Quad II ringing

[trobbins]

Originally Posted by Radio Wrangler

Because a lot of the fun things come into play well above the audible frequency range, this isn't something that can be done with an ordinary soundcard and a bit of software.

I use a circa 15 yr old USB soundcard, albeit a good one with 96kHz bandwidth, but nowadays that is pretty common with products like Focusrite 2i2. I even picked up a spare used EMU0404 on british ebay last year for 20 quid. Free software like REW (and ARTA) allows excellent spectrum analyser appreciation of noise and signal and harmonics with well over 100dB of dynamic range and full gain and phase measurement. Yes I augment that with additional spot measurements of gain and phase at milestone frequencies above 96kHz for gain peaks and dips, and n.45 phase shifts, and scope assessment of ringing frequency on squarewaves. That all helps fill in a picture like that presented in the AP plots in the 2011 Stereophile link in post #14, including the step signal level and frequency distortion plots in Fig 4 and 5, etc (noting that back in 2011 the soundcard hardware was available to get close to the AP system, but not the software).

[Radio Wrangler]

Originally Posted by GrimJosef

Long story short: this appears to me to be similar to the OP’s ‘better-behaved’ amp. It suggests that this sort of ringing is a characteristic of Quad IIs rather than some exotic feature exhibited only by the OP’s pair.

Thanks, G-J, That's about what I'd suspected. One of his amplifiers seems normal for the breed, his other has issues. It's good to have confirmation of a non-combatant unit.

It looks like there are, in a normal unit, two under-damped poles in the closed-loop response, one about an octave higher and better damped than the other. Where they come from would either mean doing a root-locus analysis backwards to extrapolate to the open loop pole positions, but better to just look at the bode plot open loop, and do some probing around. There will likely be signs of two candidate poles, though maybe a bit removed in frequency to the closed loop rings.

Looking at it open loop was a good suggestion as the outcome made it fairly certain that the resonance only appeared once the loop was closed.

I'm just hoping that there's something findable that doesn't come down to the output transformer. Are there any differences in the components of the 100v line output versions? If so is there some component value difference left over from when it was a 100v line unit?

Additionally, are there any differences between the period mainstream Quad II and the recreation version other than added C to compensate for the stray of the metal cases 'that' capacitors?

Just scratching around hoping for something other than the transformer. It couldn't be wrong phasing? no, that would surely be worse?

Swapping transformer from amp to amp might be needed unless someone can loan one already out to save disturbing the good comparison unit.

David

[trobbins]

Menno van der Veen (https://mennovanderveen.nl/images/on...download_2.pdf) appears to have made some measurements on the OPT, indicating a closed-loop resonance at 96kHz, and Fig.4.4.1 indicates a further resonance out near 200kHz. Unfortunately no measurement was made for open-loop (to indicate a 100kHz resonance) or for phase.

The feedback takeoff is from only 50% of secondary turns ratio, which may relate to why the initial 10-15us of the squarewave overshoot has ~100kHz resonance frequency that appears to be then forced to follow the ~56kHz resonance. The aberrant 56kHz resonance may well be due to an accidentally incorrect secondary connection.

Given the possible secondary resistance concern in post #28, it may be worthwhile making some turns-ratio voltage measurements (as per on Keith's info page).

[GrimJosef]

Originally Posted by Radio Wrangler

... Are there any differences in the components of the 100v line output versions? ...

I believe the only difference is that the 100V ones were fitted with an HT fuse. That might have some small resistance, but I'd be surprised if it made a difference to the HF audio performance. A quick check would be to replace it with a much higher current rated one, or just short it completely, assuming it's still in circuit.

Swapping transformer from amp to amp might be needed unless someone can loan one already out to save disturbing the good comparison unit.

With a little care (so as not to put any of the amp's weight on the KT66s) it's straightforward to run the amp upside down with the baseplate removed. If both amps were positioned this way, with the output transformers as close to one another as possible, then the wiring could be detached from the OPT in each amp. The wires in the worse-behaving amp could be extended a few inches and connected to the transformer terminals in the better-behaving amp. The worse-behaving amp could then be tested (obviously the better-behaving amp would not be powered) using the better-behaving amp's transformer. This might confirm or eliminate the transformers from the search for the difference. I probably don't need to say it, but it's critical that the wires are connected corrrectly to the test transformer. The original wiring wasn't very strongly colour-coded and it's not so hard to become confused about which one goes where.

I once had a hard-to-find problem in a Leak TL/12 Point One which turned out to be due to nonlinear behaviour of the feedback resistor from the output transformer secondary to the input valve's cathode. The consequence was that the amp performed as it should when the feedback was disconnected but badly when it was reconnected, which is what the OP here sees. In the Quad II the corresponding resistor would be the 470ohm R11. Since it ought to be quick, it might be worth temporarily replacing this resistor and seeing if that made any difference.

More fiddly to do might be to check the state of the solder joint between the leadout wire from the transformer and turret P on the terminal plate. I've very occasionally seen dry joints with this kind of turret. When everything's new and shiny the wire's physical contact to the inside of the turret is sufficient to make good electrical contact. But over the decades the surfaces corrode and, in the absence of proper solder 'wetting', a nonlinear contact can then form. That would have the same effect, obviously, as a defective R11.

Cheers,

GJ

[PJL]

What happened to the idea of running it without overall -ve feedback? R11 could be disconnected from the output transformer and connected to 0V. A scope could then locate the source of resonance.

[Radio Wrangler]

Originally Posted by PJL

What happened to the idea of running it without overall -ve feedback? R11 could be disconnected from the output transformer and connected to 0V. A scope could then locate the source of resonance.

That's in post 31. A big increase in gain but the ringing stopped. No ringing shows there is no resonance. The resonance(s) are an artefact of the gain/phase characteristic of the circuit along with the feedback acting.

David

[Richardgr]

Apart from the 100V variants there are also two types of audio OPT. The new ones are lettered from A to L (or something like that), the later ones are later letters of the alphabet. There is a note on Keith Snook's site. I have two units, one with the old and one with the new, but I don't detect a difference (but not listening too hard for one either).

Maybe you have a difference there?

Bottom of this page, a link to another document ...
Quad II info

[Valvepower]

Hello,

Just a thought.., If I remember righty the cathode of the two EF86’s are coupled together, so the feedback coming in to the first EF86 cathode will be negative whereas the feedback coming into the second EF86 cathode would be of the opposite polarity so the feedback coming into the second EF86 cathode is positive feedback.

Could this positive feedback into the second EF86 have a little to do with it. I could loosing the plot here, and barking up the wrong tree - so - I stand to be corrected.

I recon this feedback arrangement and the potential divider for the (inverter) second EF86 being returned to the bottom of the common cathode resistor and the non-earthy end of the 100Ω feedback resistor; along with the coupling between the second grids, all form part of the balancing Peter Walker speaks of in the ‘Red’ Quad brochure?

I’ll have to dig out one of my Quad II amps.

Terry.

[cathoderay57]

Thanks again to all who have taken time to consider this topic. Much appreciated GJ; the fact that you found a closely similar response on one of your Quad IIs gives me reassurance and optimism. I'll press on, and try replacement feedback resistors and solder joints first as it's easy. I might try the gain- frequency plot over the weekend unless I lose the will to live in which case I might try swapping the transformer connections between the 2 amps, as suggested. Incidentally, the 100v line transformer has a separate isolated secondary winding for the neg feedback source and just 2 terminals for the 100v line output I think. AFAIK the primaries are the same. It is somewhere in the loft so if I get chance I'll dig it out and post an image. J.

[Valvepower]

Going at a tangent regard Quad II’s…

Like many I have a soft spot for the Quad II… It was 50 years ago this year, when I got my first Quad II/22 system at the age of thirteen. This system had been removed from a house just down the road by the new owner and was originally heading for a jumble sale my Mum was helping to run. When the chap down the road spoke to my Mum about getting it to the jumble sale he said “I recon your Terry would like to play with this” – well, do bears use the toilet facilities in the woods! – Anyway, my Mum and the chap agreed a nominal jumble sale price of a few quid (yikes!) for the good cause the jumble was being run. A few days later the chap dropped it off whilst I was at school, and when I got home from school that afternoon there was a complete Quad II system with both AM and FM tuners waiting under the stairs… Well..., as regards Audio I’ve never looked back…

The Quad II has pretty much been a constant those 50 years!

Terry

[Ed_Dinning]

Hi Folks, I believe there was a "hidden component" in the potting of the output transformer. From distant memory this was a low value wire wound resistor in a paxolin former. In the cathode circuit if I remember correctly.
Was this also fitted to the 100v line transformer?

Ed

[cathoderay57]

Morning all, I followed GJ's advice and slaved in the OPT from the "good" Quad (pictured on the left of the image) onto the wires from the "bad" Quad (on the right). The output traces show that the "bad" Quad now performs virtually identically to the "good" Quad. I know the interconnecting wires are a bird's nest of stray capacitance which conceivably could be suppressing the HF oscillation but, on balance, I think the evidence is pointing strongly enough towards the OPT on the "bad" Quad being the root cause of instability. I don't have £100+ going spare to buy a replacement OPT so I will experiment with Zobel networks and maybe a Bode plot to try to stabilise the "bad" Quad a bit better. Gabriel has offered to help over Zoom, which is much appreciated. Cheers for now, J.

[GrimJosef]

Interesting.

As far as I know the only repair that can be attempted straightforwardly on the Quad II output transformer is the re-making of the solder joints between the wires from the transformer itself and the turrets on its terminal board. It's not clear to me though that a poor joint there could account for the extended ringing that was seen in the OP's 'bad' amp.

I'm also not sure how wide a range of different ringing performances is 'normal' for Quad II output transformers. For the sake of having one more data point I hooked the partner of my first Quad II to my test rig and ran a ringing test on it. The trace is attached. This amp is serial number 66960.

Apart from the initial spike being a bit bigger - the peak-to-first-trough voltage difference is now 690mV - the trace looks very similar to my earlier one with amp serial no 66913. So the behaviour of the OP's 'bad' amp, serial no 71237, does look more unusual.

Unfortunately the only square-wave source I had conveniently to hand was the Probe Compensation output on the rear of my scope, and this is fixed at a frequency of 1.5kHz, so the flat top only runs for 330us or so and I can't tell how much longer the '24kHz' ringing in my amps runs for. It does look a lot less pronounced than in the OP's 'bad' trace though.

Cheers,

GJ

[PJL]

The slow decay of the bad amps resonance is not what I would expect from feedback instability. The open-loop test was also not valid as the reduction in input amplitude would have dramatically increased the rise time at the output transformer.

It might be worth seeing if the relative resonance amplitude changes with input signal amplitude.

[Valvepower]

Hello,

If it helps, I’ve posted a couple traces at 1KHz at 16Ω from Serial numbers 56316 and 90427.

Out of curiosity what is the history of the suspected rouge output transformer?

Terry

[GrimJosef]

I suspect the details of the ringing (amplitudes of the different frequency components) will depend on the details of the square wave which excites it and, in particular, on the risetime. That risetime will need to be pretty short to get the maximum degree of ringing.

Since my square wave source (Probe Compensation signal from the back of my scope) wasn't adjustable I used a simple resistive divider to get it down to the amplitude I wanted. This raised the source impedance to the point where the capacitance of the cable between the divider and the amp could easily compromise the risetime. In the event I used a very short length of quite high impedance cable to keep the capacitance down. I measured a risetime (10%-90%) of less than 0.7us which was short enough to excite even the ~100kHz component of the ringing.

[PJL]

The concern with the slow decay is the amp appears very close to oscillation, so close that driving a different load would seem likely to push it over the edge. My suggestion is the mode of oscillation is not reinforcement from -ve feedback but is contained within the output stage and transformer and it's amplitude is hence limited.

[Radio Wrangler]

The rise time of the waveform will affect the amount of excitation of the ring.

The open loop tests giving no ring make it clear that it is a closed loop artefact. It is also at a frequency where the loop gain is rolling well, down, but not quite down to unity.

This is an area where strays are significant, bordering on dominant. Only a small change in gain or phase can make large differences like between sustained oscillation versus having some damping.... or between fairly different damping factors.

These are not precision controlled parameters, components bought with specified tolerances from reputable manufacturers. So we should expect a fair amount of variance in ringing behaviour.

David

[Valvepower]

Hello,

Here we go the signal being fed into the amplifiers. It’s taken with the signal generator connected to the amplifier with both amplifiers running.

Terry

[Valvepower]

Originally Posted by cathoderay57

If the output transformer(s) are deemed to be responsible for this symptom then what physically is likely to have gone wrong with them? J

I’ve occasionally thought about this.

What follows is not based on nothing other than my own meandering experience working with guitar amplifiers in my early days as rookie engineer. I remember working on a guitar amplifier that just didn’t sound right. I checked the power output and frequency response etc., as best I could with rather limited and rudimentary test gear I was using at the time, and it measured OK.

After checking all and sundry, I replaced the output transformer, and then the amplifier sounded OK. From memory the amplifier measured pretty much the same as it did before I replaced the output transformer.

The transformer was still functioning but there was something causing it to sound ‘bad’. Maybe, if I had better test gear, I may well have seen/found something harmonic distortion wise, however, let’s say the bloke I was working for at the time really didn’t believe in spending a lot on test gear – mind you that was well in the past – like – 46 years in the past.

I did latterly find out the amplifier had been used with incorrect loads, and I often ponder, if an output transformer is subject to enough stress with incorrect loads, but NOT to the point where it fails and breaks down causing shorting or open circuit windings, but degrading the intrinsic construction and insulation etc., of the transformer. It still functions but with compromised parameters and such like etc. One could say its ‘cooked’ but still working, or should I say, “three wheels on my output transformer but it’s still rolling along”

As I said it’s no more that my meandering experience of an incident concerning an output transformer in the past and please see it as such.

Terry