2011-06-30
Finally
The issues of the last week have been many, and mostly stem from a loose grub screw on the belt drive on the stepper motor. But there have also been issues with the hot bed sliding a little and also the stepper controllers missing (too low torque) or stalling (over temp shutdown). And then there were the fires. Feasible theories for most of them. I have spare controllers now :-)
Today has been almost as fun, but the problems have been in the firmware. Turns out it is almost as special as the hardware. Thankfully an area I am a lot happier tinkering.
I am sure there is plenty of fine tuning that can now be done to make the quality even better, but this is looking pretty good compared to anything I have printed before, so I am reasonably pleased.
Of course there is something of a TARDIS graveyard (and this is not all of them).
2011-06-29
Not impressed with AA
My wife stuck waiting for 6 hours. Madness.
What is worse is they kept changing the time estimate of when a recovery truck would arrive. I am sure if she had known it was going to be 6 hours she could have done things differently, but no.
It's the first time in, well, probably decades, that we have had to call them out to actually recover the vehicle. Now it is so late the garage will be shut. It should be a quick thing for them to fix, but will now have to be a return visit tomorrow.
I thought they gave priority to women on their own? And they were told she is diabetic with no medication on her or anything - and after this delay is not feeling that well now.
I suspect Green Flag next.
Not impressed.
What is worse is they kept changing the time estimate of when a recovery truck would arrive. I am sure if she had known it was going to be 6 hours she could have done things differently, but no.
It's the first time in, well, probably decades, that we have had to call them out to actually recover the vehicle. Now it is so late the garage will be shut. It should be a quick thing for them to fix, but will now have to be a return visit tomorrow.
I thought they gave priority to women on their own? And they were told she is diabetic with no medication on her or anything - and after this delay is not feeling that well now.
I suspect Green Flag next.
Not impressed.
2011-06-28
Quiet week?
Well, no such thing as a quiet week really, but given we are not running the training course (not enough people booked) I actually have a clear diary for a change. Well, apart from a couple of conference calls.
Anyway, new logging stuff on firebricks seems to be going well. More to do, but this means that we are doing the last of the things the FB105 did which was web view and emailed logs. Probably going to make a factory release later this week.
Bandwidth on the broadband is going pretty well since we put a huge amount extra on the 21CN links. We coped with Wimbledon with no problems, which is good. I suspect we need to look at adjusting the GB per unit for evenings on 21CN in September though as evening traffic is getting disproportionate now. We'll see how that goes.
This morning I just got some newfireworks stepper motor controllers to try in the 3D printer. Lets hope I have the logic right now. I think I may take the afternoon off and have a play...
I have been pondering making proper electronics for the 3D printer and am trying to work out if we can make that commercially viable (selling the controller boards). Whilst this is all a bit of fun, obviously I needs to make money at the end of the day somehow.
Anyway, new logging stuff on firebricks seems to be going well. More to do, but this means that we are doing the last of the things the FB105 did which was web view and emailed logs. Probably going to make a factory release later this week.
Bandwidth on the broadband is going pretty well since we put a huge amount extra on the 21CN links. We coped with Wimbledon with no problems, which is good. I suspect we need to look at adjusting the GB per unit for evenings on 21CN in September though as evening traffic is getting disproportionate now. We'll see how that goes.
This morning I just got some new
I have been pondering making proper electronics for the 3D printer and am trying to work out if we can make that commercially viable (selling the controller boards). Whilst this is all a bit of fun, obviously I needs to make money at the end of the day somehow.
2011-06-24
Pop, bang, fizzle
I have working extruder and hot end and hot bed now, so ABS should work.
Sadly still not right. I did find one of the issues is that the hot bed was gently sliding about as I only stuck it down with velcro pads, and they become slippery at 115C! That explains some of the leaning TARDIS of pisa issue. But even with that fixed and things printing a lot better the X axis was "missing" still, and causing steps and problems. Annoying as this has all been fine for a long time with the PLA.
So replaced the X controller again, and this time, well, it caught fire. In fact three controllers in a row caught fire on various axis.
I think (hope) it is down to powering up in the wrong sequence, connecting the 24V supply with the controller active perhaps. If that is the case I can just be more careful and all will be well.
So plan is to get yet more controllers (they are like £8 each) and try being very very careful on power up... So a few more days to wait, sod it.
2011-06-22
Arrrg! OFCOM
Well, once again we are having discussions with our favourite regulator.
As ever, just my personal opinions on things...
It is all very sensible that telecoms has some regulation, and most of the rules are pretty sensible. OFCOM get to make the rules within certain constraints (the "General Conditions") and they are the ones that enforce them. Fair enough.
One of the biggest issues we have with them is that when there is any sort of grey area, such as how OFCOM interpret their published rules, they will point blank refuse to answer simple questions. They say to "take independent legal advice". This is totally bullshit (IMHO) as OFCOM are the enforcement authority. They have to have a view on such matters and it is their view that matters. Now yes, there seem to be plenty of cases where their view is wrong (IMHO) and going against their view it is sensible to get legal advice, but where it is a simple question and they can say something is in fact allowed, then that is it. They are the enforcement authority and they just need to make a simple decision and tell us. They could even publish such Q&A on their web site so all operators are working to the same clarified rules. I am not even talking complex questions here.
I have, for example, asked if we are allowed to operate voicemail on our new 07 mobile numbers. This seems a daft question, you may think. It seem obvious that we should be able to as every does. But they will not say yes or no. Why is it an issue? Well, the rules are in fact quite clear that we cannot, and neither can any mobile operator. In fact, following the rules (as I am sure any independent legal advice would tell us) means nobody using mobile numbers can provide voicemail, call diverts to non mobile numebrs, or run their call centres on mobile numbers. The rules are that every call must involve a radio link to/from a mobile handset. I thought they would not treat the rules quite that literally but the first time we applied for mobile numbers they refused, stating that as there was a chance that even one call could terminate on the fixed network we would not be using the numbers in accordance with the rules! They now say we must not routinely terminate calls on a fixed network, but won't define that. Other mobile operators routinely provide voicemail, so again they are breaking these new vague rules. Is it so much to ask that a regulator actually states what is and is not allowed, please?
The latest fiasco is slightly different. The have insisted VoIP operators provide access to 999/112 (emergency services). It is questionable whether this is in fact sensible or not, but that is how it is, and we do provide it. We route the calls via a carrier we use. Of course, the rules simply say that we have to "ensure end users have access to emergency services". Nothing actually says that we have to route 999/112, and previously we made it part of our terms that users must have a fixed line or mobile as well, so as to ensure they "have access". OFCOM insist we cannot rely on instructions of terms to ensure end users have access. Even when the VoIP kit has a PSTN connection (gigaset DECT/VoIP did, and routed 999/112 via it) that was not good enough as they may have unplugged the PSTN (?!?). Now that we do route calls to 999/112 we can't in fact be compliant, as we are relying on instructions and terms to ensure people using such kit programs it to send 999/112 via us and not the (possibly unplugged) PSTN connection, but OFCOM say routing 999/112 is all they want. I.e. they do not actually follow the rules they define but make up new ones and change them as it suits them!!!
So, we route 999 and 112, good. But they now want "location information" passed to emergency operators. But hang on? The service we offer has a username and password and can be used by anyone with those details from any point on the Intrernet and change from call to call and even have multiple phones on the same number at the same time in dirrect locations.
It is a bit like us having a phone socket in Maidenhead and saying "plug your phone in here and make calls, using however many miles of extension lead as you like" and then expecting us to know where the call actually comes from. Indeed, one could argue that the service is provided at the location of the call server (data centre in Maidenhead) and everything beyond that is not part of the VoIP service we provide but a sort of virtual extension lead. The principle applies for corporate phone systems already where an office may have an ISDN line, and VoIP going out to employee's homes. The location is that of the ISDN line at the office.
So, we have no location data. It is not feasible to try and deduce location data from an IP, and even if we could (e.g. our own customers broadband lines) and even if we assumed no proxy or call server relaying the call, we have no way to get that data to emergency services in real time. The way it works is BT run a database that holds addresses for calling numbers which have to be set up in advance (not real time). Well, you can't do that for VoIP.
The best we can do is ask end users, if they happen to only use their VoIP account in one place, to provide an address that we can send to BT in advance. Sadly, BT won't talk to us for address database updates, and they won't allow the carriers we use to update the database for our numbers. And OFCOM do not seem to understand. The issue is the rules don't require BT to talk to us! We'd be happy to make reasonable efforts to get what data we have to emergency services where that is possible, obviously.
What is worse is that OFCOM are being a tad special reading their own rules, once again. We want to be compliant and technically correct anyway, and in fact we want to go ruther and provide something useful to our customers. The rules say we have to make data available. They do not dictate any specific technical means of doing that. I said (seeing as BT won't talk to us) we simply have it on a web page (with username) so emergency services can access it. That is "making available", in my view. They say that is not acceptable, but can't show me where in their rules it says that. They are just making up rules as they go along, it seems.
One of the big issues, and we saw this with directories too, is that there is a "common practice" that works for the larger telcos and usually involves BT doing some crucial role to centralise things. BT do some really useful stuff. But the rules do not fit this common practice, and the common practice does not always fit with a VoIP operator. So we end up with rules saying one thing, and OFCOM insisting on something else, and people like us stuck in the middle, complying with the rules but having discussions with OFCOM all the time.
The fact it is not technically feasible to provide accurate and reliable location data (and the rules are we have to only if it is technically feasible) seems to be ignored by them. The fact that there is no data in the telecoms network regarding the location of the caller (and the rules say that it is the data in the network relating to the location of the caller that we have to provide) is also ignored by them.
I think that, for a whole string of reasons, the rules do not require us to do anything. That means we are compliant.
I think it would be good if BT had to take data from us so that we can allow customers to tell us the location data they would like passed to emergency operators. It would be nice if OFCOM insisted that they do talk to us. Then we could do what our customers actually want. It may save someone's life one day.
I can see this one going on for a while...
Oh, and I dropped apple in it, sorry. The rules say that if I can make calls to numbers in the national calling plan then I have to be able to call 999 or 112. My iPad allows me to call any iPhone users using the mobile number of the iPhone, i.e. a number in the national calling plan, using facetime. That means my iPad, and hence apple's facetime service, has to allow calls to 999 and 112. It also has to pass location data (it has a GPS, so easy) but somehow do that in real time as it is a mobile device, and cope with the fact that my iPad calls from an email address and not a phone number (so cannot use BTs database system). I wonder if OFCOM will enforce their own rules against apple, or just quietly ignore it.
As ever, just my personal opinions on things...
It is all very sensible that telecoms has some regulation, and most of the rules are pretty sensible. OFCOM get to make the rules within certain constraints (the "General Conditions") and they are the ones that enforce them. Fair enough.
One of the biggest issues we have with them is that when there is any sort of grey area, such as how OFCOM interpret their published rules, they will point blank refuse to answer simple questions. They say to "take independent legal advice". This is totally bullshit (IMHO) as OFCOM are the enforcement authority. They have to have a view on such matters and it is their view that matters. Now yes, there seem to be plenty of cases where their view is wrong (IMHO) and going against their view it is sensible to get legal advice, but where it is a simple question and they can say something is in fact allowed, then that is it. They are the enforcement authority and they just need to make a simple decision and tell us. They could even publish such Q&A on their web site so all operators are working to the same clarified rules. I am not even talking complex questions here.
I have, for example, asked if we are allowed to operate voicemail on our new 07 mobile numbers. This seems a daft question, you may think. It seem obvious that we should be able to as every does. But they will not say yes or no. Why is it an issue? Well, the rules are in fact quite clear that we cannot, and neither can any mobile operator. In fact, following the rules (as I am sure any independent legal advice would tell us) means nobody using mobile numbers can provide voicemail, call diverts to non mobile numebrs, or run their call centres on mobile numbers. The rules are that every call must involve a radio link to/from a mobile handset. I thought they would not treat the rules quite that literally but the first time we applied for mobile numbers they refused, stating that as there was a chance that even one call could terminate on the fixed network we would not be using the numbers in accordance with the rules! They now say we must not routinely terminate calls on a fixed network, but won't define that. Other mobile operators routinely provide voicemail, so again they are breaking these new vague rules. Is it so much to ask that a regulator actually states what is and is not allowed, please?
The latest fiasco is slightly different. The have insisted VoIP operators provide access to 999/112 (emergency services). It is questionable whether this is in fact sensible or not, but that is how it is, and we do provide it. We route the calls via a carrier we use. Of course, the rules simply say that we have to "ensure end users have access to emergency services". Nothing actually says that we have to route 999/112, and previously we made it part of our terms that users must have a fixed line or mobile as well, so as to ensure they "have access". OFCOM insist we cannot rely on instructions of terms to ensure end users have access. Even when the VoIP kit has a PSTN connection (gigaset DECT/VoIP did, and routed 999/112 via it) that was not good enough as they may have unplugged the PSTN (?!?). Now that we do route calls to 999/112 we can't in fact be compliant, as we are relying on instructions and terms to ensure people using such kit programs it to send 999/112 via us and not the (possibly unplugged) PSTN connection, but OFCOM say routing 999/112 is all they want. I.e. they do not actually follow the rules they define but make up new ones and change them as it suits them!!!
So, we route 999 and 112, good. But they now want "location information" passed to emergency operators. But hang on? The service we offer has a username and password and can be used by anyone with those details from any point on the Intrernet and change from call to call and even have multiple phones on the same number at the same time in dirrect locations.
It is a bit like us having a phone socket in Maidenhead and saying "plug your phone in here and make calls, using however many miles of extension lead as you like" and then expecting us to know where the call actually comes from. Indeed, one could argue that the service is provided at the location of the call server (data centre in Maidenhead) and everything beyond that is not part of the VoIP service we provide but a sort of virtual extension lead. The principle applies for corporate phone systems already where an office may have an ISDN line, and VoIP going out to employee's homes. The location is that of the ISDN line at the office.
So, we have no location data. It is not feasible to try and deduce location data from an IP, and even if we could (e.g. our own customers broadband lines) and even if we assumed no proxy or call server relaying the call, we have no way to get that data to emergency services in real time. The way it works is BT run a database that holds addresses for calling numbers which have to be set up in advance (not real time). Well, you can't do that for VoIP.
The best we can do is ask end users, if they happen to only use their VoIP account in one place, to provide an address that we can send to BT in advance. Sadly, BT won't talk to us for address database updates, and they won't allow the carriers we use to update the database for our numbers. And OFCOM do not seem to understand. The issue is the rules don't require BT to talk to us! We'd be happy to make reasonable efforts to get what data we have to emergency services where that is possible, obviously.
What is worse is that OFCOM are being a tad special reading their own rules, once again. We want to be compliant and technically correct anyway, and in fact we want to go ruther and provide something useful to our customers. The rules say we have to make data available. They do not dictate any specific technical means of doing that. I said (seeing as BT won't talk to us) we simply have it on a web page (with username) so emergency services can access it. That is "making available", in my view. They say that is not acceptable, but can't show me where in their rules it says that. They are just making up rules as they go along, it seems.
One of the big issues, and we saw this with directories too, is that there is a "common practice" that works for the larger telcos and usually involves BT doing some crucial role to centralise things. BT do some really useful stuff. But the rules do not fit this common practice, and the common practice does not always fit with a VoIP operator. So we end up with rules saying one thing, and OFCOM insisting on something else, and people like us stuck in the middle, complying with the rules but having discussions with OFCOM all the time.
The fact it is not technically feasible to provide accurate and reliable location data (and the rules are we have to only if it is technically feasible) seems to be ignored by them. The fact that there is no data in the telecoms network regarding the location of the caller (and the rules say that it is the data in the network relating to the location of the caller that we have to provide) is also ignored by them.
I think that, for a whole string of reasons, the rules do not require us to do anything. That means we are compliant.
I think it would be good if BT had to take data from us so that we can allow customers to tell us the location data they would like passed to emergency operators. It would be nice if OFCOM insisted that they do talk to us. Then we could do what our customers actually want. It may save someone's life one day.
I can see this one going on for a while...
Oh, and I dropped apple in it, sorry. The rules say that if I can make calls to numbers in the national calling plan then I have to be able to call 999 or 112. My iPad allows me to call any iPhone users using the mobile number of the iPhone, i.e. a number in the national calling plan, using facetime. That means my iPad, and hence apple's facetime service, has to allow calls to 999 and 112. It also has to pass location data (it has a GPS, so easy) but somehow do that in real time as it is a mobile device, and cope with the fact that my iPad calls from an email address and not a phone number (so cannot use BTs database system). I wonder if OFCOM will enforce their own rules against apple, or just quietly ignore it.
2011-06-19
Extrudiate!
I have been trying to print a pink DALEK all weekend.
I now have a DALEK graveyard, and some very Dali DALEKs as well it seems. Sadly I am not going to get any more done today.
Basically this is all about getting ABS to print.
The pink is not important, just that I have pink and white ABS, that is all, though I have had a request for it in pink...
Anyway, printing ABS needs heat. Lots of heat. The hot end I am running at 245C.
I had to get the tension on the extruder right as the pink ABS is a full 3mm diameter and if squashed through the extruder then it will not feed through the 3mm nozzle in to the hot end, but if not tight enough it does not have enough force to extrude. That took a while to realise and get right, but that bit is now being very reliable.
It manages to get to 115C. The ABS sticks well and is actually hard to get unstuck later! So things print all the way, or so you would expect.
I then found that X and Y axis were slipping and so you got a spacial distortion like the poor pink DALEK shown above. I think it may have been the stepper controllers over heating as the whole thing is running a lot hotter with ABS than PLA (115C hot bed, 245C hot end, and running with fans off as well).
Unfortunately something else is wrong, and I am guessing this is the root of the problem in fact.
There was a "pop", and sparks, and proper fire...
The stepper controller is no longer looking too well!
So, time to order some more, and start with proper heat sinks from scratch this time. That will be a few days I expect.
Ho hum.
Interestingly, Kev and I wondered if there is any mileage in doing a proper design for both the electronics and the firmware to see if we can make one of these things that is actually running everything within proper margins. Hobby designs works, mostly, and need tinkering and fine tuning. Done properly the whole thing would not have most of these problems. How many boards would we sell? Good question. No idea. I suspect we'd end up needing to publish the design and software really, but tempting if ever I get some spare time!
2011-06-17
But is it art?
Making sculptures is not easy, and the finish is a tad rough, but it can be done.
This is just the FireBrick F/B logo twist extruded in openscad. It has come out quite nicely, and makes a change from printing TARDISs.
Right now I am printing all sorts, just to see if I do indeed have the machine set up reliably.
It seems I may have to clean out the hot end when I start up, but after that it just works job after job.
Of course the problems just move up a level - I can now print the box and lid I designed for a small PCB and connectors. Issue is now that the design had a tiny join between the "clip" to hold the lid on, and the rest of the lid, and so it just snapped off - I have to think about design issues now I have the printing itself sorted.
Next challenge is ABS, which means a new hot bed that gets a lot hotter than this one. Seems like some combination of aluminium and glass and 25W resistors I expect. That will take a while and I suspect I'll be cleaning up shattered glass at some point.
2011-06-16
Hot and fast...
I am working on PLA at the moment, as that is lower temperature than ABS. The ABS needs a new hot bed which I am working on and will post in due course.
Hot bed: What I have is a Prusa PCB Heatbed MkI connected to an ATX 12V power supply. It is around 1.7 ohm and so uses around 7A. Obviously, contrary to what they say on the IRC channel, just having a beefier PSU at 12V will not change that. I have a Homebase Heat Shield Mat under it as insulation (the sort of thing plumbers use to protect stuff from a blow torch). It can comfortably handle 60C which is fine for sticking down PLA but not good enough for ABS. I attached a thermistor to it to T1 on the RAMP board and make the s/w regulate the temperature for me.
Hot end: This is the standard shapercube 0.5mm diameter nozzle hot end. It is a short hot spot at the end, and a brass tube with a PEEK insulator. Works well.
You must ensure you calibrate the steps/mm and measure the filament. I have blue PLA that is nominally 3mm but actually 2.45mm! Packing density of 1 is fine on PLA. Get that right and you avoid a whole heap of issues. I also suggest using Skeingforge to make the GCODE and SCAD for the 3D models. if you get it wrong either you print stuff with loads of gaps or you get bobbles on the print and as you go up the layers it forces the hot end in to what you printed badly, causing displacement of the print and/or the hot end jamming up and not working. Get it right and a fill layer is a flat surface with no bobble and no gaps.
I have been tinkering with stepper control settings. These are controlled by links on the RAMPS board. The shapercube site says to set all for 1/8 step. For Z axis there is no point and there may be torque benefits in using less. I tried full step (no links) and that shakes the whole thing and makes a noise, so half step (just first link) is fine for Z axis. You then set the arduino code to 320 steps/mm for Z instead of 1280.
On X and Y there is no need for 1/8, why the hell not use the 1/16 which the controllers will do. Just put all three links and change the steps/mm from 40 to 80 in the arduino code.
The extruder (Wade design) is fine - but needs the springs set as loose as they go really. Basically, if all is well the filament will feed through with very low pressure. If you tighten too much you just strip the filament when something goes wrong.
Now, the trick I have been working on for two days since I got a replacement hot end is temperature (top tip, do not try and clear the nozzle on the hot end using a 0.6mm pin in a 0.5mm nozzle - you bugger them up badly).
I fitted a fan to blow across the print as well. I have an STL for a nice fan mount if anyone wants.
The solution seems to be hot and fast. I am now running reliably with 215C on the hot end, and 60mm/s feed rate. This way of working has the benefit that things print quickly as well :-)
So, that is my tips for the day for shapercube users... Have fun. Catch me on irc on #reprap if you want to know more.
High speed internet - fixed at high speed
Slight surprise from our favourite telco yesterday, and worth a mention I feel.
A customer with FTTC (Fibre to the Cabinet) service.
His line started going wrong at 9am and packed up completely by 10am. He called us up at 2pm, and by just after 4pm an engineer had visited him and replaced his VDSL modem and got him back on line. Even the diagnostics tools were sensible, for a change, and confirmed there was a fault. We did not even have to argue about SFI visits.
I suspect that is a record for an engineer fixed fault. Lets hope we see more of this type of excellent service.
Well done BT!
Why did this happen? The main reason is that for FTTC there is a BT owned and managed VDSL router on the end of the line. It means there is active equipment that BT can communicate with (or not) as part of the service. BT can tell when there is a fault or not right up to the hand over point in the customer premises. They cannot do that with ADSL hence all the SFI arguments.
Now if only BT had done this for ADSL - provided a PPPoE DSL router as part of the broadband service. That way they can tell if the line is OK or not by end to end tests. It would have to be the dumbest simplest router from a user point of view, hence saying PPPoE (with jumbo frames) so that customers have a proper choice of router/firewall. Ideally it would also have Ethernet testing built in and available to the ISP. It would be nice if the VDSL routers did this. It is a standard part of many Ethernet chips sets, and would allow testing beyond the hand-over point (Ethernet port). We'd be able to confirm no Ethernet cable, or things like a broken pair in an Ethernet cable so eliminating most of the testing. Such a system would rule out almost all of the grey areas in faults and remove the need for SFI engineers even. It could even have a pair of Ethernet ports and a loop back test option. It could constantly test the line at an ATM level and identify faults before customers see them.
We have suggested the VDSL routers for FTTC and the active NTE for FTTP should have Ethernet tests, and all tests should be available to the ISP. No idea if they will take this up.
Maybe one day we'll make a DSL modem, and build all this in. Problem is that we can't get the chip sets as cheaply as you can buy DSL routers!
But anyway, yes, on this occasion, well done BT.
And well done Terry, our latest tech support engineer to join the team.
A customer with FTTC (Fibre to the Cabinet) service.
His line started going wrong at 9am and packed up completely by 10am. He called us up at 2pm, and by just after 4pm an engineer had visited him and replaced his VDSL modem and got him back on line. Even the diagnostics tools were sensible, for a change, and confirmed there was a fault. We did not even have to argue about SFI visits.
I suspect that is a record for an engineer fixed fault. Lets hope we see more of this type of excellent service.
Well done BT!
Why did this happen? The main reason is that for FTTC there is a BT owned and managed VDSL router on the end of the line. It means there is active equipment that BT can communicate with (or not) as part of the service. BT can tell when there is a fault or not right up to the hand over point in the customer premises. They cannot do that with ADSL hence all the SFI arguments.
Now if only BT had done this for ADSL - provided a PPPoE DSL router as part of the broadband service. That way they can tell if the line is OK or not by end to end tests. It would have to be the dumbest simplest router from a user point of view, hence saying PPPoE (with jumbo frames) so that customers have a proper choice of router/firewall. Ideally it would also have Ethernet testing built in and available to the ISP. It would be nice if the VDSL routers did this. It is a standard part of many Ethernet chips sets, and would allow testing beyond the hand-over point (Ethernet port). We'd be able to confirm no Ethernet cable, or things like a broken pair in an Ethernet cable so eliminating most of the testing. Such a system would rule out almost all of the grey areas in faults and remove the need for SFI engineers even. It could even have a pair of Ethernet ports and a loop back test option. It could constantly test the line at an ATM level and identify faults before customers see them.
We have suggested the VDSL routers for FTTC and the active NTE for FTTP should have Ethernet tests, and all tests should be available to the ISP. No idea if they will take this up.
Maybe one day we'll make a DSL modem, and build all this in. Problem is that we can't get the chip sets as cheaply as you can buy DSL routers!
But anyway, yes, on this occasion, well done BT.
And well done Terry, our latest tech support engineer to join the team.
I now have a shiny new hot end
I works. Now I can get back to solving other problems. Tried printing something last night and it took 6 attempts before I realised my Z axis was slipping.
It also seems a tad silly running the Z stepper on 1/8th step making 1280/mm when it only ever moves 0.4mm at a time. I changed to 1/2 step (I tried whole step but that is noisy!). Someone (on the internet) suggested it would have more torque like that but [citation needed]. I may try changing the X & Y to 1/16 rather than 1/8 tonight as they are 40/mm which seems a tad course and they do not need anything like the same torque to work.
But right now I am printing too hot (one of the things I tried last night was printing hotter). So nearly printed a whole thing but eventually the hot end bunged up and the edges curled up (even on PLA!). I'll go back to 190C and see if that solves it.
Your hot end will stop working for various reasons. If it cannot extrude for a while (my Z axis slipping meant the nozzle was pushed up against the previous layer), or if it is too hot, you end up with the hot zone being too far up the hot end and not able to push the plastic through. So you have to clear it. That means disconnecting it, and, while hot, using a long No.4 allen key to push out the PLA through the nozzle, and then reassemble. The risk of burned fingers doing this is high, and you want to avoid it happening if you can!
One idea of using a pin to clear the nozzle was not good and one reason why my previous hot end was a bit ill, as the pin was 0.6mm and the hole was 0.5mm diameter. It seems just clearing the plastic from the hot and and starting again is all you really need to do.
Anyway, time for some real work - more 3D printing later.
2011-06-14
Too hot to handle
Good idea in principle and managed to physically assemble and get up to temperature. Yay!
Sadly there is a technical hitch...
190C melts solder.
The wires fell off.
There were sparks.
My hot end is cold again.
Bugger.
P.S. Ouch, fuck, bugger, ouch...
An iframe to save £2000 a month?
Well, our favorite telco have a publically accessable DSL availability checker at http://www.dslchecker.bt.com/adsl/
Until now we have provided a checker on our we site which uses a specific API to them to get the data and present it to customers on the web page. We use this as part of the ordering process to validate the line and offer the right services too. We also used to use it to proactively check for FTTP availablility and 21CN upgrades on lines (a low level of periodic checks over night to be nice).
Saldy, in their infinite wisdom, they started charging 60p a lookup if we exceed a pretty low level of usage! So we changed to caching replies (causing problems with out of date data) and not checking for upgrade opportunities- losing them revenue in the long run.
Seems this was not enough reduction, and so they are billing us around £2,000 a month. Madness!
So the availability checker on our web site now links to their publically available site, and even puts the results in an iframe, so looks almost exactly the same as before.
We don't see the results as it is the end users browser that actually does the lookup, not us, so if someone does then order a line we do an extra lookup, increasing the load on their checker, but at least we don't get to pay 60p every time someone uses the checker on our site now!
Best value iframe I have done yet.
Until now we have provided a checker on our we site which uses a specific API to them to get the data and present it to customers on the web page. We use this as part of the ordering process to validate the line and offer the right services too. We also used to use it to proactively check for FTTP availablility and 21CN upgrades on lines (a low level of periodic checks over night to be nice).
Saldy, in their infinite wisdom, they started charging 60p a lookup if we exceed a pretty low level of usage! So we changed to caching replies (causing problems with out of date data) and not checking for upgrade opportunities- losing them revenue in the long run.
Seems this was not enough reduction, and so they are billing us around £2,000 a month. Madness!
So the availability checker on our web site now links to their publically available site, and even puts the results in an iframe, so looks almost exactly the same as before.
We don't see the results as it is the end users browser that actually does the lookup, not us, so if someone does then order a line we do an extra lookup, increasing the load on their checker, but at least we don't get to pay 60p every time someone uses the checker on our site now!
Best value iframe I have done yet.
Still no hot end
Still can't print - this is annoying.
I do however have some aluminium. Plan is to make a hot plate around 300mm x 270mm of 8mm aluminium - and fit 25W wire wound resistors in a grid to allow me to heat it to 115C.
I have a feeling it may be too heavy for the stepper motors though.
Now to find someone that can cut 8mm aluminium, as I have 300mm x 500mm sheet here.
What fun this hobby is :-)
I do however have some aluminium. Plan is to make a hot plate around 300mm x 270mm of 8mm aluminium - and fit 25W wire wound resistors in a grid to allow me to heat it to 115C.
I have a feeling it may be too heavy for the stepper motors though.
Now to find someone that can cut 8mm aluminium, as I have 300mm x 500mm sheet here.
What fun this hobby is :-)
2011-06-13
Fed up
Well, still no new hot end, so no toy to play with.
Loads of work to do.
Yet another talk on IPv6 on Wednesday (World IPv6 forum in London).
And what is going on these days with some people. I had one customer demanding like £700 because a router was posted a day late, another demanding a refund because he did not notice or report a fault on a line for a year (actually, looks like he just had the router turned off, not a fault), another that did not bother to cease a line and paid his bills for a year and now wanting money back, and then one that had annex M installed a few days late (our fav telco messing up again) wanting loads of compensation.
We know people make mistakes, which is why we go to some lengths to define things in the terms like installation dates not guaranteed, and "prove you asked us to cease" but we'll give you a receipt when you ask, etc.
Yet still I have to deal with people being silly and just trying it on. That is just depressing. We do not normally have any of this crap, and now I have several at once. Arrrg!
Fair enough, if we messed up (and one I think we may have, shipping a router a day late) - but not like £700 compensation for a £20 router FFS. We make it clear if we don't ship on time you get the shipping costs free. We say what happens if we do screw up, in advance, and then we try hard not to screw up...
When people are nice, we are nice, and usually compensate people even when we don't have to.
And none of these are consumers. Makes me wonder if we are heading to a recession again. Businesses finding any excuse (even made up ones) to complain?
Oh well, proper rant. I feel better now.
Loads of work to do.
Yet another talk on IPv6 on Wednesday (World IPv6 forum in London).
And what is going on these days with some people. I had one customer demanding like £700 because a router was posted a day late, another demanding a refund because he did not notice or report a fault on a line for a year (actually, looks like he just had the router turned off, not a fault), another that did not bother to cease a line and paid his bills for a year and now wanting money back, and then one that had annex M installed a few days late (our fav telco messing up again) wanting loads of compensation.
We know people make mistakes, which is why we go to some lengths to define things in the terms like installation dates not guaranteed, and "prove you asked us to cease" but we'll give you a receipt when you ask, etc.
Yet still I have to deal with people being silly and just trying it on. That is just depressing. We do not normally have any of this crap, and now I have several at once. Arrrg!
Fair enough, if we messed up (and one I think we may have, shipping a router a day late) - but not like £700 compensation for a £20 router FFS. We make it clear if we don't ship on time you get the shipping costs free. We say what happens if we do screw up, in advance, and then we try hard not to screw up...
When people are nice, we are nice, and usually compensate people even when we don't have to.
And none of these are consumers. Makes me wonder if we are heading to a recession again. Businesses finding any excuse (even made up ones) to complain?
Oh well, proper rant. I feel better now.
2011-06-10
Buggered hot end
I am really getting the hang of this all now - even bright pink ABS making a TARDIS, mostly. Managed to do blue plastic earlier.
More to follow next week on the whole 3D printing saga.
2011-06-08
World IPv6 Decade is nearly over...
Oh, sorry, world IPv6 DAY. Some of us have been doing this over 9 years so you understand the confusion.
Seems to be the non-event we all wanted.
Stuff worked.
From what I can see there are various ways that some of the people testing things have ballsed up DNS slightly for the odd moment, or done IPv6 on IPv4 DNS servers (which is fine) or managed to mess up glue records a tad, but nothing that actually shows any issues with the concept - and that is what is important.
For us, I cannot even say that traffic levels are noticeably higher on IPv6 today. It is a tad spiky but often 10% of our LINX traffic (where all the IPv6 seems to go). May be more obvious this evening when residential use hits facebook on IPv6.
It was also, briefly, World blame IPv6 day this morning until we stamped on it. Every problem was "could that be because they have IPv6", even one of general (LCP monitored) packet loss. Soon sorted. They know better now :-)
Seems to be the non-event we all wanted.
Stuff worked.
From what I can see there are various ways that some of the people testing things have ballsed up DNS slightly for the odd moment, or done IPv6 on IPv4 DNS servers (which is fine) or managed to mess up glue records a tad, but nothing that actually shows any issues with the concept - and that is what is important.
For us, I cannot even say that traffic levels are noticeably higher on IPv6 today. It is a tad spiky but often 10% of our LINX traffic (where all the IPv6 seems to go). May be more obvious this evening when residential use hits facebook on IPv6.
It was also, briefly, World blame IPv6 day this morning until we stamped on it. Every problem was "could that be because they have IPv6", even one of general (LCP monitored) packet loss. Soon sorted. They know better now :-)
World IPv6 Day and nothing broken yet!
Well, today is the day. There has been some odd scaremongering though - with some reports saying that internet access will be turned off [for Ipv4] which is, of course, not true. There have even been scaremongering sales calls by people suggesting there will be some huge IPv6 DDoS. But so far no problems.
Basically, what today is about, is some of the major players all trying out IPv6 at the same time. This not only gets it out of the way in one go which is less disruptive if they do have problems, but also resolves the issue of who jumps first. People like google and facebook making their services work on IPv6 (as well as IPv4) for the day. There are a lot of people doing this from major players to small companies - I spoke to a friend of mine who was just setting up IPv6 access on his web site last night.
But today has been taken on as a focus for attention all about IPv6 and a way to get the message across that things are changing.
At A&A it is pretty much business as usual. Our traffic to google is already IPv6 and IPv4 and has been for a long time, so for one of the major players there is no change at all for us. Having offered IPv6 for 9 years now this really is run of the mill stuff. We already have traffic on LINX reaching 10% IPv6 anyway and today may just be slightly higher if that. Our customers already either have working IPv6 or have IPv4 only and not a broken installation else they would not get to our web site or mail servers, etc, so we do not expect any broken installations failing just because facebook is IPv6 now.
We are shipping an IPv6 configured broadband router and customers have them in their hands today - not sure if any will be on-line today or not.
And if you have a web site with us then you have been IPv6 for many years already - welcome to World IPv6 Day participation.
But, the day is not over yet, and people all over the world are watching and monitoring carefully to make sure things cope as they should.
Basically, what today is about, is some of the major players all trying out IPv6 at the same time. This not only gets it out of the way in one go which is less disruptive if they do have problems, but also resolves the issue of who jumps first. People like google and facebook making their services work on IPv6 (as well as IPv4) for the day. There are a lot of people doing this from major players to small companies - I spoke to a friend of mine who was just setting up IPv6 access on his web site last night.
But today has been taken on as a focus for attention all about IPv6 and a way to get the message across that things are changing.
At A&A it is pretty much business as usual. Our traffic to google is already IPv6 and IPv4 and has been for a long time, so for one of the major players there is no change at all for us. Having offered IPv6 for 9 years now this really is run of the mill stuff. We already have traffic on LINX reaching 10% IPv6 anyway and today may just be slightly higher if that. Our customers already either have working IPv6 or have IPv4 only and not a broken installation else they would not get to our web site or mail servers, etc, so we do not expect any broken installations failing just because facebook is IPv6 now.
We are shipping an IPv6 configured broadband router and customers have them in their hands today - not sure if any will be on-line today or not.
And if you have a web site with us then you have been IPv6 for many years already - welcome to World IPv6 Day participation.
But, the day is not over yet, and people all over the world are watching and monitoring carefully to make sure things cope as they should.
2011-06-06
My nozzle is too big
sOK, that is not something that I say often.
On irc the other day I said something like "If I put glue on the bed and then extrude will that bugger up my nozzle" which is one of those phrases that only works in a very specific context.
Anyway - to explain - the hot end is the bit the heats up the plastic. The extruder forces the plastic in to the hot end. My hot end is a brass tube with a fine 0.5mm diameter nozzle on the end with a heater and a thermistor.
When I started I had some issues with it and one suggestion was to clean it out - which I did. Part of this was using a 0.5mm needle to clear the nozzle. I don't have needles of that type lying around, and the suggestion of checking out the nearest crack den did not help much. So I found a very fine pin and used that. It was a tight fit but with true brute force and ignorance I managed it.
What I now know, thanks to some nice new digital calipers, is the pin is 0.60mm diameter. Oops. Since then the nozzle itself has broken off leaving a hole which is not ideal. I have a new hot end on its way from Germany this week.
In the mean time I have found what happens when your nozzle is too big, and it is a tad bizarre.
Basically you extrude too much plastic at a time or too thick or something, but the upshot is that after a few layers printing the hot end is fighting with the previously printed layer which is too high. It is pushing down on it, and well. This takes a few layers as my hot bed is a PCB and flexible, and only when it has flexed as far as it can does it really start pushing down.
One effect of this is obvious - the nozzle can push the whole piece sideways and ruin it - thankfully the 60C hot bed with PLA sticks well so that does not happen.
Secondly the plastic cannot get out as the nozzle is jammed up against the previous layer. This means the extruder cannot feed. Now this works by a knurled bolt pushing down on the filament with a sprung bearing the other side. What happens when the hot end gets stuck is it strips the filament cutting a nice groove in it meaning it does not move any more. When the hot end gets to a bit with space under it, hardly any plastic comes out. You get a few layers of fine string and then nothing - job ruined.
The solution is simple - you configure the software to know the nozzle size - it then makes necessary allowances in the layer spacing and so on, and you end up printing a courser image but it does not foul up half way through.
What is odd, is that should all be allowed for in the amount of plastic extruded, which means setting the nozzle bigger should send more plastic. So maybe it is another setting I am missing. Certainly it is a lot better running on a thicker layer setting.
When I get the new hot end I can go back to 0.5mm setting and see if that works.
P.S. It is not all about nozzle size. I had filament diameter wrong (2.8 but in fact 2.7), and packing density wrong (0.85 for ABS when PLA is 1) and E steps per mm wrong. Get these right else you get bad prints! Even a few percent wrong anbd you are screwed...
P.P.S I think my hot end is buggered so waiting for the new one to arrive.
On irc the other day I said something like "If I put glue on the bed and then extrude will that bugger up my nozzle" which is one of those phrases that only works in a very specific context.
Anyway - to explain - the hot end is the bit the heats up the plastic. The extruder forces the plastic in to the hot end. My hot end is a brass tube with a fine 0.5mm diameter nozzle on the end with a heater and a thermistor.
When I started I had some issues with it and one suggestion was to clean it out - which I did. Part of this was using a 0.5mm needle to clear the nozzle. I don't have needles of that type lying around, and the suggestion of checking out the nearest crack den did not help much. So I found a very fine pin and used that. It was a tight fit but with true brute force and ignorance I managed it.
What I now know, thanks to some nice new digital calipers, is the pin is 0.60mm diameter. Oops. Since then the nozzle itself has broken off leaving a hole which is not ideal. I have a new hot end on its way from Germany this week.
In the mean time I have found what happens when your nozzle is too big, and it is a tad bizarre.
Basically you extrude too much plastic at a time or too thick or something, but the upshot is that after a few layers printing the hot end is fighting with the previously printed layer which is too high. It is pushing down on it, and well. This takes a few layers as my hot bed is a PCB and flexible, and only when it has flexed as far as it can does it really start pushing down.
One effect of this is obvious - the nozzle can push the whole piece sideways and ruin it - thankfully the 60C hot bed with PLA sticks well so that does not happen.
Secondly the plastic cannot get out as the nozzle is jammed up against the previous layer. This means the extruder cannot feed. Now this works by a knurled bolt pushing down on the filament with a sprung bearing the other side. What happens when the hot end gets stuck is it strips the filament cutting a nice groove in it meaning it does not move any more. When the hot end gets to a bit with space under it, hardly any plastic comes out. You get a few layers of fine string and then nothing - job ruined.
The solution is simple - you configure the software to know the nozzle size - it then makes necessary allowances in the layer spacing and so on, and you end up printing a courser image but it does not foul up half way through.
What is odd, is that should all be allowed for in the amount of plastic extruded, which means setting the nozzle bigger should send more plastic. So maybe it is another setting I am missing. Certainly it is a lot better running on a thicker layer setting.
When I get the new hot end I can go back to 0.5mm setting and see if that works.
P.S. It is not all about nozzle size. I had filament diameter wrong (2.8 but in fact 2.7), and packing density wrong (0.85 for ABS when PLA is 1) and E steps per mm wrong. Get these right else you get bad prints! Even a few percent wrong anbd you are screwed...
P.P.S I think my hot end is buggered so waiting for the new one to arrive.
TARDIS
This one took five and a half hours, and went slightly wrong after around five hours so a slightly off effect near the top.
But printing for that long means I have managed to solve most of the problems I think.
I'll try again when I get blue plastic.
Obviously this is a scale model - well the outside is to scale, if I tried to do the in-fill to scale I would need a lot more plastic :-)
2011-06-04
PLA FTW
ABS I can extrude around 220C and is quite good for stopping and starting. You pretty much have to have a heat bed at 115C and a fan, and lots of crossed fingers and virgins to sacrifice... But yes, it is either extruding or not. This is good. The downside is that it contracts as it cools and is a bugger to stick down the first layer. I have, for now, given up on ABS.
PLA is different! It extrudes around 195C nicely. It is very stable in that it does not contract much, if at all, which makes for things that stay put and and print neatly. The down side is it is sticky and stringy and oozy. When you try and stop extruding it sort of carries on a bit and makes thing bits of string. It is odd stuff. So neat, but with extra bits of string. It sticks down well on a hot bed at as low as 50C and stays put well for a whole print. It is more brittle and less flexible than ABS. It is pretty solid though, and apparently bio-degradable?!
The good news is I made my first "thing"!
Now I am making a bracket to hold the new fan that is in the machine.
But yes, PLA works, it lets me print stuff, and it comes out as I designed. Yay!
Well, actually, turns out everything is mirror image due to some s/w bug, but that is easy to work around for now...
Just in case I suss the printing of a TARDIS, and I have a new 3D model of one to try, I have ordered some blue PLA :-)
"Packet loss" as a fault
Our favorite telco once again show their total ineptitude by again failing to understand the basics of IP networking.
We have reported packet loss, something we can easily test as we do LCP echos on every line every seconds.
Packet loss is normal when a link is full and the way TCP adapts to a full link, but random packet loss fools TCP so one of the main effects of packet loss is low speed on file transfers using TCP. It is not the only impact, as random packet loss can cause drops in VoIP, delayed DNS lookups, missed syslogs, slow throughput, and so on.
Our favourite telco are insisting that we identify the end users perceived problem. They cannot understand that the we have a technical end user that perceives "random packet loss" as "the problem". It is not "when I got to facebook it is slow", it is "I can see there is random packet loss". They also fail to understand that *WE* are their customer and what we as the customer perceived as the fault is measurable random packet loss on the line...
Oh what fun.
As an ISP I would far rather a customer came to me saying "I see 2% packet loss" or "I see 2% packet loss to XXX". I can test and diagnose a report like that. What I don't want is someone saying "The internets are slow". It seems out favourite telco want the latter and when you provide actual detailed reports of exactly what is wrong they cannot cope.
Arrrrg!
We have reported packet loss, something we can easily test as we do LCP echos on every line every seconds.
Packet loss is normal when a link is full and the way TCP adapts to a full link, but random packet loss fools TCP so one of the main effects of packet loss is low speed on file transfers using TCP. It is not the only impact, as random packet loss can cause drops in VoIP, delayed DNS lookups, missed syslogs, slow throughput, and so on.
Our favourite telco are insisting that we identify the end users perceived problem. They cannot understand that the we have a technical end user that perceives "random packet loss" as "the problem". It is not "when I got to facebook it is slow", it is "I can see there is random packet loss". They also fail to understand that *WE* are their customer and what we as the customer perceived as the fault is measurable random packet loss on the line...
Oh what fun.
As an ISP I would far rather a customer came to me saying "I see 2% packet loss" or "I see 2% packet loss to XXX". I can test and diagnose a report like that. What I don't want is someone saying "The internets are slow". It seems out favourite telco want the latter and when you provide actual detailed reports of exactly what is wrong they cannot cope.
Arrrrg!
2011-06-02
4D printer
Now it could just be that I still have not quite managed to get the knack of getting the first layer to stick and stay stuck.
Or it could be that it is impossible to print a TARDIS with only a 3D printer. What you need is an extra stepper motor for the 4th dimension. Well, the stepper motor is easy but the plastic part to mount it correctly can only be printed on a 4D printer and I don't know anyone that has one yet :-)
I'll manage it eventually. Getting closer though.
FB2700 to the rescue
Well, we had a slightly impromtu test of the new FB2700 3G dongle support last night when BT had some issues.
Sadly this happens, and is why we offer things like dual line (2xBT), and dual line (BT+Be) type services. DUal line tends to be a bit expensive for home users though. However the FB2700 has a USB port and over the last few days the new s/w releases have included the 3G dongle support (at last).
One of our customers has been testing this and was suitable impressed :-
"Very impressed to see the FB2700 fall back to a routed IPv4 block over 3G, when BT tripped over a cable this evening! I only noticed 10 minutes after my main connection died when I happened to glance at the CQM graphs on the FB2700."
By default the support allows simple NATted (yes, I said it) routing over a single IP on a generic 3G dongle which is ideal as a backup. However when used with A&A data SIMs it allows the backup to be for the DSL line, handling routing of the block of (legacy) IPs to the dongle when the line is down... His irc session did not even drop!
Seems this "just worked".
Thanks to "tgb" in A&A irc for confirming.
Sadly this happens, and is why we offer things like dual line (2xBT), and dual line (BT+Be) type services. DUal line tends to be a bit expensive for home users though. However the FB2700 has a USB port and over the last few days the new s/w releases have included the 3G dongle support (at last).
One of our customers has been testing this and was suitable impressed :-
"Very impressed to see the FB2700 fall back to a routed IPv4 block over 3G, when BT tripped over a cable this evening! I only noticed 10 minutes after my main connection died when I happened to glance at the CQM graphs on the FB2700."
By default the support allows simple NATted (yes, I said it) routing over a single IP on a generic 3G dongle which is ideal as a backup. However when used with A&A data SIMs it allows the backup to be for the DSL line, handling routing of the block of (legacy) IPs to the dongle when the line is down... His irc session did not even drop!
Seems this "just worked".
Thanks to "tgb" in A&A irc for confirming.
Subscribe to:
Posts (Atom)
Fencing
Bit of fun... We usually put up some Christmas lights on the house - some fairy lights on the metal fencing at the front, but a pain as mean...
-
Broadband services are a wonderful innovation of our time, using multiple frequency bands (hence the name) to carry signals over wires (us... -
For many years I used a small stand-alone air-conditioning unit in my study (the box room in the house) and I even had a hole in the wall fo... -
It seems there is something of a standard test string for anti virus ( wikipedia has more on this). The idea is that systems that look fo...