Quad II test run.

[sparkymike]

I have got as far as replacing the two ef86 and the rectifier and one output valve.
If I remove one of the ef 86 valves,I get a screeching sound from the speaker. Is this normal ?
Mike.

[cathoderay57]

Hi Mike, I would recommend you stop running the amp with one or more valves removed. It really isn't going to help because the biasing of the valves is inter-dependent and you will be running the circuit in an off-balance condition. When all of the valves are in, do you still get a hum? Is there anything connected to the audio input? Cheers, Jerry

[GrimJosef]

I'd agree with Jerry about not running with any of the valves out (OK, it's safe to run without the rectifier). It's a bit like saying 'If I leave one of the brake pads out of each of my car's calipers I find I can't stop it nearly so well and there is a loud screeching noise each time I try'.

When you said you put the rectifier in and measured 457V DC on the HT rail my heart was in my mouth. Because the whole supply is off load the voltage will be way more than normal and there is every risk of damaging the HV capacitors. It seems like you got away with it though.

I'm a bit bemused by your mention of valveholder skirts. The Quad II as produced by Quad didn't use skirted valveholders. I guess you're talking about the one you built out of parts (post #6). The Quad II circuit isn't very easy to isolate into sections, electronically. There's a bit of a tendency for the operation of every part to affect the operation of every other part. This makes it a triumph of economy (the whole amp has just 6 capacitors and 12 resistors) but unpredictable if you try to run it while disassembled.

Cheers,

GJ

[sparkymike]

Hi GJ,
I used wrong terminology re. "skirts" What I was referring to was the pressed on plates that support the plastic valve bases.
Re. running without certain valves, I understood that any push -pull output amplifier could run with either output valve removed for some time without ill effects ?
Re. the hum, would the source of this hum be visible on my scope if I work through from ef86 through the amp ? What pattern would I expect to see on the scope ? A double wave ?
Mike.

[GrimJosef]

Not true about push-pull generally being safe with only one valve in place I'm afraid. The Quad II uses a shared cathode resistor for both KT66s and it's this which largely controls the current through the valves. If you take one valve out the immediate effect will be to halve the current. The voltage will then drop across the resistor and this reduced bias will cause the remaining valve to turn on harder i.e. pass more current. A new balance point will be reached but the valve will be dissipating a lot more power inside it than before, and its half of the (precious) output transformer primary will be passing the same increased current. The valve and the transformer may well last, but it would be sad indeed if either of them didn't.

As far as the hum goes, that will be a great deal worse with only one output valve in place. The high voltage supply to the anodes of these comes from a point before the smoothing choke. So there is a good deal of 100Hz ripple on the HT there. The amp's design relies on the currents through the two KT66s being well enough matched that the current ripple, which results from the voltage ripple, is cancelled in the output transformer given that the two halves of the primary are counter-wound. If you have excessive 100Hz hum (check the frequency on your scope - you can just look at the signal across the speaker or, better, a resistive dummy load) then it will be coming from somewhere in the power supply. Mismatched KT66s could be one source.

If your hum is 50Hz rather than 100Hz then that will probably be coming from the mains wiring, or a grounding issue, or the heater circuitry including the possibility of heater-cathode leakage in one of the EF86s.

But the first thing I'd do is to measure how much hum there is. With the amp's input shorted to ground there should be around 1mV RMS (say 3mV peak-to-peak) or less across an 8ohm load, assuming the output transformer secondary is jumpered for 8ohms. Much more than that and the amp is running out of spec.

Cheers,

GJ

[sparkymike]

Hi CJ, I will try the last tip. Do I short out pin 1 to 2 or 3 on the Jones plug.
The output trans is configured for 15 ohm at the moment, so will find suitable high wattage load resistors. Then do the measuring.
Mike.

[sparkymike]

Hi CJ,
I shorted out pins 1 & 2 on the Jones plug and have 15 ohm wire wound resistor on speaker terminals. I am getting 2 volts on speaker terminals. That voltage was [possibly climbing very slowly but I thought best to switch off at that point.
both mentally and physically !!
Mike.

[Herald1360]

Wasn't CJ the "boss" at Reginald Perrin's employer?


GJ is (one of) our forum valve hifi gurus


Whatever, 2 volts of anything on the speaker output with shorted input is a definite problem! Starting points would have to be the HT filter caps and choke but output stage nasties and/or poor grounding somewhere could also figure.....


Have (more) fun!

[GrimJosef]

It would be interesting to know what the frequency was Mike. 2V RMS into 15ohms is roughly a quarter of a watt. The sensitivity of the amp should be about 1.4V in for 15W out (they say 1.3V iin the spec, but it rarely is in my experiece). So 2V out would correspond to a bit less than 200mV equivalent at the input terminal. That's a lot.

One somewhat scary scenario is that the amp has in fact become a power oscillator, perhaps at an ultrasonic frequency. The ultrasonics won't get through the output transformer very efficiently. And a multimeter might not measure up to very high frequencies. So if you measure at the speaker terminals the signal level might look very annoying but not frightening. However much larger amplitude signals may be present back in the valve stages.

I'd try to get the signal on the scope and check the frequency. If you see 2V RMS at 50Hz or 100Hz then there is a problem but it probably won't damage anything to run with it. If you see signals at much much higher frequencies, say tens or hundreds of kilohertz or even megahertz then I'd switch off quickly and try to think out what might be wrong. We'll help if we can.

EDIT While you're about it check the DC voltage across the KT66 cathode resistor, R12. The resistor should be 180ohms and the voltage 27V or so.

Cheers,

GJ

[sparkymike]

Hi GJ,
R 12 is 181 ohms and voltage is 20 volts. Weird thing, the hum did not kick in that time ? I wonder if I have disturbed something while measuring the resistor ohms ?
Mike.

[sparkymike]

Senile moment !! forgot I still have the load resistor connected !!
Mike.

[sparkymike]

Now have both output valves in and speaker connected and hum still there. Voltage on R12 is now 30 volts.
Mike.

[GrimJosef]

30V is a bit high and could indicate some continuous signal in the output valves.

Can you measure the DC volts between the output transformer terminals V and U, and then between V and W. (Careful to stay away from X, Y and Z - they're very high voltage.) You should get around 1.15-1.35V in each case. They are the cathode windings for the KT66s so they indicate the DC current flowing in each valve. If one is much greater than the other then that indicates the valves aren't well-matched. If either is more than about 1.6V then that indicates the valve really is passing too much current.

The frequency of the output signal would still be very useful to know.

Cheers,

GJ

[sparkymike]

I have dragged out my Solartron dual trace scope. (They are weighty beasts.)
Now basic question GJ, The Time/CM control ranges from .5 usec/cm to 200msec/cm. What would be approximate setting. (ie. is 1 msec 1000hz?)
Do I disconnect the input shorting link for that test ?
Mike.

[GrimJosef]

I'd use whatever input arrangement you had when you were measuring 2V AC across the load.

Start with 5ms/cm. One 50Hz cycle will be 4cm long, one 100Hz cycle will be 2cm long. Get it triggering stably then measure the cycle length. Does the waveform look like 2V RMS - roughly 6V peak to peak ?

If all you see is a broad band of light then the cycles may be so short (frequency so high) that you can't make them out. In that case keep speeding the timebase up (smaller and smaller secs/cm) until you can make the cycles out. How many msec or usec is one complete cycle ?

Cheers,

GJ

[sparkymike]

Just did a quick check and there is a band of light which seems to be a composite wave form. It is a bit hard to see at the moment as sun is streaming through the conservatory windows where I have bits set up, so will resume after dark.
Mike

[GrimJosef]

Sounds like it could have become an ultrasonic power oscillator. Prompted by the latest posts in this thread https://www.vintage-radio.net/forum/...d.php?t=164995 is there any chance that the cathode transformer tappings have got swapped over relative to the anode ones ? Or could the speaker secondary be wired the wrong way round so you've got positive feedback instead of negative ? It would be worth tracing the actual wiring from the output transformer through to the valves and making sure that everything is exactly as the circuit says it should be.

EDIT: The same problem could arise if the EF86s are driving the 'wrong' KT66s. Using Quad's original circuit numbering, V1 needs to drive V3 and V2 needs to drive V4.

Cheers,

GJ

[sparkymike]

Now it is dark in the conservatory/ aka laboratory I have better view of the CRT.
I jumped the gun slightly and have probe on output valve grid 1.(pin 5) and show the trace I am getting which I assume is very high frequency.
Mike.

[Radio Wrangler]

Yup, you have an expensive ultrasonic oscillator there. Something is badly wrong. The senses of the feedback arrangements and the inversion sense of the amop from the feedback insertion point to the output are the first things to check.

This shows why a lot of experienced engineers treat their scope as their primary bit of test gear, not their multimeter. It tells you a lot more per stab of a probe.

David

[sparkymike]

Would reversed taps on one of the output transformer windings do this ?
Mike.