Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Long life LED street lights...Mains transients?....Poor LED bind wires?

Status
Not open for further replies.
T

treez

Guest
This concerns this thread from some time back....

https://www.edaboard.com/threads/352771/

But what we dont know is how they do ORT and how do they know how much margin exists in shearing the Anode connection with the ultrasonically cold-welded micron size gold wire. THIS IS THE MOST COMMON failure mode in LEDs. It can occur for many reasons. Some designs use dual wire bond. But all must survive the extreme thermal stress from different coefficients of material expansion not get "wounded" from shock, vibration, ESD, thermal shock, transient overpower etc etc.
This quote is from post number #49 of the above thread.

…Thanks…and WOW!!!.....that is ground-breaking, and blows out of the water all the research in to finding ways of being abe to handle mains transients in order to get long-life.
By the way, do you have a paper on that?
We are now wanting to do long life outdoor street lighting.
If there is some issue with bond wires to LED substrate having to be ridiculously thin, and often, as you appear to suggest, it gets cut too thin due to machine tolerance, and that causes LED failure in the field, then there’s not much point in us investing more time in finding ways of surviving every possible mains transient that present itself to the street lights………since the bond wire issue that you speak of is going to give us failures even if we do the best ever transient protection?

Are you sure its the main cause of failure in LED lamps?
 

My street lights and traffic lights were converted to LEDs a couple of years ago. They use many fairly large and high power LEDs and a very large heatsink. Many are seen to be burnt out. Maybe they were made "over there" and sold on ebay?
 
  • Like
Reactions: treez

    T

    Points: 2
    Helpful Answer Positive Rating
This is related to the above so i hope i can post it here....

"All Long life LED streetlights must have replaceable surge protection modules which disconnect the street light when the surge protection modules fail?"

Pages 12 to 13 of the following….
https://www.littelfuse.com/~/media/..._module_design_and_installation_guide.pdf.pdf
…describe mains power system faults which result in overvoltages which last for such lengthy periods that no realistic transient protection within a power supply can withstand it. Therefore, do you agree that any individual LED street light must have an upstream Transient (surge) protection module within itself. This must disconnect the streetlight in the event of such a transient…(and indeed the surge protection module will be destroyed by this kind of transient).
This transient (surge) protection module would presumably be located in the base of the streetlight so that maintenance staff can easily access it to replace it in the event of such failure.
Even if such severe mains transients never occur, there are other mains transients which occur which eventually wear out the MOVs in the surge protection modules, such that the surge protection modules are likely to need replacing every 5 years anyway. (so says the above document on page 3).
As such , replaceable surge protection modules are absolutely essential in long life LED streetlighting, do you agree?
 

This is related to the above, so please may i post it here....

Do you know how often power system faults occur on the mains system?...especially that mains network which supplies streetlights. As seen below, these power system faults would kill any LED streetlight as its not possible for any kind of realistic protection circuit to protect against them…….

Pages 12 to 13 of the following document give details of Electrical Power System faults which can cause large mains overvoltages for periods of milliseconds to minutes. These would kill LED streetlights without doubt.

"Lighting Surge Protection Modules Design and Installation Guide" by littelfuse
https://www.littelfuse.com/~/media/..._module_design_and_installation_guide.pdf.pdf

As in the document, these include things like……

a)…Line to ground faults
b)…Loss of Neutral faults.
c)…Short circuit faults in the high to lower voltage transformers.
 

So, considering post #49 of this...
https://www.edaboard.com/threads/352771/

Is LED Bond wire failure the "dirty secret" of the LED lighting industry?

I mean, surely no LED light product can be advertised as being "long life" because the bond wires in any of the LEDs, may fail at any time?

As Sunnkyskyguy said, "This is the most common cause of failure in LEDs".

Also, Power FETs have bond wires in them too, so does the bond wire failre mode of LEDs also apply to power FETs/diodes etc etc?
 

Is LED Bond wire failure the "dirty secret" of the LED lighting industry?

I am not an expert in lighting and do not even have a fraction of the experience that many here have in that industry, but based on observations from my user's point of view, my guess is that transients of short duration and high energy level may be a relevant factor on degradation of LED devices, and it seems that many manufacturers probably do not take the necessary precautions to suppress such artifacts. What motivates me to believe this thesis is that many of the failures in bulbs for domestic use I've used in my home apparently failed after storms with atmospheric discharges, therefore, I would agree with the following statement.

Pages 12 to 13 of the following document give details of Electrical Power System faults which can cause large mains overvoltages for periods of milliseconds to minutes. These would kill LED streetlights without doubt.
 
  • Like
Reactions: treez

    T

    Points: 2
    Helpful Answer Positive Rating
Thanks but that stateent refers to "Power system faults" as oppose to "Mains transients".
As you know, the power system can be without fault but still have mains transinets on it.
But i appreciate what you mean, that power line faults/transients can kill led lights.

The thing about power system "faults" is that it is not possible to protect a led light against them, they will always kill a led light, unless you somehow switch the light off the network as the fault happens.
As you know, the way such a fault is usually dealt with is just to "crowbar" the fuse. Which kills the led light, but renders it not working
 

This is related to the above so please may i post it here...

Hello,
Page 5 of the following shows that for outdoor LED lighting, there are two regulatory mains transient tests which need to be carried out on the lamp….
https://www.littelfuse.com/~/media/..._module_design_and_installation_guide.pdf.pdf

One involves taking the input of the lamp up to 5kV for 50us.
The other involves passing 3000 Amps into the product for 20us
Do you know where we can buy the equipment needed to do these tests? Also, do you know how much it would cost?
Also, for the test that involves passing 3000 Amps into the product, -this would obviously require the voltage at the product’s input terminals to be raised until 3000 Amps passed through the product. This 3000 Amps presumably passes through the MOV of the product. How does the equipment that produces the 3000 Amps pulse know what voltage to go up to in order to get the MOV inside the product-under-test to sink 3000 Amps?
 

Not mentioned in the Littelfuse paper, but the quoted tests are basically specified in IEC61000-4-5 "Surge immunity test", you'll review the specification to know about the test setup details, and search for IEC61000-4-5 compatible test equipment respectively test houses that are providing the tests.

I believe that a "pre-compliance" test setup with similar parameters can be made in a lab. Although 3000 A and 5 kV sounds like huge energies, the test voltages and currents can be provided by handy devices.
 
  • Like
Reactions: treez

    T

    Points: 2
    Helpful Answer Positive Rating
Hello,
This is related to the above so may i post it here?
A 6kV mains transient is not like a 6kV ideal voltage source suddenly getting switched in series with the mains, but instead is less "stiff"...ie, it is just 6kV caused by charge getting induced into the wiring capacitance of the mains electricity system....as such, quenching this transient is not as bad as it sounds. Do you agree?
 

In relation to this, do you believe that an IGBT based transient protector would be a good way of providing more exact transient protection to a circuit? We have an offline circuit where the LED driver IC Bias pin connects directly to the high voltage DC bus (post recticifier)...we need to protect it but all TVS's are too wide in breakover voltage tolerance...so what about an IGBT based circuit that trips the IGBT to conduct and shunt the transient when a comparator trips due to a potential divider across the DC bus?

- - - Updated - - -

Again on this subject of mains transients if I may…. I am used to offline SMPS’s which have large electrolytic capacitors either immediatley after the rectifier (<75w)…or just the other side of the boost PFC inductor and diode (>75W). It never really occurred to me that these electrolytics obviously play a major part in quenching mains transients….but obviously they do…do you agree?
I am recently working with products which comprise utterly no Mains AC filter whatsoever, and not even a single decoupling capacitor on the post rectifier bus. Do you agree that the lack of a mains filter and post rectifier electrolytic will make it more susceptible to damage from mains transients?….i mean, they do have TVS’s, but these act above the voltage that would cause damage to the ICs.
 

Hello,
Page 24 of this………..

https://www.google.co.uk/url?sa=t&r...er.pdf&usg=AFQjCNEfWFWm6FME1p30Kv4TWHq3X6rXhw

…states that a 470pF capacitor (C1) placed at the mains input to an offline led driver can provide protection against mains voltage spikes. (schematic is on page 11).
Surely this cannot be serious?, if it was that easy to protect against mains voltage transients than you wouldn’t have the multi million dollar industry in MOVs and TVS’s and surge protectors etc?
 

Of course not!

The capacitor alone would have virtually no effect but in conjunction with the 20 Ohm input resistor and wiring inductance it helps by forming a low pass filter. It has no clamping action but it will attenuate brief pulses and spikes to some degree.

Note that the reference design is to demonstrate how it works and might not pass strict EMC regulations so there could well be a need for further filtering in the AC lines which is not shown.

Brian.
 
  • Like
Reactions: treez

    T

    Points: 2
    Helpful Answer Positive Rating
As per data sheet
Capacitor C1 (470 pF recommended) is added to improve differential mode filtering (in
combination with L1; 2.2 mH recommended). By placing this component close to the
mains input, the inductive coupling to mains is reduced. It also provides buffering against
voltage spikes.

This C1 capacitor are X type{https://www.illinoiscapacitor.com/pdf/Papers/EMI_RFI_interference_suppression.pdf } capacitor for differential mode filtering, it may attenuate the input spike but not suppers completely. You should use TVS for this application. Even MOV also not fit for passing EMC test
 
  • Like
Reactions: treez

    T

    Points: 2
    Helpful Answer Positive Rating
Thanks, it makes you wonder what percentage of mains transients that go over 450V can actually be completely snuffed out to below 400V by just a simple AC filter? I suspect many are weak.

Also, I suspect that mains transients (like the Y2K problem) are not as common as we are told. When equipment like MRI scanners etc fails its very common for the customer relations staff to make up stories of “horrendous mains transients” so as to kind of blame the mains for the failure of equipment, and placate the customer. Also, the multi million pound transient protection industry needs people to believe in mains transients.
Nobody ever seems to be able to say what is the typical magnitude of a mains transient voltage spike. Nobody has a (real, genuine) scope shot of a mains transient., in spite of sample and hold technology. There are no figures for the general average occurrence of mains transients above 450V say.
We are always told that “Inductive heavy current switching” exposes us to mains transients…. This should be simulatable, -yet try and simulate one, that will come through the AC filter of the offending equipment, and breach some other equipment somewhere else on the same mains phase, and its hard going to simulate anything that looks like a significant threat.

- - - Updated - - -

Also, if i connect the attached "mains transient catcher" to the mains in UK, Europe , or USA, then how long will it be before i "catch" a mains transient?
 

Attachments

  • Mains transient catcher.pdf
    14.4 KB · Views: 101
  • Mains transient catcher.txt
    6.7 KB · Views: 66
Last edited by a moderator:

I can only speak from experience while investigating 'clean' electrical environments used for extreme RF testing. This involved multiple screens around a room where even the air ducting had multiple screens and absorbers to make sure no signals crept in or out through the pipes. All electricity supplies passed through a bank of filters to eliminate a wide spectrum of frequencies and of course no switch mode supplies were allowed in the room!

One of the tests was to ensure no 'spikes' could penetrate the defenses so monitors for drop out and over-voltage were fitted to the incoming and internal supply. This was near Swindon in the UK but I have done similar checks on on smaller installations around the UK and USA. We found >450V spikes typically occurred on average 7 times per hour but bear in mind these were in industrial complexes where there was probably large air conditioning and mechanization motors nearby.

You have to accept that spikes are a fact of life but are mostly quite benign. When you simulate transients in your schematics you are assuming the AC source is low impedance or zero, in real life there may be hundreds of metres of cable between source and equipment and much of the energy is dissipated as a dispersed magnetic field. Place an AM radio near a mains power cable and tune to a clear low frequency and you will hear the spikes clearly, the fact that the radio picks them up with no direct connection is proof that some of the energy has been converted to EM radiation.

The other thing to consider is the duration and timing of the spikes. Suppose you get a 450V positive spike at the peak positive cycle, the total is around 340+450 = 800V but the same spike at peak negative would be -340+450 = 110V. Those are extreme cases, statistically they would fall anywhere between the two.

I'm not sure what exactly your schematic is supposed to do. I'm guessing your intention is to light the LED when the input voltage exceeds around 500V and it would probably work if you hooked it up to a DC power supply and wound the voltage up beyond 500V but it would not detect transients well. The reason is simply the response time would be too slow to catch many of the spikes.

Instead of monitoring the DC, try this simple experiment: use a 1:1 transformer, an audio isolating transformer would work well (600 Ohm type) and wire a fuse and 470pF capacitor (1KV rating or higher) in series to one side then to the mains. On the other side add a 1K resistor as load and monitor across it with an oscilloscope. Obviously take care with isolation and use an isolating transformer on the scope side if possible. You should get some 50Hz across the resistor but it will be relatively low level and will be distorted by it's harmonic contents, the higher frequencies in the transients will be less attenuated and should be easy to observe.

Brian.
 
  • Like
Reactions: treez

    T

    Points: 2
    Helpful Answer Positive Rating
use an isolating transformer on the scope side if possible.
Thannkyou, i take it this is to protect the scope from possible insulation damage due to the fact that the little audio transformer isn't very well isolated?
 

That's correct. Also use the smallest rated fuse you can find.
The object of the exercise is to see what is present on the AC mains when as much 50Hz as possible has been removed. Most small 600 Ohm 1:1 transformers for telecoms use are rated to about 1.5KV isolation so they should be safe but take no chances.

Brian.
 
  • Like
Reactions: treez

    T

    Points: 2
    Helpful Answer Positive Rating
The reason is simply the response time would be too slow to catch many of the spikes.
Thankyou for you #16 comment here, would I be wrong to infer that from this you believe that the vast majority of mains transients are very weak?. –As you know, it wouldn’t take much to turn the LED on in the schematic of post #15. So surely the majority of mains transients must be weak?, and probably most of them wouldn’t make it through the AC filter of most small <75W offline SMPS’s, let alone get to a TVS, say placed downstream of the rectifier bridge?
 
Last edited by a moderator:

I can't speak for all situations but in general, devices that create spikes will have supressors in them anyway and any other causes, such as lightning strikes or grid switching tend to be some distance away and therefore the energy is widely dispersed. Of course, it you are right under a lightning strike the spike would be very severe but you probably wouldn't live to worry about it!

Consider that in your simulations you factor in an inductance for the incoming power wires, now think how far away the nearest electricity company transformer is and therefore the addition wiring lengths and capacitance between wires and to ground. In most cases it makes a pretty good low pass filter and with low enough impedance that transients quickly get damped. Given their inductance and existing protection, it's unlikely anything significant will pass through a sub-station transformer so you are working in a fairly localized power network. When it comes to street lighting, typically, in the UK at least, a quite thin cable is used to link lights in a chain and it is buried underground. Many individual lights or small groups of lights are then fed via a timer box or an ambient light level sensor. In both situations, when switched off, the power feed is isolated so transients wouldn't reach the lighting circuit anyway. On average, but depending on the time of year, that probably reduces the risk of damage by 50%.

I'm not trying to be over optimistic, SOMETHING kills lights but I suspect manufacturing defects or overdriving them is as much responsible as outside influences. Incidentally, I've just come across a product made by a very well known lighting company that drives 3mm indicator LEDs at 80mA to get maximum brightness. The data sheet specifies a maximum of 20mA. Needless to say, all the products have failed with dead LEDS.

Brian.
 
  • Like
Reactions: treez

    T

    Points: 2
    Helpful Answer Positive Rating
Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top