Ah! So they were power tubes. Quite a different thing to a 3A5 battery tube.
Just what was the sonic benefit?
Unless you can identify it, describe it, or measure it, your claim is just hot air.
Tube sound = some even order distortion rising with signal level, some tiny microphonics (audible as very short reverberation = a "live" sound), a soft overload characteristic, some benign frequency response attributes depending on circuit design, good transient recovery provided the circuit designer knows what he is doing. Which is it?
Keit
Only 800 hours life? Less than a year at 3 hours per day? You must be joking!
300B's were professional tubes that in movie theater use lasted years and years at two full showings per day (2 features plus cartoon, about 3 -4 hours, matinee plus evening). Design life well above the normal consumer grade of 10,000 hours for AC tubes. Longer than 8 hours per day actually, the theaters I worked at switched on a hour before matinee advertised start, and didn't switch off until the evening showing finished. More like 10 hours per day.
Of course if you are going to buy Chinese clone junk, all bets are off.
If an expensive tube like the 300B does only last 800 hours, I SURE wouldn't make it any worse with extra junk like CCS feeding.
Keit
Last edited:
See various insertions by Mullard in Wireless World (published in England) from time to time in the 1950's.
See Mullard data long form datasheets for DL9x series tubes (DL9x is Euro numbering for American 1T4/1R5/1S5/3V4 series.
I quote from page 1 3rd paragraph of Mullard DL92 datasheet :-
Filament voltage, parallel operation (this refers to paralleling the 2 filaments of this double filament tube): 1.4 V
"With filaments connected in series... it is advisable to adjust the voltage across each filament to 1.3 volts."
I have an AWV (AWV = long lost Australian manufacturer that made tubes under RCA licence) application note that goes into this in detail. I see if I can find it for you. It is in fact an issue well known to those of us who worked in electronics BTTE (Before The Transistor Era).
Keit
Last edited:
Probably a lot of this was what I call 'Beetle Effect" quite common in audio.
The Volkswagen Beetle was a shocking vehicle, quite unsafe. Its pedals pivoted from below, not above like every other vehicle. It was probably ok in Germany, but here in Australia the air-cooled engines ran far too hot, causing exhaust valves to burn out. 6V electrics and cheap assembly meant dim headlights, often one very dim, even in relatively new cars.
Never the less, every Beetle owner I ever met reckoned they were the best vehicle. Psychologists know about this, the human mind seeks to justify spending money even if it was a foolish purchase.
Keit
Last edited:
Not relevant to whether CCS is good.
If I want to compare a 3A5 with a 300B etc, well I have the manufacturers curves from RCA (3A5) and Western Electric (300B). The 3A5 curves are a little less straight. But some folk like a bit of even order distortion. It's cancelled out in class A push pull operation. That is their privilege. It a bit strange to compare a miniature battery triode with a large power tube anyway. A bit like arguing which is better, a motor scooter or a Mack Truck.
Keit
Last edited:
This paragraph does not appear in any DL92 Mullard data sheet that I can see:
http://www.mif.pg.gda.pl/homepages/frank/sheets/154/d/DL92.pdf
or
http://www.mif.pg.gda.pl/homepages/frank/sheets/129/d/DL92.pdf
and it does not mention current-driven supplies at all, in either case. If fact it specifies Vf in both serial and parallel cases, which by convention, means filament is specified at a given voltage.
Please supply some actual supporting data.
Nobody is suggesting you substitute 300B for 3A5.
I am saying that 300B is a low-distortion DHT, and 3A5 is less so. Pay attention!
Yes, some of them are very bad, even Russian or European types, sad to say. WE originals are long gone. The EMLs appear to be very good quality, albeit at a high price. However, if you are too tight-fisted to lay out some cash to try some different output valves in your endeavour to get the sound you are seeking, then my advice would be: don't get mixed up in this hobby.
Make all the foolish remarks you like about current-driven heating. Others may also remember your admission that you don't have a DHT amplifier, and have never designed one, so your remarks can be taken in their proper context.
These are the short form catalog data sheets. I was referring to the long form data sheet as intended for design engineers. Pay attention.
The reference to "parallel" and "series" in these datasheets refers to operating the two filaments in the same tube in parallel or series (ie for 1.4V or 2.8 V operation) and has nothing to do with series current-fed heater strings used in battery/mains radios, an application that was not anticipated by RCA when they devised these tubes. As I'm sure you know.
About the only time you see reference to current driven filaments or heaters in datasheets is where the tube is specifically intended for use on telephone company long distance transmission repeater amplifiers, where constant current feeding from the ends of the cable/coax was the only way.
Economics made series heater strings in cheap TV sets a reality. Tubes for this application were certified for a specific warm-up time, quite a different issue.
US-made 300B's were still being produced, using old WE jigs and process sheets, at least until quite recently. It's still advertised on their website, presentation polished wooden box packaging and all! Not sure where they get their raw materials from. The US manufacturers of special tube glass (very slightly conductive, to prevent flashovers) dumet wire, high purity nickel, etc have long stopped producing same - insufficient market.
I see you are now resorting to emotional stuff getting very close to personal attack. The surest sign that you got it wrong and you know it.
Stick to the facts. If you have identified a sonic benefit, describe it, or describe how it was identified. Otherwise, your recent comments have no merit.
Interestingly, the WE website says in the 300B abridged data section, under heading "Filament rating": "The filaments of these tubes are designed to operate on a voltage basis". The meaning of this is a bit obscure, and they don't say why, but presumably it means they want you to feed it from a voltage source and not a current source.
Keit
Last edited:
OK, so you have no evidence that current-driven heating shortens the life of a filament.
That was the claim you made (in Post #36) - and is what I asked you to substantiate.
That was the claim you made (in Post #36) - and is what I asked you to substantiate.
Theres no doubt that in DHT's where the heater is also a part of the signal path ,then feeding with avoltage source will cause the heater current being modulated by the signal, this is something you completely avoid with current driven filaments.
Not only did I substantiate it, I explained why it is the case. (CVS feeding - filament temperature is in a thermal-electric negative feedback loop regulating the temperature; CCS feeding gives a positive feedback accentuating temperature variation and hot spot behavior)
In contrast, you have not explained anything, just made unsubstantiated claims that there is (an unidentified) sonic benefit with CCS.
This is quite wrong. Two things:-
1) Rod made the point that the modulation is in any case too small to show up as audible (or perhaps even measurable) distortion - which is entirely true as I agreed.
He did say there was a sonic benefit, but when asked to identify or describe, resorted to emotional attack and irrelaventcies.
2) As I also said right at the beginning, if you feed from a CVS (the normal time-honoured way), the signal current splits and leaves at both filament ends, so the modulation is reduced. Near the centre of the filament, the modulation by signal current is near zero.
If you feed from a CCS, all the signal currents leave via one pin, which increases the modulation. The modulation at the centre is the same as at one end with CVS, and at one end the modulation is doubled.
You can configure it so signal current still leaves at both ends, but then there is no improvement over CVS feeding.
Note that power is to the square of current (assuming resistance is relatively constant over the cycle, which Rod pointed out is the case), so CSS feeding is actually even a little worse than it at first appears. Without going into a lot of math, which is hard to type in this forum, the average modulation along the filament length with CCS is as 0.25 x total signal current (= anode signal current) and with CCS it is 0.5 x total signal current. But heating is as power, so heating modulation effect is as approx. twice the (0.5 - 0.25) difference.
If filament current is truly constant, including at both ends, as with a CCS that has a high dynamic impedance, then there cannot be any signal current in the filament/cathode and thus the tube cannot be working at all.
Keit
Last edited:
I've tested both, and I can assure there quite a big difference in voltage and current drive on the filaments of a DHT, even though you say the modulation is negligible. Lucky there’s more to sonics than something that is in measurement terms quite small. same as theres a big difference in signal wires that are the same, but where shield and ground is doubled.
Can you describe the difference? What does it sound like? When you made the change, did you change anything else, eg time of day, different tune being played, etc?
True, but the difference can generally be described, and measured easily with the right method. Eg 2 amplifiers with almost identical low THD figures sounding very different - this can be due to crossover distortion, easily measured with the right method.
In this case I can imagine a few modes where this might be the case. E.g., earth loop effects. Earth loops can not only introduce audible hum - they can caused non-linear distortion as the output device currents in Class B and Class AB are not sine. And they can (but not often) result in intermod distortion, to which the ear is sensitive, but not measured at all on a Distortion & Noise tester.
The moral of the story is this: If you can genuinely hear a difference, it can be explained, and it can be measured if only you measure it with the right instrument.
As an aside, I always find it amusing that the debate on the merits or otherwise of harmonic distortion testing is debated endlessly, with all manner of theories as to why it doesn't seem to correlate well with audible distortion. The motion picture industry worked it out in the 1930's that intermod testing was a much better way, and codified it in their technical standards. Not that simple intermod testing is the be all and end all either. But there was NO crossover distortion in 1930's picture theater amps using the likes of 2 300B's in Class A push pull. Nothing to measure.
Let's try again. You claim that current-driven heating shortens the life of the filament. See Post 36. This is not true.
If you believe it is true, kindly supply some verifiable literature from the manufacturers, or some peer-reviewed texts. By verifiable, I mean a screenshot, or link to material that we can check and agree is from a relevant creditable source.
To make it clear: responding with quotes from data sheets which do not appear in the publicly available copies, or prolix paragraphs of "theory" - which amount only to hand-waving, and mistaken/unproven assumptions, are of no interest to anyone here - and will be ignored, without loss.
its quite simple, when you get more its better, when things contracts gets smaller and become woolly fussy its worse. does not take a time of day or any woodoo-doll to figure out. its bit like bouncing a ball on a concrete floor or on a carpet. easy to tell which makes the ball bounce better.
i made another improvement in gyrator, increase in sound detail and treble are sharp and clean. there is 6-7v on jfet.
123
SGD
replacement for mosfet in typical circuit:
Last edited:
Attachments
Last edited:
- Status
- Not open for further replies.