Wikipedia:Reference desk/Archives/Computing/2011 August 6

= August 6 =

Battery charger
So i am trying to figure out how to build a battery charger that can recharge a battery pack containing lithium ion batteries in series totaling 7.2v 1760mA(contained in a flashlight - non removable). So far what i have done is taken a computer power supply, attached a few 4.1 ohm resistors(parallel) to it and threw it on the battery. So far this seems to be working.... but i feel the need to explore safer options, so i am asking here. How can i build a charger that will charge this battery pack? – Elliott (Talk|Cont)   04:03, 6 August 2011 (UTC)
 * I assume that you are using a 12v power supply and wiring the resistors in series to limit the current. The problem with this arrangement is that there is no continuous control of the charging, and the battery pack could easily be overcharged (and overheated).  It is much safer to use a charging circuit with some dynamic current control that is designed for that particular battery pack.  If you wish to continue with your home-made charger, then it would be wise to add an ammeter to check the charging current, and to find out the recommended charging pattern for the pack.  Possibly a timed switch-off could be added in case you forget to disconnect the power supply at the end of the charging period.  A typical charging current might be 200mA for ten hours, but if you can find out the maufacturer's recommendations, and follow them, that would help to prolong the life of the pack and ensure safety.  Lithium-ion batteries can be dangerous if mis-used, so you ought to make sure that you really know what you are doing.    D b f i r s   07:18, 6 August 2011 (UTC)
 * Yes i am aware of the risks and problems with this, which is why i am posting here asking for a better, cleaner solution. – Elliott (Talk|Cont)   16:38, 6 August 2011 (UTC)
 * The safest solution is obviously to buy the charger designed for this battery pack, but if I was building my own charger, I would use both an ammeter to ensure that the charge rate never exceeded 400 mA (and preferably keep below 200mA to avoid overheating) and a voltmeter to ensure that the voltage didn't exceed 7.2v across the pack. An alternative to sophisticated current-reducing and switch-off circuitry is vigilance in monitoring progress of the charge.  I would be happy using this method for NiCad and NiMh cells, but extra care is needed with Li-ion cells.  The safest procedure might be to stop at 70% of maximum charge, when the pack reaches its rated voltage, because overcharging reduces battery life and, as you know, can have more dangerous consequences.  I have successfully charged Li-ion batteries by this ad-hoc method in an emergency, but I would hesitate to do it regularly because my memory has deteriorated to the point where I might forget and leave the pack on charge for too long.  A simple timer on the mains side of the PSU would solve this problem, but I don't know enough about Switched-mode power supply circuits to determine what the reverse leakage current would be on an unenergised circuit, or even if a back-emf of 7.2v might cause problems in the circuitry.  Perhaps an expert could comment on this?    D b f i r s   21:35, 6 August 2011 (UTC)
 * The computer switched mode power supply usually needs a load on it to function. With your series resisters you will get a higher current when the cells are flat and so may over heat them, with consequences of a lithium fire. Graeme Bartlett (talk) 22:32, 6 August 2011 (UTC)
 * Yes, a high-current variable resistor would be more flexible than your collection of 4.1 ohm ones to limit the current. I used to make my own from old electric blanket wire.  As Graeme warns above, there are serious risks in overheating the battery pack, so it would not be wise to attempt amateur recharging unless you know how to approximately mimic the charging pattern of the "official" charger.  As you will see from this linkcharging Li-ion cells correctly is much trickier than for any of the other rechargeable cells.  If I was charging your battery pack and it was almost flat, I would charge at 200mA for eight or nine hours (avoiding the risks in fast charging) and disconnect when the voltage rises to just over 8 volts, never allowing it to go over 8.4 volts.   D b f i r s   07:28, 7 August 2011 (UTC)
 * I agree. Nearly every single manufacturer of li-ion batteries recommends a CC/CV charging pattern. If you want to DIY and have some knowledge and experience with electronics but don't really know what you are doing (which to be blunt, if you have to ask you don't) my recommendation would be to use one of the many ICs designed for the purpose. You could perhaps skip the CV stage altogether although that would obviously reduce the amount of charge. Another concern is that it sounds like there are 2 cells in series. That being the case, you ideally should be balanced charging. If you're flashlight doesn't provide leads/connections for balanced charging (i.e. a connection between the cells) then I guess there's nothing you can do but if it does I would strongly recommend it be used. The alternative to DIY if you can't get or don't want to use a charger desiged for the flashlight is to use a decent hobby charger or some other charger which is already properly designed for charging li-ions in series (edit: e.g. .) (Hobby chargers themselves carry risks if misused and have been the cause of some fires [some due to damaged batteries but I think some due to misconfiguration] but if you are careful with your set-up and connections it's likely safer then a DIY when you don't know what you're doing.)
 * BTW it's not clear to me if the battery pack (or flashlight) has built in protection. If it doesn't, then you should also make sure the batteries are not overdischarged. If the voltage falls below 6V after the voltage is allowed to stablise after use, I would consider discarding the battery pack. And definitely below 5V. As   (primary lithium not li-ion) show, having a flashlight explode is not a fun experience.
 * Nil Einne (talk) 10:02, 8 August 2011 (UTC)

Again, sata drives are being strange
http://en.wikipedia.org/wiki/Wikipedia:Reference_desk/Archives/Computing/2011_July_23#Sata_power_connector PSU has been replaced, and yet it still doesn't see the drives, nor do either of them spin up. The strange part is, I plugged in my own HD into it, and that's not recognized. And plugging in the new HD to my working computer doesn't work either. I'm completely out of ideas. Anyone know what's happening? KyuubiSeal (talk) 04:21, 6 August 2011 (UTC)


 * I've seen cases where a bad PSU experienced an isolation failure, and delivered mains voltage to the low-voltage components it was supposed to be powering. So it killed all the PC's components (motherboard, drives, expansion cards), leaving very evident scorching and obvious evidence of ICs having been cooked. If your PSU was very bad (and your previous post suggests it was) then it may have damaged the drive's controller too. -- Finlay McWalter ☻ Talk 15:27, 6 August 2011 (UTC)


 * But then why would a working HD not spin up using the new PSU? EDIT: The (working) drive seems to be manufactured by Toshiba, so maybe it's something proprietary. Just a wild guess though. KyuubiSeal (talk) 16:32, 6 August 2011 (UTC)


 * More confusion. When testing the OD in my computer, it worked. So is the problem the PSU again? KyuubiSeal (talk) 21:01, 6 August 2011 (UTC)

Dell loves to nickel-and-dime us. What computer manufacturer gives us a better customer experience?
(Foreword: Don't include Apple. Their products are historically overpriced, and they seem more fragile than glass tables. In fact, they seem to nickle-and-dime users more than Dell, if anything.

Also, a brand/manufacturer that you recommend should allow me to choose my own component parts; have a build-your-own-PC system like Dell's when I order one online.)

So, what brand/manufacturer would I be happier with? How so? --70.179.165.67 (talk) 04:26, 6 August 2011 (UTC)


 * I wouldn't be able to offer much in the way of advice here. But what i do suggest is the same thing i do, i manage my own computer, i build my own, i make sure to keep any important information stored off my computer (external hard drives, cloud services etc) so when my computer starts to act up i just wipe it and start a-new. Now, if you don't know how to do this then my advice is a complete waste of your time, and for that i apologize. – Elliott (Talk|Cont)   04:36, 6 August 2011 (UTC)

I don't care much for brands either because I was once in a situation like yours and a friend recommended Cyberpowers PC to me. I think this is going to sound like an ad but they really are good. I have been using them for the last decade or so and having bought maybe six desktops from them, I have only become more loyal with each purchase. They are simply the best. They allow all sorts of customizations and are pretty cheap. You tell them what you want. They put it together and ship it to you. The only single thing that ever went wrong was that on my last machine that I got, they messed up the power connectors on the motherboard and the machine wouldn't start...took me about an hour to figure it out and then fix it...and in all fairness it was probably because I got it rushed. Just google them and check them out. 174.29.65.10 (talk) 06:21, 6 August 2011 (UTC)


 * Lateral question from a country where we have different names for our coins... What does "to nickel-and-dime us" mean? HiLo48 (talk) 06:31, 6 August 2011 (UTC)
 * It means to charge you a large total amount in many small fees. (As opposed to simply advertising the total price.) APL (talk) 08:35, 6 August 2011 (UTC)
 * Ah thanks. That makes sense. HiLo48 (talk) 08:51, 6 August 2011 (UTC)

This is an opinion question that the reference desk seems uniquely unqualified to answer. This is not a forum. Shadowjams (talk) 09:39, 6 August 2011 (UTC)
 * I'm perfectly happy using a few-years-old computer someone gave me. Buying or begging a second hand computer and installing linux and/or freeware would be the way to avoid having to over-pay. 92.28.252.178 (talk) 15:26, 6 August 2011 (UTC)


 * I agree with the above. I suggest you might want to get in touch with the IT directors from local schools, and ask them to give you a call when they upgrade their computers. I've gotten a few perfectly operating CPU's, only a couple of years old, in a "salvage sale" for ~$150-$250.  Of course, any type of warranty, return policy, or tech support is pretty much nonexistent in this scenario...but for $200, you can wipe it and install whatever you want. If it goes out on you in a year or two, just get another one.  In my experience, the larger, better funded schools (usually private grade schools or colleges) will upgrade every 2 or 3 years.  But don't forget to send the IT director a bottle of Scotch or something at Christmas time so he/she will be sure to "remember" to call you :) Quinn &#10048; BEAUTIFUL DAY  15:51, 6 August 2011 (UTC)
 * You could also try freecycle or ebay. Its not difficult to learn how to increase the memory or install a read-write DVD drive if needed, again using second hand components. 92.24.133.68 (talk) 14:57, 7 August 2011 (UTC)


 * Just to stick up for Dell's business practice here, I quite like the way they let me customise many things rather then quote different prices for subtly differently specified PCs based on guesses of "what is popular with other customers". I choose a chassis and CPU model, then up the spec on other components until I'm happy with the overall spec and price.  On the other hand, I do find Dells insistance on asking first which type of customer am I quite problematic - it restricts choice such that I have had to pretend to be a business just to be able to pick the laptop I want.  That said, I have seen manufacturers such as Toshiba list many subtly different models at a variety of fixed prices, but I don't know what their customer service is like.  Astronaut (talk) 22:52, 8 August 2011 (UTC)

LAN - WIRELESS CONNECTION
I have a LAN using Novell Netware 4.2 and have 30 users. The software using for my business purpose is developed in Clipper 5.0 and run on DOS background. Two of the users are about one kilometer away from the server. The connection is established through lan extenders and a dedicated rented dataline from BSNL. But many often the connection get interrupted or the line become failure and the billing process at the two nodes became suspended. Hence I wish to know whether there is any other networking method available to connect the two distant nodes uninterrupted. Also I may ask whether any LANEXTENDER with wireless facility will solve the problem. Any kind of help to activate the two nodes is expected. — Preceding unsigned comment added by 117.241.57.149 (talk) 09:24, 6 August 2011 (UTC)


 * Is it possible for all parties to get to the Internet ? If so, perhaps communicating that way would be both less expensive and more reliable. StuRat (talk) 16:06, 6 August 2011 (UTC)


 * Using wireless Ethernet could work if you had high gain antennas and a line of sight. But it is risky and liable to have interference or be disrupted.  You could also get unlit optical fibre installed but this could cost more than $20000.  Alternately you could get redundant data li.nks put in from another supplier, that would be less likely to fail at the same time.  Instead of a LAN extender, you may be better off with a higher quality router that can handle Novell protocols.  Older Novell protocols will not travel on the itnernet, but a suitable router could tunnel them over IP.  Graeme Bartlett (talk) 22:19, 6 August 2011 (UTC)


 * Given that you're using rather old protocols on a small network, I assume you're using relatively light amounts of bandwidth. I think having both nodes connected to the internet and running a VPN between the two (which could be encrypted if that was a concern) would be the simplest and cheapest. If you have to physically connect these nodes you're going to need something other than twisted pair (I read somewhere that Cat7 could support 1 km runs but that seems questionable), probably fiber.
 * Basic wireless (802.11x) is not a realistic option for 1km. There are guys at defcon in the desert who can span huge ranges sure, but that's not a practical option for a production system. If you want a wireless link you'll need to do something more professional.
 * If the distance was indoors you could string together cables and connect them with any level 2 device and it would work as a repeater. At least if I understand your setup from your description. But leveraging an already existing connection, like the internet, would be the easiest option imo. Broba (talk) 22:51, 6 August 2011 (UTC)
 * If there is some reason why you cannot use the internet, then it is possible to use a microwave link.  I don't know the regulations or suppliers in the USA, but you do need line of sight.  This reply is being sent via a series of such microwave links spanning about 10 km, and weather seems to be only a minor problem.  The cost is probably only about $200 if you can find a supplier in the USA.    D b f i r s   10:12, 8 August 2011 (UTC)

AntiSec's law enforcement hack
Recently hacked data was released from various law enforcement agencies, and the group responsible, Operation AntiSec, posted a lengthy hacklog of how it was done. It makes for a very comical read, but not having up-to-date knowledge of computer security myself, I'm curious: does this log illustrate that the hackers were very smart, or that the law enforcement sites' security was very stupid? Is there someplace where this kind of message is taken apart piece by piece, line by line, and the errors (or exploits) explained so that any idiot can prevent this happening to their system (or take advantage of someone whose system hasn't been updated)? Oh yeah, and last but not least... they couldn't really "call release inmate a few times" and have any success... right...? Wnt (talk) 22:47, 6 August 2011 (UTC)
 * It does detail some of what may have happened, but you have to understand php/mysql/command line in order to follow it. It does show that something strange was going on (shell_exec on an unsanitized query is simply asking for problems)...release inmate simply changes the date as to when inmates are to be released.Smallman12q (talk) 00:58, 7 August 2011 (UTC)

OpenSSL API
I'm trying to use some newly added functions from openssl that aren't available in the command line tool. My understanding is that the openssl libraries are accessible through an API itself that doesn't use the command line tool at all. Is that true? How does that interact with EVP (http://www.openssl.org/docs/crypto/evp.html)? Finally, do the wrappers written in different languages need to define which functions they give access to, or are there libraries that give access to all of the openssl functions (for example if something was added very recently) without needing to know those functions explicitly? Broba (talk) 23:07, 6 August 2011 (UTC)


 * 1) Library: yes, the openssl source distribution includes source for the ssl library (libssl) and its support cryptography library (libcrypto). It also includes the sources of binary programs like the openssl executable (in the apps folder) and some simpler examples (in the demos folder). Almost all programs that use openssl for cryptography will use libssl themselves.
 * 2) EVP: It looks like EVP is built into libcrypto, and is used by libssl, or you can call it yourself.
 * 3) Wrapper: Typically when you wrap a C or C++ library with an interface wrapper in another language, you have to write custom foreign function interface bindings that mirror the functions and datastructures of the underlying code. Usually this is done manually  (with things like Java's JNI or Python's ctypes, for example).  I've seen cases where people try to automate this, by analysing the C source files and emitting the appropriate glue logic (SWIG does this, for example). Usually the C/C++ API is deliberately kept pretty stable, and when it changes the wrappers all have to be changed to match.
 * -- Finlay McWalter ☻ Talk 00:36, 7 August 2011 (UTC)


 * Unsurprisingly, in the openssl distribution the libcrypto code is in the "crypto" folder and the libssl (that's the SSL/TLS protocol stuff) is in the "ssl" folder. But as a normal consumer of the code you need only use a binary distribution of openssl, with the appropriate headers and libraries. Finlay McWalter ☻ Talk 00:52, 7 August 2011 (UTC)