Martin said......
First of all, Im glad to see some activity here on old Usenet. Most people seem to be in web based forums these days.
And the web based forums don't work as well as the Usenet forums. BUT, the Usenet forums are non moderated which means all the arsoles can terrorize everyone who does not agree with the arsoles' opinions, so most newbies and ppl who didn't like the slightest challenge to whatever they typed fled in terror to the smaller moderated groups where everyone is forced to be nice to each other. The result is like a lot of benign zombies attempting communication in short almost content free 1 sentence postings, and that's why I've stuck to this place. If people don't like the heat, get out of the kitchen. If a terrorist enters the kitchen, tip a pot of boiling oil on him, kick him out, and set fire to him then close the door. Nobody gets away with terror with me. I'm here to discuss technical issues, not to pose as smarter than anyone else, but to share my experiences regardless of whether I am in error or not.
In my description of the DC supply, I meant that instead of using the existing 16+16uF cap, I placed two separate caps in the circuit - one 33uF before the choke, and a 47uF after. I did not add another choke or anything.
In Quad-II, the original had C1 16uF + C2 16uF in one box cap, plus 20H choke with Rw about 600r. The B+ for OPT CT was taken from C1 16uF where the Vripple = approx 19Vrms, 100Hz, saw tooth wave. The L + C2 form a filter to the fixed screen supply for KT66, also for input tube B+ and any external tuner or Quad22 preamp. The 20H + 16 uF reduce the Vripple at Eg2 to be about 25mVrms, low enough to keep noise at output low enough.
Using GZ34 with 33uF will be OK but only halve the Vripple at OPT CT. The use of 47uF for C2 will further reduce Vripple at Eg2 to about 3mVrms, a considerable betterment.
BUT, if you use 8r0 speakers with amp strapped for 9 ohms, the RLa-a anode load is about 3k5, and above what is a low class A threshold, the amp works in class AB where the Vripple at OPT CT is in series with the "ON" half of the OPT load and its tube on every 1/2 wave cycle so the THD and noise increases hugely.
To force the Quad-II to work better with low load values, most especially when 4ro speakers are used, and, they often are, despite ppl saying oh no they don't,
the B+ supply at OPT CT should be far better filtered and be a low impedance to 0V, so I like a modern 470uF cap, rated for 450V, easy and cheap. But 470uF used for C1 in CLC would kill any tube rectifier because the peak charge current rises way above the rating which is often less than 0.2A.
SO, I get rid of the tube rectifier. Then I install a pair of 1N5408 in series on each 1/2 primary of HT winding and the B+ at C1 will rise to around +400Vdc if wanted, but better is to use say 150uF for C1, then replace the existing 20H choke with a 4H choke with Rw 50r, and have the 470uF after the L. additional R can be added to between HT winding ends and diodes, say 47r, 68r, 100r to reduce the working B+ to where you want it, or to about +375Vdc, ( +5%, -10%.)
The Vripple at OPT CT will be 3mVrms, and then an R&C filter with say 1k0 + 220uF for Eg2 will reduce Vripple to 0.021mV, and utterly negligible.
With individual R&C cathode biasing networks with 470r + 470uF, the amp will give better bass, and the amount of Vdc change during transients becomes neglible, and the noise injected in class AB is minimized.
A lot can be done to make Quad-II amps far better than the original designers intended.
Phil, thanks for sharing your experiences while I was working on the amps. Since I do not own a dozen of them I could not check what would be the normal behaviour of the circuit.
Patrick, I have read your interesting site about tube electronics. I understand your angle, and I think it is how one should think about new production amps. But in these Quads I was more concerned with safety and longevity than altering the circuit.
But the mods I routinely do to old original junk like Quad-II does make them last longer, ie, KT66 life is longer, and their safety is made better. Both issues require circuit changes that exceed what you are doing just by changing the two main electros in PSU.
I saw a Quad system in a museum recently, together with an early EMT turntable, SPU cartridge, a Revox G36 and so on. Special, quirky but high quality stuff that must to most people seem like steam engines these days. Call me sentimental but I personally dont want to alter such machines too far from how they were made in their time.
OK, you are in love with the crummy old quality of the past. Maybe you would not mind using hand operated shears to shear a sheep, but I'd prefer the powered wool cutter. I really like old steam engines, and the locals in Canberra have a Garratt articulated loco almost restored, and there are no mods. OK, quite a sight'n'sound to see running. 262 tonnes of power from 1960. Wow. But I'd hate to own the darn thing. Same goes for a Spitfire, or, rather like a Quad amp, a Morris Major. Now I just cannot think of a 1950s amplifier that excites me like say a D Type Jaguar, or a Vincent Rapide. Back in 1950s, amps were useful slaves, seen and not heard, tucked away out of sight, turned on by a switch on a preamp on a console.
The amps sold to the public and to BBC were pretty bloody awful, compared to what we can do now because of better diodes and caps, and because decent wire and iron for PTs and OPTs costs far less in real terms than in 1950, when most of the good materials had been sent to the bottom of oceans or blown all because men who ruled human kind are blind, vain, stubborn, vile, ruthless, greedy, and arsolic fuctards.
I understand the problem with the common cathode resistor, and I wonder if maybe this is one of the reasons many Quads have failed with melted tar all over. In my amps I swapped tubes around and measured voltage differences for a little bit better DC balance. But separated cathodes sounds like a good idea.(I built a Williamson inspired amp once, using huge wirewound pots between the two output tube cathodes and ground. Was easy to balance but maybe not the last word in highend audio
Leak only used individual biasing of the 2 output KT66. To get the most years of trouble free use with output tubes, there MUST be individual biasing, and the MUST lower value grid bias resistors used to prevent the inevitable positive grid current causing grids to go positive even in the idle condition. Quad used 680k, and its LUNACY. Quad should have used a 6CG7 or 12AU7 LTP which have LOW Ra and then Rg could be 150k max. Aging KT66 will develop "reverse grid current", ie, you'll see perhaps Eg1 at +5Vdc, indicating 7.4 micro amps which sounds like SFA, but that +5Vdc causes about 20 to 30mA of extra Ia to flow, and tube gets too hot, and other becomes cold with Ia too low because tubes don't age at the same speed. So the Idc balance in OPT becomes poisonous to the music because of unwanted dc magnetization of an un-gapped core. This imbalance is reduced with individual R&C bias networks, and the use of Rg = 150k would reduce
Eg1 from say 5Vdc to 1Vdc, and imbalance is then tolerable, and you get a few more years from KT66. I've watched this happen in many amps I have serviced and where owners would cry tears of blood before ever consenting to buy new tubes for their bloody horrid old junk amps!!!
It would be interesting to hear more ideas about how to make Quads safe for future use and longevity. I understand it is not a good idea to put a fuse between the HT center tap to ground?
Chassis MUST be grounded. So each power amp chassis must have IEC connector to take standard 3 wire leads with IEC connector at one end and the mains 3 pin plug at other for wall sockets.
The IEC chassis plug can have a mains fuse in it, so that the mains cable MUST be removed from amp to remove the mains fuse.
To minimize hum, the 0V rail must have 5r0 between it and 0V rail, and if there is a B+ fuse, best place is between HT CT and 0V, NOT to chassis "ground."
Original Quads do have a 5r0 between 0V and Chassis. Now have a real good look at your original Quads' earthing method. What do you find?
The trouble with fuses is that they may take time to blow, and PTs can heat up and potting mix will melt and run out before the fuse blows, and I've seen this a few times because of arcing ESL57 speakers. You can have one KT66 operating with red hot anode, the owner is oblivious, and amp limps along for awhile damaging itself, maybe towards a very expensive repair. SO, the ONLY remedy is to have some active circuit which detects excessive Idc flow in output tubes and this turns off the amp and a red LED lights up which tells an owner "I will NOT work until YOU fix a problem". The active circuit works far better than any HT fuse. I have installed such circuits many times into new and old amps and owners became so glad I did.
Tube amps can be hell to live with. I've repaired and rebuilt tube amps which have nearly burned houses to the ground. Get real, Old junk is always old junk unless proved otherwise, and much must be done to save them from themselves and from dopey owners so that THEN the past can be really enjoyed, with added benefit of better modern parts. Needless to say, all old Quad amps, especially the Quad 22, need total R&C replacement, and with a few wise other mods I mention at my website.
I'm preparing a page about how I rebuilt a pair of Quad-II-Forty. These Chinese made souped up versions of Quad-II have many of the same horrid features to be found in original Quad-II.
Patrick Turner.
Martin