||[Nov. 15th, 2013|07:45 pm]
Facebook sucks; I honestly can't understand its popularity.
It has so many gratuitous and useless bells and whistles that I'm reminded of those busy box toys for toddlers (not to be confused with the very useful software package whimsically called busybox).
I still find ordinary email far better for non-real-time communications with friends, family and colleagues. When I have something to publish to the world, I put it on my own website. When I want to rant about something, I'll come here (though it should probably also go on my own website). And I'll use various web discussion boards to debate specific topics such as news stories.
Facebook seems to be designed for exhibitionists addicted to sharing every trivial detail of their lives. I use it only because a few people don't seem to be available anywhere else, such as acquaintances from school and more distant relatives. But it's a major pain in the ass, especially when I trip one of their "security" mechanisms.
This happened again yesterday. While already logged in from home, minding my own business, I was told of Facebook's real-identity policy and presented with pictures of people I was supposed to identify to prove I'm really me.
I got a wide shot of a conference room full of people, with none of their faces visible. I got baby pictures. How in hell am I supposed to identify them?
I got maybe one out of three, but that wasn't good enough for Facebook's security Nazis; skip two and you're completely locked out for an hour!
For all I knew, these pictures were of high school classmates I've not seen in 40 years and didn't really know that well even then. They might have been distant relatives I've never even met in person -- or even their family members. In any event, this is totally ridiculous. I've been locked out three times now. The last time this happened, I retried on and off for days until I got lucky enough to recognize enough pictures in one session.
When I complained to Facebook I was told that to even process my inquiry I would have to scan and provide my birth certificate, passport, drivers' license and other sensitive identifying information to prove I was a real person using my real name. Fuck that! Other than having a few credit cards stolen, which was aggravating enough, I've never been a victim of serious identity theft and I'm not about to become one now by giving that kind of information over insecure email to some anonymous flunky at Facebook.
People really ought to boycott Facebook. Use email. Use the web. Use chat boards. Use instant messaging. They all work fine, at least if you're an adult or near-adult. Leave Facebook to the toddlers who can't figure out how to use the grownups' Internet.
|Giving up on AT&T U-verse
||[Oct. 10th, 2013|08:34 pm]
I'm finally giving up on AT&T's U-verse service after four years of frustration.|
Around October 1 AT&T pushed down a firmware update to my "residential gateway" (the AT&T-provided VDSL modem and router I must use) that supports IPv6. That would seem like a good thing, but it's not. IPv6 is disabled, I have no way to turn it on, AT&T's support staff has no idea what IPv6 is or when they'll enable it, and the gateway now blocks ALL IPv6-in-IPv4 tunneled packets, including those I had been tunneling myself. That is, it blocks IPv4 protocol 41.
In other words, AT&T completely broke IPv6 for me without any warning. They did it when I was out of the country, so I had to remotely diagnose the problem, disable IPv6 on all my hosts and remove my AAAA DNS records.
The really funny thing is that just a month ago an AT&T chat agent talked me into accepting a new residential gateway with the IPv6 firmware. I hooked it up, discovered the same problems related above, put my old gateway back in service and returned the new one to AT&T. Now they have made my old gateway equivalent to the new, broken one I rejected.
AT&T's disregard for their customers is stunning. Not one of the bugs I documented shortly after I got the service in November 2009 has been fixed. This is the last straw.
I originally got U-verse because it was the fastest internet service available in my area at the time: 3 Mb/s up and 24 Mb/s down. Times have changed; Time Warner has rolled out DOCSIS3 and I've signed up for their fastest service: 5 Mb/s up and 50 Mb/s down.
Yet another promising technology ruined by corporate incompetence. I'm almost embarrassed to say that I used to work for AT&T (through Bell Laboratories) 30 years ago.
|Recommendation to NTSB on the grounding of DC power system neutrals
||[Mar. 25th, 2013|04:10 pm]
To the NTSB, 787 Battery Investigation:
I have been following this investigation as an electrical engineer interested in power systems.
I believe this incident highlights the drawbacks of the apparently standard practice of bonding one side of aircraft DC power buses to aircraft ground, i.e., of grounding the DC system neutral. The NTSB should officially question this practice. What follows is my rationale.
It is clear that large ground fault currents flowed at some point in the 787 battery failure sequence as evidenced by overcurrent damage to the green ground bonding jumper wire and to the stainless steel shield on the J1 signal cable, and by the arcing between Cell 5 and the battery box.
The ground fault current arced from Cell 5 to the battery box, flowed out in parallel along the ground bonding jumper and the shield of the signal cable, through the aircraft grounding system to the (presumably single point) connection to the aircraft (-) 28V DC bus, back to the battery via its (-) lead, and back into Cell 1 via the battery box's (-) bus bar.
I see no over-current protective devices to interrupt this fault current. Even the battery disconnect contactor would not have helped as its contacts were only in the battery (+) bus bar to the (+) terminal of Cell 8.
You report that the battery cases nominally float with respect to each terminal, so this ground fault must have occurred after an earlier fault caused an internal electrode in Cell 5 to short to its case.
Nevertheless, I believe it should be noted that even with these internal cell and cell-to-box faults, no damaging ground fault currents would have flowed had the (-) side of the 28 VDC aircraft power bus not been intentionally bonded to aircraft ground.
I believe it is time to re-examine the apparently standard practice of bonding the (-) side of aircraft DC power systems to aircraft ground.
This presumably follows the long-standard practice of grounding one side (the neutral) in terrestrial AC power systems to protect against lightning strikes on the power lines, although even here I think we should carefully re-think the practice. But aircraft DC power systems are isolated, so I see only downsides to this practice -- as shown by this incident.
In contrast, the high voltage DC systems in the current generation of electric and hybrid vehicles are NOT grounded to the vehicle body.
Please note that I am only questioning the bond between the negative current-carrying DC power bus conductor and aircraft ground. There is no question that every exposed metal component of an electrical device must be grounded to protect against electrical shock.
If aircraft practice parallels terrestrial AC practice, where every device has separate neutral and ground wires, all DC loads already isolate both sides of the DC supply from aircraft ground to prevent objectionable return currents (ground loops) through the aircraft ground system (and body in a metallic aircraft) under normal conditions. That's why there's only one connection between the (-) DC bus and the aircraft ground. Because of this isolation, this single bond could probably be removed with no impact on or redesign of existing 28 V DC aircraft equipment.
But the DC bus should not just be left to float. To drain any static charges the low impedance bond should be replaced with a high impedance bond (i.e., a resistor) monitored for voltage drops by an alarm. Under normal conditions, no current would flow through this resistor and there'd be no voltage drop. Should a ground fault occur anywhere in the aircraft DC system -- generators, batteries, loads or wiring -- a potential would appear across this resistor and sound an alarm. (The potential difference could also be logged by the aircraft flight data recorder.) Yet the DC bus and equipment would continue to operate normally until the fault is located and repaired by maintenance personnel.
Continued operation in the event of a ground fault is no small advantage considering that much of this equipment is undoubtedly critical to safe aircraft operation. Indeed it's why batteries are present in the first place. Continued normal operation with an alarm is preferable to having the DC system fail due to a circuit breaker or fuse opening, and vastly preferable to uncontrolled ground fault currents flowing from a battery capable of supplying extremely large currents.
Thank you for your time.
San Diego, CA
|Ecotality's brain damage continues
||[Dec. 11th, 2012|04:15 pm]
You can see from a previous Livejournal entry just what I think about Ecotality and their (On The) Blink network of electric vehicle chargers.
Their brain damage continues unabated.
About the only useful feature of their bloated software is to delay charging to an off-peak electricity rate period. This could have been a "delay" button like on our dishwasher: each push delays charging by another hour. "Delay until midnight" would be useful as that's when the super-off-peak rate begins.
Of course, Ecotality can't do anything that absurdly simple. To delay charging you have to enter a programming menu, enter your PIN (not needed for immediate use) and then increment or decrement the charging start time, first selecting hours or minutes. If you select 12:01 AM, you have to remember to increment the date (carefully choosing month and day) or you'll select the midnight already past!
I didn't count my keystrokes but it was easily several dozen. Remember this is all on a small touch screen with "buttons" smaller than my fingertips.
A while ago I programmed it to always delay charging to midnight. This worked until one evening when I decided to charge immediately. There's a menu option for that, but when I pressed it the screen turned red and informed me there was some sort of self-test error. Naturally, it "failed unsafe" and refused to provide power. This happened several times until I gave up and cleared out the delay through the programming menu.
In a few months my agreement ends and the charger becomes mine. I can't wait. I will remove the control computer and force-enable the PIC that handles the actual J-1772 interface to the car so it will simply work.
I just got an email from Ecotality inviting me to "extend your Blink charger warranty for free" until the end of next year. I immediately declined and I urge every technically-inclined EV driver do the same. Every problem with my charger has been one Ecotality created itself with their bloated control computer, and getting rid of it will make a far more reliable unit that simply won't need their "free warranty".
|A silver lining to Hurricane Sandy
||[Nov. 18th, 2012|02:46 pm]
Here's a fascinating story about the damage Hurricane Sandy did to Verizon's cable infrastructure in lower Manhattan:
The Broad Street cable vault was filled with salt water for two days. Their ancient copper cable plant is completely destroyed, but fiber came through fine.
So they're replacing all the destroyed copper with fiber. Every customer previously served with copper gets FiOS, Verizon's last-mile fiber system.
Wow. So it only took the most destructive hurricane in the history of the northeast to finally get a phone company off its ass and install fiber.
Too bad we don't have hurricanes here in southern California.
|Credit card insurance for rental cars? Don't believe it!
||[Nov. 3rd, 2012|08:08 pm]
Note: VISA has finally paid! See the followup at the end.
Everyone knows that the various insurances ("collision damage wavier", etc) that rental car companies sell you are seriously overpriced and represent almost pure profit. This became so well known that years ago the major credit card companies began to offer equivalent insurance for free if you charge your rental to their card and decline the rental companies' CDW.
Well, don't believe it. At least not from VISA.
In late July I made a business trip to Silicon Valley. I rented a car from my then-favorite provider, Enterprise, for two days and charged it to my AT&T Universal VISA card. On the first evening, July 26, as I was eating dinner with a large group in Mountain View, the rear window was smashed and my laptop bag was stolen.
The car was a hatchback. The cover over the rear storage compartment was missing, so the thief could see my bag through the rear window. Lesson learned: rent only sedans with trunks, not hatchbacks or vans, and if you can't get a sedan, carry everything or make sure anything you leave in it is very well hidden from view.
It could have been worse. My laptop was a 4+ year old, pre unibody, Apple MacBook Pro that had finally died that afternoon (that's why I left it in the car). I'd been planning to replace it anyway, and all of the data on it was fully encrypted and fully backed up; I didn't lose a bit.
I still lost medications and accessories that had to be replaced, along with the SSD I'd installed in the laptop a year or two earlier that was probably worth more than the laptop itself. And my time and aggravation in dealing with the aftermath has been the biggest loss of all.
I called the police and also Enterprise, who I asked to tow and replace the car as in my opinion it was unsafe to drive with all the broken glass. I had to wait hours for both to show up, but eventually they did. Other than the wait I can't fault the service I got either from the Mountain View Police Dept or Enterprise, who gave me a new car.
The fun started after I returned home. Naturally I expected a bill from Enterprise for the broken window, which I got several weeks later. I told them that I'd charged the car to Visa, and that they would pay according to their insurance agreement.
Three months later, VISA still hasn't paid. Although their insurance claims contractor, "VISA Enhanced Services" asked for a ridiculous amount of supporting documentation, some of it was actually halfway reasonable (police report number, etc). I gave them everything they asked for.
Everything, it seems, but one: "The monthly billing statement that you provided did not include the card number on which the transaction was made. Please provide the charge receipt and the monthly billing statement that shows the full card number".
As I have explained to them at least a half dozen times now, I could not provide the charge receipt because it was in the laptop bag that was stolen, and the card numbers did not match because, as a precaution, I had called VISA immediately after the burglary and requested that my previous card be cancelled and a new one issued just in case the thief obtained my previous card number from the receipts on my bag. The rental charge appeared on my August bill with my new card number, not the one to which I originally charged it. The account, of course, is the same, and my name and mailing address haven't changed either.
This is where it sits. VISA Enhancement Services says they're unable to get the information they need from VISA, despite being their own contractor. I've made multiple requests to VISA Customer Services to send a letter explaining the situation. Each was quickly answered with promises to do so, yet no letter ever seems to make it either to VISA Enhancement Services or to me. (I have, however, gotten several forms asking me to list fraudulent charges even though I had never complained about any. Again, I had changed my card number simply as a precaution.)
It's beginning to dawn on me that this is, in all likelihood, a deliberate ploy by VISA to drag their feet until I eventually give up in disgust and pay the claim myself.
I'm reminded of those "collect the pieces" games that fast-food restaurants often run where you collect all in a group and win a big prize. It is, of course, easy to get every piece except for one that's extremely rare -- and you aren't told which one that is. VISA required a long list of items of documentation without saying which they'd refuse because of some bullshit excuse completely outside my control. (They probably pick whatever item you give them last.)
It's especially infuriating that VISA Enhancement Services could easily get the information they want themselves, inasmuch as they're all supposedly part of the same company. They all use the same VISA logo on their letterhead.
I see now that it was a major mistake to play it safe by having my card number changed just in case the previous number had been compromised. The alternative would have been to contest any fraudulent charges that may have appeared. But I've been there too. Shortly before Katrina, we had a credit card number stolen on a visit to New Orleans, probably in a restaurant. Even though by law I was never personally liable, it took the better part of a year and many, MANY phone calls to resolve the situation.
The bottom line is this: credit card companies are more than happy to get the extra business you give them in good faith because of their "cardholder benefits" but don't expect to actually get those benefits. Not without a long fight that will leave you wondering whether it was worth it -- if you get them at all.
Followup: After at least three (3) requests, VISA finally sent me the letter listing both old and new card numbers. I forwarded it to VISA Enhancement Services, and within a few days they accepted it and began processing the claim. On November 12, they cut a check to Enterprise for the amount of the damage.
So it only took 3.5 months to pay a simple claim because they needed information they could have gotten for themselves. It's silly for any number of reasons. Not only is it inefficient, time consuming and tedious, but insecure too. Without verification, how do they know I'm not forging something? And if they can verify it, why not just get it directly in the first place?
I still think the system is designed to be as difficult as possible so that at least some of their customers will eventually give up in disgust.
|More Ecotality braindamage and incompetence
||[Jul. 13th, 2012|01:19 am]
The web server on my Blink EVSE has been broken for the past several months. (EVSE = Electric Vehicle Supply Equipment, a fancy term for the now-standard-but-special extension cord used to charge an electric car.) Whenever I tried to connect to it, I got a "500 Server Error" message.|
(Note added in response to a comment: no, this is not a joke. The extension cord that charges my electric car really does have a web server -- only worse. Read on...)
Thinking maybe the network configuration got screwed up, last night I went to the keypad to reprogram it.
This took only the better part of an hour. Maybe the touch screen has gotten dirty, but I had a tough time getting it to see my finger. And when it did, it would usually sense it twice, very rapidly. The "buttons" seemed smaller than with the last software version, and the calibration seemed to have drifted; I had to offset my finger a little to get the right letters.
It took me a while to realize this as I had been trying to enter my rather lengthy WiFi network password, and like most computers today it blots out password echoes "for security". This ensures that I'm the only person who can't possibly see it. Have you ever tried to type your password on a malfunctioning keyboard?
A half hour later, I finally could go back inside to my web browser. Same thing: "500 Server Error". Applying the Fundamental Theorem of Electronics Repair (99% of all equipment problems can be cured by cycling the power) I did just that. I cycled the 40A breaker on my garage subpanel supplying power to the unit.
Big mistake. Now the unit was completely and totally dead. Nothing on the display. No power to my car. Since it had at least charged my car normally, this was definitely a worsening trend. Fortunately it had already finished charging so it wasn't yet an urgent situation. I cycled power a few more times with no benefit. Later I noticed the screen was still slightly warm so I knew it was getting power. (Even when not in use, the Blink EVSE burns something like 15W continuously powering the computer, WiFi and display. How symbolic for a "green" technology like the electric vehicle.) I left it on to see if it would recover. When I got home tonight it was still dead. But in the dark I could see that the backlight was on, confirming that it's getting power.
Around noon I called Ecotality for support and waited 10 minutes or so for an agent who assured me that my repair would be a high priority. She promised me a return phone call, which I have yet to get. (See Followup below.)
What a surprise. Ecotality is still totally incompetent and their EVSE is still a piece of crap.
It is simply astounding how much effort (and government money) they have wasted taking something that is literally nothing more than a heavy-duty extension cord and turning it into a bloated, overdesigned, overpriced, oversized, unreliable and energy wasting monstrosity. It's simply ridiculous.
Supposedly all of this complexity was necessary to gather EV charging statistics for their DoE contract. But that very same information could be gathered with another conventional interval-type (time of use) electric utility meter just like the one SDG&E installed next to my main utility panel a few years ago. That meter would probably melt if it drew 15W to power itself. It has never failed in any way, and it certainly hasn't interrupted all power to my house. Ecotality's own software has been so chronically buggy that they'd get much better data from these standard utility meters, with no development budget at all.
But I guess it was much more important to tell the world that they exist by deploying big futuristic-looking boxes with touch screens and the company logo everywhere. Or perhaps their engineers (if they still have any) considered a simple, reliable, low-tech extension cord too small a challenge. Whatever, their incompetence is becoming a real problem. I'm beginning to fear that it will create a public perception that charging an electric car is inherently difficult, expensive and unreliable. And these kinds of initial impressions are both totally unnecessary and very difficult to shake.
My contract with them specifies that I get to keep the EVSE at the end of the term. I can't wait until then to go into my unit and disconnect the computer and touch screen that's given me so much grief (and wasted so much energy!) Maybe I'll finally have a dependable (if not more reasonably sized) way to Just. Charge. My. Car.
Followup: Friday morning (a day later) I got a call from an Ecotality technician and his assistant. They arrived a half hour later to take a look at my problem, and it was apparent they'd seen it before. A lot. In April Ecotality had pushed out a software upgrade but it didn't "take" and the data on the SD (flash memory) card had been corrupted. The computer in my EVSE had actually crashed at that time and they hadn't heard from it since! It kept working because the PIC that actually talks to the car was still enabled but when I cycled power the unit wasn't able to reboot and re-enable the PIC.
He swapped the card and my EVSE came right back up. Then we wasted the better part of a half hour trying to get it to peer with one of my WiFi base stations; it simply wouldn't do it, and it wouldn't say why.
On a hunch I turned off all but one of our four base stations; that did it. But why was a mystery. They verified operation with my car, buttoned up the unit, and left. I went back inside and turned on the rest of my base stations, came back out and looked at my unit.
They'd left only 10 minutes earlier. I clenched my teeth and cycled the breaker. The unit came back to life but then took at least 10 minutes installing a flurry of system updates. Since the controller inside runs Linux I even recognized a few. During this time it "apologized" for not being available for use.
I was so fed up with the flakiness of its WiFi that I spent the next hour running a hardwired Ethernet cable to the thing just to make those problems go away.
Sorry, but all this just doesn't cut it. There is simply no excuse for this thing not working whenever it has power. After all, as I keep saying, it's nothing more than a glorified extension cord. That's it! It would have been absolutely trivial for the software to simply lift the pin to the PIC microcontroller that enables it to work, and that would have let the EVSE supply power while the computer was doing its thing. That, and the fact that the software update had corrupted the flash memory in the first place (and had apparently done so on many other EVSEs) simply shows the carelessness or incompetence of Ecotality and its programmers. Every embedded systems programmer is (or should be) concerned about the reliability of his update procedures; they should always be designed to keep the old firmware and fall back to it if for some reason the new firmware is incompletely received, fails checksum, etc. It wasn't like the functionality of the computer was actually needed during this time (or at any other time for that matter).
Interestingly enough, when I later got on the blinknetwork web site I saw that all the usage data for the past several months was there; it hadn't been lost. That explains what I had seen inside of the unit while it was being worked on, a CX-10000-3 standalone watt-hour meter that performs the actual logging of energy use. Their Linux system simply sends the data to them. In other words, they were already doing what I had previously suggested, using a dedicated watt-hour meter to collect their usage data. That gave them even less of an excuse to not function at all times. The detailed usage data that is so precious to them is still there.
The only function provided by that firmware that is of any use to the customer at all is the ability to program the EVSE to supply power to the car at a programmed time, and to turn it on and off remotely (when it's working). I've never found that particularly useful as I'm not on a time-of-use tariff yet, and probably won't be until there are significant nighttime discounts to make it worthwhile. Besides, my car has its own delayed-charging timer, so that feature in the EVSE is largely redundant.
Even its energy usage logs aren't especially critical. It's on our regular house meter, which SDG&E replaced a while ago with an electronic unit that logs usage every 15 minutes and makes it available through their web site. It's obvious from that data when the EVSE is in use, and it always draws the same power whenever my car is charging.
I still can't wait until my contract with Ecotality is over and I can do to this thing what Dave Bowman did to Hal in 2001: A Space Odyssey. This is actually a pretty apt metaphor; Dave didn't turn off Hal completely, he just disabled Hal's higher brain functions leaving his autonomic functions intact. In the Blink EVSE I will disable (or reprogram) the Linux computer that controls the charger but leave the PIC microcontroller that actually runs the J1772 protocol with the car.
Note: I should say that the technicians who came to my house were helpful, friendly and competent. I have no gripe with them; my gripe is entirely with Ecotality the company and its bad design decisions.
|SOPA is horrific
||[Dec. 19th, 2011|03:26 am]
The proposed Stop Online Piracy Act is truly horrific. Please sign the petition shown here:
|The problem of overpriced electric vehicle charging equipment
||[Oct. 19th, 2011|02:20 am]
This is an excellent article but it omits the root cause of the problem: the authors of the National Electrical Code (NEC). They placed all sorts of unique "safety" requirements on EVSE (Electric Vehicle Supply Equipment) that were never placed on extension cords -- which is all these EVSE things really are. For example, they require that the cord not be energized until it's connected to the EV, and they prohibit the use of 240V dryer or range outlets for charging EVs even though they've been used safely for decades for RV hookups.
Gasoline pumps seem far more dangerous than extension cords. But they're not required to sense the presence of a car before letting gas flow. Nothing but common sense keeps you from squeezing the handle and spraying the gas on the ground.
The use of 240V vs 120V isn't a safety issue because, in North America, a "240V" circuit is really just two 120V circuits. It's only 240V between the two live pins, which you are very unlikely to touch at the same time. Nearly all shocks occur when a grounded person touches a hot conductor. The voltage to ground from each conductor of a 240V circuit is only 120V. Whatever shock hazard there might be is completely mitigated by ground fault interrupters (GFIs), which have been required on nearly all kitchen, bathroom, garage, basement and outdoor circuits for years.
The SAE J1772 standard for EVSE does have some features that are actually useful, such as telling the car how much current it can draw. And it interrupts the power flow before you actually remove the connector from the vehicle so it won't break under load and arc at the connector contacts. But these features currently come at a high price; even the portable EVSEs, which also implement J1772, cost much more than they should but at least Nissan bundles them with the car.
The article is right that the permanent EVSE units -- which are much more expensive than the portable units -- are way overdesigned. The WiFi connection does provide some minimal user remote control as an obvious afterthought (it's rather poorly designed) but the NYT article doesn't mention the real reason it's there. It's so Ecotality, the company that got that huge grant from the feds to install these things, can collect detailed usage data. The collection of this data is part of the user contract, so it's obviously valuable to them. But they don't compensate you in any way for this invasion of your privacy; the only way to avoid it is to pay full retail for a unit and forego the federal subsidy. Seems to me that either the federal government drove a poor bargain, or somebody there is in bed with Ecotality. Or both. Wouldn't be the first time.
Fortunately, the data collection is only to continue for another year or so. After that I fully intend to block this data from reaching their servers and to show everyone else how to block it too.
||[Oct. 6th, 2011|03:54 pm]
I never met the man, but I still feel his passing as a lost symbol of our generation. Back in the 1970s I was building homebrew computers just as Jobs, Wozniak and many others were. All of us could see the enormous potential in those primitive machines. We all had vague notions that someday they'd be everywhere and ordinary people would use them. But Jobs knew how to make that actually happen.|
I think future generations of kids will read about Steve Jobs in the same way that we read about Thomas Edison when we were kids. Larger than life, idealized, given credit not only for what he personally did but for some things that should rightly go to the people who worked for him -- but nonetheless somebody who really did completely change the world for the better with technology and who will become a symbol of that for many generations.
||most recent entries