Hacker News new | past | comments | ask | show | jobs | submit login
Say Goodbye to the Last Vacuum Tube Product (electronicdesign.com)
112 points by tlb on Oct 22, 2014 | hide | past | favorite | 97 comments



This story really buries the lede. You can get 250W of radiated RF power out of a transistor! That is pretty neat. Back in college I got to hear how the magnetron would not ever be feasible in a solid state device, the free electron path length was too short. Another impossible thing, now available from Digikey. Also a bit scary as it makes some previously impractical weapons more so (like the MASER).

If they are cheap enough (and I suspect they are) they might make a good bird abatement device as well.


It'd be interesting to build a phased array of them (which you can't do with magnetrons) and try to aim at individual birds flying. Except, swarms of mosquitos instead of birds.


I think that's what Goji is doing with these power transistors. Go to the Freescale site in the Electronic Design article, or go here:

http://www.gojifoodsolutions.com/

Their claim is that they cook food more rapidly and uniformly (no hot spots).

The site says little about how the tech works, but guessing, I think their oven beam-forms and scans the food. By looking at the reflected power as a function of the scan position, you should be able to calculate the spatial power absorption profile of the food. That would let you control beam geometry and power to eliminate hot spots.

Would be fun to try an oven that has this tech. Beam forming at microwave frequencies has been common in military systems for decades, but I didn't think I'd see it get cheap enough to use in my kitchen in my lifetime. Yum!

Now about that poached salmon...


A press release[1] indicates the price is $26 - $37 each in 10k lots.

[1] http://www.businesswire.com/news/home/20141005005052/en/Free...


That puts the cost at over $100 for a 1kW oven though; seems like it will only be in high-end ovens at first.


Basically what happened with LCD panels and televisions. Once Freescale ramps up and someone else, like Triquint, enters the market the price will come down.


And also what happened with magnetrons in microwave ovens. They were so very expensive 6 decades ago. Now I can go to Costco and get a 1kw microwave source for less than $100. It's great for home-brew plasma deposition systems.


My samples are on their way, I can't wait. :-)


> Also a bit scary as it makes some previously impractical weapons more so (like the MASER).

Well, Raytheon has a crowd-control microwave gun already, but that's only available to very big buyers.


That's probably this thing:

http://en.wikipedia.org/wiki/Active_Denial_System

It operates at 95GHz, which is firmly in vacuum tube territory (gyrotron, crossatron, etc.) for some time to come.


Would this make desalination more efficient, and therefore cheaper?


No. A simple resistance heater is more efficient if all you want is heat.

But the best desalination units use a membrane not heat.


But with a well-controlled compact microwave emitter, you can build a compact, fast-reacting heater for shower. Finally.


You can, but there are easier ways. A low mass resistance heater for example.


Yeah but the pipes are still cold!


Are you going to Microwave the bird with that?


I have two tube headphone amps on my desk, one of which I just bought new a couple months ago right in front of me, and 3 other people in my office have similar tube amps. Probably not the norm, but still the death of tubes has been greatly exaggerated.

edit: photographic evidence: http://jdon.at/15kV9


A great many music products (compressors, amplifiers) still contain vacuum tubes.

Simply including a "tube" is often feature enough to help sales.


I can't find the link right now, but I've seen sales/marketing efforts _so_ determined to include "vacuum tube" in their features list, that they designed an amplifier with only opamps and transistors in the audio stages, but a single tube as a rectifier in the power supply...


Of course, because the tube rectifies more smoothly and makes the sound more pleasant as a result. They even light up the rectifier with DC to reduce humming noise...


Indeed, even order harmonics in your power supply ripple create measurably more musicality and a higher resolution soundstage... (I hate myself for even _thinking_ that...)


And tubes in UHF broadcast stations are in no way anachronistic.


But high-powered analog UHF transmitters themselves starts to be, considering that only major application is analog TV.


Is that true? I was under the impression that ATSC transmitters used tubes as well.


I think the message is that this is the last common product where a tube was required.


That was the first thing I thought also, but I guess you can't argue that tube amps (for headphones, guitars, whatever) is a pretty niche product these days.


I use one of those, http://hughes-and-kettner.com/products/tubetools/ and it's pretty sweet!


For guitars they are nothing but niche. In fact they are probably more guitar tube amps produced today than they were 20 years ago.


Less niche than they were 10-20 years ago. Lots of really good stuff coming out of China that isn't at all spendy.


They are not at all hard to find.


Not much is, anymore.


They have never been hard to find. Virtually every guitar store has carried a wide variety of tube amps since rock music went mainstream.


You know what electronic part is hard to find these days? Variable capacitors! At the caps needed to make radios. Sure, you can buy them in bulk or at ham fests, but I've found it frustrating as a hobbyist.


Because they are fiddly, hard to manufacture and expensive, while solid state voltage or digitally controlled filters and oscillators are cheaper and more reliable.


You can fairly easily build your own.


You're right. I've followed instructables on how to do it (e.g. some cds and some foil). However, I am mesmerized by the beauty of some old school variable capacitors. Wish they still made 'em.


We used to call them meat mincers :)

If you want good quality variable capacitors then your best source is flea markets and thrift stores, look for old tube radios that you can scavenge them out of. And if the radio is nice enough then maybe restore it!


I assume most news articles are really just reworded press releases with some background information, and this one seems to be from Freescale. There is even a link in the article. However, this article lacks a quote by a person from the company, an expert, or a related company in the field.


Nice! I just recently got a I+ too. How's your experience been so far?


Pretty happy overall. It's a little more receptive to noise than I'm used to, if my phone is anywhere near it I get a sort of low hiss. I adjusted the ohms for my Q701s but that seemed to make the noise problem worse. Despite all that, like I said pretty happy.


The Bravo Audio V2 has been sitting on my Amazon wishlist for a while now.


Any semi-serious guitarist knows vacuum tubes are alive and well.


My perspective, as someone with some experience with truly high power microwave tubes (multi-MW, not sub-kW): It's great that that a quarter kilowatt can be done in solid state, but it's not a death knell for microwave tubes. 58% efficiency at 2.45GHz and 63% efficiency at 915MHz sound reasonable but aren't a breakthrough, especially for a narrowband application like RF heating. The big savings come from ditching the strong magnets and high voltage power supplies. Size may be a significant improvement, but these devices still need to dissipate more than a third of their power as heat, so it's not as drastic as pictures may imply.

Solid-state devices have been gradually encroaching on vacuum tubes for a long time, but vacuum tube tech has been advancing, too. I'm not sure that cathode lifespan is much of a concern anymore at the power levels that solid-state devices can reach. The upper limit of how much power you can get out of a single tube is still growing, as is the efficiency. There's a lot of low-hanging fruit for the design of microwave tubes now that we can actually do a 3d simulation without a supercomputer.

Solid-state devices haven't surpassed any of the fundamental limitations on vacuum tube devices except for the entry costs of having a high-temperature high-voltage cathode, a vacuum chamber, and a strong magnetic field.


I kind of wonder what someone with your background thinks of the following articles:

http://www.lowtechmagazine.com/2014/06/thermal-efficiency-co...

http://www.lowtechmagazine.com/2014/07/cooking-pot-insulatio...


Whoa it may be the last mass market vacuum tube is going away but that doesn't mean all of them. Hams use linear amplifiers that have tubes in them. Pretty sure the military still uses them as well.

http://en.wikipedia.org/wiki/Linear_amplifier

Most vacuum tube production these days is in Eastern Europe and Russia.


These sexy things don't seem to be going away: http://en.wikipedia.org/wiki/Vacuum_fluorescent_display

A fluorescent lamp also qualifies as a vacuum tube:

http://en.wikipedia.org/wiki/Fluorescent_lamp#Principles_of_...

http://en.wikipedia.org/wiki/Fluorescent_lamp#Cold_cathode_l...

"Most fluorescent lamps use electrodes that operate by thermionic emission, meaning they are operated at a high enough temperature for the electrode material (usually aided by a special coating) to emit electrons into the tube by heat."

Thermionic emission is how vacuum tubes work, like the 12AX7's and 6L6's or whatnot in your guitar amp.


[flagged]


I think this implies that men cannot be sexy. While I'm not attracted to men I have it from a reliable source that this is false.


Can-be-construed-as x-ist is a weasel pattern used to project a straw-person's sensibilities onto common discourse.


Ahem... the last vacuum tube product?

http://www.samsung.com/us/2012-vacuum-tube-amplifier/


The last useful vacuum tube product.

Tube amps distort the sound more, and clip more gradually. This basically means that a digital amplifier can perfectly emulate a vacuum tube amplifier as long as you aren't maxing out the range and clipping the digital amplifier. (You may need a slightly higher rating on the digital amplifier to achieve the same volume.)


I take it you're not a guitarist. Tube amps can generate tones and sounds that are impossible to emulate accurately via solid state. I've owned some really nice solid state amps that can do amazing things, but even a shitty tube amp can do overdriven fuzz better than any solid state amp.


> impossible

Not impossible. Hard, expensive, or currently not available but not impossible.


I'll concede that "impossible" might be the wrong word choice, sorry. How about "impractical to the point of near impossibility"?

But I only have 22 years experience playing guitar and bass across genres as diverse as jazz, blues, gospel, metal, modern rock, and folk, so hey what do I know?


Well, allow me to go into a bit of science detail. The two polar opposites in amplification technology today are Class A and Class D amplifiers. Class A is the classic one, and any audiophile and sound engineer will swear by them on the quality metric. Class D is modern, based on transistors (usually MOSFETs), and while it dumps the linearity, it has a huge advantage in energy efficiency. Only in recent years has Class D become competitive in quality with the lower end of Class A technology, as the linearity of amplification makes a huge difference in high quality systems at higher amplification where the details pop out more.

Class A uses tubes, and works in the simplest and most accurate way by amplifying the complete signal. It maintains the most important characteristic needed to perfection - linearity of waveform amplification. When you try and get extra juice out of it, instead of the annoying clipping you get from digital amplification, you get a nicer sounding and more gradual reaction. The main drawback is energy efficiency. There's a theoretical peak efficiency of 50% to Class A amps, and what's worse is that it's usually less than half that. It makes tube amps that aren't small require massive heat sinks, and increases the BOM.

Class D is totally different. It dumps linearity out the window in favor of energy efficiency. It basically outputs a bunch of pulses, and then applies a low-pass filter to remove the annoying artifacts. While you might say that it's close enough - it isn't for high fidelity systems, especially in production studios, live performance, and audiophile cases. While competitive in quality with the low end of amplifiers (under $200), if you are looking for accurate detail reproduction there is no replacing linear amplification yet. The advantage is that even at the worst use scenarios, Class D amps have over 50% energy efficiency, and usually manage over 90% for higher volumes, making them need far less in the way of heat sinks which allows for leaner designs, and a better price/performance ratio.

The Class D amps will eventually win out. Class A amps have decades of R&D as an advantage, but Class D is improving by leaps and bounds. Eventually, Class D will become "good enough" and replace Class A altogether because of the advantages in price and energy efficiency. But that day has not yet come.

That being said, I use a class D amp for my home set up. It's close enough, and my power bill thanks me monthly.


>Class A uses tubes...

Class A amplifiers can be made in solid state as well, very easily. As a truly simple, DIY example: http://headwize.com/?page_id=31

There's nothing about solid state that requires a circuit that "outputs a bunch of pulses, and then applies a low-pass filter" -- true linear amplification exists in the transistor domain just as well.

And even if solid state did require quantization (beyond the shot noise that vacuum tubes suffer from also), given that we can do direct digital synthesis up to the 100's of MHz regime pretty easily these days (i.e. beyond the bandwidth of most audio tube amplifiers), with 24+ bits of resolution, there's no reason that a solid DSP system couldn't emulate a vacuum tube response well enough to be electrically indistinguishable from the actual tube.

But, as the grandparent post notes, it's one heck of a lot easier/cheaper to just use vacuum tubes. Which is why I'm staring at a tube-based headphone amplifier on my desk right now. Though it does use solid state buffers, which incidentally are way more linear than the typical ways -- capacitors, transformers, etc. -- of removing the DC offset inherent to tubes...


There's a bit of misinformation in your post, but someone else already addressed it.

Personally I like Class D amps for basic uses like computer speakers and such. That said, I have a very basic class D receiver/amp that I use as a preamp for a larger 2.1 system with studio monitors for computer sound, as I like my FPS games and music to be clear and accurate while still being loud enough to "feel" things like explosions or timpanis.

Technically the Class D amp can push those monitors well enough, but at more than 1/3 volume there is some distortion and clipping, whereas on my discrete amp I have yet to find any distortion within comfortable listening levels.


Line 6 does a pretty good job at it.


True, but for some people "pretty good" isn't enough. As I said before, I've had some very nice solid state amps that I loved playing through. It's just that there are some sounds that have yet to be emulated properly via modern circuitry. Granted, it's all subjective; why does a Fender sound better or worse than a Gibson for a particular style of music? It all comes down to what kind of sound the musician wants to get out of his equipment, and if a solid state amp doesn't give him what he wants but a tube amp does, then I'd say the tube amp is still relevant.


Exactly, when the tube gets warmed up and you crank up the pre-amp gain you get a warm-sounding natural distortion that afaik is impossible to emulate with a digital amp.

I used to have a Marshall JCM 2000 Dual Super Lead, that thing sounded amazing.


> afaik is impossible to emulate with a digital amp

Absolutely not impossible. If you can define it you can design it.


Sounds like you've never tried an Axe-Fx. Their simulations are absolutely mindblowing.


The whole point of the continued popularity of tube amps is to drive them too hard and get interesting noises ... people have been trying to duplicated that in digital for decades, and failed.


> people have been trying to duplicated that in digital for decades, and failed.

Because decades ago we simply didn't have the processing power to that kind of DSP.

Today, however, you carry that kind of processing power in your pocket.

It used to be really hard, because (amongst other things) the oversampling required to avoid digital aliasing. This is not really a big deal anymore, today. Also, it so happens to be the kind of calculation that parallelises very nicely.

I can really recommend this video to dispel the myths about what we supposedly can't to today in the digital domain: http://xiph.org/video/vid2.shtml (the guy also presents the matter really well, I found it very enjoyable to watch, even if I already knew over half the things he explains). To really make his case, he feeds signals to both a digital and an analogue frequency analyser.


They keep failing because they keep creating integrated "modeling amp" products which put the original amount of wattage in the back end, and that destroys the model at volume: you start hearing the break up of the back-end amplifier.

They are doing that to save costs: real tube amplifiers continue to be posed as higher end at higher prices.

A decently modeling amplifier would probably need several hundred watts, if not into the thousands, just to replicate a 50W tube combo at volume. The speaker cabinet and drivers would have to be different: high fidelity units, not guitar speakers.

I think people do get good results when they plug their modern amp modeling digital signal processors directly into a loud PA system, where the models find a decent amount of headroom and fidelity to reproduce what they are doing at loud volume.


People report good results with the AxeFx rigs.

If you just want "interesting noises" the VST plugin FuncShaper from Met-RS works great for adding distortion. It's a lifesaver on direct bass tracks. Pipe that thru a convolution of a good impulse response from a nice amp and you can't tell the difference easily.


According to (some) tube afficinados, you can simulate pre-amplifier can emulate a tube pre-amplifier: small-signal elements based on tubes can be substituted with silicon. However, you probably can't emulate that big signal stuff that goes on between the tube power amp, its output transformer and a speaker, and the sound that ultimately emerges. That is to say, not in the same amount of wattage, and with the same speaker. Maybe a 5,000 watt rig (amplifier and speakers) could faithfully emulate what comes out of a fully cranked 50W tube amplifier, its output transformer, and a specific 50W loud speaker.

As a related illustration of the idea, it is common to see microphones in front of small tube amplifiers, which then feed into a large PA system. The small amplifier acts as the "signal processor" which produces the tone, and the PA just makes it bigger. It has the headroom to do that.

In that scenario, the tube amplifier with a microphone in front of it can be replaced by digital modeling: something which produces the same transfer function as what occurs between the signal from the instrument, and the signal going into the console's microphone input.


A Class T amplifier[1] is the closest you can get to a tube. One of the amplifiers, the TA2020 (1998), was named one of the twenty-five chips that 'shook the world" by the IEEE Spectrum magazine.[2]

[1] http://en.wikipedia.org/wiki/Class_T_amplifier

[2] http://spectrum.ieee.org/semiconductors/processors/25-microc...


They're also very common in linear RF amplifies.


"Maybe the magnetron is really not the last vacuum tube product"

Not by far. A more obvious example than the ones the article gives would be tube amplifiers.


I dunno, I think a microwave with a vacuum tube will make my food sound warmer and trued to the original analog food I remember as a kid.


Am I wrong, or is this a press piece for Freescale's new transistor?


I think this article is focused on consumer products? Last time I checked, large scale items such as radar were still using vacuum tubes. I know there were huge breakthroughs in solid-state radar back in the 1990s, but I don't think that technology has seen a wide commercial roll-out.


I went to college with this guy who was on a scholarship to study tube design. (This would have been 1981 or 1982.) I said, "What are you studying that for? Semiconductors are going to make tubes obsolete." He said, "You have 6000 amps you're trying to control. Try that in a transistor." I mean, yes, there are schemes that try to use multiple high-current transistors in parallel, but... 6000 amps is a lot of current for semiconductors, even today.


There's been a lot of progress since then.

One IGBT pair. 1200 amps. 4500 volts. 140x190 mm housing:

http://www.infineon.com/dgdl/Infineon-FZ1200R45HL-DS-v03_01-...


Impressive, except that the times are listed in units of microseconds, so it's completely irrelevant to microwave use.


And even if it came to the end of tubes, servicing obsolete hardware (and software) can be very lucrative. Kids don't learn COBOL anymore.


I think more than a couple of people still buy guitar amps!


They often cook unevenly and typically lose power over time as the magnetron filament and cathode deteriorate. At last it is possible to replace the magnetron with solid state devices.

My microwave is several decades old now and its magnetron hasn't ever needed replacement, and I haven't noticed any degradation in power level either. I wonder if these new "all-solid-state" ovens will last nearly as long - I've read plenty of reports about the "inverter" types (essentially an HV SMPS) that use many more components which are often operated near their maximum limits and fail sooner, compared to the older design with a simple (but more expensive) large HV transformer.


Magnetrons are rated for thousands of hours of operating time, and the degradation is gradual. Unless you use your microwave oven to the exclusion of all other heating sources, a 40 year old oven should still be doing alright if it was well built to begin with. It's only industrial uses that have to worry, and even so the amortized cost is tiny. (I've seen an estimate of $1/hour for a 100kW industrial heating magnetron with a rating of 8000 hours, which is quite a bit smaller than the operating cost.)


Don't forget about the krytron, http://en.wikipedia.org/wiki/Krytron If you need to switch thousands of amps in under a nanosecond you'll probably need one.


I know I made the mistake of ordering these and Beryllium sphere at the same time. Oopsie!


Yea, sounds like it's a bit early to title this post "How I Learned to Stop Worrying and Love the Semiconductor."


Vacuum Tubes are very popular in high fidelity audio spectrum.

http://www.head-fi.org/t/655961/pictures-of-your-high-end-sy...


There are some very interesting stories in the history of vacuum tubes. One is the 416A planar triode for the Bell Labs TD-2 microwave relay network: http://long-lines.net/tech-equip/index.html (I like especially the story of the terrorist attack on the microwave relay station by the "American Republican Army")

Another is John Pierce on Klystrons and TWTs: http://www.smecc.org/john_r__pierce____electron_tubes.htm


TWTs are going to be around for a long time. There are only a handful of companies in the world that make them, and there are no real alternatives to producing high power microwave signals.


Well, TWTs are just one member of a family of vacuum tubes that can do things solid state devices can't. There are also IOTs, klystrons, and BWOs, and their cousins from the gyrotron side of the family. There's a lot of territory for solid-state devices to conquer before vacuum tubes are obsolete, and solid-state devices can't even operate in most of those regions, let alone outperform vacuum tubes.


Even within the article under "Related" there's a link to a story titled "Long Live Vacuum Tube Amps" from earlier this year. Perhaps a more accurate title would be "...Last Mass-Market Vacuum Tube Product". Because even just last year I bought a nice little desktop tube guitar amp, and it looks like the Fender Greta is still being sold. (Which I'd recommend if you're looking for a tiny amp to noodle around with and still get crunchy tube goodness.)


I think more than a couple of people still buy guitar amps.


Consider the array of new tube combo guitar amps carried by Guitar Center right now:

http://www.guitarcenter.com/Tube-Combo-Guitar-Amplifiers-Com...


There are plenty guitar amplifiers still made with tubes. Look up Fender Twin for example.



Are you kidding me? People who are serious about music swear by an all-analog mastering process with tube amps. The even harmonics make for a warmer sound than the harsh odd harmonics of transistors.


The only reason you get any harmonics is if you drive them to nonlinear distortion though.

If you have enough headroom, you can get near-zero transistor distortion and then apply your tube modelling digitally.

Audiophiles are often stuck in the 1970's with this stuff.


Aren't vacuum tubes still used in photomultipliers?


A photomultiplier is a special kind of vacuum tube, but it's a pretty niche market. Plus avalanche photodiodes are making inroads.


Virtually every kitchen has a built in microwave? I'm not so sure about that monsieur.


Um, vacuum tube diodes?




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: