-
#25 Reply
Posted by
Monkeh
on 13 Dec, 2017 16:32
-
Of course faults can occur. No need for truth to spill out about that, it's no secret.
Also, as far as 'line-line' sockets go.. you can't inadvertantly do that. If you can: That's a faulty installation or a faulty idiot with a screwdriver.
-
#26 Reply
Posted by
grumpydoc
on 13 Dec, 2017 16:39
-
Thanks,
I think the truth is spilling out now all over the place!....the UK mains (and other country’s) can go well high…certainly up to 292VAC….and all because of power system faults!
Yes,
faults can cause voltages to increase - but that wasn't the thrust of your initial post was it? You implied in that that it was almost
normal for UK mains voltage to ride at 292V for sustained periods because, well, we "don't care about regulation".
How much fault voltage you want to design your products to withstand is, of course, your own choice but the UK tolerance is (IIRC) 230 V +10%/-6%. Anything persistently outside that is a fault and can be reported as such (and will get fixed with luck).
-
#27 Reply
Posted by
dmills
on 13 Dec, 2017 16:41
-
Neutral faults on three phase systems are notable and memorable when they happen to you, but are a long way from being commonplace, they are for the most part one of those things that you might know somebody it happened to (or somebody who knows somebody), but it is not something you really expect to ever experience.
PME/MEN style distribution wiring makes the things even less commonplace, albeit somewhat more dangerous when they happen.
Nobody designs single phase product to survive lost neutral or 400V line to line, and for the most part nobodies stuff blows up when you plug it in (And when it does, it is seldom because of a major mains brewup), generally a lost neutral causes you MUCH bigger issues then a few blown light fittings.
Failures due to gross sustained over voltage are a zebra, you have hoofprints and in the first instance should be thinking of something far more common like say a horse or a product design cockup.
You will also find that if you track down the standards for electrical supply to street furniture (And the ones for grounding of same) that they are quite detailed.
Old theatre air handlers are a law unto themselves, but the motor would almost certainly have been three phase, probably with no neutral, so not particularly relevant, but with old nightclub wiring who knows?
Line to line is usually a RED CEE connector, not something that it is easy to mistake for anything else.
Regards, Dan.
-
#28 Reply
Posted by
wraper
on 13 Dec, 2017 17:06
-
No, you have failures because of crap products. A bunch of other treads you started brilliantly show how your products are designed
. Overall it seems you try to find justifications to all issues rather than fix the mess your company is which is the root cause of most problems.
-
#29 Reply
Posted by
wraper
on 13 Dec, 2017 17:44
-
In addition to that, IMO your products have so "big" design margin that cannot survive additional 10 volts of mains voltage UK has compared to the rest of Europe. Although it's supposed to be 230V, it's still good old 240V in most cases.
-
#30 Reply
Posted by
coppice
on 13 Dec, 2017 17:55
-
No, you have failures because of crap products. A bunch of other treads you started brilliantly show how your products are designed . Overall it seems you try to find justifications to all issues rather than fix the mess your company is which is the root cause of most problems.
From the other threads he started about his product failures you have to wonder if they did any EMI/EMC/ESD testing at all.
-
#31 Reply
Posted by
dmills
on 13 Dec, 2017 18:09
-
Treez does seem to have more EMC/EMI/ESD and general reliability pain then any other three engineers I know of.
I think part of it may be excessive focus on BOM cost to the detriment of doing it right for reliability, make it as simple as possible **but no simpler**, there is an art to that.
Now I don't play in highly cost constrained markets, but I cannot but think that improving design and testing to cut customer returns is something that pays big dividends until the return rate is very, very low, and that this is probably more true as margins reduce.
Regards, Dan.
-
#32 Reply
Posted by
wraper
on 13 Dec, 2017 18:47
-
I think part of it may be excessive focus on BOM cost to the detriment of doing it right for reliability, make it as simple as possible **but no simpler**, there is an art to that.
BOM cost may be very cheap if design is well thought. One of my designs has around 4-5 times lower component cost compared to competitor. Yet it does not have any of noise immunity, ESD susceptibility, overall reliability issues that competitor device is plagued with. And even has better performance. Making things expensive does not give any guarantee they will be any good.
-
#33 Reply
Posted by
jc101
on 13 Dec, 2017 18:59
-
I had a couple of years worth of logs from my UPS until recently (disk failure), with the min, max, and average of the voltage at 1min intervals. Only once did I get up to 254v on the incoming supply and the power network people turned out to take a look. They called a short while later to ask if things were normal, which they were, and muttered something about the HV boys trying out some new gear that was running a bit high. As I'm close to the local substation my voltage was just over the limits. All sorted in a couple of hours.
I have seen faults internally to an install kill all sorts of kit, I remember seeing a neon switch from a power strip fly horizontally in front of me before the smoke from the kit led me to leg it out the room. That was a 4 pole switch on the incoming supply, the neutral contact was iffy. It was replaced with a 3 pole and a solid bolted neutral connection and the problems all went away.
-
#34 Reply
Posted by
Gyro
on 13 Dec, 2017 19:24
-
Funny, I've seen very similar things happen in an office block where the neutral connection had worked loose and was arcing, I watched PC after PC fail with a mushroom cloud of smoke pouring out the back of them, finally managed to convince the office manager to get everyone to unplug their machines.
Thanks, this says it all….terribly high overvoltages can after all happen on the UK mains… 
.all it takes is a badly wired neutral. 
I wonder who is going to pay much money for eg a load of streetlights if this sort of shenanigans can kill them in no time. 
There’s also the fact that some customers might inadvertently plug eg a load of streetlights into a line-line outlet, which , as you know, is 415VAC.
Maybe this is the reason that Harvard lighting never roled out their leafnut streetlighting system as far as was predicted? 
http://www.harvardtechnology.com/solutions/leafnut/
…some 10 years ago, Leafnut hit the UK market and was supposed to dominate the entire UK streetlighting market, but then it all seemed to go very quiet for Leafnut....the role out was never anywhere near as big as predicted.....maybe the mentioned power system faults keep companies from wanting too much exposure to LED streetlights?
No, sorry - you're really clutching at straws now!
A 3-phase Neutral fault can happen in any country. How many of your outside lights are installed in tower blocks?
A Neutral fault in a normal domestic street is an extremely rare event, it would affect many other pieces of consumer equipment and would not show up as a significant product failure rate.
You have a product problem that you need to investigate. Making wild assumptions about massively out of legal spec mains voltages is distracting you from focusing on this!
P.S. There are a very large number (majority now I think) of LED streetlights in the UK, there is no reason why they would have to adopt Leafnut though.
-
-
Making things expensive does not give any guarantee they will be any good.
No guarantee at all. However, I was once the unfortunate owner of an Ariston washing machine and after four replacement controller boards in three years I bought a Bosch and haven't had a problem. One of the guys who fitted one of the controllers said they were crap and you need to fit a surge protector. At the time the house was supplied by overhead lines and I think brown outs were more of a problem and I even needed a call out one winter to fix a dodgy connection, nothing else failed only the crappy washing machine.
-
#36 Reply
Posted by
dmills
on 13 Dec, 2017 20:26
-
Oh indeed, no argument there, but there is a point at which BOM squeezing starts to do bad things, the point can be lowered with clever architecture, but it is always there.
Regards, Dan.
-
#37 Reply
Posted by
coppice
on 13 Dec, 2017 20:35
-
Oh indeed, no argument there, but there is a point at which BOM squeezing starts to do bad things, the point can be lowered with clever architecture, but it is always there.
Regards, Dan.
Spending on the BOM to achieve EMI/EMC/ESD goals is largely a substitute for NRE. Spending is no guarantee of success in EMI/EMC/ESD tests, but a little competent engineering effort and some additions to the BOM will usually get you through. From there you spend on stripping the BOM to an extent that is appropriate to the production volume. If the volume justifies some serious effort, you can usually get the BOM quite close to its value before the EMI/EMC/ESD began.
-
#38 Reply
Posted by
rstofer
on 13 Dec, 2017 20:37
-
There can always be a LOCAL problem like a floating neutral. But it certainly won't be nationwide. Further, it would be easy to find with a DMM or perhaps a logging (min/max) DMM. And other stuff would be burning out as well. If that were the case, I would expect someone to have called the electric company. That's kind of what they do for a living.
-
-
I don't know, stick the bloddy thing on a bench connect a scope to it and wind the variac up, testing is part of development. You must know what failed in the units in the field that went pop and that would be a good place to start. Did something overheat, is it overvolts on a switching device, something went short because of a transient spike or maybe a control loop went ape shit. You've got to test for ALL of these things and a hell of a lot more before you put it on the market, otherwise it will only come back and bite you on the arse. I've just got a teeny tiny little flyback converter working as it should thanks to some good silicon and a lot of hard work and good advice from a lot of other engineers. It will work quite happily at 300V maybe even 320V but not yet tested to 320V. My next job is to kick the electrical shit out of it and then it can go into a product and that's what engineers do.
-
#40 Reply
Posted by
tggzzz
on 13 Dec, 2017 22:28
-
At a guess, it feels like treez' development team has drunk too deeply of the XP Kool Aid that is prevalent in the software world: TDD, or test driven development. While there is some validity to that in limited and controlled circumstances, unthinking application of the XP dogma often causes problems akin to building castles on sand.
Whereas TDD can work within a solid well thought out foundation, simply rushing into the simple "happy daze" functional tests often leads to running foul of subtle "corner case" requirements at a very late stage in the development. The software solution often involves large scale refactoring.
Now large-scale refactoring at a late stage is sometimes practical with modern strongly statically typed languages, but that doesn't work with hardware at a smaller scale than the minimum replaceable unit (usually a PCB).
Given the questions and statements from treez in this and other threads, it tells like they have a long slow path ahead of them.
-
#41 Reply
Posted by
IanB
on 13 Dec, 2017 23:50
-
At a guess, it feels like treez' development team has drunk too deeply of the XP Kool Aid that is prevalent in the software world: TDD, or test driven development. While there is some validity to that in limited and controlled circumstances, unthinking application of the XP dogma often causes problems akin to building castles on sand.
Whereas TDD can work within a solid well thought out foundation, simply rushing into the simple "happy daze" functional tests often leads to running foul of subtle "corner case" requirements at a very late stage in the development. The software solution often involves large scale refactoring.
I think I would like react to this view with a slightly different perspective.
One shouldn't follow a process because it is dogma, one should follow it with a view to how it can help you get better results.
I think part of the TDD process is to start by asking the question, "What the heck is this thing supposed to do, and how would I test that?"
Then with the answers to that question, you would figure out what test methods, processes or test jigs you will need to verify that it does the right things. And you use that knowledge to inform the design, and test early as you go along, rather than thinking about the testing at the end. This is the "test driven" part of test driven design.
If you leave testing to the end, you may find a test case that you didn't account for in the design, which would force you to go back and do rework.
-
#42 Reply
Posted by
tggzzz
on 14 Dec, 2017 00:04
-
At a guess, it feels like treez' development team has drunk too deeply of the XP Kool Aid that is prevalent in the software world: TDD, or test driven development. While there is some validity to that in limited and controlled circumstances, unthinking application of the XP dogma often causes problems akin to building castles on sand.
Whereas TDD can work within a solid well thought out foundation, simply rushing into the simple "happy daze" functional tests often leads to running foul of subtle "corner case" requirements at a very late stage in the development. The software solution often involves large scale refactoring.
I think I would like react to this view with a slightly different perspective.
One shouldn't follow a process because it is dogma, one should follow it with a view to how it can help you get better results.
I think part of the TDD process is to start by asking the question, "What the heck is this thing supposed to do, and how would I test that?"
The with the answers to that question, you would figure out what test methods, processes or test jigs you will need to verify that it does the right things. And you use that knowledge to inform the design, and test early as you go along, rather than thinking about the testing at the end. This is the "test driven" part of test driven design.
If you leave testing to the end, you may find a test case that you didn't account for in the design, which would force you to go back and do rework.
All very rational, and it is how I develop
on top of a sound foundation. The sound foundation implies and ensures that the awkward scenarios have already been dealt with in the fundamental structure.
Never ever forget the old
engineering (cf amateur hacking) maxim: "you can't
test quality into a product". (It has to be
designed in!)
Unfortunately XP and TDD zealots often get too enthusiastic and dogmatic, to the exclusion of common sense and using the appropriate combination of techniques.
When confronted with someone that believes too much in TDD, I ask them how they will
test - ACID transaction properties, both selection of the correct subset, and that it is correct for their application.
- cross-clock domain synchronisation
- distributed FSM behaviour (think telecom systems) in the presence of missing and duplicated events
- resilience to and recovery from ESD events
and so on. Too often I simply get a blank look indicating such considerations are beyond their comprehension.
-
#43 Reply
Posted by
David Hess
on 14 Dec, 2017 00:11
-
I thought solar inverters had voltage regulators? Or do the cheapest ones have completely unregulated output?
It is not that simple. The grid tied inverter can regulate the voltage where it is located but by removing the load which would normally be present, the voltage between the grid tied inverter and the power plant rises. The result is poorer regulation in general no matter what the grid tied inverter does. Solutions include lowering the resistance of the distribution system which the power company is obviously not going to look favorably on since it costs them money to invest in infrastructure just so customers can purchase less power from them.
The last house I lived in, had constantly high voltage of around 249v, also my UPS was reporting an "overloaded neutral", called the supplier out, they left a monitor for a week, but they said it was all within tolerance, if on the high side.
This happened to me earlier this year. My UPS started reporting power faults a couple times an hour leading me to do a general inspection and discover that a tree had knocked the line between the house and the pole down breaking the neutral. My temporary solution was to connect my whole house 240 volt AC phase splitter shown below until the power company could come out and fix the line.
-
#44 Reply
Posted by
coppice
on 14 Dec, 2017 00:19
-
I thought solar inverters had voltage regulators? Or do the cheapest ones have completely unregulated output?
It is not that simple. The grid tied inverter can regulate the voltage where it is located but by removing the load which would normally be present, the voltage between the grid tied inverter and the power plant rises. The result is poorer regulation in general no matter what the grid tied inverter does. Solutions include lowering the resistance of the distribution system which the power company is obviously not going to look favorably on since it costs them money to invest in infrastructure just so customers can purchase less power from them.
It is fairly simple. Any grid tied inverter has to track the grid voltage, to be able to feed into the grid. Many older inverters did this in a fairly dumb way, and could track upwards a long way. The latest ones just stop feeding the grid if the measured grid voltage rises above a safe threshold. If the voltage rises too far you might find a lot of solar inverters suddenly drop off the grid, and the reduction in capacity might then lead to a blackout. You won't see a dangerously high voltage, though.
-
#45 Reply
Posted by
BrianHG
on 14 Dec, 2017 00:24
-
This happened to me earlier this year. My UPS started reporting power faults a couple times an hour leading me to do a general inspection and discover that a tree had knocked the line between the house and the pole down breaking the neutral. My temporary solution was to connect my whole house 240 volt AC phase splitter shown below until the power company could come out and fix the line.
After a severe summer storm about a decade ago, on of the main phases wired into my house slowly became more and more iffy as well... If I weren't an electrical engineer, I would have not realized what was truly going on and would have not known to call the power company risking a fee if the problem was determined to be inside my house instead of outside with the connection to the pole transformer.
Small world....
-
#46 Reply
Posted by
Jr460
on 14 Dec, 2017 00:38
-
Had dropped phase once in the underground service to the house.
From the looks of things the wires had been nicked 20+ years ago when the phone company ran too close to them. It was just a matter of time for the ground settle just right and heavy rain the day before and to problem was in a low lying area.
What we saw was some light not working, then some were fine and others were dim.
Within an hour or so they came out, fund the area of the fault, dug down a few feet and spliced the blown apart wire.
-
#47 Reply
Posted by
David Hess
on 14 Dec, 2017 01:11
-
It is not that simple. The grid tied inverter can regulate the voltage where it is located but by removing the load which would normally be present, the voltage between the grid tied inverter and the power plant rises. The result is poorer regulation in general no matter what the grid tied inverter does. Solutions include lowering the resistance of the distribution system which the power company is obviously not going to look favorably on since it costs them money to invest in infrastructure just so customers can purchase less power from them.
It is fairly simple. Any grid tied inverter has to track the grid voltage, to be able to feed into the grid. Many older inverters did this in a fairly dumb way, and could track upwards a long way. The latest ones just stop feeding the grid if the measured grid voltage rises above a safe threshold. If the voltage rises too far you might find a lot of solar inverters suddenly drop off the grid, and the reduction in capacity might then lead to a blackout. You won't see a dangerously high voltage, though.
*You* will not see a dangerously high voltage but someone between you and the power station may. Your grid tied inverter supplies power which you otherwise would draw from the power line and in addition, may feed power back consistent with the local conditions it measures which is usually only voltage. I suppose in theory it could double the variability without overloading the circuit but I assume other limits like local voltage and line resistance would limit the power driven back into the power distribution system.
The power distribution network may be modeled as a series of resistors between power sources and power sinks. The voltage drop across the non-zero resistances degrades the load regulation. Previously there were only loads but if the various loads become opportunistic sources, it increases the variability of the loads so the regulation must become worse. Whether it becomes dangerously worse depends on how much impedance margin (conductance margin?) exists in the power distribution network.
-
#48 Reply
Posted by
ocset
on 14 Dec, 2017 06:19
-
Thanks, from the above it definetely seems that disconnected neutral faults, and mains voltage rise due to all the grid tied inverters could possibly be a real issue causing damaging overvoltage...eg , up to 292vac.
-
#49 Reply
Posted by
wraper
on 14 Dec, 2017 08:45
-
Thanks, from the above it definetely seems that disconnected neutral faults, and mains voltage rise due to all the grid tied inverters could possibly be a real issue causing damaging overvoltage...eg , up to 292vac.
There is no widespread damaging overvoltage. And how solar power in UK is supposed to be so different from other countries?