[[email protected]]

Hi all,

I posted before while I was fixing up a couple of Quad II amps. After working fine for a number of weeks, one of them just started oscillating like mad before I even could put on a record. The "starting up a harley davidson"-type of oscillation, ending with a squeal when I shut it off. Now, the circuit is almost original, but a lot of things and wires have been resoldered and I have replaced the OPT with another original one, which I strapped for 9 ohms.
Swapping out all the tubes for known good ones did not help.

Last time I checked, all the original resistors were ok and less than 10% from spec. I have raised the shared cathode R to 220 ohms to be able to run 6L6GC. The C's are all renewed: C1 is a Wima pp, C2/C3 are NOS Good-All glass sealed PIO with metal case, C4/C6 are replaced with modern 105C 33 and 47uf. C5 is also modern, but raised to 100uF/100v.

I know I will soon have to fire up the scope and do my own work, but have any of you had this happen before? Is the most likely cause a bad solder joint, or something like a shorted resistor in the global negative feedback loop?

Thanks!
Martin

[Phil Allison[_3_]]

I posted before while I was fixing up a couple of Quad II amps. After
working fine for a number of weeks, one of them just started oscillating
like mad before I even could put on a record. The "starting up a harley
davidson"-type of oscillation, ending with a squeal when I shut it off. Now,
the circuit is almost original, but a lot of things and wires have been
resoldered and I have replaced the OPT with another original one, which I
strapped for 9 ohms.
Swapping out all the tubes for known good ones did not help.

Last time I checked, all the original resistors were ok and less than 10%
from spec. I have raised the shared cathode R to 220 ohms to be able to run
6L6GC. The C's are all renewed: C1 is a Wima pp, C2/C3 are NOS Good-All
glass sealed PIO with metal case, C4/C6 are replaced with modern 105C 33 and
47uf. C5 is also modern, but raised to 100uF/100v.

I know I will soon have to fire up the scope and do my own work, but have
any of you had this happen before? Is the most likely cause a bad solder
joint, or something like a shorted resistor in the global negative feedback
loop?


** If all the small components are OK - then suspect a broken wire, bad
solder joint, bad ground or an unintended short to ground that affects the
feedback loop.

On ohm meter and your eyeballs are all you need.


..... Phil

[Patrick Turner]

On Sunday, 8 June 2014 07:22:16 UTC+10, wrote:
Hi all,



I posted before while I was fixing up a couple of Quad II amps. After working fine for a number of weeks, one of them just started oscillating like mad before I even could put on a record. The "starting up a harley davidson"-type of oscillation, ending with a squeal when I shut it off. Now, the circuit is almost original, but a lot of things and wires have been resoldered and I have replaced the OPT with another original one, which I strapped for 9 ohms.

Swapping out all the tubes for known good ones did not help.



Last time I checked, all the original resistors were ok and less than 10% from spec. I have raised the shared cathode R to 220 ohms to be able to run 6L6GC. The C's are all renewed: C1 is a Wima pp, C2/C3 are NOS Good-All glass sealed PIO with metal case, C4/C6 are replaced with modern 105C 33 and 47uf. C5 is also modern, but raised to 100uF/100v.



I know I will soon have to fire up the scope and do my own work, but have any of you had this happen before? Is the most likely cause a bad solder joint, or something like a shorted resistor in the global negative feedback loop?
Thanks!
Martin

Phil could be right, you just need to eyeballs and a multimeter.

BUT, so many ppl do not have huge experience with repairing electronics of many varieties, so they will struggle to see anything as they gaze at an amplifier upside down on the bench without any single question appearing in their mind.
You may have a crook output tube with loose wire inside a pin perhaps a socket has a loose gripper, perhaps a resistor somewhere is intermittent, or a capacitor.

Unless you replace ALL the old R as well as the C which you have replaced, expect troubles with intermittent connections which can be a royal pain in arse until you assume **** will happen with old junk unless serious time is spent doing all you can to keep it road worthy. I'm no fan of paper in oil PIO caps.
I've seen to many of them fail in the past, and Wima polyproylene MKP rated for 630V are the best.

I've never heard a Quad-II imitate a Harley starting during turn on and then squeal at turn off. The grumbly start sound indicates something reluctant to connect until some voltage is across the connection long enough to make tiny little arc to cause a little local heat and the connection is made, which then conceals the whereabouts of the fault. So you need to be able to quickly check rising Vdc everywhere just after turn on. And try putting 2uF from each output tube grid to 0V then turn it on when cold. If it grumbles&squeals, the problem is in the output stages, not in EF86s. Try change of output tubes, it that makes not difference, then you have a dodgy old resistor.
The old resistors may not be more than +/-10% away from correct value, but that doesn't mean they have connection integrity. The old R used in Quad-II fail randomly after 55 years. Modern 1 Watt rated metal film should be used and in another 55 years they will probably have exactly the right value and not be faulty, unless they have suffered many heat cycles.

The only thing which makes Quad-II oscillate continuously is a 0.22u cap across the output when strapped for 9r, and the oscillation is at F well above 20kHz so you won't hear it.

Does the amp grumble&squeal with a pure R load, say a 10W 8r2 resistance? Does it happen only with a speaker?

Once turned on, how do the Vdc everywhere around the circuit compare with Vdc in the other channel? I assume you have 2 Quad-II amps. Is there a damn Quad 22 control unit involved? Is the problem in a preamp?

Fact is, your plea for help here raises more questions than I have time to poke a stick at.

I always found life was to be figured out, and unless I befriended Unkel Doubt, I could figure nothing out.

Patrick Turner

[Phil Allison[_3_]]

"Phil Allison"


** If all the small components are OK - then suspect a broken wire, bad
solder joint, bad ground or an unintended short to ground that affects
the feedback loop.

On ohm meter and your eyeballs are all you need.


** FYI:

I tried shorting the 470 ohm feedback resistor = no oscillations.

I tried removing the filter electro after the choke = no oscillations.

I tried linking the output direct to the input = loud noise like an
un-muffled two stroke.

The symptoms you describe are typical of reverse feedback, like when anode
wires are reversed.



..... Phil

[[email protected]]

Thanks guys, but now the problem is GONE. All I could find was that the first EF86 socket seems slightly wiggly, perhaps it lost contact with one of the pins.

The other possibility is that I had a longer speaker cable to the channel that oscillated, combined with the fact that the home studio monitors I was using recently were *6 OHMS* and not 8 as I stupidly assumed. When I swapped the speakers between channels it was OK again. An issue with capacitance vs impedance? An issue with one of the second hand irons? I have only played at low to moderate living room levels so an impedance mismatch should AFAIK not cause harm. I hope.

Thanks Phil for checking what a shorted GNFB R does. Once I put trimpots in an old amp to learn different effects of feedback. Past a certain pont it would start growling like an old engine.
And thanks Patrick for posting your thoughtful advice. Yes I see the point of using hassle free components I see you updated your site? Great reading.

[[email protected]]

Just a followup to my original question in case others have the same problem. I managed to localize an intermittent contact somewhere around the 680k resistors (R7, R9) by lightly tapping on every R with a stick while the Quad II was in operation. Replaced both resistors plus the 2.7k (R8) inbetween, and resoldered everything else on the tag board. I am not sure how this fault could produce an oscillating squeal, but now things are stable inside the amp.

Martin

[Patrick Turner]

Just a followup to my original question in case others have the same problem. I managed to localize an intermittent contact somewhere around the 680k resistors (R7, R9) by lightly tapping on every R with a stick while the Quad II was in operation. Replaced both resistors plus the 2.7k (R8) inbetween, and resoldered everything else on the tag board. I am not sure how this fault could produce an oscillating squeal, but now things are stable inside the amp.

Martin

I can't answer exactly why your amp squealed.
The minimum improvement to Quad-II is to
1. Replace all old caps and resistors installed in Quad-II in 1950s-60s.
2. Use Welwyn 1.2W metal film for most R.
3. Use at least use 390r 5W wire wound bypassed with 470uF at each KT66 cathode to CFB winding end.
4. Ground CT of CFB winding.
5. Coupling caps all should be Wima MKP 630Vdc rated.
6. Electros in PSU can be 33uF.
7. Use GZ34 instead of GZ32.
8. While resoldering new R&C parts to bent brass lugs on old Quad terminal board, be prepared to find that lugs just break off, because the bend in the lug is so sharp, and this fatigues the metal over 50 years. Sometimes I have deliberately broken off all lugs, and relied on soldering to the riveted part in the board.
9. Install an RCA input socket.
10. Make sure the chasssis of all units used are connected to EARTH via green/yellow wire amoung 3 wired from wall socket cabble.
11. Replace all R&C in Quad-22 control unit, if used, especially those under one switch bank which relate to RIAA eq.

The cathode biasing of each KT66 ensures much better Iadc balance so OPT Ia is balanced so iron related distortions and hum is less.

For best class A performance, don't use less than 16 ohm speakers when OPT is strapped for 9 ohms. This seems strange, but with 9 ohms used the RLa-a is about only 4k, way too low for best class A.
These old amps were best driving sensitive 1950s - 60s speakers, many of which were 16 ohms nominally, eg, Tannoys et all.
Patrick Turner.