I use the eWon, and MBConnect devices all the time, one or the other goes in to every machine we build. They are VPN gateways with secure login so we can remotely work on a machine instead of having to immediately travel to it to check the slightest thing.
None of our customers leave the internet side of the device plugged in. Unless we are on the phone with them, and they are by the machine, it is unplugged. As an additional level of security, the device has a keyswitch connected to it that must be turned on to allow it to connect to the internet, just in case it gets plugged in.
Most devices are managed through the respective manufacturers applications via the cloud, so we just have to download their application, and log in, and it handles getting the keys, and establishing the secure VPN tunnel. It is possible to manage your own infrastructure, but I don't know of anyone who is large enough, or chooses to do it.
I put the eWon app on my brand new work PC, now I have to check if I got pwned the first day got my new Lappy:( The remote access apps are one of the few things that does not get installed on the VM. Connecting to the VPN, through the VM can really be a pain!
The MBConnect devices are really cool, they can even verify the entire system, and reload anything that does not match what is stored inside itself. Besides providing a huge obstacle for anyone wanting to Stuxnet the system, they allow a customer to replace a PLC with a spare, reboot, and have everything come back to normal, and they allow for easier updating of a whole system by passing the program to the MBConnect device, and having it apply the update locally.
Nothing more scary than flashing a PLC remotely, and rebooting it. If it doesn't come back online, you might have to take your Lappy, and leave on an immediate road trip!
Many things can happen when going into production. Perhaps they always intended to go into production with the cheaper controller, but they had a problem with the firmware they were struggling with, and so they used a more expensive controller that was a slam dunk to buy themselves more time to perfect the cheaper controller, and not miss their target delivery dates.
I have 4G now, and it is still as slow as 3G, which is as slow as 2G, which is as slow as 1Xrtt when everyone is using their phones and the pipe to the tower is full. I often see 10 - 30 Kbps during peak times.
During the middle of the night, 1 bar will get me 1.3 - 1.9Mbps on 3G, and 3 - 5 Mbps on 4G, but during the day, I struggle to get 100Kbps on 3G or 4G, even with 5 bars.
I can watch my download speed increase as everyone goes to bed. It's funny (sad) to graph my download speed and see it jump up on the hour, and jump a little less on the half hour as the pipe opens up.
Everyone who can, will jump ship to Theo's version ASAP. Being strong enough to demand the best, and accept nothing less is a good thing when it comes to software security.
Many of these libraries are essential, and mostly taken for granted.
How many people thought that it was already an OpenBSD project, and had Theo's scrutiny already?
There are some true leaders within the OSS world, and we are lucky to enjoy what is made of their efforts. As the TLAs become increasingly invasive in our daily lives, having well written clearly documented textbook code is the only thing you can count on to provide any level of security.
Version 4.0 of Linux will be the same way. Linus will take his kernel, and go where things are sane again, and there will be no compromises. Take it or leave it. This is how things are done. Correctness is the law of the land.
Or you will be taken of your privacy, and computer security at every possible turn.
I worked on a hybrid demo that used 14,400 of them.
25 of them connected in parallel. (3.7V @ 75Ah)
12 groups of the 25 in series. (44.4V @ 75Ah)
16 of the 25 X 12 sets in series. (710V @ 75Ah)
3 of the 16 X 25 X 12 sets in parallel. (710V @ 225Ah)
Very scary to work with, especially in the bilge of a boat!
We went way overboard with the charge monitoring, but when we had cells fail while we were testing, they could be easily identified, and swapped out, then the rest of the pack was returned to service. I think the same could be done here.
My guess is there will be rebuilders that spring up to disassemble and test the cells, and replace the few bad ones with other used ones of similar vintage and capacity, and sell the used / repaired packs. A few weak cells can really hurt the pack as a whole.
When I used to race R/C everyone had their secret methods of treating cells to boost Voltage and or Capacity, and I'm sure there is some money to be made if someone can recondition the cells to get some of their lost capacity back. Perhaps take the packs apart, hook them up to a windmill, and cycle them through a few test and re-conditioning charge cycles to bring some of their life back while making a little coin as grid storage.
1. collect old EV batteries 2. disassemble 3. test (PROFIT!!!) 4. sort 5. recondition (PROFIT!!!) 6. reassemble 7. PROFIT!!!
Many fuel injected performance cars already do that. When the ECM detects you are at WOT, it will shut off the A/C compressor to get a few extra horsepower.
My 5.0L Mustang had A/C cut. I used to drag race it (:Yes with the A/C on:) and 12 seconds later, when I got to the "Big End" of the 1/4 mile it would be blowing warm air out of the registers, but as soon as I let off the gas, it would cool down again!
I tried running it with the A/C turned off as well, but it made no difference in my elapsed times. Changing to a shorty belt that bypassed the A/C and power steering was good for a tenth of a second or so, but it was not worth the hassle of changing the belt all the time, and sweating my butt off while I was having fun.
You are correct, with a CV joint the rotational angle in = rotational angle out, this is not the case with a traditional U joint. If you've ever been in a 4wd truck, and taken a tight turn, you can feel the lurching as the U joints change the wheel speeds. This is not what is at play in this case.
The issue is due to the geometry of the front wheels, and the fact that they want to "flop" outwards when torque is applied. Normally one wheel would counteract the other, but the uneven angles in the drivetrain cause an additional imbalance that can be felt in the steering wheel.
Wikipedia to the rescue again...
The main component of torque steer occurs when the torques in the driveshaft and the hub are summed vectorially, giving a resultant torque vector around the steering pivot axis (kingpin). These torques can be substantial, and in the case of shafts making equal angles to the hub shafts, will oppose one another at the steering rack, and so will cancel. These torques are strongly influenced by the position of the driveshaft universal joint (CV joint) in relation to the steering axis, however due to other requirements such as achieving a small or negative scrub radius an optimum solution is not generally possible with simple suspension configurations such as Macpherson strut.
Many front wheel drive vehicles already have a torque imbalance going to the front wheels because the CV axles are different lengths, and therefore are at different angles. This causes torque steer under acceleration. I'd venture to guess that it must not be that big of a deal because cars continue to be built that way.
Most front wheel drive "performance" cars have equal length CV axles to eliminate the issue.
If you had a failure of a wheel motor, you could just reduce the max power you allowed the other wheels to make so the car was drivable, and not too unwieldy during acceleration.
I agree that wheel motors are too much unsprung weight, and impractical for a number of other reasons though too. I worked on a diesel hybrid project that had large rare earth magnets mounted behind the flywheel, and that thing was downright dangerous to work on. Pulling tools right off the bench, and out of your hands. If you held on to them, you learned very quickly that it was a bad idea! Blood blisters galore! The amount of force in a large ring of rare earth magnets is incredible. We had to be very careful assembling it because if it got away from you, someone was getting hurt! You could stand them on edge, and feel them trying to jump off the wood cart and smash your tool box 4 feet away. Walk by an I beam in the shop, and it would leap right off the cart. It then took a small army with non-metallic pry tools to get it off the beam.
Not something you want to work on in the driveway! You don't just put it back together like installing a brake drum.
The cheap brushless motors used in fans are switched by the magnetic field of the commutator, so putting more current to them will cause them to reach a higher peak voltage in the windings, and spin faster. They are very much like a DC brushed motor in that respect.
The larger higher powered motors are synchronous to the field applied to the windings, and putting more current to them will produce more torque, but the RPM is dependent on the frequency being sent by the drive electronics. If the frequency does not change, the RPM will not change either (besides perhaps 1-2% less slip).
If the motor has position feedback, you can use the encoder to tell the drive when to switch the field based on the position of the armature. In this mode, the large motor, will behave much like the small fan, or a DC motor, and vary the frequency, and resulting torque according to the power being sent to the armature, and the load on the output shaft. The RPM is simply the result.
There are also hybrid modes that allow for keeping very tight control on RPM, while allowing you to adjust the torque output for applications such as tension control, where you want one motor to pull something at the same speed as another motor, but want to keep a certain amount of slack or tension between them.
Example: unwinding a roll of paper, cutting it to width, and rewinding it. The drives keep the outer surface speeds of the two rolls matched as they change diameter, and keep perfect tension on the paper through the change in torque due to the changing diameters so the slitting operation works properly.
There are so many ways to make a motor turn, you can get exactly the properties you want from it.
My ramblings were only about the synchronization of motors.
I'd agree on the differential being simpler (although not the modern computer controlled torque splitting design) but I'd also expect Tesla to have some top notch motor controls.
It's funny, I've been sent in to fix problems with our control systems to find the only problem to be a multi million dollar machine running for 2+ years without maintenance. I've often asked the plant managers if they'd treat their cars like that:) Production lines are there to make money, and shutting them down for maintenance (for some people) doesn't seem to make sense. They just run them, till they won't run anymore, then blame the electronics because the drive has a fault, but the drive is only faulted because it is trying to turn a gearbox that has gear teeth floating around in (what's left of) the gear lube.
When I was 17 I built a go-kart with a snowmobile engine, and a hydraulic (swash plate) transmission, ooohhh how I wish I could get one in my car... Lots of maintenance, if you consider twisted axles maintenance, which I soon did;)
I've seen lots of people struggle with their hydrostatic lawn mowers though, and we had a forklift with a hydrostatic transmission that has claimed it's fair share of walls too:) although Case should never have put the shifter pedal where the creeper pedal is on every other fork lift in the world...
Then how do all of the millions of CD and DVD players out there manage to properly play the discs, or VCRs play tapes without the TV losing sync?
Matching motor speeds is easy, even without encoder feedback. If a motor is turned at a different speed than it is driven it distorts the waveform going to the motor. Modern drive electronics can sense that, and adjust accordingly.
I've worked on machines that produce fabrics at 10 - 15 feet per second, and they contain multiple motors that are perfectly synchronized. If not, you would quickly have piles of fabric on the floor, or huge gobs of fabric getting wound up in the various stages of the machine. Getting them all tuned so they change speed at the same time takes a little bit of know how, but once they are properly set, they run trouble free.
Modern control systems are that good.
I've seen systems demoed with motors that have large triangular, square, and pentagon shaped metal plates mounted on them with slots cut in them barely wider than the plates. They start and stop so the plates can pass through the slots in perfect synchronization. You watch hypnotized as the different sided plates ratchet through each other at about 10 RPM, then the thing ramps up to a blur, and you watch in amazement as the other half of your brain tells you to run like hell in case one of those plates misses the slot, and the whole thing goes BAM, and comes flying apart.
Mitsubishi had a multi-master synchronized motor demo with some small servo motors mounted facing shaft to shaft about 3 inches apart driving parallel metal discs with pieces of lead from a mechanical pencil stuck through tiny holes in both of them. They had little levers mounted above the discs so you could push down on them to slow down or stop the discs. The goal was to try and get them out of synch enough to get the pencil lead to break while they went through a little cycle of ramping from about 50 to 1000 RPM. If you broke the lead, you got a real nice Mitsubishi mechanical pencil. I didn't see anyone manage to get one though.
Trade show demos are the coolest:)
I'd imagine you would have to allow them to NOT be in synch enough to go over bumps without creating additional jerkyness in the ride.
Where I used to work had a policy like that, and you are right, the number of post-it notes with !t$Feb2014 or similar you could find stuck around was incredible.
I have had that same thought. I'm sure ATM's could record the numbers, as well as the counting machines that are used when you bring in lots of bills to the bank.
My old bank put in an automated counter that dispensed your with withdrawal to the teller, and it didn't give me the correct amount and the cashier called in back and they verified the transaction and gave me my missing money.
I asked her previously why they didn't count the money back when they gave it to me out of the machine, and she said they didn't have to anymore.
They didn't have to come count what was left in the machine, so they must have reviewed what went out to me and counted every bill it dispensed to find out. The part of the machine that dispensed the money thought it did the right thing, so there had to be another way for them to remotely verify my transaction that was independent from the counter mechanism. What better way than to independently take a picture or record the serial numbers of each bill dispensed. With how terrible the BETA site is, I mean after what we've learned from Edward Snowden, just imagine what that system is all connected to!
I'm sure there's a list of bills that are known to still exist from a time before they could record everything that get heavily scrutinized when they finally show back up at the bank. Who had these and how did they get them, because there's 99999 more of those $100 bills have been hiding (likely in sequential order) somewhere for 20+ years.
Slashdot Mirror
Build a giant hubless flywheel in one of those old particle accelerator rings.
Store all the extra wind gusts underground for a whole city
IBM has DIMMs with flash memory already.
www-03.ibm.com/systems/x/options/storage/solidstate/exflashdimm/
I use the eWon, and MBConnect devices all the time, one or the other goes in to every machine we build. They are VPN gateways with secure login so we can remotely work on a machine instead of having to immediately travel to it to check the slightest thing.
None of our customers leave the internet side of the device plugged in. Unless we are on the phone with them, and they are by the machine, it is unplugged. As an additional level of security, the device has a keyswitch connected to it that must be turned on to allow it to connect to the internet, just in case it gets plugged in.
Most devices are managed through the respective manufacturers applications via the cloud, so we just have to download their application, and log in, and it handles getting the keys, and establishing the secure VPN tunnel. It is possible to manage your own infrastructure, but I don't know of anyone who is large enough, or chooses to do it.
I put the eWon app on my brand new work PC, now I have to check if I got pwned the first day got my new Lappy:( The remote access apps are one of the few things that does not get installed on the VM. Connecting to the VPN, through the VM can really be a pain!
The MBConnect devices are really cool, they can even verify the entire system, and reload anything that does not match what is stored inside itself. Besides providing a huge obstacle for anyone wanting to Stuxnet the system, they allow a customer to replace a PLC with a spare, reboot, and have everything come back to normal, and they allow for easier updating of a whole system by passing the program to the MBConnect device, and having it apply the update locally.
Nothing more scary than flashing a PLC remotely, and rebooting it. If it doesn't come back online, you might have to take your Lappy, and leave on an immediate road trip!
My Bittorrent client can do that:)
Dangit... You just made me jump to my bootloader...
You definitely need to get yourself on the waiting list for the upgraded PRO version with the quantum solvers then...
Perfect for exactly these chicken and egg type problems, where both sides are seemingly full of shit simultaneously!
(and the upgraded blue LEDs are friggin' awesome)
Many things can happen when going into production. Perhaps they always intended to go into production with the cheaper controller, but they had a problem with the firmware they were struggling with, and so they used a more expensive controller that was a slam dunk to buy themselves more time to perfect the cheaper controller, and not miss their target delivery dates.
I have 4G now, and it is still as slow as 3G, which is as slow as 2G, which is as slow as 1Xrtt when everyone is using their phones and the pipe to the tower is full. I often see 10 - 30 Kbps during peak times.
During the middle of the night, 1 bar will get me 1.3 - 1.9Mbps on 3G, and 3 - 5 Mbps on 4G, but during the day, I struggle to get 100Kbps on 3G or 4G, even with 5 bars.
I can watch my download speed increase as everyone goes to bed. It's funny (sad) to graph my download speed and see it jump up on the hour, and jump a little less on the half hour as the pipe opens up.
Cheers
Everyone who can, will jump ship to Theo's version ASAP. Being strong enough to demand the best, and accept nothing less is a good thing when it comes to software security.
Many of these libraries are essential, and mostly taken for granted.
How many people thought that it was already an OpenBSD project, and had Theo's scrutiny already?
There are some true leaders within the OSS world, and we are lucky to enjoy what is made of their efforts. As the TLAs become increasingly invasive in our daily lives, having well written clearly documented textbook code is the only thing you can count on to provide any level of security.
Version 4.0 of Linux will be the same way. Linus will take his kernel, and go where things are sane again, and there will be no compromises. Take it or leave it. This is how things are done. Correctness is the law of the land.
Or you will be taken of your privacy, and computer security at every possible turn.
Because of Minix.
I worked on a hybrid demo that used 14,400 of them.
25 of them connected in parallel.
(3.7V @ 75Ah)
12 groups of the 25 in series.
(44.4V @ 75Ah)
16 of the 25 X 12 sets in series.
(710V @ 75Ah)
3 of the 16 X 25 X 12 sets in parallel.
(710V @ 225Ah)
Very scary to work with, especially in the bilge of a boat!
We went way overboard with the charge monitoring, but when we had cells fail while we were testing, they could be easily identified, and swapped out, then the rest of the pack was returned to service. I think the same could be done here.
My guess is there will be rebuilders that spring up to disassemble and test the cells, and replace the few bad ones with other used ones of similar vintage and capacity, and sell the used / repaired packs. A few weak cells can really hurt the pack as a whole.
When I used to race R/C everyone had their secret methods of treating cells to boost Voltage and or Capacity, and I'm sure there is some money to be made if someone can recondition the cells to get some of their lost capacity back. Perhaps take the packs apart, hook them up to a windmill, and cycle them through a few test and re-conditioning charge cycles to bring some of their life back while making a little coin as grid storage.
1. collect old EV batteries
2. disassemble
3. test (PROFIT!!!)
4. sort
5. recondition (PROFIT!!!)
6. reassemble
7. PROFIT!!!
I know... I hate having to use those darn line numbers..
And don't even get me started on the three letter variable names!!!
Qbasic FTW!
Better be careful what kind of tape you use!
Fail.
His code checked input 2 twice.
Many fuel injected performance cars already do that. When the ECM detects you are at WOT, it will shut off the A/C compressor to get a few extra horsepower.
My 5.0L Mustang had A/C cut. I used to drag race it (:Yes with the A/C on:) and 12 seconds later, when I got to the "Big End" of the 1/4 mile it would be blowing warm air out of the registers, but as soon as I let off the gas, it would cool down again!
I tried running it with the A/C turned off as well, but it made no difference in my elapsed times. Changing to a shorty belt that bypassed the A/C and power steering was good for a tenth of a second or so, but it was not worth the hassle of changing the belt all the time, and sweating my butt off while I was having fun.
Cheers!
Ahhh the fun I've had in a Chinook...
You are correct, with a CV joint the rotational angle in = rotational angle out, this is not the case with a traditional U joint. If you've ever been in a 4wd truck, and taken a tight turn, you can feel the lurching as the U joints change the wheel speeds. This is not what is at play in this case.
The issue is due to the geometry of the front wheels, and the fact that they want to "flop" outwards when torque is applied. Normally one wheel would counteract the other, but the uneven angles in the drivetrain cause an additional imbalance that can be felt in the steering wheel.
Wikipedia to the rescue again...
The main component of torque steer occurs when the torques in the driveshaft and the hub are summed vectorially, giving a resultant torque vector around the steering pivot axis (kingpin). These torques can be substantial, and in the case of shafts making equal angles to the hub shafts, will oppose one another at the steering rack, and so will cancel. These torques are strongly influenced by the position of the driveshaft universal joint (CV joint) in relation to the steering axis, however due to other requirements such as achieving a small or negative scrub radius an optimum solution is not generally possible with simple suspension configurations such as Macpherson strut.
From: http://en.wikipedia.org/wiki/T...
Cheers!
No she's a spreadsheet.
Many front wheel drive vehicles already have a torque imbalance going to the front wheels because the CV axles are different lengths, and therefore are at different angles. This causes torque steer under acceleration. I'd venture to guess that it must not be that big of a deal because cars continue to be built that way.
Most front wheel drive "performance" cars have equal length CV axles to eliminate the issue.
If you had a failure of a wheel motor, you could just reduce the max power you allowed the other wheels to make so the car was drivable, and not too unwieldy during acceleration.
I agree that wheel motors are too much unsprung weight, and impractical for a number of other reasons though too. I worked on a diesel hybrid project that had large rare earth magnets mounted behind the flywheel, and that thing was downright dangerous to work on. Pulling tools right off the bench, and out of your hands. If you held on to them, you learned very quickly that it was a bad idea! Blood blisters galore! The amount of force in a large ring of rare earth magnets is incredible. We had to be very careful assembling it because if it got away from you, someone was getting hurt! You could stand them on edge, and feel them trying to jump off the wood cart and smash your tool box 4 feet away. Walk by an I beam in the shop, and it would leap right off the cart. It then took a small army with non-metallic pry tools to get it off the beam.
Not something you want to work on in the driveway! You don't just put it back together like installing a brake drum.
Cheers!
The cheap brushless motors used in fans are switched by the magnetic field of the commutator, so putting more current to them will cause them to reach a higher peak voltage in the windings, and spin faster. They are very much like a DC brushed motor in that respect.
The larger higher powered motors are synchronous to the field applied to the windings, and putting more current to them will produce more torque, but the RPM is dependent on the frequency being sent by the drive electronics. If the frequency does not change, the RPM will not change either (besides perhaps 1-2% less slip).
If the motor has position feedback, you can use the encoder to tell the drive when to switch the field based on the position of the armature. In this mode, the large motor, will behave much like the small fan, or a DC motor, and vary the frequency, and resulting torque according to the power being sent to the armature, and the load on the output shaft. The RPM is simply the result.
There are also hybrid modes that allow for keeping very tight control on RPM, while allowing you to adjust the torque output for applications such as tension control, where you want one motor to pull something at the same speed as another motor, but want to keep a certain amount of slack or tension between them.
Example: unwinding a roll of paper, cutting it to width, and rewinding it. The drives keep the outer surface speeds of the two rolls matched as they change diameter, and keep perfect tension on the paper through the change in torque due to the changing diameters so the slitting operation works properly.
There are so many ways to make a motor turn, you can get exactly the properties you want from it.
My ramblings were only about the synchronization of motors.
I'd agree on the differential being simpler (although not the modern computer controlled torque splitting design) but I'd also expect Tesla to have some top notch motor controls.
It's funny, I've been sent in to fix problems with our control systems to find the only problem to be a multi million dollar machine running for 2+ years without maintenance. I've often asked the plant managers if they'd treat their cars like that:) Production lines are there to make money, and shutting them down for maintenance (for some people) doesn't seem to make sense. They just run them, till they won't run anymore, then blame the electronics because the drive has a fault, but the drive is only faulted because it is trying to turn a gearbox that has gear teeth floating around in (what's left of) the gear lube.
When I was 17 I built a go-kart with a snowmobile engine, and a hydraulic (swash plate) transmission, ooohhh how I wish I could get one in my car... Lots of maintenance, if you consider twisted axles maintenance, which I soon did;)
I've seen lots of people struggle with their hydrostatic lawn mowers though, and we had a forklift with a hydrostatic transmission that has claimed it's fair share of walls too:) although Case should never have put the shifter pedal where the creeper pedal is on every other fork lift in the world...
Cheers!
Then how do all of the millions of CD and DVD players out there manage to properly play the discs, or VCRs play tapes without the TV losing sync?
Matching motor speeds is easy, even without encoder feedback. If a motor is turned at a different speed than it is driven it distorts the waveform going to the motor. Modern drive electronics can sense that, and adjust accordingly.
I've worked on machines that produce fabrics at 10 - 15 feet per second, and they contain multiple motors that are perfectly synchronized. If not, you would quickly have piles of fabric on the floor, or huge gobs of fabric getting wound up in the various stages of the machine. Getting them all tuned so they change speed at the same time takes a little bit of know how, but once they are properly set, they run trouble free.
Modern control systems are that good.
I've seen systems demoed with motors that have large triangular, square, and pentagon shaped metal plates mounted on them with slots cut in them barely wider than the plates. They start and stop so the plates can pass through the slots in perfect synchronization. You watch hypnotized as the different sided plates ratchet through each other at about 10 RPM, then the thing ramps up to a blur, and you watch in amazement as the other half of your brain tells you to run like hell in case one of those plates misses the slot, and the whole thing goes BAM, and comes flying apart.
Mitsubishi had a multi-master synchronized motor demo with some small servo motors mounted facing shaft to shaft about 3 inches apart driving parallel metal discs with pieces of lead from a mechanical pencil stuck through tiny holes in both of them. They had little levers mounted above the discs so you could push down on them to slow down or stop the discs. The goal was to try and get them out of synch enough to get the pencil lead to break while they went through a little cycle of ramping from about 50 to 1000 RPM. If you broke the lead, you got a real nice Mitsubishi mechanical pencil. I didn't see anyone manage to get one though.
Trade show demos are the coolest:)
I'd imagine you would have to allow them to NOT be in synch enough to go over bumps without creating additional jerkyness in the ride.
Cheers!
(&BuckFeta)
Where I used to work had a policy like that, and you are right, the number of post-it notes with !t$Feb2014 or similar you could find stuck around was incredible.
Sounds like they need a little piece of duct (500 MPH) tape to put over the vent!
I have had that same thought. I'm sure ATM's could record the numbers, as well as the counting machines that are used when you bring in lots of bills to the bank.
My old bank put in an automated counter that dispensed your with withdrawal to the teller, and it didn't give me the correct amount and the cashier called in back and they verified the transaction and gave me my missing money.
I asked her previously why they didn't count the money back when they gave it to me out of the machine, and she said they didn't have to anymore.
They didn't have to come count what was left in the machine, so they must have reviewed what went out to me and counted every bill it dispensed to find out. The part of the machine that dispensed the money thought it did the right thing, so there had to be another way for them to remotely verify my transaction that was independent from the counter mechanism. What better way than to independently take a picture or record the serial numbers of each bill dispensed. With how terrible the BETA site is, I mean after what we've learned from Edward Snowden, just imagine what that system is all connected to!
I'm sure there's a list of bills that are known to still exist from a time before they could record everything that get heavily scrutinized when they finally show back up at the bank. Who had these and how did they get them, because there's 99999 more of those $100 bills have been hiding (likely in sequential order) somewhere for 20+ years.