Quad II Mods/Tweeks

[cerrem]

Johan....
I am fully aware of the QUAD II OPT and partly agree with you about the UL.... It is a bit different regarding the G2 tapping than just simply applying the "tap" to the cathode...
I do see a problem with the winding layout sheet... It shows connection from U to Z and W to X ..... The CATHODE winding should be DC isolated from the Plate winding...since these are only AC coupled...

Chris

[cerrem]

I looked over some QUAD winding sheets I have for those interested...
If you were to look at the FULL primary of the QUAD output transformer you would see 3200 turns....This included the cathode winding.... You would get a Plate load of 3.786K into secondary of 16 ohms, 9 ohms and 4 ohms....
If one examines KT66 plate curves for the correct screen and plate voltages you will see this plate load is a perfect textbook fit...
The "distributed load" seperated for the cathode is 10% .....
So this means 2880 turns are in the plate winding and then 320 turns in the Cathode.... The Plate and Cathode windings are seperate and isolated..only AC coupled.....
I disagree with the # 36 gauge wire for the entire primary....that was only used in the cathode winding BUT paralleled by 3 to bring the Circular Mills/amp up to 75 which is acceptable....
The core was a EI-100 lamination...
I would ask Johan how the core was stacked on the transformer he has seen.... Since I have seen it two ways...I have seen the E-I lamination stacked 1:1 and another one was stacked at 2:2 groupings.....

Chris

[cerrem]

One other note....
The open loop power output would be 25W calculated with 10% cathode feedback is applied... Then with global feedback loop the power output is reduced to 22W ......Which is about the MAX measured power output of most QUADS...
ALthough I did mention a 4 ohm tap in the previous thread...it is "un-official" since PIN Q on the output transformer is a feedback tap located at the 4 ohm .... The wire gauge would not support the current for a 4 ohm load at pin Q .....

Chris

[Johan Potgieter]

ALERT - ERROR!

I am astonished at my own inaccuracies. Please everybody: In my previously attached diagram of the Quad output transformer there were errors as pointed out by Cerrem:

1. There are NO connections between points X-W and U-Z.
2. The sections VU and VW are not 160 turns x 3, but only 160 turns x 1 each, i.e. one layer. (The others are 160T x 3.)

I profusely apologise for this and am indebted to Cerrem for pointing this out. There is no excuse but perhaps an explanation would be considered: The print of my drawing came out light, especially some lines which were quite thin. The friend who scanned this for me drew in some of the lighter lines and accidentally also connected the above. This was then e-mailed to me and I did not look carefully at this again before pressing the button to attach.

The Quad thus uses 10% cathode feedback, indicating a 10% UL tap. I have just checked this on a Quad in my possession.

The data was written down 2 decades ago, and if you have an update on the #36 wire, Chris, please inform us. That is what I have noted at the time and can unfortunately not recall details now. Neither, unfortunately, can I recall the exact assembly of the core, but a local friend rewound such a transformer about a year ago, and I will mail him for details.

I agree with your calculations, Chris, only am not quite sure what you mean by "parallelled by 3" for each cathode winding, unless you took my erroneous indication of "160t x 3" as meaning 3 layers in parallel. That was not the case according to my notes, as said above. By the way, I have measured the primary d.c. resistances as follows for those interested:

XY = 119 ohm, ZY = 178 ohm, VW = 16,4 ohm, VU = 16,8 ohm

I am also not with you when you say that "It is a bit different regarding the G2 tapping than just simply applying the tap to the cathode". As I see it in the equivalent a.c. diagram, one "breaks" the primary winding at the tap, calling the [anode-G2] part Y-Z (per the previous sketch) and the [G2-B+] part V-U; V and Y being the same point for a.c., i.e. the G2 tap of the basic UL configuration. How am I misunderstanding you?

Once again apology for the errors and my profound thanks to Chris.

Regards.

[EC8010]

Calm down, Johan - to err is human. I'm wondering about your 10% cathode feedback ratio. I measured the voltage ratio between the windings and obtained 9.375:1 rather than 9:1. I'm reasonably confident about my (single) measurement, so I'm wondering where your turns figures came from. I'm not saying I'm right, but it would be nice if we could all agree.

[cerrem]

EC8010...
Are you using two seperate AC meters both using the same RMS processing??? Are both meters calibrated??
If using one AC meter for both measurements, loading errors will exist...
WHich winding are you exciting for this measurement???? Cathode excitation current is significant to cause such error..
I have seen the windings in a number of QUAD OPT's and they come up to 9:1 for the distributed load...
The Cathode winding is 3 paralleled sections for each side of cathode...each of 3 sections is uniformly sectionalized between plate windings to reduce leakage and increase coupling with plate winding which is extremely critical for Hi-Fi.....
I have redesigned the QUAD OPT on EI-112 with lower turns and better band-width....
Johan...
Do you remember how the EI laminations were stacked on the OPT you took down????

Chris

[Johan Potgieter]

Thanks EC8010.

Only when "to err is stupid" rather seems to sum the situation up, perhaps the boot should be applied to a certain area of the anatomy.

I went and switched the Quad on for the second time tonight. Measured under signal conditions I got ratios of 8,45 for both tubes at 300 Hz; 9,16 and 8.95 respectively at 2 KHz; and 9,09 and 8,89 respectively at 7 KHz. (This was not the same amplifier I had repaired then.) I do not have an amplitude (cursor) readout scope and has to judge by eye. I got the number of windings back then by unwinding and counting. Could a mistake have crept in? - possible, but the sections were all equal, so I cannot explain the discrepancy. The most I can do is enquire locally whether others also have data. I would be surprised if I were the only one to have ever had to rewind a Quad output transformer.

Cerrem seems to have data; hopefully also figures for number of windings? The discrepancy will not make much of a difference in practice, but as you say, it would be nice to have agreement.

Regards.

[Johan Potgieter]

Cerrem

We seem to be mailing simultaneously. But you now have me worried, I can definitely not recall the 3 cathode sections you mention. Could mine have been a very early model?

Also, having thought about it, the very low cathode impedance bypassing the cathode winding could be what you meant by the earlier mentioned difference. (That would not exist were such a winding from G2 to B+.) I have also made my own designs for higher output and used 20 - 25% cathode windings without experiencing difficulty, but that was for class AB and I had to sectionalise the primaries quite some more for proper performance. (I used fewer secondaries.) In one instance the cathode windings were bifilar to parts of the (anode) primary, but I did not find that to be noticably better than sections.

Do you have a windings count?

[cerrem]

Johan...
I did provide windings count in one of my previous threads, this was based on data I have from 8 years ago....
3200 total primary windings....2880 for the Plates, 320 for the cathodes.... Lamination was EI-100 ......
I modified the design slightly from the original in that I wind this OPT with better coupling between cathodes and plates...I felt the original OPT had inadequate coupling between plates and cathode windings.... I use 6 cathode layers that are uniformly interleaved within the plate windings, thus lower leakage inductance between all the primary sections... ALso use better di-electric materials with lower dielectric constant....
The problem with varnish impreganated paper is that the di-electric constant is far from constant, it has a slope based on applied voltage gradient..for example I measure the full frequency bandwidth of the QUAD from 5W to 22W in increments and the plots show the high frequency pole droping with applied increasing power output, since the leakage is truly constant as a function of winding geometry, thus the increased winding capacitance...
The newer materials I use show a more stable freqeuncy response over wide range of power....
The original OPT had a peak inductance of 135H .....The 3200 turns @ 20W @ 20Hz given this core size, would normaly push the flux density a bit higher than conservative Hi-Fi design...but by stacking the core in very large groups...your effective gap increases enough to buy you another 1000 Guass of usable headroom, so all is well...the inductance is still way up there and no need for worry... At 135H the low frequency -3dB pole is roughly 6Hz.... i usually wind up with over 200H

Chris

[Tony]

johan, chris,

just to make sure i understand, are the plate windings and cathode windings in phase or not?

by large groups stacking, you mean in groups of say 8 lams or more? instead of the 1 to 1 or 2 by 2 method?

thanks