Quad II Mods/Tweeks

[Johan Potgieter]

David,

I have already posted a lot on your other Quad threads so beg patience from those who already read that, but your quiery here opens the field slightly.

Firstly you can replace the KT66 with a 6L6GC without any bother if you have trouble in getting the former. They should be easy to get and are so similar that SOVTEK uses the same innards for both! (If folks will measure they will find that the SOVTEK KT66 takes 0,9A heater current and not 1,27 as specified for KT66.) The 6L6GC also has a max. plate dissipation of 30W compared to 25W for KT66. One advantage of KT66s is that they do look very sofisticated!

But here I am obliged to sound a warning from dire experience - and which will be difficult to control when ordering. Some manufacturers are rather free with their designations. More serious than calling a 6L6 a KT66, is the fact that they take the old 6L6GB (23W) and glibly call it a 6L6GC. The visible difference is comparatively large cooling fins on G1 and the innards mounted directly on pins going to the outside pins for the GC, as contrasted to small or no cooling fins on G1 and the innards mounted on a glass tongue inside the tube with thin wires running to the outside pins as per old practice, for the GB. (The "direct" mounting method conducts more heat to the outside world - in this case the sockets.) Also avoid the Coke-bottle shape - the hot glass envelope is too close to the choke/power transformer and cooks it!

Other than that I would not change any values except power supply electrolytics. As said elsewhere I found some resistors to be off by up to 40% and I replaced with 0,5W 1% metal film types. I also mentioned moving the hot 180 ohm cathode resistor away from directly under the choke, and away from the cathode bypass capacitor, and increasing the value. You would have read this elsewhere. You could change the coupling capacitors on the KT66 sockets to polyester types, but I would advise just a check with a square wave and scope if you are able. The original "Hunts" capacitors contained in a metal tube have about 18 pF to earth. Without this your square wave might just show a little overshoot, which you could correct with outside caps. I have found some phase lead compensation by way of a 1,5 nF capacitor across the 470 ohm feedback resistor to be advantageous at the high frequency end in this regard.

I see no reason to decrease the feedback - it is after all part of the performance. The rather high input signal required should be no problem for modern ancillaries. I did on occasion replace the Jones input connector with an RCA, in which case some construction is required to cover the hole. (This is for separate pre-amplifiers.) I have also converted to the IEC power system.

Then there is every reason to change to a GZ34 rectifier - the heater current is the same. Power supply caps - I have changed to 100 + 100uF without a problem - one gets a decent dual from JJ. It is true about the max capacity for GZ32, but this also depends on the plate supply impedance. For the GZ34 any capacitance could be used provided there is at least 150 ohm plate-plate in series. The dc resistance of the Quad h. v. winding is already 150 ohm, the impedance at 50 Hz would be slightly higher. The old 16 + 16 uF is rather low. Lastly, be aware that the input filter capacitor needs to be able to handle a minimum ripple current. (That is why the original had a plain foil type at the input.) Do not just put the first small size available there.

I am not "drawing rank" (in this case experience and age!) here, but I have done a number of these over the decades, rewound power and output transformers, measured and listened, as others have. I would respectfully plead that you do not make unnecessary changes.

With that I think enough from me for the present.

Regards and success!

[405man]

Johan
Would you have details of the Quad II output transformer, I have had mains transformers wound but could not find details to get output transformers wound

Thanks

Stuart

[Johan Potgieter]

Stuart,

Yes. Give me a day - I am out tomorrow (Monday) but will post afterwards.

Johan

[Johan Potgieter]

Stuart,

The Quad II output transformer is rather complicated, comprising 8 primary sections interleaved by 6 secondary sections. I will have to post a sketch for all to be clear, and Murphy has intervened. I cannot scan at the moment and will do that somewhere else and then post. Some patience please.

Regards.

[Johan Potgieter]

Stuart,

I see that we have 2 Quad threads - hope you are reading here. I am now able to attach details about the Quad output transformer. I apologise for the long delay.

I must emphasize as I state on the attachment, that this data came from unwinding one transformer. I would not be aware of possible changes in different models.

Regards.

[Tony]

Johan,

interesting details, can you post the core details, like center leg and stack depth?

about the gimmick, your sketch indicate it in series with the secondary coils, is this so?

thanks

[cerrem]

405man and Tony....

I do wind QUAD II and MacIntosh output transformers from time to time....
I have analyzed the original design and in some cases made small changes to increase the bandwidth... Sometimes I have inserted screen taps...
If there is some info you need let me know...

Chris

[Johan Potgieter]

Tony,

This was some time ago, and I unfortunately did not note the complete dimensions of the core, seeing as how I thought one would rewind on an existing core. I know of somebody that might have that, and will try to get it for those who want to try and have one wound from scratch.

I did note the following at the time:

Window: 22 mm
Tongue: 23.8 mm
Stack: 26.5 mm (i.e. 1 inch)

Generally speaking, if a Quad core is not available, any core of approximately the right size should give a satisfactory product. If it is significantly different, one could recalculate the wire gauges to fill the window.

The "gimmick" is where shown in the sketch. I have read some comments that it is a resistance of an ohm or so to "balance" things. That still does not make sense to me. For the 8 ohm connection the secondaries are balanced in the sense that leakage reactance would be balanced. I can also recollect reading that this gimmick was not present in some transformers. But all that was 2 decades ago. Perhaps someone else can comment.

Regards.

[cerrem]

I am at work right now

When I get home in a few hours..I will look in my winding log book and provide the QUAD II transformer info....
From what I remember, they used standard Imperial core size at that time... I believe it was EI-112 ....but let me verify later on..

Chris

[Johan Potgieter]

Chris,

We seem to have posted simultaneously.

I must point out that the Quad transformer is already a distributed load (UL) design if one looks at the equivalent diagram. (The part of the primary between screen and B+ is simply put on the cathode side.) In this design the G2 "taps" are at 25% to place reasonable demands on the phase inverter. I am not saying though that you could not use extra screen taps, making the screen-B+ portion say 33% without having it all on the cathode side. But that would obviate the convenience of being able to use a lower screen voltage.

Since this I have always used this topology in tube amplifiers. In a recent 100W job I used a 33% screen-B+ section, all on the cathode side. That required a formidable 350Vpp maximum grid signal. I have since decided to stay with 20 - 25% in future!

Regards.