypass cap, with no effect on this marked slope-wise reduction of treble. In fact I took all the first stage components off the ef86 and rewired the stage with the first stage AC15 circuit, as I'm more familiar with how that behaves and this seems similar to it in that it is a very high gain ef86 circuit. Again, no change to the rolloff.

OK I think that's all the suggestions answered. If I find more later I'll answer them too.

Right, now to the current sitch. As I wrote the issue is observable directly on the input, and this slope-off of treble occurs in a similar way all the way through the amp. But it disappears from the input and goes back to a fully linear response when I pull the ef86 out, scope probe still connected to the input, the other side of the input cap from the grid. I've tried several ef86, all Mullards, and even a Telefunken ef806s (yes, from my pension fund valve collection). No difference. I've tried lifting everything up from the input socket, in case there was some internal capacitance. Nope, no change.

Here's an oddity. With the valve out, the now-linear amplitude of the signal on the input is like that at the top of the treble range with the valve in, ie about 20% of the bass signal. The effect of the valve being plugged in is thus not so much to roll off the treble as to amplify the bass! This suggests some kind of positive feedback. Well I'm none the wiser as to the cause of that.

I would have decided that I am looking at testing artefacts somehow if it wasn't for the very clear treble rolloff (or bass boost, as it appears) effect on the input. Your input is welcomed!

[Mick Bailey]

Bass boost/treble cut is a feature of an RIAA equalised amplifier in order to recover the bass which is intentionally lost during the recording process. When the vinyl master is cut, the cutter head excursion on bass notes could cause the groove walls to collapse so an eq curve is applied to de-emphasise the bass. Also the treble is boosted. The playback stylus has a much easier time in reproducing the resulting track.

The playback amp should have the inverted curve. The way to thoroughly test the amp is to run your signal generator through an RIAA filter. Most of the filter designs are for playback - you need one to generate the origination curve. In theory this should then scope out to a level response.

https://www.lathetrolls.com/viewtopic.php?t=5848

[Helmholtz]

As long as we don't know if the cartridge (and its stylus) is good, we can't be sure there's an amp fault.
Maybe test with an external turntable.

My other question was if you disconnected the cartridge when testing with the signal generator.

With the valve out, the now-linear amplitude of the signal on the input is like that at the top of the treble range with the valve in, ie about 20% of the bass signal. The effect of the valve being plugged in is thus not so much to roll off the treble as to amplify the bass! This suggests some kind of positive feedback.

Are you saying that a bass signal at the input increases when you pull the tube??
Are you scoping before or after the 68k ?

Anyway, I'm sure something is missing from your schematic, like some additional component(s) connecting to the input grid or before.

[g1]

Here's an oddity. With the valve out, the now-linear amplitude of the signal on the input is like that at the top of the treble range with the valve in, ie about 20% of the bass signal.

Can you clarify this, and where you are checking? To me it does not sound linear. Which would make me want to suggest further verification of the generator.

[Helmholtz]

Maybe too large an input signal?
With grid conduction the input impedance will drop and become frequency dependent.
Input signal should be in the 100mV range.

[Mick Bailey]

I think the input signal needs to be around 10mV or less to simulate a magnetic cartridge. The actual output is about half of this or less.