Quad II - PT autopsy begins...

[dougietamson]

Some more unwinding photos and numbers

HT wind is 875t-ct-875t of 0.21mmOD wire

Primary mains is 630 turns of 0.5mmOD wire for the 0-240v
The 220v tap is at 47 and the 200v tap 106t

Waxed pressed paper between each layer and extra layers between the pri and sec. Not sure what the black strips are made of under and on top of the copper strips the leads soldered to. The margins 4mm.

[kalee20]

The winding in these photos, in sharp contrast to the heater windings, looks neat and workmanlike.

However, the gauge of wire for the HT winding seems rather light! At well over 100mA of HT demand (I don't know offhand what the Quad amplifier draws so this is a guess), 0.21mm seems rather inadequate. I have heard that Quad mains transformers are pone to burning-out, is this the reason?

Originally Posted by dougietamson

Not sure what the black strips are made of under and on top of the copper strips the leads soldered to. The margins 4mm.

Might be Elephantide / Presspahn?

[dougietamson]

Originally Posted by kalee20

The winding in these photos, in sharp contrast to the heater windings, looks neat and workmanlike.

However, the gauge of wire for the HT winding seems rather light! At well over 100mA of HT demand (I don't know offhand what the Quad amplifier draws so this is a guess), 0.21mm seems rather inadequate. I have heard that Quad mains transformers are pone to burning-out, is this the reason?
Might be Elephantide / Presspahn?

Now I think about it I may have made an error measuring it, I'll check it again.

[dougietamson]

I made a mistake with the HT wire OD, it's 0.25mm.

[Radio Wrangler]

That's ten-thou wire so it's being pushed hard at 236mA which amounts to 3000A/sq inch.

A third of that 1000A/sq inch was an old rule of thumb, but became considered to be a bit too kind! HP used design figures towards the gentle end beacuse our stuff was intended to keep its reputation of taking a direct thermonnuclear hit and still come up smiling. Quad have always tended to play things rather close to the limits and many rimes over them.

That wire should work, but there isn't much comfort margin.

Thiss relates purely to current rating of the wire. As far as heat generation within the winding stackm that's another matter. A thicker wire gives less resistance and donates less heat to the stack. But the thicker wire takes more of the bobbin winding profile area and robs other windings of cross-sectonal area, creating more heat in them.

There is a degree of artistry in planning a good compromise for a transformer, and it'a already been said that the Quad II mains transformer is somewhat miserly in size, so good compromising is needed to get decent reliability. My transformer designing was all far to the carefully conservative extreme, well away from the knife-edge compromise.

I'd suggest you take advice from Kalee, Ed Dinning and Mike Barker, they've all had more experience at the sharp end of transformers close to the limits.

That Quad IIs keep cropping up with OC HT windings tells us just which winding is closest to its limits. Having a fuse in EACH end of the HT winding is a jolly good idea.

Having a transformer re-wound and re-impregnated as a one-off job is not cheap. Quad, I think, may still have finished transformers in stock for their reprised Quad II, and as they were bought in quantity, the price of a spare from Quad might be worth checking.

David

[Craig Sawyers]

David - as I mentioned the heater and HT windings also have to supply the preamp, am tuner and fm tuner - a further 13 valves. The Quad II transformer is pushed really hard if peripherals are added.

Craig

[dougietamson]

More info, the core area is 35x52mm = 1820m2.

I'm tweeking the coil winder for the auto-traverse with the old wire.

I'm off to the mens shed now to make a wood block support for the bobbin.

Photo attached of the winder.

[kalee20]

Originally Posted by Radio Wrangler

That's ten-thou wire so it's being pushed hard at 236mA which amounts to 3000A/sq inch.

I don't know where you got 236mA David, but yes that would worry me hugely!

I based my comment on the Quad II being a 15W amplifier, having a GZ32 rectifier (max current 125mA) so guessing the HT current is close to the max rating. And since the rectifier is operating into a capacitor-input filter, RMS current is about 1.6x the DC output current giving 160mA... each half-secondary passes current half the time so each section sees 160mA / root-2 = 113mA. And that's a lot for a wire of 0.21mm OD (even when revised to 0.25mm by the OP). I don't have my wire tables to hand, but assuming that 0.25mm OD wire has conductor size 0.22mm, the current density is 2.97 A/sq mm which is a bit high.

For minimum total loss, the current density in each winding should be roughly the same. Any reduction in losses that you get by using a bigger gauge for any winding is outweighed by the increase in losses in other windings because there's less space for them. However, practical considerations also apply - and luckily the minimum is a shallow minimum so departures from this ideal aren't necessarily a showstopper.

Quick calculations indicate, 630 turns on a 35 x 52 sq mm core (so allowing 95% fill factor that's 1729 sq mm of iron) and hit by 240V 50Hz, gives a flux density of 0.99T.

That does seem very low! Maybe Quad wanted to keep the hum from the transformer itself super-inaudible, likewise the radiated field... looking critically at the design, there would seem scope to reduce turns by 5% and using the space saved to increase wire gauge. Both the reduction in turns, and the increase in diameter, of course reduce resistance, so significantly reducing copper losses and hence operating temperature.

Down side is that this is quite a recalculation exercise, hopefully getting a winding right-first-time, but in practice I have found that more often than not I do need to strip and rewind at least one winding for any new design.

I can't help feeling that Quad got their design finished, it worked (and well!) so rather than faff about with final optimisation they moved on to designing accessories, preamps, radio feeders, for the Quad II. So rather than treating the design as sacrosanct, there are improvements to be made - but they do suck time.

I'm interested in how the winding progresses, obviously wishing the OP good luck - and despite the above gloomy comments if he copies the design as-is, he'll at least have a modern rebuild which is guaranteed to work even if not quite being able to survive the thermonuclear blast of David's former company!

[Radio Wrangler]

I'd come across 300A/Sq in as an aggressive current rating for wire and did the arithmetic for the diameter of the original to get a comparison figure for an absolute upper end for current. So whatever the power amp, preamp and tuners take, the HT must be less than that, and other considerations may demand it be lower.

"A further 13 valves" Even if the HT demand is kept light, reckoning on 300mA heaters..... Thinking about it, an accessory power supply unit to power preamps and tuners, taking their strain off of a Quad II could be a saleable product from a reliability point of view. Further, if double insulated, it could have floating outputs for preamp, FM tuner and maybe AM tuner. This could break up hum loops quite nicely. Flog ready made harnesses for the non-technical? DC heat?

I honestly don't see the attraction of valved preamps other than as historical artefacts. It's one area where transistors and ICs have advantages, but if it has to be valves, then this is one area with proper engineering justification for adding an external PSU

David

[trobbins]

Perhaps setting up a PSUD2 file with the related transformer details (secondary voltage and effective resistance) and with rectifier and filter part values, and anticipated loading, would get a good estimate of the winding wire rms current. That may reduce the margin of uncertainty.

[Richardgr]

Majestic Transformers in Poole are namechecked by Keith Snook as being suppliers of replacement transformers for Quads.

[Ed_Dinning]

Hi Folks, just wondering on the original transformer; was the core a stalloy or a Unisil type. This would have a bearing on flux density and price.
Assuming double insulated wire is now used, but without interlayer insulation, this may give more space for wire diameter, but as Kalee says, much calculation involved.

Ed

[dougietamson]

Originally Posted by Radio Wrangler

I honestly don't see the attraction of valved preamps other than as historical artefacts. It's one area where transistors and ICs have advantages, but if it has to be valves, then this is one area with proper engineering justification for adding an external PSU
David

I've seen a few Quad II amps converted to stand-alone amps, eg fitted with gold plated RCA input/speaker terminals and IEC mains.
Add your low noise pre-amp/source of choice.

[turretslug]

Quad did at least make use of low heater current valves in the FM tuner, and the AM tuner is fairly straightforward with a combined IF amp and detector valve, but the high tally of extra consumption in a comprehensive set-up is indeed still distinctly worrying- especially when the transformer is a tad tightly specified when running the amp alone anyway. It is a bit surprising that there wasn't at least the option of a "pass-through" PSU option for the accessory chain with appropriate attention to mains safety/signal commoning concerns as mentioned above. It certainly sounds like a wise course of action.

[dougietamson]

Originally Posted by Ed_Dinning

Hi Folks, just wondering on the original transformer; was the core a stalloy or a Unisil type. This would have a bearing on flux density and price.
Assuming double insulated wire is now used, but without interlayer insulation, this may give more space for wire diameter, but as Kalee says, much calculation involved.

Ed

I'll count the number of E/I plates and measure the dimensions.

Would make it a lot easier to not add paper to each layer.

I was planning on using the exact same number of turns as the original.

[GrimJosef]

Originally Posted by kalee20

... I based my comment on the Quad II being a 15W amplifier, having a GZ32 rectifier (max current 125mA) so guessing the HT current is close to the max rating ...

The circuit https://keith-snook.info/schematic/QUAD-II-Schematic.pdf shows 26V (quiescent) across the output valves' common 180R cathode resistor, so that's 144mA. At the EF86s' cathodes there's 2.2V across 780R which adds another 3mA making a total of 147mA HT current for the amp on its own. Any peripherals will draw a bit more, although none of them gets remotely close to the KT66s of course.

ISTR hearing somewhere that Quad wound their own, quite special, output transformers but bought the mains transformers in. So it might not be fair to blame them for any shortcomings directly, although it should have been incumbent on them to check the quality of their suppliers' work to some degree. The mains transformers do fail from time to time but in my experience that's usually due to the output stage having been driven into thermal runaway by leaky Hunts interstage capacitors. I guess as long as the transformers lasted 40-50 years Quad might feel their customers had not been treated too badly. And these amps are 60-70 years old now.

Cheers,

GJ

[trobbins]

Originally Posted by kalee20

I don't know where you got 236mA David, but yes that would worry me hugely!

I based my comment on the Quad II being a 15W amplifier, having a GZ32 rectifier (max current 125mA) so guessing the HT current is close to the max rating. And since the rectifier is operating into a capacitor-input filter, RMS current is about 1.6x the DC output current giving 160mA...

PSUD2 indicates about 173mArms for each half HT secondary winding, based on effective secondary resistance of 78 ohm, and 310Vac unloaded secondary, and using a 5V4G diode model into 16uF with 147mAdc loading.

[kalee20]

Originally Posted by GrimJosef

The circuit https://keith-snook.info/schematic/QUAD-II-Schematic.pdf shows 26V (quiescent) across the output valves' common 180R cathode resistor, so that's 144mA. At the EF86s' cathodes there's 2.2V across 780R which adds another 3mA making a total of 147mA HT current for the amp on its own. Any peripherals will draw a bit more, although none of them gets remotely close to the KT66s of course.

That blew me away - though on closer inspection of the GZ32 ratings I learn it can do 250mA with 350V AC input; the 125mA I quoted was for 500V AC input.

Originally Posted by GrimJosef

ISTR hearing somewhere that Quad... bought the mains transformers in... The mains transformers do fail from time to time but in my experience that's usually due to the output stage having been driven into thermal runaway by leaky Hunts interstage capacitors. I guess as long as the transformers lasted 40-50 years Quad might feel their customers had not been treated too badly. And these amps are 60-70 years old now.

Yes. That's as I'd assumed - Quad probably knew the mains transformer was an Achilles heel in their amplifier (and not the only one!), but it was still 'good enough' and R & D effort could be better spent elsewhere.

Originally Posted by dougietamson

Would make it a lot easier to not add paper to each layer.

I think with modern Grade 2 enamel you can get away with that. Just insulate the leadouts well, take great care not to get any kinks in the wire, wash your hands immediately before handling it, and keep the winding build-up even. You don't want a turn from towards the end of the winding coming into contact with a turn from near the beginning of the winding.

Not having paper interleaving will improve heat transfer through and out of the winding, so that's quite a good plus.

Originally Posted by dougietamson

I was planning on using the exact same number of turns as the original.

I think I'd do the same, in your position. Of course, you could calculate the space saved by omitting the paper, and go up the appropriate wire gauge (adding external limiting resistance so the rectifier sees no difference), if you feel like trying a low-risk improvement.

Originally Posted by Radio Wrangler

Thinking about it, an accessory power supply unit to power preamps and tuners, taking their strain off of a Quad II could be a saleable product from a reliability point of view. Further, if double insulated, it could have floating outputs for preamp, FM tuner and maybe AM tuner. This could break up hum loops quite nicely.

Now that would be really good! It's the way towards no-compromise engineering excellence.

Further, the extra independent rectifiers adds a few more glowing bottles - and puts the set-up higher in the street-cred Top Trumps league!

[Radio Wrangler]

Hmm, a couple of 6X4s ought to make HT rectifiers for two isolated HT outputs, sticking with thermionics on religious grounds.

Two 6.3v windings. One for preamp, one for tuner. Can heat the 6X4s as well.
Two 250-0-250 HT windings for the HT supplies. Electrostatic screen to mains ground, case mains-grounded. Two impressive earth posts for those who want them. (Mithril-plated?)

Extra-generous core size, with generous turns/volt to give low magnetising flux, for low acoustic noise. Also helps to an extent with inrush current.


Two isolated HT outputs (Maybe some big antiparallel silicon diodes as catchers for safety)

Two isolated heater 6.3v feeds.

Fuses on all and mains input.

Reservoir-resistor-smoother on each HT circuit.

Fuses, case IEC receptacle, 6X4s and holders, capacitors resistors are all commodity items. It hinges on the transformer. Done as a DIY project, the transformers are likely to have a good turnover if it catches on.

Another thought, is that Quad is still where the buck stops for their mains transformer even though it may have been farmed out. They set the can size with their design of the chassis and that forced a smaller core than would have been best on whoever did those transformers.

Just musing,

David

[kalee20]

Originally Posted by Radio Wrangler

Hmm, a couple of 6X4s ought to make HT rectifiers for two isolated HT outputs, sticking with thermionics on religious grounds.

That's exactly the valve I had in mind!

Originally Posted by Radio Wrangler

Two 6.3v windings. One for preamp, one for tuner. Can heat the 6X4s as well.
Two 250-0-250 HT windings for the HT supplies. Electrostatic screen to mains ground, case mains-grounded.

I'd actually have separate heater windings for the 6X4's - I prefer to rely on transformer insulation than red-hot mineral oxides at HT-sort of volts...

Originally Posted by Radio Wrangler

Extra-generous core size, with generous turns/volt to give low magnetising flux, for low acoustic noise. Also helps to an extent with inrush current.

As does the use of 6X4's as it both delays and slows the HT capacitors charging by 15 seconds or so - us religious-grounds folk can justify our beliefs!

Originally Posted by Radio Wrangler

Maybe some big antiparallel silicon diodes as catchers for safety)

Originally Posted by Radio Wrangler

Another thought, is that Quad is still where the buck stops for their mains transformer even though it may have been farmed out. They set the can size with their design of the chassis and that forced a smaller core than would have been best on whoever did those transformers.

Yes - and coming back on-topic that's why I was surprised at the just-under 1T operating flux density. That seems incredibly low. There is considerable 'scope for reducing the turns for each winding; increasing wire diameters; the copper losses would reduce and robustness to overloads increase. (For the record, I tend to use 1.2 - 1.3T with silicon-iron laminations, and there's no audible hum - and a recent laminated design which really does push the boundaries was 1.44T. Toroids and C cores can go to 1.7T).

The curves - or at least the blue, magnetising current curve, as per my Post #23 would be most revealing for the Quad mains transformer. Hopefully our OP, dougietamson, will do this when he's rewound his transformer - I'd bet that the current wouldn't start to rise disproportionally till the input hit 320V.