Tube Question, Ancient Technology.

[Bob Scott]

A long, long time ago, my Dad told me that tubes were voltage controlled voltage sources (VCVS). Seeing as how tubes are obsolete, I didn't really care but I believed him. Regular bipolar transistors are current controlled current sources (CCCS) and FETs are voltage controlled current sources (VCCS).

Just recently I got interested in tube theory because I had just seen a bunch of old schematic diagrams of tube electronics in a bunch of HAM books I bought off Ebay. (Also I have been spending a lot of time the last couple of weeks reading 1940's Popular Science c/w tube schematics.) Even in the schematics of Hi-Fi equipment in the 1975 RCA tube manual, I don't see any negative feedback loops anywhere. Even the cathode resistors are all bypassed by capacitors. It seems that all those tubes are running at maximum gain all the time with no thought given to the gain loss of aging tubes.

Then I looked at the transconductance curves of the 6AQ5. It's a pentode that can be wired as a triode. There are transconductance curves for both cases in the tube manual. The curve of the triode wired tube is that of a voltage controlled resistance, not a VCVS. All "curve" lines are pretty much straight lines eminating from the zero point like a starburst.

The pentode wired 6AQ5 curves look identical to the transconductace curves of a power MOSFET, meaning that if the 6AQ5 is connected as a pentode, it is a voltage controlled current source VCCS. (see curves of IRFZ44 MOSFET) Ah, my Dad the electrical engineer was not often wrong but I think he was on this point.

So, getting back to these amps with no feedback in the schematics... They do have feedback after all, but it looks to me like it is inside the triode amplifiers. With no screen or supressor grid voltages to shield electrons passing past grid #1 from the plate voltage, the gain of the triode is reduced when the plate voltage swings lower, then there is less attraction of electrons to the plate, and less output. So the triode does have built-in negative feedback. But it's so primitive!

And to all those audiophools who despise negative feedback but love their triode tube amps, I am smiling like Mona Lisa!

[CeaSaR]

Aha! Pure knowledge rules the day! Thank you, Spock.

As an aside, I was on Instructables the other day (I give N&V equal opportunity over there) and there
was a question about whether or not tubes were preferable to transistors in music "creation" situations,
ie. guitar amps etal. So by your discovery, it seems that MOSFETs are the most similar to tubes in their
"function" (for lack of a better term). Very interesting. Do you have any idea or proof of what the major
harmonics of MOSFETs are, odd or even? I'd love to know.

Does this mean you can troll the audiophool forums with even more bait?

CeaSaR

[Bob Scott]

Aha! Pure knowledge rules the day! Thank you, Spock.

As an aside, I was on Instructables the other day (I give N&V equal opportunity over there) and there was a question about whether or not tubes were preferable to transistors in music "creation" situations, ie. guitar amps etal. So by your discovery, it seems that MOSFETs are the most similar to tubes in their "function" (for lack of a better term).

Only pentodes, I think, and not EXACTLY. Grid voltages are in a different range than Gate voltages. ie: You cant plug a FET into a tube socket, but the basic function is the same.

But at voltages under 100mV, a FET acts as a variable resistor. I think that would act as a triode, but at a way lower voltage than a tube uses.

As for music creation, the musician chooses what tools or instruments he uses to create his art. Only his opinion counts.

Do you have any idea or proof of what the major harmonics of MOSFETs are, odd or even? I'd love to know.

MOSFETS own harmonics? Maybe you refer to the non-linear response of IR hexfets. Odd or even depends on the circuit. Class A amp using hexfets - even harmonics. Symmetrical push-pull, just odd harmonics.

There are Japanese mosfets that have linear responses also eg: 2SK1057, 2SJ161, so much less distortion.

Does this mean you can troll the audiophool forums with even more bait?

I could post that in the diyaudio.com tube amp section to separate the phools from the philes, to see who is which, if I were so inclined at the time, but I'm not tight now

[k7elp60]

Bob,
When I started out in electronics (early 1950's) I built a lot of stuff with tubes, and I remember a few things. The tetrode(cathode, control grid, screen grid, and plate) was interesting as the screen grid(closest to the plate) could control the electron flow dramatically. Since it was normally at a + potential there was some screen current. On some tetrodes if the screen has no + voltage there was no plate current.

On the pentodes(cathode, control grid, screen grid, suppressor grid, and plate) the suppressor grid was sometime tied internallly to the cathode, and as I remember it was designed to attract the electrons that bounced off the plate.

I used tubes that had battery operated filaments as well as the ones that were supplied with ac.
A typical table model superhet radio had 5 tubes. All the filament voltages generally added up to 117V and they were connected in series and it used the line voltage for power. There was even a time when tubes in auto radios were developed that worked with 12V on the filaments and 12V on the plates all DC of course.

About the biggest project I ever built was a regulated power supply to run an amateur radio 200W tube tranciever. My power transformer was way overvoltage, but I had lots of tubes so I used tubes as series regulators for HV(750V) and 250V( medium voltage) plus supplied the bias about -150V.
Ned

[Bob Scott]

Hi Ned!

Bob,
When I started out in electronics (early 1950's) I built a lot of stuff with tubes, and I remember a few things. The tetrode(cathode, control grid, screen grid, and plate) was interesting as the screen grid(closest to the plate) could control the electron flow dramatically. Since it was normally at a + potential there was some screen current. On some tetrodes if the screen has no + voltage there was no plate current.

That second grid was handy, like a second gate in a FET, for modulating two signals. Apparently not handy enough. I don't think that two gate FETs are made either anymore either.

On the pentodes(cathode, control grid, screen grid, suppressor grid, and plate) the suppressor grid was sometime tied internallly to the cathode, and as I remember it was designed to attract the electrons that bounced off the plate.

Yes, something like that. As the tube manual says, the suppressor grid actually repelled electrons back to the plate. Those electrons had bounced off the plate and were called "secondary emissions" if I recall correctly.....but you knew that already. We just find some old memories a bit dusty after a few decades of storage in the cranium.

I used tubes that had battery operated filaments as well as the ones that were supplied with ac. A typical table model superhet radio had 5 tubes. All the filament voltages generally added up to 117V and they were connected in series and it used the line voltage for power.

Ah, the AA5 radio. The "All American Five". I wonder how many hundred million were made.

There was even a time when tubes in auto radios were developed that worked with 12V on the filaments and 12V on the plates all DC of course.

I have never seen one of those. The older automobile tube radios I have seen had "vibrators" to boost the voltage through a transformer.

About the biggest project I ever built was a regulated power supply to run an amateur radio 200W tube tranciever. My power transformer was way overvoltage, but I had lots of tubes so I used tubes as series regulators for HV(750V) and 250V( medium voltage) plus supplied the bias about -150V.
Ned

It must have been warm in the room when that PS was running. I would have liked to see the station running back then in the heyday of amateur radio, when Hams knew how to build their own stuff, not just operate premade boxes.

[Dean Huster]

12.6v and 6.3v tube heaters were naturals for auto radios. Generally, vibrators were used in a "switching" supply to generate the B+ plate voltage. One of the electronics hobbyist mags (I think it was Popular Electronics) had an article for (I think) a vacuum tube oscillator that was a "starved" circuit, using rectified heater voltage for the plate voltage.

Bob, don't forget that unbypassed cathode resistors develop negative feedback and are used often in audio circuits while bypassed cathode resistors develop cathode bias voltage, similar to the way it's done in transistor circuits -- actually, the transistor circuits do it in a way similar to the way it was done in vacuum tube circuits, don't they?

Some audio amps used output transformer taps as feedback into earlier parts of the circuit, off the secondary as I recall.

[Robert Reed]

I have never seen one of those. The older automobile tube radios I have seen had "vibrators" to boost the voltage through a transformer.

Actually they were fairly common towards the end of the tube auto radios. I saw some pop up in the Maritime Service transceivers (2-4 Mhz DSB band) about that time. Of course the Tx final and its driver still operated on high voltage supplys. I can't recall the period but they were only around a couple years until they all went solid state. I never did understand how they could do that unless the element spacings were made super close.

[Bob Scott]

Hi Dean. It's good to hear from you.

One of the electronics hobbyist mags (I think it was Popular Electronics) had an article for (I think) a vacuum tube oscillator that was a "starved" circuit, using rectified heater voltage for the plate voltage.

Wow. I have not heard of using A+ heater voltage for the B+ plate. Your mention of "starved circuit" reminds me of an amplifier design of that name "starved amplifier" used in Electronics Illustrated magazine projects in the early '60s. I was just a 13 year old beginner hobbyist and did not eppreciate or understand the circuit at the time, but I think they generated massive voltage gain by using a very large value plate resistor, and used very small plate current so that the output would not saturate. I am assuming from the limited memory I have of those magazines.

Bob, don't forget that unbypassed cathode resistors develop negative feedback and are used often in audio circuits while bypassed cathode resistors develop cathode bias voltage, similar to the way it's done in transistor circuits -- actually, the transistor circuits do it in a way similar to the way it was done in vacuum tube circuits, don't they?

Yes Dean, but if you reread my first post I said:
"Even the cathode resistors are all bypassed by capacitors. It seems that all those tubes are running at maximum gain all the time with no thought given to the gain loss of aging tubes."

Yes, depletion mode JFETs can be biased just as if they were mini-tubes. You bias the grid/gate at ground or reference voltage. You know that.

Some audio amps used output transformer taps as feedback into earlier parts of the circuit, off the secondary as I recall.

I've seen that, using one output tap for feedback, and also some hifi amps had a separate winding just for feedback.

Ah, the olden days. I wish I could do life all over again, like "have another ride".

[Dean Huster]

Oops! I missed the "bypassed caps" description for some reason -- possibly age or near-death surgery.

Seems to me that bypassing all the cathode/emitter resistances is just asking for higher THD and it's contributing flaws of additional non-linearity and near-clipping.

Of course, we can get totally ridiculous and run everything between input source and speaker Class A. I'm not a tree-hugger, but 200 watts of Class A power? Eeeeesch!

[MrAl]

Hi,


So what is this thread all about, comparing MOSFETs to Tubes then?

Just to note, long time ago it was harder to resolve say 0.1v on a panel meter so the bipolar transistor
was usually said that it was a current controlled current device, but more recently the alternate view
that the transistor is also a voltage controlled current device has gained some ground too. The control
voltage is just smaller than usual, that's all. This view is also evident from some of the formulas that
have come up for the bipolar transistor. Myself, i still think of it as a current controlled device but i
keep aware of the fact that in some applications or for some other reasons it may be more useful to
think of it as a voltage controlled device.
Now that we more commonly have meters that can resolve down to 1uv we can look at things a bit
differently, plotting the output current vs input base emitter voltage and see a relationship that holds
for bipolars just like other relationships like beta. This relationship is often referred to as the
bipolar's transconductance (gm) also: gm=dIc/dVbe .
Personally, i've never needed this except when involved with the theory of the transistor.