Tubes4e4,
You are strictly correct in your isolating of the two modes of design. but I used the terms 'pentode' and 'beam tube' in their general concept. As indicated above, the usual meaning of 'beam tube' refers to a tube (power in this context) with control and screen grids aligned (i.e. of the same pitch then) plus beam forming electrodes. On the other hand the classic power pentode has a control grid of different pitch as the screen, so that one can find adjacent screen wires not 'protected' from the direct cathode emission. This has nothing to do with lack of precision in construction; it must simply occur in any pentode. While your description certainly allows four combinations if that was your intention, I have never noticed two of them in practice, thus one uses the two best known concepts mentioned anove. But true, descriptions are not exact.
Apology if my statement appeared strong; perhaps to have said the infrequent glowing of a screen winding in a pentode from collecting too many electrons may be a problem. To explain, I had this with a few EL34s out of many, and one or two EL84s. - and no, no specs were exceeded (UL operation within normal specs, no oscillation of any kind). The phenomenon disappeared when replacing with other tubes
Last edited:
Keit,
I have not noticed that (regarding suppressor wire alignment) - but I have not 'disected' that many pentodes. I would not be surprised at at least the same pillars being used; goes to economy
But in the ECL82 'pentode' part, the screen wires are aligned with every second turn of the control grid - at least in the tube I inspected (cannot remember the brand now). By that I mean a full screen turn the same 'pitch' as g1, (to get alignment), then a gap of one turn, then another full turn. I can't recall anything appearing compromised in the transfer graphs. It was a beam tetrode, with beam forming electrodes as normal, no suppressor grid. As said I vaguely recall the same structure for an ECL86, but that was decades ago.
But I have steered matters OT, sorry.
I have not noticed that (regarding suppressor wire alignment) - but I have not 'disected' that many pentodes. I would not be surprised at at least the same pillars being used; goes to economy
But in the ECL82 'pentode' part, the screen wires are aligned with every second turn of the control grid - at least in the tube I inspected (cannot remember the brand now). By that I mean a full screen turn the same 'pitch' as g1, (to get alignment), then a gap of one turn, then another full turn. I can't recall anything appearing compromised in the transfer graphs. It was a beam tetrode, with beam forming electrodes as normal, no suppressor grid. As said I vaguely recall the same structure for an ECL86, but that was decades ago.
But I have steered matters OT, sorry.
EL84/6BQ5 are just ill/sounding mainly imo haha, while pretty different than 6v6... hell even 6F6s sound better than EL84/6Bq5s.... EL84/6BQ5s always/really sound rather compressed/bland imo....
and 6v6 sounds closer to EL34 like as well, imo. and EL84/6BQ5 can't be abused/voltages either they go to hell real fast haha, while many 6v6s can tolerate waaay over-spec conditions same with some 6F6s haha, that's when they start sounding better.
and 6v6 sounds closer to EL34 like as well, imo. and EL84/6BQ5 can't be abused/voltages either they go to hell real fast haha, while many 6v6s can tolerate waaay over-spec conditions same with some 6F6s haha, that's when they start sounding better.
Last edited:
AudioFreak88,
Not to be unkind, but I have always wondered about the 'sound of a tube' (different for same number tubes) per se, almost as if it has a personality - sorry.
Such depends entirely on the circuit, except where obviously the tube is not within spec. In such cases it is unfair to judge a tube specifically by number full-stop. I have read that guitar playes comment about this, but in all cases which I saw where the experimenter took the trouble to measure the tube-under-test or in-circuit characteristics, 'different sounding' tubes were also technically different - some to a factor of 60% so! That indicates tubes out-of-spec, in which case the argument is obviously void.
Not to be unkind, but I have always wondered about the 'sound of a tube' (different for same number tubes) per se, almost as if it has a personality - sorry.
Such depends entirely on the circuit, except where obviously the tube is not within spec. In such cases it is unfair to judge a tube specifically by number full-stop. I have read that guitar playes comment about this, but in all cases which I saw where the experimenter took the trouble to measure the tube-under-test or in-circuit characteristics, 'different sounding' tubes were also technically different - some to a factor of 60% so! That indicates tubes out-of-spec, in which case the argument is obviously void.
If I was really worried about the difference, I'd build up an amp circuit that had sockets for both the 6BQ5 and the 6V6, each running pushpull, with parts optimizing each tube, no negative feedback anywhere, except an optional ultralinear mode test, and then listen to it for a while on each tube, same songs. All my guitar amps have no feedback and sound relatively Hi-Fi when nothing is overdriven. In the end, for a hi-fi power amp you'll probably want to have at least 12dB of negative feedback in order to have a very low output impedance driving the speaker and its source impedance sensitive crossover. But if you don't hear a difference with no feedback, then I'd just go after availability and reliability. The 6V6 might be more reliable, and I think the guitar amp industry will keep both tubes in production for many years to come. .
6P14P-EV run fine both in triode and UL at 350V. More voltage than that I don't think is going to be so useful with these tubes.
I actually still have a small PCL86 spud amp in my annexe which I run to listen to the radio while I am at my workshop. These tubes have also been running at 350V/23-24 mA triode-strapped into 8K for 4000-5000 hours at least. Those PCL86's are branded Polamp (Polish Philips factory). 2 full watts of Pout at 5% THD and work absolutely fine.
Using too little NFB decreases lower order harmonic distortion but increases higher order. Using ample amount of NFB reduces both. If you aim for low distortion figures then use sufficient amount of NFB. ...Which unfortunately always isn't an option with tube-based designs.
Actually, that's wrong in two aspects.
Firstly, NFB creates harmonics where none existed before. Its isn't just a matter of making existing higher harmonics higher in level, NFB makes new harmomics, and makes low harmonics before it makes high. And makes them more and more as you increase feedback.
Consider an amplifier with a single active stage containing a device whose output voltage is at least in part proportional to teh square of the input voltage. With high school trigonometry, you can show that the output will contain the input fundamental and some second harmonic. There are no other harmonics. Vacuum tubes closely approximate such devices, as do junction FETs.
Now, apply some negative feedback. The feed back adds the 2nd harmic to the input fundamental, generating sum and difference frequencies - which in this case is DC plus 3rd harmonic. So you get a less 2nd in teh output, but some 3rd. Now the 3rd gets back in the input and generates 4th and 5th. ...and so on. Where NO 3rd, 4th, and 5th existed without feedback.
What happens is essentially this: As you add a little bit of feedback, the 2nd goes down and the 3rd rises up. Still more NFB and the 2nd goes down more, the 3rd starts to go down, and the 5th starts comming up. Still more feedback, and the 2nd goes down an bit more, teh 3rd goes down as well, the 5th starts going down agian, and the 7th and 9th start increasing in level. And so on, until with lots of feedback you get 2nd, 3rd, 4th, 5th, 6th, ....20th, 21st, 22nd, etc etc all at similar low levels.
The second thing wrong with your opinion, Teemuk, is this: Applying ANY neg feedback ALWAYS makes a practical amplifier sound better, added harmonics notwithstanding. That's because while the ear is somewhat sensitive to harmonic distortion (harmonic distortion is also generated within the middle ear, and at high sound levels, at the ear drum), the ear is VERY sensitive to intermodulation distortion.
And NFB allways reduces intermod distortion, in proportion to the amount of feedback, because NFB straightens out the input/output transfer characteristic in proportion to the amount of feedback. That's why millions of 5-tube table radios, radiograms, battery tube radios, etc were made with a only few dB of neg feedback. Because it made them sound better, without sacrificing too much gain. More feedback would be better still, except that the cheap output transformers would have caused stability isues, and because there was not enough gain with just a triode voltage amp and a 6V6 or 6AQ5 tetrode output stage anyway.
Last edited:
Other things being equal, the effect of NFB on IM is the same as its effect on harmonics. This is because IM and harmonics arise from the same source: nonlinearity. Second-order IM arises from two different frequencies mixing together in a second-order nonlinearity; second harmonic (plus some DC) comes from one frequency mixing with itself in a second-order nonlinearity. So NFB creates, and then reduces, higher order IM in exactly the same way as it does for harmonics.Keit said:
Really? Have you done the math? Have you done the measurements? I think not. Other things aren't equal. There are many circumstances in practical/real amplifiers where THD is unrelated to IM, and many circumstances where the effect of NFB is different. And those simple 5-tube table-top radios made in the millions in the 1940's and 1950's, with their 3 or so dB of feedback (way below what is typical in a high quality stereo amp) are a set of such circumstances.
Those that claim small amounts of feedback, (which do add harmonic energy not present in the basic audio stage) is worse than none at all should ask why manufacturers did it. If it didn't make the radio sound better, manufacturers would have simply saved some money and left it out. Now, ask yourself WHY did those radios sound better with feedback?
You have your terminology somewhat scrambled anyway. The order of an IM product is the sum of the order of the apparent harmonic components. E.g., for two input signals f1 and f2, the result correspoinding to a frequency of 1f1 +2f2 (ie fundamental of f1 plus 2nd harmonic of f2) is referred to as 3rd order. A frequency of 2f1 + 2f2 is termed 4th order. The reason for this terminology is that the level of a product in a simple active device rises with signal level in proportion to the order: 3rd order frequencies rise three times the rise in signal level for example.
Last edited:
Yes. Have you? I am from the generation which did lots of algebra and trigonometry as a teenager so I am well aware of how harmonics and IM arise from non-linearity. Harmonics are just a special case of IM, where the two (or more) input frequencies happen to be the same.Keit said:
I would say some circumstances, not many circumstances. I did say "other things being equal". The main source of both harmonic and IM distortion is ordinary non-linearity, which is what NFB is meant to improve.
In Europe most of these had no feedback at all. I suspect the same is true of most US models.
No scrambling. f1+f2 is second order, as is f1-f2. If f1=f2=f then you get second-harmonic (f+f) and DC (f-f).
Take a pure second-order device. Then f1 and f2 in the input give f1+f2 and f1-f2 in the output, so IM is present without feedback. Add feedback, then you will also get
(f1+f2)+f1 = 2f1+f2,
(f1+f2)-f1 = f2, *
(f1+f2)+f2 = f1 +2f2,
(f1+f2)-f2 = f1, *
(f1-f2)+f1 = 2f1-f2,
(f1-f2)-f1 = f2, *
(f1-f2)+f2 = f1, *
(f1-f2)-f2 = f1-2f2
* are components already in the input - these usually give rise to compression, as their total level now varies non-linearly with input level
No. Feedback was very common. In Europe, partly because Europe invented the pentode and Philips controlled key patents, Pentode output was almost universal in European radios, while beam tetrodes were virtuall all you ever saw in America and elsewhere. It happens that distortion is a greater problem with pentodes than it is with beam tetrodes. So what happend is that most American 5-tube sets have a few dB of feedback, but the very cheapest, especially 4-tube sets (which had barely enough gain) often did not.
In Europe, pre-war, it was common for pentode distortion to be slugged down with shunt capacitors across the output transformer primary (which also killed the treble response), but post-war feedback became pretty much universal. Probably not a coincidence that the audio bandwidth of AM transmitters in Europe and UK was usually 50 Hz to 5 kHz, whereas in the USA and elsewhere (including Australia) it was 50 Hz to 10 kHz. Not much point in building good treble response into a radio if there is no treble broadcast. Post war there was FM, which changed the picture. With Europe's pentode output stages, feedback with FM's 50 Hz to 15 kHz bandwidth was then essential. A few dB of feedback makes all the difference between objectionable and nice.
No. your terminology was scrambled. Taking your example listing of products which can arise in a device having a square law transfer response and feedback, all in your list are 3rd order, which I'm sure you know, so I can't see what you are getting at.
You terminlogy in your earlier post was scrambled because you said that the IM order is the order of the device transfer. A different thing.
And yes, I have done the math, AND the measurments. Lots of measurements. I started out in my career in the early 1960's as a radio & TV repair tech, but obtained an Honours degree in electronic engineering some years later. The course taught all aspects of audio equipment engineering, including distortion. I have designed professional grade audio equipment (studio and radio staion equipment) in part of my career. Done commisioning and purchaser engineering representation (ie signing off that what the factory sent meets the specification) as well. However, the electronic manufacturing industry in Australia started to die and just about dissappeared by 1985 or so, so I made a career change into electrical power generation.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.