The inexorable march of technology

[-]

Kirkwood1994@reddit

the vintage audio fanboy in me wants to know what those sound like in audio gear. Assumping they look like 12A(X/U)7?

[-]

They used to be available for a dollar each, by the thousands. Eventually the audiophiles convinced themselves that these digital tubes sound good. The number is 6211, optimized for long life doing on/off service not amplifying audio signals.

[-]

MechaGoose@reddit

I still don’t understand how vacuum tubes work in audio world

[-]

Rezrex91@reddit

Vacuum tubes operate just like (and they are the precursors of) transistors. Transistors and vacuum tubes in audio are used mainly as signal amplifiers. They can be used as such because they're acting like electronically controlled dimmer switches or valves. So you can control them with an audio signal (a time-varying AC voltage) and put a much larger DC voltage to their input, and your output will be a much larger DC biased AC voltage (meaning that it has a DC component still that needs to be removed, but that's neither here or there, just wanted to make clear that it doesn't make clean AC from the DC input) with the same waveform but higher amplitude than your input signal.

[-]

m-in@reddit

Vacuum tubes specifically operate very much like junction FETs (not MOSFETs), up to and including the grid and cathode acting like a diode — just like the gate-channel junction acts in a FET.

[-]

flatfinger@reddit

One thing I was curious about: in the Falstad simulator (which treats tubes as having a magically powered filament with no connection), tubes seem like they're capable of producing power from the non-filament pins. At first I thought that seemed ridiculus, but then I realized that since tubes do have a separate source of energy (the filament), it would seem plausible that tubes might be able to convert some of that energy into electrical energy elsewhere; probably not a usable amount, but measurable.

Do you know if tubes actually work that way?

[-]

m-in@reddit

I don’t quite know what you’re trying to say. Draw a circuit of how the tube is connected and where you measure it. Then we can talk :)

[-]

flatfinger@reddit

Falstad's simulator offers unlimited energy, which is obviously not realistic, but the essense is that when the voltage on the grid is higher than the voltage on the cathode, this will drive current to the plate even if the plate voltage is slightly below the cathode voltage. This would seem plausible, to a limited extent, but with the caveat that electrons leaving the cathode would cool it down, and the energy required to overcome that heat loss would exceed the energy that could be harvested from the plate current.

[-]

Rezrex91@reddit

You're right of course, but that would've been a more in-depth explanation than what I was going for, and to be honest, I myself aren't particularly clear on the differences between the various types of transistors and tubes, so I didn't want to write something wrong.

I was trying to focus on the operating principles of the most basic transistors and tubes, meaning that they are a valve for electricity, controlled by electricity, and how this can be used for amplification by using a higher voltage/current for the input than what's used for the control.

[-]

OddbitTwiddler@reddit

Just like...well grandads vacuum tube pacemaker was the size of a refrigerator and took 100Amp 220V. The good news is the how's was warm in winter.

[-]

Screwthehelicopters@reddit

Essentially, vacuum tubes are similar to transistors; three-legged devices with one controlling current flow between the other two.

In a tube, current flow is between an anode and a cathode and these are contained in a vacuum tube. When the anode voltage is high and the cathode is heated (by a separate heating coil), then current flows as electrons are "boiled off". That flow can be varyingly blocked by a voltage-controlled grid/plate. The voltage could be an input signal then controls the main flow to provide amplification of that signal.

Vacuum tubes are more bulky due to the glass, size of plates and the separate heater which are needed to exploit the physics of how they work.

[-]

MechaGoose@reddit

Amazing explanation, thank you!

[-]

Rezrex91@reddit

You're welcome!

[-]

Screwthehelicopters@reddit

Essentially, vacuum tubes are similar to transistors; three-legged devices with one controlling current flow between the other two.

In a tube, current flow is between an anode and a cathode and these are contained in a vacuum tube. When the anode voltage is high and the cathode is heated (by a separate heating coil), then current flows as electrons are "boiled off". That flow can be varyingly blocked by a voltage-controlled grid/plate. The voltage could be an input signal then controls the main flow to provide amplification of that signal.

Vacuum tubes are more bulky due to the glass, vacuum, size of plates and the separate heater which are needed to exploit the physics of how they work.

[-]

MechaGoose@reddit

Interesting! Thanks for explaining

[-]

Screwthehelicopters@reddit

In a way, transistors are harder to understand because there is nothing to see other than the three external connectors. But at high level, they work in a similar way.

[-]

m-in@reddit

Most tube amps are shit from signal quality perspective. People like how they color the sound for some reason. It’s all a circlejerk. I have no problem with everyone listening the way they want to. I do have a problem with people scavenging historic test & measurement and computer gear for tubes. Buy your own new ones and pay for it. They are still made.

[-]

mimavox@reddit

Well, they certainly look cool.

[-]

jdx6511@reddit

The 6211 has a mu of 27, so more than a 12AU7 (20), but substantially less than a 12AX7 (100).

[-]

justananontroll@reddit

And the bottom case is really only that big to make room for the solder pads.

[-]

voxadam@reddit

This needs a Wafer-Level Packaging example.

[-]

nixiebunny@reddit (OP)

It needs the 1-micron-square portion of a modern chip required to duplicate the 704 circuit Z

[-]

m-in@reddit

More like a 50x50nm square and that’s being super generous. 1 micron square would have been the case maybe in early ‘00s.

[-]

NightmareJoker2@reddit

Yeah, except you guys forgot about the capacitors, resistors and inductors on die that surround the actual logic on the chip to facilitate interfacing with it at the electrical level the chip supports at the package pins. There are limits (physical) to how small you can make those, and you need them to prevent the chip from burning out. And since you’re doing that anyway, you might as well make the logic traces a bit larger to need less of that. NXP still makes 74-series logic, and they hold a careful balance between wafer space, power efficiency, and acceptable voltage ranges in the 74ACT series. Which is fully compatible with any 74-series part of the same number, and can operate at 250MHz, possibly more.

[-]

m-in@reddit

I laid out a few GPIO pins by hand and there were no resistors per se, nor capacitors nor inductors. The only time I had to stick a capacitor into a digital layout - admittedly long time ago - was for the substrate bias charge pump.

[-]

NightmareJoker2@reddit

Some designs put them outside the chip. If you are going as tiny as nanometers, though, you need to have something to step down input voltages and step up outputs. Especially considering 74-series calls for up to 20mA per output pin to drive an LED and then some in a prototyping circuit.

[-]

nixiebunny@reddit (OP)

There is a lot of circuitry in the passives section below the tubes. Consider each tube to be a register, and all the gates are below, implemented in diodes and resistors. But yeah, it’s not much stuff.

[-]

m-in@reddit

Each tube is not a register. The nice bit about cmos is that all those passives are not needed.

[-]

Useful_Resolution888@reddit

You could just pretend that that's in the photo.

[-]

phire@reddit

Same with the through-hole chip. The actual die is tiny, sitting in the very middle.

[-]

crimaniak@reddit

So this is what the mycelium of radio tubes looks like.

[-]

Artful3000@reddit

That’s a huge byte.

[-]

Girderland@reddit

Is this basically the "same" part?

Same component, same function?

[-]

Skycbs@reddit

No not at all

[-]

spektro123@reddit

Most likely not but they could be. The naming isn’t consistent here. IBM 704 and CDC 6600 are mainframe computers. What we see here are cards for those. One is valve based and the other is transistor based. AFAIK those do one simple function like ICs described later. 74LS is series of logic ICs and what we see here is a DIP/DIL, through the hole package. 74LS is newer than older 74(nothing) series, which was introduced in 1964 (the same year CDC 6600 was). They (74 logic chips) usually do one thing like are being an and gate, flip flop or whatever. SOIC is a just smaller surface mount package type. 74 chips can be bought in this packages as well.

[-]

nixiebunny@reddit (OP)

Not the same but close enough. My friend refers to it as a computron, i.e. a basic unit of computing. The later decades got so small that you can pretend they are shown, but the camera can’t resolve anything that small.

[-]