"you don't want it open even a little bit once compression starts."
You do if you're running Miller cycle. The point being that effective compression stroke is longer than effective exhaust stroke in this mode.
Being able to dynamically alter the engine mode from Otto to Miller would be an efficiency win.
Holding valves open to disable a cylinder is unlikely to assist. From memory when 8-6-4 engines came along one of the points made in the engineering magazines was that this was tried and found to be less efficient than leaving the valves closed and letting the deactivated cylinder act as an air spring.
"If a valve gets stuck or the timing belt snaps then the piston will hit and bend the valve."
This is because the camshaft (or tappet and bucket assembly) prevents the valve being pushed back into the head.
It may well be that in a pneumatically actuated system the valve would get shoved back, but unlikely, given that valve stems are seldom perpendicular to the piston top.
On the other hand the biggest labour cost of replacing bent valves is down to the timing assembly needing to be removed/replaced/recalibrated, so a bent valve may not be such a big deal.
poppet valves can be driven by desmodronic cams, but the problem there is that it's almost impossible to change the timing in any meaningful way.
Fiat are already producing engines where the camshaft doesn't directly drive the valves (the Twin Airs) - it controls penumatic valves with electronic augmentation and I don't have any doubt they'll move fully electronic eventually.
The largest win of moving away from cam-driven poppet valves is reduction of internal engine friction. Everything else is pollution reduction.
On the other hand, if an engine is single speed, single load then it can be highly optimised and drive a generator which charges batteries with electric drivetrain. This concept has been around for a long time and works. The penalty is mostly battery weight, as losing a gearbox and transmission system is about matched by the mass of the motors and generator. There's an efficiency penalty at full load steady speed but most cars simply don't do that for 95%+ of their operational lifetimes.
We can write data to these tiny glass disks and read it back again.
Using equipment the size of a house. Slowly.
The acid test will be fitting all that gubbins into a 5.25" single height case (or something smaller) and getting read/seek rates high enough to be useful.
The original communist manifesto was for a post-capitalist world where increasing automation meant that large parts of the population were un (or under-) employed and was viewed as a natural progession, not something achieved by violent revolution.
Viewed in that context it may well be the way we need to go.
Socialism is an intermediate step and works quite well in the western countries which have tried it.
Unfortunately there have also been (and are) systems which call themselves communist or socialist that are nothing of the sort - but that's the type of society we associate with those words now. An Oligarchic/plutocratic/autocratic commend economy structure may be many things but it bears little relationship to Marx' original document.
Firstly: the lasers being used are 500mw+ devices -t eh kind which can pop balloons - they're not legally on sale in the UK, but are widespread. One was used from a nearby park against my house, resulting in the window going opaque green and the room lighting up more than adequately to read a book.
Secondly: They don't just flashblind. I've been lased whilst driving, It HURTS - and I observed other drivers lased on the same piece of road swerve when hit. This has higher potential to kill or injure than a pilot hit whilst autolanding (or taking off)
Thirdly: As above - the fuckwits doing this are also lasing cars and trains.
Enforcement is _hard_ - the worst offenders simply find a pub when the police helicoptor shows up and go inside. The ones where people are caught are almost always someone staggeringly naive. (There's a particular block of flats in Edinburgh which is infamous for lasing incidents - again, whilst there have been many police raids noone has ever been caught, let alone convicted.
Offensive weapon laws in the UK are used for public posession, but the real problem is that the penalties for endangering transport aren't severe enough or enforced enough. There's a need for mandatory minimum sentences to make sure judges don't make the offence a minor one (which is what's been happening). These same penalties could also be used to deal with the UK's epidemic of copper theft - by putting 15 year penalties on interfering with railway safety systems it'd discourage the thieves in a way that a large fine won't.
Newspapers were (and are) 100% paid for by advertising.
The cover price has historically been pretty close to that of the paper+printing process itself.
The problem isn't the advertising. We already skipped around all that in print media. The problem is the proliferation of abusive, intrusive advertising and malware.
TV stations mostly stopped running loud obnoxious adverts once remote controls became common. Advertisers need to improve their game so that they don't annoy people, not demand that adblockers be banned.
"If they brag about hitting 40 on a residential street that dead ends after 20 houses (10 each side), why would I expect them to slow down due to shared space?"
Dashcams. Specifically, your dashcam.
Having evidence of stupid driving behaviour does a lot to help your insurance company decide it's not going to allow the injury claim AND tell the other company it's not taking "knock for knock"
"The right balance is well marked speed cameras in problem zones combined with mobile speed cameras that could be anywhere, combined with fines with teeth"
Alternatively: Speed measuring units that trigger a red traffic light just up the road. Instant slowdown, no fine needed.
These are HIGHLY effective where they're deployed, but as they don't generate revenue countries which use "speed camera partnerships" and other revenue-driven systems won't touch them with a bargepole.
"You really want to make the road safer? 100% divided, with 100% gated fencing to keep kids and wildlife from wandering into the road. "
And on such roads (in the UK we call 'em "motorways") the speed limit is 70mph with most people actually driving closer to 100.
The problem is that on the other roads, the dangers you describe already exist and all the paint, roadside furniture and other shit lulls the average driver into a false sense of security, so they driver faster than is safe for the conditions - safe for everyone on the road including that deer and those kids playing on the footpath, not just safe for that driver.
Specifically, the problem is asshole drivers hitting 40-50mph on urban/suburban streets. The liklihood of pedestrian death if it approaches 100% - and there are associated noise and stress issues to go with the cars. It's clear that some people woen't be reasonable in their driving, so they need to be forced to do so - and speed humps simply aren't working (they tend to get treated as part of a slalom course with the assholes accelerating hard after they pass each one.)
Roads are _not_ the domain of cars, despite 100 years of attempts by manufacturers to lobby governments for rules pushing things that way.
Removal of paint and other moves are mostly being driven by the Living Streets initiative http://www.livingstreets.org.u... coupled with Transport for London's policy of dramatically reducing pedestrian deaths and injuries by 2020 https://tfl.gov.uk/corporate/s... - which is concentrating on undoing changes aimed at prioritising vehicular traffic at the expense of quality of life for non-drivers (and for the most part haven't done that much to improve traffic flow in any case), but also removing "safety" changes such as pedestrian fencing and excessive parking restrictions which have statistically proven to have the opposite effect to what was intended.
Slashdotters would do well to look it up and consider the
The problem is (and this comes primarily from US studies, but is backed by UK ones) that the more paint you put on the road and rigidly define your lanes, the faster people drive. They compartmentalise lanes as "safe" when they're only a marking on the road and pay less attention to the environment. It's sometimes referred as lane-related tunnel vision.
Removing the paint and roadside furniture which creates a non-existant "demarcation line" helps break drivers of the tunnel-vision habit and pay more attention to actual road and roadside conditions, which in turn often means they slow down to reasonable speeds.
Additionally: in almost every part of the world, you are required to be able to stop your vehicle in the visible distance of clear road ahead when there's a centreline. If there is no centreline, the allowed distance is reduced to HALF the visible clear road (the justification being that no centreline == narrow road and 2 oncoming vehicles need to be able to pull up if need be). Posted speed limits are an upper bound on speed and you can easily be cited for speeding or reckless driving if travelling too fast for the conditions.
In the long run this may not matter. Automated vehicles and other changes mean that fewer people will be driving _and_ there will be fewer cars in urban areas within 20-30 years. We're not at "peak car" worldwide but it's certainly close to it in developed countries.
Have you looked at the power consumption of a TGV?
300+mph has been trialled, but the rolling stock suffered severe coning effects (look that up) which nearly resulted in the train oscillating off the track, the track, ballast, pantographs and (more importantly) the overhead wiring all suffered major damage and the energy consumption for the test rig (a short, specially modified train) precluded doing it on a normal TGV - most TGV sets are over 1/4 mile long and at these speeds on these kinds of trains most friction is on the sides of the carriages no matter how well polished they may be - the aerodynamics up front is primarily to ensure the nose of the train stays grounded and carriage lower fairings are more about stability than reducing friction. The practical limit for railed vehicles on conventional tracks seems to be about 250mph.
On top of that a TGV track is a very expensive piece of high precision heavy engineering with a roadbed going down 10-12 feet in most places and a very low tolerance for any kind of trash near the tracks (the winds generated by passing trains are severe and magnetic eddy currents from the motors will pick up and throw any loose bolts/nuts. FOD inspections are almost as rigorous as on airport runways and that means there's an (expensive) army of maintenance workers associated with TGVs compared with lower speed passenger lines.
The construction costs of Hyperloop are low by comparison, which mean that it can run a lot more places and offer greater frequency of service. On longer routes you can expect that pods will link up to form trains, so it won't be "1 pod whizzing past every 30 seconds", it'll be "20-30 entrained pods every so often" - which makes route switching and other logistics much easier to handle.
I really hope this works. If it does then shorthaul flying will be pretty much eliminated. As it is I much prefer to take the train (TGV) from my house near London to Paris/Amsterdam/Nice than to fly - and to Paris/Amsterdam the doorstep-to-doorstep time is usually significantly lower than trying to fly. To Nice it's longer - but far more comfortable.
"To be worth it they would have to send a lot of pods down the tube, with minimal spacing"
Or link them into trains once they're underway.
My personal feeling is that the tube diameter is too small.
The real moneyspinner is freight and if the tubes can't handle a pod carrying an unmodified intermodal shipping container (that's the kind seen on ships or the back of trucks, not the ones used in aircraft) then the extra manual handling of what's being shipped will kill the economics of using it.
With regards to the Japanese Maglevss the huge advantage is energy consumption. Even with all that fancy aerodynamics on it, each maglev requires tens of MW to punch through the air. Partially evacuating the tubes reduecs power consumption dramatically.
With regard to safety: if the tube pressurises, the pods will slow down quickly to zero (but not uncomfortably so - remember they're air-suchion vehicles) and if the power goes off, passive safety systems can ensure that the tube is vented to atmospheric pressure in a safe manner.
Switching is one of the issues to be addressed but there are a number of ways of approaching this.
"I think this hyperloop is going to crash into the harsh realities of dealing with a vacuum."
This has always been the bugbear of evacuated tubeways and what killed the 1960s propsals.
Hyperloop doesn't pull a hard vacuum. It's more like 1-2psi than 0.02 and Whilst proposals are for induction motors for forward motion, the units are intended to fly on an aircushion, not use magnetic levitation (A suitable lifting fan would provide a fair bit of forward motion in a tube anyway, so the induction kickers can be on the tubes at intervals, not in the pods.)
"Until SSDs are less than 50% more expensive than spinning disks, spinning disks will still have a place."
In the consumer arena the margin is 300%, but for anything below 1TB SSD has already won. The new generation of M2 PCIe 1TB devices are so cheap that they'll be the norm in less than 6 months.
On the other hand I just put 24TB of SSDs into a server for scratchpad use, partly because of the speedups but mostly because we're pretty sure that the drives will have less downtime than HDDs.
"Your BEST spinning disk has IOPs in the range of 700-800"
Sequential IOPS. Random is lucky to achieve 130.
"Your average SSD is now in the range of 100,000 IOPs"
That's random _or_ sequential - and even if there's a write slowdown most current SSDs drop to "only" 50-75,000 IOPs
Then there's the warranty. Consumer SSDs mostly have 3-5 year warranties with some having 10 years. Consumer HDDs get 12-24 months at best. Enterprise HDDs get 3-5 years but that brings the price up into SSD territory for performance which can't match SSD.
"I've heard of lots of SSDs die, none of which would have come close to approaching the flash write limit. "
The most common failure modes here have been down to lead-free solder (seriously) and only on the low cost ones.
That said, I've seen SSD drives which lock up after 49 days uptime (anyone remember IBM deathstars?) and some which would lock up if left powered on for a week and stay dead unless put on a shelf for a couple of days (presumably until internal capacitors fully discharged). Those same drives used in laptop environments were 100% reliable.
The market is maturing and that kind of shite is mostly in the past, but it's a reminder that this stuff happens. For the most part however, SSD reliability (and warranties!) trumps HDD cheapness until you're in the 2TB+ range.
I should also comment that we see virtually no reliabiliity difference between enterprise or consumer drives from any given manufacturer. In fact the _least_ reliable drives seen in the last 6 years have been Seagate Constellations, with a 180% replacement rate over their warranty period (Yes, that does mean almost 2 replacements per originally-purchased drive).
That is a higher rate than our replacement rate for Seagate's Baracuda DM series (in desktop systems), which were all replaced whilst their DL predecessors soldiered on. DM's were notably the worst single drive in Backblaze's stats and I believe the rates they saw. The average service life in a desktop was 12 months and in the end we didn't bother with warranties despite having 3-5 year support on purchased systems. The manhours lost simply weren't worth it.
This kind of failure rate was the final straw which pushed us to put 1TB SSDs into desktops as a standard feature.
TLC 3D may well be good for 20-30,000 writes. The cell size is much larger than previous-gen 2d MLC so there's more room for deterioration.
So far the ONLY SSD I've killed via excess writes was a 128Gb Samsung 840pro - it got 200TB written to it in 2 years as a ZFS L2ARC drive and gave plenty of warning that it was going to die.
Slashdot Mirror
"you don't want it open even a little bit once compression starts."
You do if you're running Miller cycle. The point being that effective compression stroke is longer than effective exhaust stroke in this mode.
Being able to dynamically alter the engine mode from Otto to Miller would be an efficiency win.
Holding valves open to disable a cylinder is unlikely to assist. From memory when 8-6-4 engines came along one of the points made in the engineering magazines was that this was tried and found to be less efficient than leaving the valves closed and letting the deactivated cylinder act as an air spring.
"If a valve gets stuck or the timing belt snaps then the piston will hit and bend the valve."
This is because the camshaft (or tappet and bucket assembly) prevents the valve being pushed back into the head.
It may well be that in a pneumatically actuated system the valve would get shoved back, but unlikely, given that valve stems are seldom perpendicular to the piston top.
On the other hand the biggest labour cost of replacing bent valves is down to the timing assembly needing to be removed/replaced/recalibrated, so a bent valve may not be such a big deal.
poppet valves can be driven by desmodronic cams, but the problem there is that it's almost impossible to change the timing in any meaningful way.
Fiat are already producing engines where the camshaft doesn't directly drive the valves (the Twin Airs) - it controls penumatic valves with electronic augmentation and I don't have any doubt they'll move fully electronic eventually.
The largest win of moving away from cam-driven poppet valves is reduction of internal engine friction. Everything else is pollution reduction.
On the other hand, if an engine is single speed, single load then it can be highly optimised and drive a generator which charges batteries with electric drivetrain. This concept has been around for a long time and works. The penalty is mostly battery weight, as losing a gearbox and transmission system is about matched by the mass of the motors and generator. There's an efficiency penalty at full load steady speed but most cars simply don't do that for 95%+ of their operational lifetimes.
I've seen devices which don't bother with the message. They just shut down after 4 hours of no remote control signals.
We can write data to these tiny glass disks and read it back again.
Using equipment the size of a house. Slowly.
The acid test will be fitting all that gubbins into a 5.25" single height case (or something smaller) and getting read/seek rates high enough to be useful.
"hence the fuss about vitirification of nuclear waste"
Some of us are opposed to vitrification because that "waste" is mostly useful material for future reactors.
The original communist manifesto was for a post-capitalist world where increasing automation meant that large parts of the population were un (or under-) employed and was viewed as a natural progession, not something achieved by violent revolution.
Viewed in that context it may well be the way we need to go.
Socialism is an intermediate step and works quite well in the western countries which have tried it.
Unfortunately there have also been (and are) systems which call themselves communist or socialist that are nothing of the sort - but that's the type of society we associate with those words now. An Oligarchic/plutocratic/autocratic commend economy structure may be many things but it bears little relationship to Marx' original document.
Firstly: the lasers being used are 500mw+ devices -t eh kind which can pop balloons - they're not legally on sale in the UK, but are widespread. One was used from a nearby park against my house, resulting in the window going opaque green and the room lighting up more than adequately to read a book.
Secondly: They don't just flashblind. I've been lased whilst driving, It HURTS - and I observed other drivers lased on the same piece of road swerve when hit. This has higher potential to kill or injure than a pilot hit whilst autolanding (or taking off)
Thirdly: As above - the fuckwits doing this are also lasing cars and trains.
Enforcement is _hard_ - the worst offenders simply find a pub when the police helicoptor shows up and go inside. The ones where people are caught are almost always someone staggeringly naive. (There's a particular block of flats in Edinburgh which is infamous for lasing incidents - again, whilst there have been many police raids noone has ever been caught, let alone convicted.
Offensive weapon laws in the UK are used for public posession, but the real problem is that the penalties for endangering transport aren't severe enough or enforced enough. There's a need for mandatory minimum sentences to make sure judges don't make the offence a minor one (which is what's been happening). These same penalties could also be used to deal with the UK's epidemic of copper theft - by putting 15 year penalties on interfering with railway safety systems it'd discourage the thieves in a way that a large fine won't.
Newspapers were (and are) 100% paid for by advertising.
The cover price has historically been pretty close to that of the paper+printing process itself.
The problem isn't the advertising. We already skipped around all that in print media.
The problem is the proliferation of abusive, intrusive advertising and malware.
TV stations mostly stopped running loud obnoxious adverts once remote controls became common. Advertisers need to improve their game so that they don't annoy people, not demand that adblockers be banned.
"If they brag about hitting 40 on a residential street that dead ends after 20 houses (10 each side), why would I expect them to slow down due to shared space?"
Dashcams. Specifically, your dashcam.
Having evidence of stupid driving behaviour does a lot to help your insurance company decide it's not going to allow the injury claim AND tell the other company it's not taking "knock for knock"
"The right balance is well marked speed cameras in problem zones combined with mobile speed cameras that could be anywhere, combined with fines with teeth"
Alternatively: Speed measuring units that trigger a red traffic light just up the road. Instant slowdown, no fine needed.
These are HIGHLY effective where they're deployed, but as they don't generate revenue countries which use "speed camera partnerships" and other revenue-driven systems won't touch them with a bargepole.
"You really want to make the road safer? 100% divided, with 100% gated fencing to keep kids and wildlife from wandering into the road. "
And on such roads (in the UK we call 'em "motorways") the speed limit is 70mph with most people actually driving closer to 100.
The problem is that on the other roads, the dangers you describe already exist and all the paint, roadside furniture and other shit lulls the average driver into a false sense of security, so they driver faster than is safe for the conditions - safe for everyone on the road including that deer and those kids playing on the footpath, not just safe for that driver.
"slowing down" is the goal in residential areas.
Specifically, the problem is asshole drivers hitting 40-50mph on urban/suburban streets. The liklihood of pedestrian death if it approaches 100% - and there are associated noise and stress issues to go with the cars. It's clear that some people woen't be reasonable in their driving, so they need to be forced to do so - and speed humps simply aren't working (they tend to get treated as part of a slalom course with the assholes accelerating hard after they pass each one.)
Roads are _not_ the domain of cars, despite 100 years of attempts by manufacturers to lobby governments for rules pushing things that way.
Removal of paint and other moves are mostly being driven by the Living Streets initiative http://www.livingstreets.org.u... coupled with Transport for London's policy of dramatically reducing pedestrian deaths and injuries by 2020 https://tfl.gov.uk/corporate/s... - which is concentrating on undoing changes aimed at prioritising vehicular traffic at the expense of quality of life for non-drivers (and for the most part haven't done that much to improve traffic flow in any case), but also removing "safety" changes such as pedestrian fencing and excessive parking restrictions which have statistically proven to have the opposite effect to what was intended.
Slashdotters would do well to look it up and consider the
The problem is (and this comes primarily from US studies, but is backed by UK ones) that the more paint you put on the road and rigidly define your lanes, the faster people drive. They compartmentalise lanes as "safe" when they're only a marking on the road and pay less attention to the environment. It's sometimes referred as lane-related tunnel vision.
Removing the paint and roadside furniture which creates a non-existant "demarcation line" helps break drivers of the tunnel-vision habit and pay more attention to actual road and roadside conditions, which in turn often means they slow down to reasonable speeds.
Additionally: in almost every part of the world, you are required to be able to stop your vehicle in the visible distance of clear road ahead when there's a centreline. If there is no centreline, the allowed distance is reduced to HALF the visible clear road (the justification being that no centreline == narrow road and 2 oncoming vehicles need to be able to pull up if need be). Posted speed limits are an upper bound on speed and you can easily be cited for speeding or reckless driving if travelling too fast for the conditions.
In the long run this may not matter. Automated vehicles and other changes mean that fewer people will be driving _and_ there will be fewer cars in urban areas within 20-30 years. We're not at "peak car" worldwide but it's certainly close to it in developed countries.
Have you looked at the power consumption of a TGV?
300+mph has been trialled, but the rolling stock suffered severe coning effects (look that up) which nearly resulted in the train oscillating off the track, the track, ballast, pantographs and (more importantly) the overhead wiring all suffered major damage and the energy consumption for the test rig (a short, specially modified train) precluded doing it on a normal TGV - most TGV sets are over 1/4 mile long and at these speeds on these kinds of trains most friction is on the sides of the carriages no matter how well polished they may be - the aerodynamics up front is primarily to ensure the nose of the train stays grounded and carriage lower fairings are more about stability than reducing friction. The practical limit for railed vehicles on conventional tracks seems to be about 250mph.
On top of that a TGV track is a very expensive piece of high precision heavy engineering with a roadbed going down 10-12 feet in most places and a very low tolerance for any kind of trash near the tracks (the winds generated by passing trains are severe and magnetic eddy currents from the motors will pick up and throw any loose bolts/nuts. FOD inspections are almost as rigorous as on airport runways and that means there's an (expensive) army of maintenance workers associated with TGVs compared with lower speed passenger lines.
The construction costs of Hyperloop are low by comparison, which mean that it can run a lot more places and offer greater frequency of service. On longer routes you can expect that pods will link up to form trains, so it won't be "1 pod whizzing past every 30 seconds", it'll be "20-30 entrained pods every so often" - which makes route switching and other logistics much easier to handle.
I really hope this works. If it does then shorthaul flying will be pretty much eliminated. As it is I much prefer to take the train (TGV) from my house near London to Paris/Amsterdam/Nice than to fly - and to Paris/Amsterdam the doorstep-to-doorstep time is usually significantly lower than trying to fly. To Nice it's longer - but far more comfortable.
"To be worth it they would have to send a lot of pods down the tube, with minimal spacing"
Or link them into trains once they're underway.
My personal feeling is that the tube diameter is too small.
The real moneyspinner is freight and if the tubes can't handle a pod carrying an unmodified intermodal shipping container (that's the kind seen on ships or the back of trucks, not the ones used in aircraft) then the extra manual handling of what's being shipped will kill the economics of using it.
With regards to the Japanese Maglevss the huge advantage is energy consumption. Even with all that fancy aerodynamics on it, each maglev requires tens of MW to punch through the air. Partially evacuating the tubes reduecs power consumption dramatically.
With regard to safety: if the tube pressurises, the pods will slow down quickly to zero (but not uncomfortably so - remember they're air-suchion vehicles) and if the power goes off, passive safety systems can ensure that the tube is vented to atmospheric pressure in a safe manner.
Switching is one of the issues to be addressed but there are a number of ways of approaching this.
"I think this hyperloop is going to crash into the harsh realities of dealing with a vacuum."
This has always been the bugbear of evacuated tubeways and what killed the 1960s propsals.
Hyperloop doesn't pull a hard vacuum. It's more like 1-2psi than 0.02 and Whilst proposals are for induction motors for forward motion, the units are intended to fly on an aircushion, not use magnetic levitation (A suitable lifting fan would provide a fair bit of forward motion in a tube anyway, so the induction kickers can be on the tubes at intervals, not in the pods.)
"Until SSDs are less than 50% more expensive than spinning disks, spinning disks will still have a place."
In the consumer arena the margin is 300%, but for anything below 1TB SSD has already won. The new generation of M2 PCIe 1TB devices are so cheap that they'll be the norm in less than 6 months.
On the other hand I just put 24TB of SSDs into a server for scratchpad use, partly because of the speedups but mostly because we're pretty sure that the drives will have less downtime than HDDs.
"Your BEST spinning disk has IOPs in the range of 700-800"
Sequential IOPS. Random is lucky to achieve 130.
"Your average SSD is now in the range of 100,000 IOPs"
That's random _or_ sequential - and even if there's a write slowdown most current SSDs drop to "only" 50-75,000 IOPs
Then there's the warranty.
Consumer SSDs mostly have 3-5 year warranties with some having 10 years.
Consumer HDDs get 12-24 months at best.
Enterprise HDDs get 3-5 years but that brings the price up into SSD territory for performance which can't match SSD.
"I've heard of lots of SSDs die, none of which would have come close to approaching the flash write limit. "
The most common failure modes here have been down to lead-free solder (seriously) and only on the low cost ones.
That said, I've seen SSD drives which lock up after 49 days uptime (anyone remember IBM deathstars?) and some which would lock up if left powered on for a week and stay dead unless put on a shelf for a couple of days (presumably until internal capacitors fully discharged). Those same drives used in laptop environments were 100% reliable.
The market is maturing and that kind of shite is mostly in the past, but it's a reminder that this stuff happens. For the most part however, SSD reliability (and warranties!) trumps HDD cheapness until you're in the 2TB+ range.
I should also comment that we see virtually no reliabiliity difference between enterprise or consumer drives from any given manufacturer. In fact the _least_ reliable drives seen in the last 6 years have been Seagate Constellations, with a 180% replacement rate over their warranty period (Yes, that does mean almost 2 replacements per originally-purchased drive).
That is a higher rate than our replacement rate for Seagate's Baracuda DM series (in desktop systems), which were all replaced whilst their DL predecessors soldiered on. DM's were notably the worst single drive in Backblaze's stats and I believe the rates they saw. The average service life in a desktop was 12 months and in the end we didn't bother with warranties despite having 3-5 year support on purchased systems. The manhours lost simply weren't worth it.
This kind of failure rate was the final straw which pushed us to put 1TB SSDs into desktops as a standard feature.
Backblaze's stats tally closely with experience in our (much, MUCH smaller) datacenter.
If you're writing that much to a HDD it'll probably shake itself to pieces before the SSD expires.
That's certainly been the experience at $orkplace when people abused nearline arrays as scratchpads.
TLC 3D may well be good for 20-30,000 writes. The cell size is much larger than previous-gen 2d MLC so there's more room for deterioration.
So far the ONLY SSD I've killed via excess writes was a 128Gb Samsung 840pro - it got 200TB written to it in 2 years as a ZFS L2ARC drive and gave plenty of warning that it was going to die.
They haven't done this kind of chicanery with the APUs. It'd get them so much bad press that it's simply not worth it.