KT66 gm?

[peterp1]

Some my old KT66s have been tested by a local electronics man. Some of the KT66s have tested 3gm/4. How low can the gm be before the valve is no good. Does the operation of a low gm valve spell danger to the rest of the Quad 2 valve amplifier?

[Craig Sawyers]

Not too sure what you mean by 3gm/4.

But the KT66 datasheet says gm = 7 in tetrode connection. The AVO manual says gm = 6.3 for the operating point they chose.

So the value depends on the operating conditions on the valve tester used.

What did your local electronics man use to test the valves?

Craig

[Heatercathodeshort]

From OSRAM VALVES Maintenance types for broadcast receivers and amplifiers. December 1944.
There are three pages in all but I think this is all you require. John.

[Diabolical Artificer]

It doesn't mean anything, the best test for a valve is in a circuit read this " As I have explained, the Emission/Gm test is the last test you will make, and often it will be the least useful — even though most people incorrectly put the most emphasis on it. In summary (and assuming that a test score is even RELEVANT to your tube type!), a higher test score does NOT mean that tube will have LESS NOISE or BETTER SOUND or anything else that most people ASSUME that the “test score” means. Really! A tube testing “100” is no more likely to “sound better” or “work better” than a tube testing “80”, or vice-versa. Technicians know this, but the buying public does not." Copied from https://tubesound.com/tube-testing-h...a-tube-tester/

Andy.

[Radio Wrangler]

But it does give those who don't know better a warm and happy feeling. People are prepared to pay a lot of money for warm and happy feelings.

It doesn't answer the question of 'will this valve bias-up OK in my set? and give the right amount of gain?'

David

[trh01uk]

Originally Posted by peterp1

Some my old KT66s have been tested by a local electronics man. Some of the KT66s have tested 3gm/4. How low can the gm be before the valve is no good. Does the operation of a low gm valve spell danger to the rest of the Quad 2 valve amplifier?

Firstly, "3gm/4" has no meaning that I recognise. Either its a typo or your testing man is telling you nonsense. The gm might be "3 to 4" which is below the typical spec as Craig noted, but will depend on the actual conditions in the circuit.

Low gm just means the gain of the stage the valve is in is low. Ultimately, that would mean less output from the amplifier. Will it cause danger to the rest of the amp? No. Except that eventually if you took no other action, you might be so fed up with the brute not providing enough volume that you chuck it in a skip.....

Richard

[GrimJosef]

Answering the OP's questions in reverse order:

Problems with gm itself won't damage the amplifier. There is, however, some correlation between low gm and low Ia, where Ia is the (quiescent) anode current. As long as the two KT66s have reasonably similar Ia it won't lead to damage, even if it is somewhat low. But if you were to have one KT66 with low Ia and one with high the latter one would find itself being worked hard by the shared biasing arrangement. It's not out of the question that that would cause it to fail in some way and that could do some damage, although the devil is very much in the detail. If the Ias are OK though and the gms are where the problems/differences lie then you needn't worry about damage.

Low gm can affect the amp's performance - particularly in terms of distortion. The amp's circuit applies several different types of negative feedback in several different places. This feedback tries to correct for any peculiarities in individual valves' performances. But it works on a percentage basis*, in a sense. Which is to say that it can reduce the problems to, say, 10% of what they would have been without the feedback. So if we start out with poor quality valves the feedback can't correct the performance to what it would have been with good quality valves. Good quality valves will always work better than poor ones.

Turning to your first question, the long-form version of this valve's datasheet, available here https://frank.pocnet.net/sheets/084/k/KT66_GEC.pdf, says on p3 that "A valve is considered to have reached the end of life when gm [is] <5.5mA/V". As Craig has said, a new valve would be expected, under the same test conditions, to show a gm of 7mA/V. So M-OV reckon the valve has had it when its gm has fallen to 5.5/7 or 79% of the new value. Of course if the circuit is undemanding then valves which are more tired than this might still work OK. And anyway, the valves would have had a spread of values of gm when they were new. So making a single measurement now can't tell you whether you've got one which started out strong but is now on its last legs, or whether it's always been a bit on the weak side and isn't much worse now than when it came out of the box.

Cheers,

GJ

*Actually it's more complex than that, because using poor quality valves will raise distortion levels and if the valves are low-gain then the feedback system will also have less 'resource' to fix that distortion.

[Craig Sawyers]

The most important parameter in a valved output stage is emission. ie does it produce the correct anode current for the grid voltage *in the actual circuit*. gm is of secondary importance.

The amplifier actually has local feedback applied via windings on the output transformer to the cathodes, meaning that the gain of the output stage is determined more by the local feedback than the valve gm.

And then global feedback is applied too.

It is a cunning design from 1952 by Peter Walker and DTN Williamson. https://worldradiohistory.com/UK/Wir...ld-1952-09.pdf page 357 onwards.

Craig

[Heatercathodeshort]

Remaining data from the GEC/Osram spec.

A large number of service departments had valve testers gathering dust in a remote corner of the workshop. They were viewed with distrust and it was far cheaper to stock a large range of new replacement valves to make substitution tests which is after all the ultimate and reliable test.

My dust covered example was a Mullard 'High Speed Tester' It would do the job testing a wide range of valves giving a good bad or indifferent result. I have a superb example of the AVO characteristic MK3 that can be fun to use especially when I am presented with a gift of a large number of 'pulls' but even then nothing is guaranteed.

Other than that I would not trust any valve tester to give 100% accurate readings especially when conditions within the valve change after a period of time.

They were mostly money earners in the 30's 40's and 50's when radio shops tested customers valves writing off thousands of perfectly serviceable valves in the hope of selling them replacements. Hope the data is useful. John.

[Craig Sawyers]

Another datasheet point worth making is a lifetime of at least 10,000 hours at Pa+Pg2 of 21W. The QUAD amp has 22W standing dissipation, so it is reasonable to expect >10,000 hours out of the KT66's, or over a year if turned on continuously.

Craig

[Craig Sawyers]

Also worth noting that if the OP (where is he - one post and then nothing) had his valves tested on an AVO, the gm for the set-up conditions of Vg = -15V, Va/Vg2 = 250V is Ia = 65mA and gm = 6.3mA/V.

But it would be helpful if the OP could elaborate, because we are advising in a vacuum.

Craig

[Radio Wrangler]

GJ has answered the original poster's question:

The valve manufacturer considers the valve to be no longer good once Gm has fallen to 5.5mA/V or less.

A competent designer will have toleranced his design so that with any valves better than that value, the full amplifier ought to meet its specs.

David

[ionburn]

Originally Posted by peterp1

Some of the KT66s have tested 3gm/4.

In my opinion:

Obviously it would not be sensible to make assumptions but, as figures are involved, my take on this would be to use the normal mathematical meaning hence 3gm/4 would be identical to 0.75 x gm or three quarters of the stated new book value. The only exception would be if the original gm for a particular valve were stated but, if so, the particular valve would need to be identified. Obvious other complications occur when a range of value is stated in manufacturers data as a spot figure is then a bit misleading.

[David Simpson]

Please peterp1, any chance on some info on the circuitry in which your KT66's are being used? A Circuit Diagram would be great.
Also, perhaps you could ask your local electronics man about which Valve Tester he used, & obtain info on the Anode Currents (Ia) achieved, Anode & Screen Voltages used, and the Grid Voltages used ? Or the Valve Data Book he used ?
For example, if he used the AVO Valve Data Manual(which quite a few of us use), Va = 250V, Vs = 250V, Vg = -15V, and expected Anode Current (Ia) is 65mA. Gm = mA/V & should be near enough 6.3 mA/V. Assuming the valve is in good condition. Some Valve Testers(or Tube Testers) just give a coloured meter display. I.e. RED zone = Fail, WHITE zone(middle of the meter) = maybe just usable, and GREEN zone = Pass.

Regards, David

[GrimJosef]

He asked about damage to a Quad 2 David. Strictly speaking that would be a Quad II of course. Circuit here, for example https://www.keith-snook.info/amplifi...-Schematic.pdf.

Cheers,

GJ

[trh01uk]

Presumably, GrimJosef, if Ia tracks gm as you said above, then we could have one KT66 taking much more current than the other. In which case, I would expect the output transformer to be pushed away from the linear region of its magnetic curve, and the distortion will become severe - well beyond anything that the feedback system can cope with.

Richard

[GrimJosef]

The correlation between Ia and gm is quite loose. I imagine they're linked mostly by the cathode emission (see Craig's post #8).

The design of the Quad II, including its various feedbacks, is subtle and the amp is supposed to be able to withstand a significant mismatch in Ia between the two KT66s. The specification, taken from the manual, says "Valve mismatching up to 25% (introducing 2nd harmonic) not to cause distortion to exceed 0.18%" this distortion being measured at 700Hz and 12W output.

Whether or not you think that 0.18% distortion is 'severe' and an indication that the feedback system isn't coping is, I guess, a personal judgement. In my experience a mismatch as large as 25% might well cause the background noise specification of -80dB referred to 15W to be breached, due to poor rejection of the 100Hz ripple on the HT supply. If the speakers are at all sensitive then this might well be more irritating to the listener than a small fraction of a percent distortion

Cheers,

GJ

[trobbins]

A mismatch of 25% would also lower the primary side winding inductances and hence likely affect the 20Hz 15W low frequency specification.

[Radio Wrangler]

Originally Posted by GrimJosef

He asked about damage to a Quad 2 David. Strictly speaking that would be a Quad II of course. Circuit here, for example https://www.keith-snook.info/amplifi...-Schematic.pdf.

One interesting note added on that version of the schematic. This amplifier has, of course, two 'THAT' capacitors (those capacitors?) and that the design used the stray capacitance from the outer foil to their grounded metal case as a significant contributor to the open loop pole needed for the feedback system to be stable.

Something to watch out for when doing a precautionary replacement.

Little subtleties like this tend to get lost as time goes by.

It can be overcome by adding a pair of small capacitors, but first you need to know of the need.... then you need to know the value. It might be helpful if someone with an original amplifier in working condition measured it.

David

[Craig Sawyers]

That is an interesting observation, David. Perhaps the way out of this conundrum is to gut the old old Hunts, hide a modern component inside, and make a connection from whatever end to the original metal shell. Or dare I say it - use an "audiophile" paper in oil capacitor with a metal can.

Craig