Hacker News Reader

Comment by ptha | Hacker News Reader

Hard to believe the Soviets were using vacuum tubes instead of transistors In hindsight, the MiG 25, which the West had been so worried about, turned out to be a ‘paper tiger’. Its massive radar was years behind US models because instead of transistors it used antiquated vacuum tubes (a technology that did, however, make it impervious to electromagnetic pulses from nuclear blasts). The huge engines required so much fuel that the MiG was surprisingly short-ranged. It could take-off quickly, and fly in a straight line very fast to fire missiles or take pictures.

View on HN

The factoid about resistance to EMP gets repeated endlessly online and in print, but it doesn't really make sense, and I doubt it's true. EMP damages electrical connections which are:

a) long. b) thin. c) far from a ground plane.

In more detail: the induced current depends directly on the length of the conductor; a thin is more easily overheated and damaged than a thick one; and conductors running near ground planes make poor antennas for picking up pulses.

Now, all three of these factors describe vacuum tubes more than solid-state electronics. Vacuum tubes have to have very thin heated cathode wires, and the dimensions of vacuum tubes result in a substantial distance to the ground plane and in long conductors. Compared with solid-state electronics, which, especially in modern design, can practically live on or between ground planes, this makes for a lot of sensitivity to EMP and external interference.

One sanity check is to consider electromagnetic noise emitted by the device. Any device which is emitting substantial RF noise from its internal components is also going to be vulnerable to external EMP or interference, since antennas work both ways - anything which is good at emitting will also be good at receiving. Now, vacuum tube devices emit more RF noise than modern solid state devices, and generally need more shielding.

The components in the MiG-25 radar had more to do with what was available to the Soviet designers in the 1950s. The radar of the MiG-25 came from the Tu-128 interceptor which first flew in 1961, which means that the radar was designed in the mid-1950s at the latest. Wouldn't surprise me if the US radar of the same vintage also used vacuum tubes.

The difference is that tubes saturate differently and they won't die due to high instantaneous voltages.

Like I said - people say this all the time, but a lot of vacuum tube equipment did die in EMP testing (in the USSR in 1964), and I don't see why vacuum tubes should be less vulnerable than solid state electronics to either short or long pulses (nuclear EMP contains both). A glass vacuum tube will pick up orders of magnitude more interference than a small transistor sitting inside its grounded metal can (and it's weight-prohibitive to build a metal shield around all your vacuum tubes), and the high voltage induced can damage either the vacuum tube itself, or something else inside the device. Unfortunately, very little has been published on this (although we do know that militaries test their modern solid-state devices with EMP, and they generally pass), and any online search produces a deluge of statements to the effect of "vacuum tubes are impervious to EMP" without anything to back that up. If there was ever a direct test of vacuum tubes vs. solid state, that would be very interesting to find.

Transistors of the 1950s and 1960s were particularly vulnerable to high voltages. Modern transistors and ICs have some resistance to high voltage discharges such as static electricity, but back in those days, you could easily destroy entire circuits with your hand. Early transistors were made with materials and processes we would consider sub-standard today, which is particularly significant because they were almost exclusively using bipolars instead of MOSFETs (which are much more durable), and the manufacturers of the 60s could not and did not add protection diodes. ESD control was one reason why certain American companies got so far ahead of the Soviets (and the less-adaptable competition such as Phillips, whose workforce refused to adopt ESD control practices).

True, and electrostatic discharge from your finger is no joke - it can be tens of kilovolts and carries significant, if not huge, energy. However, a discharge like that going right into a very small device, and the actual die of the transistor is tiny, is very different from interference/EMP damaging said device, where the small size actually offers protection. After all, a completed transistorized device from the 1950s is not vulnerable to static discharge - otherwise, you couldn't pick it up. With everything connected to ground where it should be, it is quite robust.

Modern ICs are much more durable and resistant to ESD than devices from those days; you can run an electronics manufacturing facility (not a semiconductor fab) without much ESD protection these days, whereas that would've been a pipe-dream in the 60s. They are sending consumer and commercial-grade semiconductors into space on satellites with service lives over 12 months these days; back in the 1960s, all satellites required rad-hard components (and had very short service lives). Semiconductors are much better these days, due to improvements in materials, processes, testing, and inspection equipment.

People underestimate the energies produced by nuclear EMP. Car magnetos were dying in the trials, power lines and telegraph wires were getting shot through. Your nice tube amp would have no chance.

Do you have a reputable source for any of that? EMP effect is measured by field strength in volts per meter. If you don't have an antenna, or something to act like an antenna (telegraph/telephone/electric power lines) connected to a piece of equipment, the induced voltage in the equipment cannot be extremely large.

Solid state electronics are extremely susceptible to even mild overvoltages. Tubes are not. And magnetos are not. Low-tech magnetos ignition systems are considered to be practically invulneable to EMP.

> I also must re-emphasize the fact that during Soviet high-altitude nuclear tests over Kazakhstan in 1962, rugged diesel generators having no solid state parts were burned out by E1 EMP.

http://www.futurescience.com/emp/vehicles.html

Magnetos have induction coils, that's sufficient.

Typical E1 pulse component has up to 50KV/m at ground level and is a high-frequency spectrum, you don't need a long antenna for that. EU railway field immunity test is 10 volts per meter and huge portion of industrial electronic equipment fails that already.

Stuff that burns power lines are lower frequency (and lower yield) E2&E3 components.

https://cdn.selinc.com/assets/Literature/Miscellaneous/EMP-T...

Not reputable. Does the author have any conception of diesel engine theory? I do. Prior to computer controls (circa 1990s), the only active electrical component involved in keeping a diesel like the one in my Audi 5000 running, once started, is the fuel shutoff solenoid. Jam that open mechanically, and you can throw the alternator and battery away.

Your Audi has an alternator, just as the diesel generator mentioned there. It has no active components, just stator coils and possibly a rectifier, yet it was cooking up.

Yes, it has an alternator, and as I already explained, if you push-start the car and jam the fuel cutoff solenoid, you don't even need the alternator (or the battery). OK, you need them at night, for the headlights.

Those windings and diodes will carry 100 amps continuously. The E1 pulse isn't going to touch them. It doesn't last anywhere near long enough for them to even begin heating up. The E2 and E3 pulses aren't even a factor.

I can't provide a citation, but a few months ago I spoke with a retired EE who had specialized in large-scale/high-power systems[1], and he believed that tube electronics would better resist an EMP. i.e. if that belief is incorrect, it's something even experts in the field believe.

[1] one of his example projects was city-block-sized capacitor arrays that can temporarily provide power to an entire city.

You're fundamentally wrong on this.

What you are talking about is true only for inductive coupling caused by fluctuating EM waves.

As someone who works in Embedded Electronics and Software for Aerosopace let me drop a little knowledge bomb.

The problem isn't with induction, all the units can be easily shielded from this and they are in EMC chambers every day. A simply faraday cage, some bypass and reservoir capacitors a choke here and there an opto-isolator on box boundaries and bobs your uncle.

The real problem with High Energy Particles (Neutrons, Gamma particles, Protons, etc) is that they displace atoms in your semiconductor lattice and change the doping densities and behaviours of your junctions. Causing your semiconductor to fail... HARD.

When we buy in components, we tend to buy rad-hard devices which have special considerations taken into account when designing the semiconductors because when you're in space, you're bombarded with LOADS of high energy particles all the time, and its not just enough to have memory TMR, you need RAD-HARDENED semiconductors.

Vacuum tubes and transistors have different failure modes. If you apply high voltage to a vacuum tube, it will arc and warm up, but not sustain much damage. Do the same with a transistor and it will promptly turn into a resistor.

Source: burned plenty of transistors and vacuum tubes while studying electrical engineering.

We were _really_ preparing for nuclear war. While US military adopted the MAD doctrine, considering their nuclear weapons just the ultimate deterrent, Soviet generals thought that nuclear war can actually be fought — and won.

I am quite surprised that we are not living in Fallout by now.

Pretty sure the US and NATO had "flexible response" where they would have options other than MAD.

NB I'm having a binge of reading novels about WW3 scenarios in Europe at the moment, my favourite being "Chieftans" by Bob Forrest-Webb, mind you it doesn't end well.

[NB Rather hoping WoT gets to Challengers/Chieftans/M1s vs T-80s at some point...]

Maybe after the Cuban Crisis. But Kennedy basically had to tell his advisors to shut up, because they were slavishly following RAND's escalation policy.

Kennedy repeatedly dismissed other opinions and had to change his positions after being proven wrong. He accused Eisenhower of allowing the Soviets to accumulate more missiles than the USA, and found out he was wrong after getting elected. Kennedy then supported the Bay of Pigs assault, but pulled back at the crucial moment when the rebels needed support, and the whole affair was a disaster. Kennedy even refused to admit that he had agreed to all of Kruschev's terms in relation to de-escalating the Cuban missile crisis, and pretended that it was a victory of American diplomacy and intimidation.

Castro was also a troubled character who actually advocated using nuclear weapons against the USA, and had to be persuaded that it was a bad idea because the fallout would hit Cuba.

" Kennedy then supported the Bay of Pigs assault, but pulled back at the crucial moment when the rebels needed support, and the whole affair was a disaster."

Kennedy was handed the Bay of Pigs by Ike. I've wrote up a timeline of the Bay of Pigs (citations included) about 3 weeks ago ~ https://news.ycombinator.com/item?id=12324273

"Castro was also a troubled character who actually advocated using nuclear weapons against the USA, and had to be persuaded that it was a bad idea because the fallout would hit Cuba."

Maybe, the decision to use tactical nukes in the battlefield was left to commanders in the field.

Their tank expert is an actual tanker:

http://www.militarytimes.com/story/veterans/best-for-vets/20...

He's got a Youtube channel that has fascinating reviews of WW-II era tanks, Watching him try and fit his over-6' height into some of them is hilarious.

https://www.youtube.com/user/TheChieftainWoT

Something to take note of was that analog systems could drive far more power than solid state systems at that time, especially given the USSR's relative backwardness in solid state electronics. It really is worth noting how montrously powerful that MiG 25's radar is. It could pump out 600kW. Western fighter radars top out in the low 10s of kW. (they went with raw power to burn through jamming).

I think the whole EMP robustness thing is overplayed everytime something like this pops up - I think that at most the EMP robustness is just a side benefit to the analog systems.

Analog is not the opposite of solid state. You can build analog systems with solid state components.

To further this point, it is often said that "at high speed, everything is analog".

Or "Digital circuits are built from analog components"

It might not be extremely out of date. According to some random wiki[0],

"In 1963 Westinghouse was awarded a US Navy contract for the AN/AWG-10 to provide fire control for guns and Raytheon AIM-7 Sparrow and Raytheon AIM-9 Sidewinder air-to-air missiles on the F-4B Phantom II. The first radars were delivered in 1966 and became standard equipment. The AN/AWG-10 was claimed to be the first interceptor radar to feature transistorised circuitry."

Meanwhile, the MiG-25 first flew in 1964, which would suggest that at that time it was developed western interceptors used tube-based radars too.

[0] http://wiki.scramble.nl/index.php/Westinghouse_AN/APQ-72

Most aircraft make their initial flights without their first avionics suite (including radar and weapons control systems). Since the MiG-25 was introduced into service in 1970, and the radar it used was a development of another first used on an aircraft that entered service in 1964, I would hazard to guess that the Foxbat's radar wasn't delivered until 1965-1969.

Reply on news.ycombinator.com