Time to blow off some steam regarding the network provider, Three. These guys supply both my broadband and my mobile needs. Last month, I renewed my Broadband contract with them for another 18 months, the carrot being the upgrade of my modem/router to the Huawei B525 for only €49. During this process, an additional €1.56 was mysteriously added to my forthcoming monthly bill, for "Data" used that was outside of my monthly allowance, so I queried them about it via their online 'Chat' service.
The person I spoke to couldn't explain where it came from, as I hadn't exceeded my monthly allowance, and told me not to worry about it, it would be taken care of. Fast forward a few weeks to when my bill was issued - the additional €1.56 was still there! As my direct-debit payment had yet to come into effect, on point of principle, I get onto Three again via Chat, to get this corrected, wasting almost 20 minutes this time, for a lousy €1.56. Again, I was assured that this would be corrected. Having just checked, the direct-debit went through and included the €1.56.
But whinging over €1.56 is not the purpose of this post. As I mentioned, they are also my mobile provider. About 2 months ago, I broke the LCD of my trusty but ancient Samsung D-500 mobile phone. As buying a new phone was cheaper than the cost of replacing the LCD, I went for, on Amazon, the cheapest phone I could find, a Awow 1802C. Mistake. This is an awful phone - almost no internal memory (making a memory-card essential), terrible Bluetooth audio, terrible radio reception, terrible camera and buggy firmware, where if an Alarm-alert is issued, it is often without sound, seemingly arbitrarily. But getting to know all this took time, more than the two weeks grace-period you get from Amazon before they will no longer accept returns. So I was stuck with it, but as I don't use a mobile much anyway, I figured I could live with it - alarm problem excepted; I've been late to work more than once because of this!
Until today anyway. I have 'topped-up' my credit twice since buying this. First time around, I was mystified as to how my credit had gotten used up so quickly, but as I needed to make a call, topped up online again. Today, I get a text from Three alerting me to the fact that my credit had all been used up - despite my having made just one single call of just 30 seconds since my last top-up.
So I contact Three today, again via Chat, and learned that this shitty Awow phone has been sending "data" without my knowledge, and Three has been only too happy to charge me for the service! My old phone, with the same sim-card, never pulled a stunt like this in the decade that I had it. But I think I was more pissed off to learn that the data packets that were being sent/recieved were only 0.2kB in size, about 10 in total, and for each, Three charged me €2!!! That's just criminal. I also learned that the €2 actually covers up to 200MB of data, all of which must be used on the same day - but not much use with a phone that has no memory, no web-browser, and when you don't even know that you're being charged 'data-rates' to begin with! And as a "Valued Customer" - the term the Three Rep. I contacted, used several times today! - Three sure as hell wasn't going to tip me off about it either. Ironically, in my effort to save money, Three has ended up ripping me off for more than the new phone has cost me. Heavy sigh.
But that's nothing compared to the scam they appear to have going with "Third-party Services" - or at the very least, 'aiding & abetting' these con-men. About two years ago, I had €95 worth of credit stolen from the same card-account, and all with Three's blessing!!! The organisation in question was called Prizevault, who somehow got hold of my mobile number and were charging me to be entered into weekly 'prize-draws' - all without ever having sent me a single email or text message apprising me of the fact. And they continued to do so until all of my credit had been used up! When I realised what had happened I immediately contacted Three, only to be told in no uncertain terms that as it was a "Third-party", the Three network was not liable! The Three Rep. I spoke to seemed genuinely sympathetic, even admitting that she (therefore, Three) was well aware that it was going on, and advised me to contact the Network Regulator here. This I did, and within about two weeks, received a cheque for €95 from Prizevault, but without explanation or admission of any wrong-doing on their part.
The fact though that Three allows these kind of scams to proliferate on their own networks speaks volumes about where their interests lie. It's also the reason I felt queasy today being labelled a "valued customer", when their above actions would suggest the exact opposite.
6 August 2017
5 August 2017
Huawei B525 4G Router/modem - mini-review...
I've been using the Huawei B525 for almost a month now, so I thought I'd post a 'first-impressions' review of it here.
First off, I consider this one ugly design. I thought so before I got it, and one month down the road, that opinion hasn't changed - I mean, what was Huawei thinking! Having only two control buttons, these extend across the entire top of the case, even though in use, you must press near the centre to get them to 'click' - another horrible design decision imo.
But that's about its only negative. Performance-wise, it is impressive, particularly when compared to the Huawei E5220. Although I mentioned in a previous post that due to the poor reception in my neck of the woods, I would be buying an external antenna, so far this seems unnecessary. The two (ugly) chunky rabbits-ears antennas that it comes with appear more than adequate for now. Usually, the five-bar signal strength LED's on the unit show between 2 - 3 bars, when the router is placed right next to the window. The E5220 on the other hand would get 1 - 2 bars, as seen from the browser - the unit itself had only a red-orange-green LED to indicate signal strength. So, at least for now, spending €20-€25 on an antenna seems an unnecessary extravagance.
Although I've been using dual-band wireless cards for years, this is the first dual-band router that I've owned, so wasn't sure of the setup. For instance, I wasn't expecting that two WiFi network signals would be available - I figured I'd still have a single network, through which the router would multiplex the 2.4 & 5Ghz bands. Apparently not! In fact, both networks cannot be selected simultaneously, either with Windows or Linux, which I find decidedly odd! Used in conjunction with Intel 7260-AC wireless cards, this is precisely how I thought it would operate - both bands used simultaneously, to achieve bit-rates in the order of 700-800Mb/s. So yeah, right now, I'm confused.
[3 years down the road, I've just noticed that the dual-networks displayed, has mysteriously disappeared, now with just a single network present. Also, now under the 'WiFi Basic Settings' for the B525, sub-option '5Ghz preferred', there is the explanation, "The 2.4 GHz and 5 GHz signals are combined and displayed as one. The
faster 5 GHz network is preferred when their signal strength is equal.
Turn off this switch to use them separately." Apparently the B525 has auto-updated its firmware in the not too distant past. This single-network makes a lot more sense. Under Linux's Network Manager->Connection Information, I've seen the 'Speed:' indication showing figures in the 500-625Mbps range (with a Intel 7260-AC wireless card) - which is another first!].
But impressed! Even using single-bands, my local network transfer-rates have gone from 1.4MB/s to 8-9MB/s. Surprisingly, even with the limited testing that I've done, transfer rates on the 5Ghz band are only marginally faster than rates on the 2.4Ghz band - I was expecting wonderous things from the 5Ghz band! Moreover, while faster, the 5Ghz band seems hardly usable when using it to stream video, even at bitrates as low as 200KB/s - the video stuttering every couple of minutes. Switching to the 2.4Ghz band, the stuttering stops. Although the 8-9MB/s is typical, often it can be as low as 5MB/s, rising momentarily to between 12-14MB/s - impressive!
Internet-wise, things are equally impressive. My download speed has gone from a once-in-a-blue-moon high of 1MB/s (average-high normally being 700kB/s) to a usual download rate of 1.5MB/s. Testing online often sees speeds of 14-16Mb/s, which blows away my ISP's average of 7Mb/s, and just tops the national average here. Upload speeds are equally impressive, being at least 2 - 4 times my old average. I downloaded an ISO via torrent just after getting it, and the download speed never dropped below 2MB/s. Unheard-of performance for me, and kinda scary - those speeds could eat through my 60GB monthly allowance in no time, and the out-of-allowance transfer charges here are just crazy. So yeah, I'm fairly happy with the performance. :)
The Huawei's firmware is also a lot better than that of the E5220, allowing much greater configurability.
One wrinkle that's become apparent is that the Linux wireless is performing under-par. This involves the 7260-AC wireless cards, not the Huawei B525. Within a day or so, the network in operation will fail, either disconnecting completely or becoming unresponsive. I then have to switch to the other network (2.4Ghz or 5Ghz). I've left Win7 running for days on my other PC and have never witnessed a dropped network, so the problem seems to be down to the Linux drivers that I'm using. But I think I'll save that for another post.
Overall, cosmetics excepted, I'm very happy with the Huawei B525's performance.
Edit:
I've been playing with my old TP-Link modem's upgraded OpenWrt firmware, and I gotta say that in retrospect, the Huawei B525 firmware really is garbage in comparison, feature-wise at least. I've only just realised that you cannot even set the TX power levels with this junk - that really is the pits!
I was warned recently about the dangers of emitted RF from these devices, by someone that is adversely effected by it. My immediate response was "Why don't you just turn down the TX power that's being emitted!" My first router had the option, the TP-Link had the option, the OpenWrt firmware has the option (on the devices it supports) - and all the time I presumed that all of the three Huawei modem/routers that I've owned had the option - they don't! That's perverse.
The B525's user-interface (again, in retrospect) seems particularly condescending in what features it allows the user access to. And this 'dumbed-down' approach appears a recurring theme as far as Huawei is concerned. I've just spent the last 30+ minutes searching the Web for information on this and all I'm being presented with is Huawei articles with the word "Basic" as part of their titles - does Huawei think that all of its customers are capable of understanding only the most rudimentary of information on the devices they've purchased?
As someone who hasn't had any health problems because of Wifi, it doesn't effect me, but having spoken to someone that is adversely effected by it, you would think that Huawei would be covering their arses from a class-action lawsuit by at least providing effected users of their equipment with a means of mitigating the problem? I would have thought that the 'ambulance-chaser' side of the legal profession would have been all over this by now.
Time I started thinking of getting a new non-Huawei modem, and maybe even a new ISP. Three have within the last six months upgraded their 4G network here, so that I now have gone from having a 1-bar signal strength, to a constant 4-bars - on the face of it, great! What sucks is the service they're providing. From 2-6am in the morning, I can get a blistering-fast (for here) 15-18mbps download speed, but during normal waking hours the speed rarely goes above 1.5mbps - the 'contention-ratio' Three are employing must be huge! Those 'normal' speeds are pathetic.
30 July 2017
Kingo: "This copy of Windows is not genuine" error...
As this blog no doubt reflects, practically all of my computing is done on Linux boxes. Any Windows 'assistance' that may be called for is usually more than catered for via VirtualBox - a VB-Windows XP in my case.
I do have a Windows 7 installation on one PC that I boot into occasionally, when circumstances demand, and yesterday, one such occasion cropped up. I had left my Android tablet on-charge for a day or so and voltage spikes from an intermittent charger lead had hopelessly corrupted the OS, with the result that it would no longer boot properly - just like my Mint 17.3 woes in fact! Anyway, I elected to reinstall Android 'Lollipop' from scratch, and for that, Windows and a USB cable was a prerequisite.
Once this had been achieved, I needed to 'root' the device in order to make it usable. Like the first time round, I went the 'Kingo' route (sic), downloading the software straight from their site. Avira, the (free) Win7 anti-virus software immediately flagged the download as containing Malware and quarantined it. Since I had used it previously with no ill-effects, I disabled Avira and tried again. When I ran the Kingo software, I had second thoughts about continuing when it started downloading stuff in the background. When it then produced a EULA box, I thought, fuck it, I don't remember any of this the first time round, so hit 'Decline' to end the installation. But it continued nonetheless! Next, what I assumed was the usual Windows warning of an 'unknown' driver being installed came up, to which I selected 'Cancel' - but that had no effect either! It just sat there until, running out of patience, I hit 'Install' , whereupon installation immediately recommenced . Upon completion, I got the usual Windows prompt that the computer needed to be restarted.
On rebooting, I was presented with the infamous black desktop with its, "This copy of Windows is not genuine" message. I had seen this a couple of times in the past, and it was always after doing a Windows update. But removing the dreaded 'KB971033' update had always 'fixed' it, and from past experience, I normally have updates disabled anyway. After Kingo had successfully rooted the tablet, I set about correcting the '...not genuine' situation. I was surprised to find that there was no KB971033 file installed, and even more surprised to discover that the 'Daz' Bootloader installation no longer had any effect! Given that this invariably 'reset' things in the past, I'm now inclined to think that M$ is somehow in cahoots with Kingo, out to disable as many Win7 boxes as possible, and by any means necessary. Right now, my Win7 installation is back on a 30-day Trial, with none of the usual 'fixes' having any effect whatsoever. Is M$ really this desperate, particularly with an OS that is practically End-Of-Line?!?!? Maybe, as I read not long ago that the uptake for their shitty Win10 offering was still only round 25%, with more than 50% of Windows users having chosen to stay with Windows 7.
I'm debating whether or not to reinstall a fresh copy, and Updates be damned! While I'd be more than happy to replace it with a copy of Win-XP, my new & improved 7260-AC wireless cards are not compatible with it. That, and the fact that it is also unable to access my external USB 3TB (GPT) hard-drive.
So, just a heads-up for anyone thinking of installing Kingo rooting software on their M$ boxes - you might get more than you bargained for!
Edit:
Although I swore to myself I'd never have anything to do with Kingo again, circumstances dictated otherwise. My car failed to start one morning and needing the check its computers fault-codes, I needed a rooted tablet.
Long story short, one week after installing Kingo on Win7 in order to root the tablet, the dreaded black wallpaper returned, along with the "Windows not genuine error".
Turned out to be the least of my problems as well. While setting about reinstalling Win7 from a backup on an unattached SSD, I presume I accidentally shorted something on the motherboard with a dangling USB cable, because after a restart, it failed to power up again. All I can say for certain is that the SSD I was attaching is dead. The motherboard probably is as well, but I won't know for certain till I can check the CPU. As penance for my stupidity, I've ordered from Amazon what is probably the last & cheapest AMD AM1 motherboard being sold, to do just that. Time will tell...
Fuck Windows, but more so, that Kingo garbage. This has proven to be one expensive lesson.
I do have a Windows 7 installation on one PC that I boot into occasionally, when circumstances demand, and yesterday, one such occasion cropped up. I had left my Android tablet on-charge for a day or so and voltage spikes from an intermittent charger lead had hopelessly corrupted the OS, with the result that it would no longer boot properly - just like my Mint 17.3 woes in fact! Anyway, I elected to reinstall Android 'Lollipop' from scratch, and for that, Windows and a USB cable was a prerequisite.
Once this had been achieved, I needed to 'root' the device in order to make it usable. Like the first time round, I went the 'Kingo' route (sic), downloading the software straight from their site. Avira, the (free) Win7 anti-virus software immediately flagged the download as containing Malware and quarantined it. Since I had used it previously with no ill-effects, I disabled Avira and tried again. When I ran the Kingo software, I had second thoughts about continuing when it started downloading stuff in the background. When it then produced a EULA box, I thought, fuck it, I don't remember any of this the first time round, so hit 'Decline' to end the installation. But it continued nonetheless! Next, what I assumed was the usual Windows warning of an 'unknown' driver being installed came up, to which I selected 'Cancel' - but that had no effect either! It just sat there until, running out of patience, I hit 'Install' , whereupon installation immediately recommenced . Upon completion, I got the usual Windows prompt that the computer needed to be restarted.
On rebooting, I was presented with the infamous black desktop with its, "This copy of Windows is not genuine" message. I had seen this a couple of times in the past, and it was always after doing a Windows update. But removing the dreaded 'KB971033' update had always 'fixed' it, and from past experience, I normally have updates disabled anyway. After Kingo had successfully rooted the tablet, I set about correcting the '...not genuine' situation. I was surprised to find that there was no KB971033 file installed, and even more surprised to discover that the 'Daz' Bootloader installation no longer had any effect! Given that this invariably 'reset' things in the past, I'm now inclined to think that M$ is somehow in cahoots with Kingo, out to disable as many Win7 boxes as possible, and by any means necessary. Right now, my Win7 installation is back on a 30-day Trial, with none of the usual 'fixes' having any effect whatsoever. Is M$ really this desperate, particularly with an OS that is practically End-Of-Line?!?!? Maybe, as I read not long ago that the uptake for their shitty Win10 offering was still only round 25%, with more than 50% of Windows users having chosen to stay with Windows 7.
I'm debating whether or not to reinstall a fresh copy, and Updates be damned! While I'd be more than happy to replace it with a copy of Win-XP, my new & improved 7260-AC wireless cards are not compatible with it. That, and the fact that it is also unable to access my external USB 3TB (GPT) hard-drive.
So, just a heads-up for anyone thinking of installing Kingo rooting software on their M$ boxes - you might get more than you bargained for!
Edit:
Although I swore to myself I'd never have anything to do with Kingo again, circumstances dictated otherwise. My car failed to start one morning and needing the check its computers fault-codes, I needed a rooted tablet.
Long story short, one week after installing Kingo on Win7 in order to root the tablet, the dreaded black wallpaper returned, along with the "Windows not genuine error".
Turned out to be the least of my problems as well. While setting about reinstalling Win7 from a backup on an unattached SSD, I presume I accidentally shorted something on the motherboard with a dangling USB cable, because after a restart, it failed to power up again. All I can say for certain is that the SSD I was attaching is dead. The motherboard probably is as well, but I won't know for certain till I can check the CPU. As penance for my stupidity, I've ordered from Amazon what is probably the last & cheapest AMD AM1 motherboard being sold, to do just that. Time will tell...
Fuck Windows, but more so, that Kingo garbage. This has proven to be one expensive lesson.
1 June 2017
My Wireless Experiences - a rethink...
The fickle finger of fate has had me burning the midnight oil once again trying to improve my computers wireless connectivity. Contrary to my somewhat damning previous conclusions regarding Intel cards, I now believe, that it is probably better the Devil you know...
Not to put too fine a point on it, I'm currently of the opinion that Broadcom wireless cards suck huge! Up until recently, my work with the last Broadcom cards that I had bought - one BCM43142A0 and a BCM94322 - involved the new(ish) mini-PC that I had put together, and that revolved around the AMD Athlon 5350. Before its arrival, I was using an integrated quad-core Intel Atom 330, Zotac motherboard. This was bought pre-built and had come with a full-length Atheros wireless card. Having it left running 24/7 for 3.5 years, I can say with some confidence that wireless on this setup was lousy. I in fact have never come across one Atheros-based card that has performed well, and I've had 2-3 of them. Anyway, with the AMD motherboard taking only a single half-length wireless card, I needed to go a-shopping for anything other than Atheros - enter Intel and Broadcom.
Given how few peripheral slots are available on mini-ITX boards, a single wireless/bluetooth solution seemed always the way to go, hence my little obsession with getting this setup to work properly. My experience with Intel's N-7260 had left me somewhat jaundiced by the whole affair admittedly, but work it did, after a fashion, at least with Linux. I have had the 7260 in my Dell laptop, running 24/7 for almost a year now, and say what you will about Intel, their wireless cards rock! Hardly anything to complain about, despite earlier rantings of mine in a previous post. Bluetooth excepted of course - this continues to suffer from the minor annoyance of either failing to start at boot, or when Bluetooth is switched off. Suspending/Resuming the computer a few times will eventually see it start up again.
The N-7260 running with Windows 7 is another matter entirely, at least where Bluetooth is concerned. Wireless is pretty much flawless, just as good as with Linux. Bluetooth, despite another exhaustive attempt to get to the bottom of the matter, refuses to work at all, which is just plain bizarre. After having downloaded and tried every single relevant Intel driver, I recently downloaded & tried drivers from both Dell and Lenovo, in the futile hope that one of them contained the much sought-after magic sauce. All in vain of course. Intriguingly, I googled a reference by someone that said that their 7260 problem was caused by an installed SSD that he was using, so yesterday, I opted for the onerous chore of installing Win7 on a hard-drive to see if that would fix things - the mini-ITX/Athlon 5350 setup has two SSD in it. Unfortunately, shitty Windows refused to continue with the installation when it discovered an ex4 Linux Mint partition on the same drive, claiming that it was a GPT partition and could not continue. So the jury is still out on that one.
But onto my Broadcom rant. Like I said, wireless had always been lousy with my previous Intel Atom/Atheros setup, so when I upgraded to the AMD 5350/Broadcom setup, it took a while to discover that I seemed to have similar problems with a completely different setup, but both use(d) the same all-metal case. Hence my suspicions grew that maybe it was the computer case/antenna setup that was causing the problem, not the wireless cards themselves, and there I left the matter until recently, when wireless once again had gotten so bad with the mini-ITX that I could not even stream a movie from it without it either stuttering badly or the connection failing completely. I therefore opted to investigate in depth the problem and was appalled by what I discovered about the Broadcom cards - namely, they suck!
Putting my (now!) trusty old N-7260 in the mini-ITX computer, I was amazed to find that I had a blisteringly fast & reliable wireless connection. Up until then, I had never really checked the wireless, having been obsessed with the Bluetooth problem. So now I knew the issue had nothing to do with either the metal case or the antenna - two things less to consider! I then went checking the Broadcom cards. Of the two, the one with integrated Bluetooth (43142) was also by far the worst. [Note: This family of cards has such a poor reputation, that (evidently) Linux doesn't support them 'out-of-the-box' - drivers must be manually installed from the repository (bcmwl-kernel-source)] With it located just 1-2m from the router, and Win7 indicating it (unsurprisingly) had 'Excellent' signal strength, it would routinely stutter and stop entirely every couple of minutes. I watched this in disbelief for hours. It performed similarly when in the laptop, whether beside the router or at distance in another room. So this card was completely useless, at least as far as wireless was concerned. Its Bluetooth worked perfectly however, though not (at all!) on Linux, but more on that later. The other Broadcom card's wireless performance, the 4322, was much better, but that still wasn't saying much. It was after all what had prompted the investigation in the first place.
My one and only question is, how can Broadcom even exist as a business entity selling wireless cards that perform this badly? Of the three broadcom cards I've bought, none could be remotely considered as being thought of as a 'good' wireless card.
Given that currently the N-7260 must remain in the mini-ITX computer, and that my Dell laptop would have been Bluetooth-less without it, I attempted to get the BCM43142's bluetooth working on Linux - and succeeded! Mainly thanks to the following user-posts that I googled;
https://www.spinics.net/lists/linux-bluetooth/msg47879.html
https://forums.linuxmint.com/viewtopic.php?t=224099
https://github.com/winterheart/broadcom-bt-firmware/blob/master/brcm/BCM43142A0-04ca-2006.hcd
The first, refreshingly describes the steps carried out by the Broadcom card in order to initialise Bluetooth, and where to look for said information in Linux. The second, though not a fix in my case, shows you the necessary steps needed for removing/reinitialising Bluetooth and where to go looking. The third link points to the BT firmware needed by the card itself in order for it to work. Without this, the BT adapter will appear to be working, but scanning for BT devices will reveal nothing. Note that this firmware is for my Broadcom wireless card, other cards may/will use different firmware. You will need to check the Linux "kern.log" file in order to find what Broadcom 'Patch' firmware the BT adapter is trying & failing to load. Do a google for this, copy it into wherever your flavour of Linux keeps this stuff (for Mint, it's '/lib/firmware/brcm') then remove/reinitialise Bluetooth. Hey Presto! - everything should then magically work, or at least in my case it does!
Thankfully I have two wireless slots available on my laptop. One now contains the BCM43142 card for its Bluetooth, which works perfectly, unlike the N-7260. Its wireless is also available of course, but is for the most part, unusable - Shame on you Broadcom! The other slot contains my other trusty Intel card, the 5100.
Right now I'm thinking of upgrading my modem/router from the old but reliable Huawei E3220, to a Huawei E5577. The 4G feature is nice, but where I live, not really an option for the foreseeable future, given the 3G reception itself is poor most of the time. But it does have a dual-band WLAN, whereas the E3220 is limited to just the 2.4G band. This will also (hopefully) finally allow me to see a significant increase in network speed. Right now for practical purposes it's limited to about 1.4MB/s which in this day & age, just doesn't cut it.
That will also mean that I will be considering some more wireless card purchases, and despite my previous utterances, it seems that Intel is the only way to go. Unfortunately the 7260-AC cards are the most highly rated (and expensive) that they currently offer, and given my nightmarish time with the N-7260, I'm reticent about pulling the trigger on one of these. The 'AC' part is also of little use to the E5577, which is not AC-enabled. Decisions, decisions. One thing I know for certain - I will not be buying any more Broadcom cards!!!
Edit1:
Some time later... Windoze finally installed after I had formatted the hard-disk beforehand. No difference whatsoever whether it's on a SSD or a hard-disk, as far as Bluetooth is concerned. I even tried two flavours of Windows 7, Win 7 Home Basic & Win 7 Ultimate - no difference! I then stumbled across another post where someone said that clearing the CMOS info fixed it for him - not for me it didn't! I finally updated the Bios, from version 1.2 to 1.5, no difference there either.
As a last resort, I started downloading & installing Win8.1 & Win10 drivers. I was exhilarated initially to discover that with early Win8.1 drivers, after a reboot, the drivers installed without errors!!! Unfortunately, that is all they did. When I tried changing & saving Bluetooth settings, it failed, saying the device might be unplugged. There was also no Bluetooth tray icon or Control panel settings to indicate that Bluetooth even existed. Naturally, trying to 'Add' a Bluetooth device detected nothing.
So after fighting the good fight, I am admitting defeat on this. Although as I have already said, the N-7260 works fine with Linux, one interesting thing I found was that when installed on the mini-ITX motherboard, I could not disable it from the Bios - I still have a wireless connection when it's supposedly disabled! That is suggestive of a hardware problem somewhere, and given the grief this card has put me through, it's probably the cause.
After swearing to never get another Intel card, I have just gone and ordered 2 7260 AC cards! This means that I will not be getting the E5577 router after all, going more upmarket and getting a Huawei B525. I can get one of these babies for just €50 from my ISP as long as I sign up for another 18 months with them. Given I'm with them well over a decade already, that's no biggie. These also allow for an external antenna, which given the lacklustre 3G reception at this location, means that that's one more thing I'll have to order. Fun, fun, fun.
Edit2:
My curiosity got the better of me, so I tested my N-7260 with Windows 10 - the same problem exists! At this stage, it's looking very likely that it's a hardware fault with the card, and nothing to do with the drivers. Along with the usual 'Cannot start (code 10)' error, one additional piece of info that is shown on Windows 10's Device Manager is a line saying that the Bluetooth adapter is suffering from a 'Power_failure'. I'd probably have taken a hammer to it ages ago but for the fact that N-7260's Bluetooth works fine with Linux. I also discovered why the N-7260 still had internet access when disabled in the Bios - I had put a sliver of cellotape over one of the pins ages ago (a Youtube 'fix' I suspect) and had forgotten about it! When removed, the card once again honoured the Bios settings.
With regard to Windows 10, that was the first opportunity I had to play with it. Wow, I was underwhelmed - I can see why Microsoft has been trying to sneak it onto people's PC's while they are asleep. I cannot understand why they abandoned the 3D graphics look - that 2D desktop looks terrible, like something from the Windows 3.1 era. Windows 7 is a far more attractive desktop.
Edit3:
Haha, even Broadcom's BCM43142 Bluetooth is crap! While it does work, and is usable when listening to music, watching a movie etc. is another matter entirely. The audio stream is full of short intermittent breaks, resulting in the audio going out of sync with the video after just a minute or so. I never had a problem like that with the Intel's N-7260. Yep, Broadcom sure sucks!!!
Not to put too fine a point on it, I'm currently of the opinion that Broadcom wireless cards suck huge! Up until recently, my work with the last Broadcom cards that I had bought - one BCM43142A0 and a BCM94322 - involved the new(ish) mini-PC that I had put together, and that revolved around the AMD Athlon 5350. Before its arrival, I was using an integrated quad-core Intel Atom 330, Zotac motherboard. This was bought pre-built and had come with a full-length Atheros wireless card. Having it left running 24/7 for 3.5 years, I can say with some confidence that wireless on this setup was lousy. I in fact have never come across one Atheros-based card that has performed well, and I've had 2-3 of them. Anyway, with the AMD motherboard taking only a single half-length wireless card, I needed to go a-shopping for anything other than Atheros - enter Intel and Broadcom.
Given how few peripheral slots are available on mini-ITX boards, a single wireless/bluetooth solution seemed always the way to go, hence my little obsession with getting this setup to work properly. My experience with Intel's N-7260 had left me somewhat jaundiced by the whole affair admittedly, but work it did, after a fashion, at least with Linux. I have had the 7260 in my Dell laptop, running 24/7 for almost a year now, and say what you will about Intel, their wireless cards rock! Hardly anything to complain about, despite earlier rantings of mine in a previous post. Bluetooth excepted of course - this continues to suffer from the minor annoyance of either failing to start at boot, or when Bluetooth is switched off. Suspending/Resuming the computer a few times will eventually see it start up again.
The N-7260 running with Windows 7 is another matter entirely, at least where Bluetooth is concerned. Wireless is pretty much flawless, just as good as with Linux. Bluetooth, despite another exhaustive attempt to get to the bottom of the matter, refuses to work at all, which is just plain bizarre. After having downloaded and tried every single relevant Intel driver, I recently downloaded & tried drivers from both Dell and Lenovo, in the futile hope that one of them contained the much sought-after magic sauce. All in vain of course. Intriguingly, I googled a reference by someone that said that their 7260 problem was caused by an installed SSD that he was using, so yesterday, I opted for the onerous chore of installing Win7 on a hard-drive to see if that would fix things - the mini-ITX/Athlon 5350 setup has two SSD in it. Unfortunately, shitty Windows refused to continue with the installation when it discovered an ex4 Linux Mint partition on the same drive, claiming that it was a GPT partition and could not continue. So the jury is still out on that one.
But onto my Broadcom rant. Like I said, wireless had always been lousy with my previous Intel Atom/Atheros setup, so when I upgraded to the AMD 5350/Broadcom setup, it took a while to discover that I seemed to have similar problems with a completely different setup, but both use(d) the same all-metal case. Hence my suspicions grew that maybe it was the computer case/antenna setup that was causing the problem, not the wireless cards themselves, and there I left the matter until recently, when wireless once again had gotten so bad with the mini-ITX that I could not even stream a movie from it without it either stuttering badly or the connection failing completely. I therefore opted to investigate in depth the problem and was appalled by what I discovered about the Broadcom cards - namely, they suck!
Putting my (now!) trusty old N-7260 in the mini-ITX computer, I was amazed to find that I had a blisteringly fast & reliable wireless connection. Up until then, I had never really checked the wireless, having been obsessed with the Bluetooth problem. So now I knew the issue had nothing to do with either the metal case or the antenna - two things less to consider! I then went checking the Broadcom cards. Of the two, the one with integrated Bluetooth (43142) was also by far the worst. [Note: This family of cards has such a poor reputation, that (evidently) Linux doesn't support them 'out-of-the-box' - drivers must be manually installed from the repository (bcmwl-kernel-source)] With it located just 1-2m from the router, and Win7 indicating it (unsurprisingly) had 'Excellent' signal strength, it would routinely stutter and stop entirely every couple of minutes. I watched this in disbelief for hours. It performed similarly when in the laptop, whether beside the router or at distance in another room. So this card was completely useless, at least as far as wireless was concerned. Its Bluetooth worked perfectly however, though not (at all!) on Linux, but more on that later. The other Broadcom card's wireless performance, the 4322, was much better, but that still wasn't saying much. It was after all what had prompted the investigation in the first place.
My one and only question is, how can Broadcom even exist as a business entity selling wireless cards that perform this badly? Of the three broadcom cards I've bought, none could be remotely considered as being thought of as a 'good' wireless card.
Given that currently the N-7260 must remain in the mini-ITX computer, and that my Dell laptop would have been Bluetooth-less without it, I attempted to get the BCM43142's bluetooth working on Linux - and succeeded! Mainly thanks to the following user-posts that I googled;
https://www.spinics.net/lists/linux-bluetooth/msg47879.html
https://forums.linuxmint.com/viewtopic.php?t=224099
https://github.com/winterheart/broadcom-bt-firmware/blob/master/brcm/BCM43142A0-04ca-2006.hcd
The first, refreshingly describes the steps carried out by the Broadcom card in order to initialise Bluetooth, and where to look for said information in Linux. The second, though not a fix in my case, shows you the necessary steps needed for removing/reinitialising Bluetooth and where to go looking. The third link points to the BT firmware needed by the card itself in order for it to work. Without this, the BT adapter will appear to be working, but scanning for BT devices will reveal nothing. Note that this firmware is for my Broadcom wireless card, other cards may/will use different firmware. You will need to check the Linux "kern.log" file in order to find what Broadcom 'Patch' firmware the BT adapter is trying & failing to load. Do a google for this, copy it into wherever your flavour of Linux keeps this stuff (for Mint, it's '/lib/firmware/brcm') then remove/reinitialise Bluetooth. Hey Presto! - everything should then magically work, or at least in my case it does!
Thankfully I have two wireless slots available on my laptop. One now contains the BCM43142 card for its Bluetooth, which works perfectly, unlike the N-7260. Its wireless is also available of course, but is for the most part, unusable - Shame on you Broadcom! The other slot contains my other trusty Intel card, the 5100.
Right now I'm thinking of upgrading my modem/router from the old but reliable Huawei E3220, to a Huawei E5577. The 4G feature is nice, but where I live, not really an option for the foreseeable future, given the 3G reception itself is poor most of the time. But it does have a dual-band WLAN, whereas the E3220 is limited to just the 2.4G band. This will also (hopefully) finally allow me to see a significant increase in network speed. Right now for practical purposes it's limited to about 1.4MB/s which in this day & age, just doesn't cut it.
That will also mean that I will be considering some more wireless card purchases, and despite my previous utterances, it seems that Intel is the only way to go. Unfortunately the 7260-AC cards are the most highly rated (and expensive) that they currently offer, and given my nightmarish time with the N-7260, I'm reticent about pulling the trigger on one of these. The 'AC' part is also of little use to the E5577, which is not AC-enabled. Decisions, decisions. One thing I know for certain - I will not be buying any more Broadcom cards!!!
Edit1:
Some time later... Windoze finally installed after I had formatted the hard-disk beforehand. No difference whatsoever whether it's on a SSD or a hard-disk, as far as Bluetooth is concerned. I even tried two flavours of Windows 7, Win 7 Home Basic & Win 7 Ultimate - no difference! I then stumbled across another post where someone said that clearing the CMOS info fixed it for him - not for me it didn't! I finally updated the Bios, from version 1.2 to 1.5, no difference there either.
As a last resort, I started downloading & installing Win8.1 & Win10 drivers. I was exhilarated initially to discover that with early Win8.1 drivers, after a reboot, the drivers installed without errors!!! Unfortunately, that is all they did. When I tried changing & saving Bluetooth settings, it failed, saying the device might be unplugged. There was also no Bluetooth tray icon or Control panel settings to indicate that Bluetooth even existed. Naturally, trying to 'Add' a Bluetooth device detected nothing.
So after fighting the good fight, I am admitting defeat on this. Although as I have already said, the N-7260 works fine with Linux, one interesting thing I found was that when installed on the mini-ITX motherboard, I could not disable it from the Bios - I still have a wireless connection when it's supposedly disabled! That is suggestive of a hardware problem somewhere, and given the grief this card has put me through, it's probably the cause.
After swearing to never get another Intel card, I have just gone and ordered 2 7260 AC cards! This means that I will not be getting the E5577 router after all, going more upmarket and getting a Huawei B525. I can get one of these babies for just €50 from my ISP as long as I sign up for another 18 months with them. Given I'm with them well over a decade already, that's no biggie. These also allow for an external antenna, which given the lacklustre 3G reception at this location, means that that's one more thing I'll have to order. Fun, fun, fun.
Edit2:
My curiosity got the better of me, so I tested my N-7260 with Windows 10 - the same problem exists! At this stage, it's looking very likely that it's a hardware fault with the card, and nothing to do with the drivers. Along with the usual 'Cannot start (code 10)' error, one additional piece of info that is shown on Windows 10's Device Manager is a line saying that the Bluetooth adapter is suffering from a 'Power_failure'. I'd probably have taken a hammer to it ages ago but for the fact that N-7260's Bluetooth works fine with Linux. I also discovered why the N-7260 still had internet access when disabled in the Bios - I had put a sliver of cellotape over one of the pins ages ago (a Youtube 'fix' I suspect) and had forgotten about it! When removed, the card once again honoured the Bios settings.
With regard to Windows 10, that was the first opportunity I had to play with it. Wow, I was underwhelmed - I can see why Microsoft has been trying to sneak it onto people's PC's while they are asleep. I cannot understand why they abandoned the 3D graphics look - that 2D desktop looks terrible, like something from the Windows 3.1 era. Windows 7 is a far more attractive desktop.
Edit3:
Haha, even Broadcom's BCM43142 Bluetooth is crap! While it does work, and is usable when listening to music, watching a movie etc. is another matter entirely. The audio stream is full of short intermittent breaks, resulting in the audio going out of sync with the video after just a minute or so. I never had a problem like that with the Intel's N-7260. Yep, Broadcom sure sucks!!!
19 February 2017
'www.oceanfree.net' email server back in operation.
'www.oceanfree.net' has been on life-support for many years, allowing only the receiving of emails, and no possibility of creating new email accounts. It had in fact at one time given its users a date by which it would scrap the service entirely. Never happened for some reason.
About a week ago, I realised that all emails from Oceanfree had ceased, and figured that they had finally decided the put this mut out of its misery. Nope! Some kind of 'reset' was after been done though, because when I signed in online - doing so from a email client no longer had any effect - I discovered that I could once again receive emails. Thinking something might be afoot, I also tried sending emails from my oceanfree account and was elated to find that this also worked - first time in at least 8 years! Unfortunately, while receiving emails using an email-client is once again possible, it refuses to allow the sending of emails from anything other than web-browsers.
Getting excited about an pop3 email server is kinda sad I admit, but this was my first email address and the only one that I have used consistently over the years, Microsoft's excepted. Since it was mainly used as a junk-email dump, it would have been no great loss either, but then, there's the Nostalgia Factor to consider. :)
The company behind oceanfree, British Telecom, seems to be out to take another stab at the Irish broadband market as it is adding an advert - "from €20 broadband" - to all emails going through its servers. Seeing as I'm paying twice that to Three, another British-based company, I thought it worth checking out, but couldn't find any info on what you would get for that money. Just looking to generate traffic it seems.
Edit:
Despite the above rhetoric (nearly 2 years ago), an indication of how little time I have for this Oceanfree email account is that I've only just discovered that my account's password has been hacked - a couple of months ago, in fact! - and some clown is attempting to extort money, both by threatening to release 'saucy' details contained therein to my family/associates etc. or else triggering an embedded virus to wreak havoc on my system. Good luck with that. This actually gave me a giggle, given that, a) I'm Linux-based, and b) it's been virtually unused for nearly two decades and contains nothing but junk mail! The only reason they managed to hack this account is that it is not on a secure https server - also something I've only just realised...
Which renders all of the above, null & void! If BT were looking to resurrect the Oceanfree email service, the very first thing they would have done was stick it on a secure server - which they didn't, obviously 'cos the notion existed only in my head! Confirmation of this was found when I went to change the email account's login password - I got a 404 'Page not found' error!!!
All of this doesn't explain why sending emails suddenly started working again. Not that any of this matters now, given what has happened.
Final edit:
A couple of months on, and someone has finally drove a stake through this dog's heart - the 'www.oceanfree.net' mail server is no more, with the link now redirecting to a paid-for web-services site. The end of an era, if only in my own head, it being my first email address and all. It will at least put paid to one wannabe extortionist. They'd been spamming this account for over two years with this nonsense, varying the amount demanded by a few hundred dollars, but always with the proviso that this would be their final demand - pay now or suffer the consequences! It would be funny but for the fact that there must be a sizeable percentage of people that fall for this stuff & pay-up.
Oceanfree - RIP.
About a week ago, I realised that all emails from Oceanfree had ceased, and figured that they had finally decided the put this mut out of its misery. Nope! Some kind of 'reset' was after been done though, because when I signed in online - doing so from a email client no longer had any effect - I discovered that I could once again receive emails. Thinking something might be afoot, I also tried sending emails from my oceanfree account and was elated to find that this also worked - first time in at least 8 years! Unfortunately, while receiving emails using an email-client is once again possible, it refuses to allow the sending of emails from anything other than web-browsers.
Getting excited about an pop3 email server is kinda sad I admit, but this was my first email address and the only one that I have used consistently over the years, Microsoft's excepted. Since it was mainly used as a junk-email dump, it would have been no great loss either, but then, there's the Nostalgia Factor to consider. :)
The company behind oceanfree, British Telecom, seems to be out to take another stab at the Irish broadband market as it is adding an advert - "from €20 broadband" - to all emails going through its servers. Seeing as I'm paying twice that to Three, another British-based company, I thought it worth checking out, but couldn't find any info on what you would get for that money. Just looking to generate traffic it seems.
Edit:
Despite the above rhetoric (nearly 2 years ago), an indication of how little time I have for this Oceanfree email account is that I've only just discovered that my account's password has been hacked - a couple of months ago, in fact! - and some clown is attempting to extort money, both by threatening to release 'saucy' details contained therein to my family/associates etc. or else triggering an embedded virus to wreak havoc on my system. Good luck with that. This actually gave me a giggle, given that, a) I'm Linux-based, and b) it's been virtually unused for nearly two decades and contains nothing but junk mail! The only reason they managed to hack this account is that it is not on a secure https server - also something I've only just realised...
Which renders all of the above, null & void! If BT were looking to resurrect the Oceanfree email service, the very first thing they would have done was stick it on a secure server - which they didn't, obviously 'cos the notion existed only in my head! Confirmation of this was found when I went to change the email account's login password - I got a 404 'Page not found' error!!!
All of this doesn't explain why sending emails suddenly started working again. Not that any of this matters now, given what has happened.
Final edit:
A couple of months on, and someone has finally drove a stake through this dog's heart - the 'www.oceanfree.net' mail server is no more, with the link now redirecting to a paid-for web-services site. The end of an era, if only in my own head, it being my first email address and all. It will at least put paid to one wannabe extortionist. They'd been spamming this account for over two years with this nonsense, varying the amount demanded by a few hundred dollars, but always with the proviso that this would be their final demand - pay now or suffer the consequences! It would be funny but for the fact that there must be a sizeable percentage of people that fall for this stuff & pay-up.
Oceanfree - RIP.
7 January 2017
Linux Mint 18.1: Mini review...
Apart from Ubuntu 10.04 LTS, Linux Mint is my next most-used Linux distro. Deepin Linux of course has a place in my heart as well, but right now, the file-managers available are just too 'primitive' for my liking. The Deepin developers themselves seemed to have already decided that the Gnome3-based file-manager ('Files') was a bit of a dog, so started developing one of their own. This is great to see, but right now, the Deepin file-manager itself is fairly limited, and as a result, I rarely use my Deepin installation anymore - I'm waiting for the Beta of the next version to be released.
So right now, Mint 17.3 is my main OS - with the Mate desktop of course! When Mint 18 was released, one of the main Mate developments was the phasing out of the 'old' GTK2 libraries, to make way for their GTK3 equivalents. Given how much I detest the Gnome3 desktop and its associated components, I took this as a bad omen as far as Mate's future was concerned. Therefore I gave Mint 18 a pass. When Mint 18.1 was released however, curiousity got the better of me. Booting 18.1 from a USB drive, I was pleasantly surprised by what I saw.
First surprise was that Compiz seemed to work perfectly, right out of the box. Next was how 'unchanged' the Mate desktop appears, despite its use of GTK3 - well done Mate developers! It also ran nice and quickly. Things were looking good.
Given these favourable first impressions, I decided that I may as well replace Mint 17.3, but backed it up beforehand - lucky that I did too!
First problem that became obvious was that 18.1 fails to 'validate' my Transcend SSD. It 'detects' it ok, but when it fails to 'validate' it, it just disables it completely - major bummer! This ruled out any possibility of replacing 17.3, at least by performing a normal installation. Not to be outdone, I then decided to install it on a Hitachi hard-drive, make an image-file of the installation, then overwrite the 17.3 SSD partition with this. After doing all of this, imagine my disgust when Mint 18.3, after booting from the SSD, then proceeded to disable the same SSD when it couldn't complete the 'validation' process! Fighting the good fight, lastly I tried installing a few different kernel versions in the hope that one of them could validate the SSD - nope. Having run out of ideas, I switched to the Hitachi HD for a week of Mint 18.1, to see if it really was worth pursuing any further.
It wasn't. There are just too many bugs present, some old, some newly introduced. The SSD thing itself is a deal-breaker. Then there's Compiz, which despite my earlier hopes, is just too unstable, regularly crashing. Unbelieveably, the audio-preview 'Caja-freezing' bug is still alive & well. Audio-previews, which are disabled by default, tend to suggest that the Mint developers are aware of this problem as well but are unable to fix it. Thing is, when I compile Caja from source, on 17.3 anyway, the problem disappears! A next bug involves audio as well. Speaker audio works fine. However when you plug in a headphone/external speaker plug, audio is always muted, requiring that the user calls up Audio preferences and toggles from Speaker to Headset or vice-versa - doesn't matter which, it just needs to be toggled to start audio getting piped externally. Bloody annoying. Then there's the Bluetooth fiasco. First off, I couldn't get the fully-featured 'Blueman' to work at all. After many hours of trying to get Bluetooth working through 'Blueberry', the feature-lite but new bluetooth manager, I eventually figured out the 'sequence' of actions needed in order to get it running. Problem is, once you disconnect the bluetooth headset, you need to go through the same rigmarole in order to re-connect, which is just not acceptable for everyday use.
There it stands. Another Mint that was worth a look, but with my hardware at least, is just too buggy to be usable. Thanks to the 17.3 partition image that I had created, switching back was a relatively painless procedure. I was impressed though with Mate in general, so much so, that I actually donated €10 towards the project! The developers are doing a great job staying true to the Gnome2 desktop model, despite having had to go the GTK3 route. Although I submitted a bug-report on the SSD problem with Mint, I later tried the same SSD on a HP Pavilion laptop with Mint 18.1 running from the USB, and Mint had no problem detecting & validating the drive. So it's a kernel/hardware combination incompatibility that is causing the problem - after all, on the same Dell laptop, Mints from 16 -> 17.3 work perfectly with the Transcend drive.
So right now, Mint 17.3 is my main OS - with the Mate desktop of course! When Mint 18 was released, one of the main Mate developments was the phasing out of the 'old' GTK2 libraries, to make way for their GTK3 equivalents. Given how much I detest the Gnome3 desktop and its associated components, I took this as a bad omen as far as Mate's future was concerned. Therefore I gave Mint 18 a pass. When Mint 18.1 was released however, curiousity got the better of me. Booting 18.1 from a USB drive, I was pleasantly surprised by what I saw.
First surprise was that Compiz seemed to work perfectly, right out of the box. Next was how 'unchanged' the Mate desktop appears, despite its use of GTK3 - well done Mate developers! It also ran nice and quickly. Things were looking good.
Given these favourable first impressions, I decided that I may as well replace Mint 17.3, but backed it up beforehand - lucky that I did too!
First problem that became obvious was that 18.1 fails to 'validate' my Transcend SSD. It 'detects' it ok, but when it fails to 'validate' it, it just disables it completely - major bummer! This ruled out any possibility of replacing 17.3, at least by performing a normal installation. Not to be outdone, I then decided to install it on a Hitachi hard-drive, make an image-file of the installation, then overwrite the 17.3 SSD partition with this. After doing all of this, imagine my disgust when Mint 18.3, after booting from the SSD, then proceeded to disable the same SSD when it couldn't complete the 'validation' process! Fighting the good fight, lastly I tried installing a few different kernel versions in the hope that one of them could validate the SSD - nope. Having run out of ideas, I switched to the Hitachi HD for a week of Mint 18.1, to see if it really was worth pursuing any further.
It wasn't. There are just too many bugs present, some old, some newly introduced. The SSD thing itself is a deal-breaker. Then there's Compiz, which despite my earlier hopes, is just too unstable, regularly crashing. Unbelieveably, the audio-preview 'Caja-freezing' bug is still alive & well. Audio-previews, which are disabled by default, tend to suggest that the Mint developers are aware of this problem as well but are unable to fix it. Thing is, when I compile Caja from source, on 17.3 anyway, the problem disappears! A next bug involves audio as well. Speaker audio works fine. However when you plug in a headphone/external speaker plug, audio is always muted, requiring that the user calls up Audio preferences and toggles from Speaker to Headset or vice-versa - doesn't matter which, it just needs to be toggled to start audio getting piped externally. Bloody annoying. Then there's the Bluetooth fiasco. First off, I couldn't get the fully-featured 'Blueman' to work at all. After many hours of trying to get Bluetooth working through 'Blueberry', the feature-lite but new bluetooth manager, I eventually figured out the 'sequence' of actions needed in order to get it running. Problem is, once you disconnect the bluetooth headset, you need to go through the same rigmarole in order to re-connect, which is just not acceptable for everyday use.
There it stands. Another Mint that was worth a look, but with my hardware at least, is just too buggy to be usable. Thanks to the 17.3 partition image that I had created, switching back was a relatively painless procedure. I was impressed though with Mate in general, so much so, that I actually donated €10 towards the project! The developers are doing a great job staying true to the Gnome2 desktop model, despite having had to go the GTK3 route. Although I submitted a bug-report on the SSD problem with Mint, I later tried the same SSD on a HP Pavilion laptop with Mint 18.1 running from the USB, and Mint had no problem detecting & validating the drive. So it's a kernel/hardware combination incompatibility that is causing the problem - after all, on the same Dell laptop, Mints from 16 -> 17.3 work perfectly with the Transcend drive.
Subscribe to:
Posts (Atom)