Kato platform lighting

[cteno4]

Nice solution and hides the strip and wires to boot.

 

Jeff

[Khaul]

The filters work well and they do hide both the strip and the wire connections nicely. But I used more white paint in one of the filters than in the other so now the light is a bit different. I will make new ones soon. I wonder if the LEDs would shine through Evergreen sheet.

[cteno4]

Might take a look around the craft store to see if there is some partially opaque thin plastic film that you could use strips of instead of paint.

 

Jeff

[cteno4]

Wonder if a light masking tape like this might work

 

http://www.ebay.com/itm/White-Rolls-Striping-Nail-Art-Tips-Tape-Line-Sticker-DIY-Guide-UV-Gel-Tips-/401136060376?hash=item5d659287d8:g:sCgAAOSwmtJXXf90

 

Also might try just using fine sand paper to opaque the clear acetate. Also there is the window films that you can use to frost windows. Some places will send samples free or for a small charge.

Jeff

Edited June 7, 2017 by cteno4

[kvp]

In the paper and craft store i usually get my foamcore boards, they have clear and various tinted (including semi transparent white) plastic sheets used as covers for comb binded booklets. The clear ones are great as glass for scratchbuilt buildings and they come in nice A4 sized pages. I think the semi transparent white one could be usable as a diffuser.

[Khaul]

Excellent ideas!

[inobu]

Kinda, what post #66 was eluding to. If you use the dots you can fill in the areas where it emits too much light and open the areas where you need more light. Its the density of the dots that allows you to control the lighting.

 

Inobu

Edited June 7, 2017 by inobu

[Khaul]

My lights got burnt. Well, not the lights but the resistor. Initially I thought there was a problem with my covers being too close to the LED strips but then I realise the problem also affects to the lights in the station entrance, which have no filters. I think it is just the cheap, nasty lights I got. I'll check the voltage in the power supply, just in case.

 

Edited June 16, 2017 by Khaul

[inobu]

Khaul,

 

I think this is an issue that should be commented on as its a safety issue. I believe that the problem is in the dynamics of your components. The LEDs that you are using is designed for 12v and its usage is calculated in watts per foot. It is meant for longer runs and higher lumination.

 

Although it has marking for inches it is meant for 1ft + inches. Our modeling needs are low level, low voltage and low lumination. This means in the area of 3v.

 

The 3 LED appears to be too much for the 1 resistors as the resistor is meant to absorb the current and is calculated based on a foot run which would include a number of (1 foot) LED's and resistors to share the load.

 

Resistors have a wattage level which defines the amount of current it can handle. I think the resistor does not have the rating to support 3 12v LEDs by itself hence the over heating.

 

I don't know what the specs are on your LED strips but I think this is the issue.

 

Try this, Calculate the resistor value needed to operate 3 LED,s and compare it to the resistor on your strip. If the values are way off then you will know for sure.

 

Inobu

[cteno4]

Inobu,

 

On these strips each of these little circuits of 3 leds and resistor are independent and thus not bound by how many are used. Each resistor is just dropping the current for its three leds. The strip basically has a power bus that each led circuit of three leds in series and its current dropping resistor feeds off of. No difference to each resistor if one set or 100 are there (except if the power source doesn't have the amperage to run 100 sets).

 

In this case if white leds at like 3.5v ea and 20ma then it would be 1.5v difference therefor a 75R resistor and it would be dissipating 30mw. Most of these smd resistors are usually 125mw, some a bit lower but should be over 60mw for a safety margin. Granted this is running the leds full out which is not good, but I expect they run them pretty high like this.

 

If like 18v were applied then the resistor would be dissipating more like 150mw and could fail before the leds blow from excess current.

 

I agree it's better to run leds much lower than their full output used in regular lighting situations. using lower voltages means less power usually being dissipated thru resistors so less chance of failure. Usually better to use more, dimmer leds, than one bright one in lighting.

 

Jeff

[Khaul]

I think Jeff is right, each set of 3 LEDs has it's own resistor to drop the voltage to adequate levels. The strips are marketed to run at 12V so there should be a safety margin to run them at 12V. I'll run them at a lower voltage next time, just in case.

[kvp]

Many of these strips tend to burn out if not mounted on a heat sink material. This is because some components are undersized and won't operate reliably in a partially enclosed area mounted on a thermally isolating surface like wood or plastic. Using thick copper or aluminum strips as a heat sink could help. For power, it's important to have a smooth 12V DC supply as non filtered 12V adapters might have an average of 12V and a peak of almost 17V. Also checking for AC signals (or AC components, like turnout relay spikes in case of a shared aux supply) is always a good idea. Running them slightly above the turn on voltage, like 10V for a 3 white led chain might also help.

[inobu]

I think Jeff is right, each set of 3 LEDs has it's own resistor to drop the voltage to adequate levels. The strips are marketed to run at 12V so there should be a safety margin to run them at 12V. I'll run them at a lower voltage next time, just in case.

The issue is not the LED's configuration. It is 3 LED's and 1 burnt resistor.Bottom line the resistor could not support the 3 LED's.

 

I read some where that the number of LED's on the on the run and the distance between them plays a roles. Also the 2835 LED have different specs.

 

In any case becareful those LED have adapters running 12v @6amps and thats hot.

 

Inobu

[Khaul]

I think I have to check my supply, which is independent of the power pack. I assumed it gives a flat 12v dc but that's probably assuming too much. I don't want to buy an oscilloscope, though! The heatsink looks like a good idea, and it could also work as a better monting surface for the strip.

[mrp]

Maybe add a simple PWM dimmer like this one?

 

https://www.aliexpress.com/item/EWS-PWM-Dimming-Controller-for-LED-Lights-Ribbon-Strip-12-24-Volt-12V-24V-8-Amp/32788937390.html

 

(Note: I haven't used this item myself.  It seems suspiciously cheap at $1.97 plus 39 cents shipping - but it does have a 5.0 star rating...)

[cteno4]

Inobu,

 

if the strip uses 2835 leds those are 60ma and if at that level then yes the resistor would be dissipating like 150mw and thus need 1/4w resistors, which I'm sure they would use if pumping at 60ma.

 

Longer length will require more amps from the power supply to run them all, but even a single circuit won't matter if on. 6a or a 0.2a power supply as the circuit only draws the power it runs.

 

If there were a power spike the resistor would most likely blow first as leds do have a short peak amp they can sustain for a bit.

 

Since more than one look to have blown at once it would either be they all have deficient resistors or there was a power spike to them all.

 

All that said I do agree its always best to engineer things with fat saftey margins. I don't really use these strips for modeling. I've played with them a lot to see what I might use them for and used them for regular lighting and not yet had any failures like this in the cheapo ones I've gotten from ebay. But I prefer to make my own led circuits and I do like over engineering so stuff like this won't happen! I also like to use 3-5v for the power supply for better safety and less power waste in resistor heat.

 

Cheers

 

Jeff

[inobu]

I think I have to check my supply, which is independent of the power pack. I assumed it gives a flat 12v dc but that's probably assuming too much. I don't want to buy an oscilloscope, though! The heatsink looks like a good idea, and it could also work as a better monting surface for the strip.

The heatsink only treats the symptoms of the problem. You have to address the root cause. The heatsink is designed to draw heat to itself. That means it will get hot. If it is near the plastic it will melt or warp it.

 

I would not use the strips nor recommend them after seeing this but thats me.

 

 

Ichi

[Khaul]

For the record, this is the DC supply I used:

 

 

[kvp]

That's a nice unregulated and unfiltered adapter that could put out way more volts if the 240V power supply has a spike and it has an unfiltered output, so the voltage could vary between 0 and 17V with only a time average of 12V. My advice is to never use an unfiltered, unregulated power supply for anything with solid state electronics in it (like leds). A filtered supply should have a very smooth DC level, both unloaded and under full load, without any ripples.

 

A few other points:

 

Heatsinking is a proper way to give more heat dissipation surface to a component. If the heat absorbed by the sink is in balance than what it could give off at its designated operating temperature then it will be stable. So if the heatsink has warmed to its maximal operating temperature with the given load and its still below the melting/warping point of the plastic, then it will be stable. Heatsinks could actually help spread the heat from point heat sources like leds to a lower overall value so instead of melting a plastic cover at one spot, the whole cover gets a bit warmer. A word of caution though, the operating tempreature where the balance point is dependent on the temperature and humidity of the cooling air (besides the power output of the heat source and the surface area of the heatsink) So proper heatsinking requires careful thermal calculations or you could just go way above the requirements to be really sure. Getting the power of the leds down to allow self cooling is another way, but has the same logic behind it.

 

These ribbons are designed as a parallel chain of short 3 led strips, where each 3 led group is in series with a single series current limiting resistor. If you cut them where the scissor marks are printed, you get independent 3 led strips. This means if you use them in any lenght, you just have to count the number of strips and since the are in parallel, you just add up the power consumption. With a stable input voltage, current requirements will be determined by the number of 3 led substrips. As Jeff has said, giving it a larger current capacity supply wont hurt it will use as much current as it needs. The voltage must be very stable though as a higer voltage will cause the circuit to operate at a higher current and the weakest component will blow up sooner or later. On the other hand a voltage below the cumulative turn on voltage of the 3 leds will cause them to turn off and no current will be used. In this case, each white led needs around 3.3V, so 3 leds means 9.9V. For 12V power, the leds have to regulate current by allowing through the selected current (probably 10 to 20 mA) at the remaining voltage (12-9.9=2.1V). If the power supply peeks at 17V, the same resistor will get 17-9.9=7.1V, which means you'll get a peek current more than 3 times. That's still within the temporary overload of the leds, but the poor resistor will get really hot if it was not selected with a 4 times larger power rating than the absolute maxium of the circuit. Usually this absolute maximum is around 14.4V, meaning a 20% spare to allow instabilities in a regulated but not fully filtered power supply. This low tolerance i what makes these led strips very cheap.

[inobu]

The length of the strip is directly proportional to the length of its heatsink. These strips will operate better by the foot which becomes a problem for our application.

 

Adding other LEDs will help stablize the current draw and adds to the length of the heatsink reducing thermal runaway. The LED's used on these strips have a higher wattage/rating therefore prone to thermal runaway on short runs.

 

Add this to the low tolerance and you have a burnt resistor and potential fire.

 

Your station looked good with the diffusers now you need to find the right LED's and you will be good to go.

 

 

Inobu

[kvp]

No Inobu, these strips are made from self contained 3 led substrips. You need a stable power source and then the number and distribution of 3 led substrips are indifferent as long as the total power consumption is below the power supply rating. The key is a stabilized and filtered power supply as the wasted power and used current only depends on the voltage. Any heatsinking should be calculated and done on a substrip basis as the power scales exactly linear with the number of substrips.

[cteno4]

Inobu,

 

Sorry I don't agree, the dissipation should be linear, each block should dissipate the same amount of heat. If anything single modules should dissipate heat better as they have the ends then as well open and not a unit dissipating heat right next to it when in multi unit string. So I don't understand your logic here.

 

When I've run these playing with then with a decent unfiltered 12v power supplies they never got hot and never fried. I left some on for a few days straight as I was concerned they might get hot. I've done this with whole 5m roll for a few days just to see what the failure rate might be and none fried.

 

Mrv's suggestion of a simple Pwm dimmer is great as it will take the high shine off the strips and help mitigate power spikes some by reducing the time the components are getting the increased voltage.

 

Again I'm not a fan of the led strips for modeling for visual and practical reasons, but I've not had them go catastrophic in my playing with them. I do think the most likely culprit is a power spike since all the resistors seemed to fry at once. A good test would be to see if khaul could put a few feet of these leds on the power supply for a few days to see if more fry and if they do, then try some with with another power supply.

 

Personally I don't use the strips in modeling and I like to roll my own on this stuff and give higher protection and lower margins for extra power in my design. But I've not heard of these strips catastrophically failing a lot out there and lots of folks are using them in many different ways these days.

 

Allowing components that get warm to float to get air circulation around it or adding a heat sink for it to dissipate heat better is a good practice. Smd components are a problem as they tend to be tiny so very little mass to suck up heat but they do have a higher surface area to volume ratio to dissipate heat. Smds are designed for the pic board to be a contact heat sink and then open air above to dissipate heat as well. They are also designed to fail quickly at over power so it can protect the rest of the circuit as well as limit how much energy is there heating things up before going poof and breaking the circuit.

 

Again I like to have things designed so there is as little heat being dissipated as possible from the whole circuit, better lighting effect, better lifetime performance, less chance of a poof, and less power waste.

 

While any resistor meltdown could potentially start a fire, I seriously doubt that this little smd resistor failure is going to catch the layout on fire. Ive seen much more serious heat damage to plastics and wood that looked much closer to a fire starting with old grain of wheat bulbs being burned at 18v and never heard of large numbers of layout fires occurring from them.

 

Jeff

[inobu]

ok, So why did resistor fry?

 

 

Inobu

[kvp]

ok, So why did resistor fry?

Inobu

Probably overheating caused by overcurrent, caused by overvoltage from the unstabilized power supply. The leds took it better and the weakest components blew.

 

Unstabilized power supplies have fluctuating output voltages based on line voltage and load, while unfiltered but stabilized supplies have noisy but mostly load independently stable outputs. Stabilized and filtered is the best for long led life and no flicker.

 

Stabilized supplies are a must for leds but filtered and stabilized ones are better.

[inobu]

Probably overheating caused by overcurrent, caused by overvoltage from the unstabilized power supply. The leds took it better and the weakest components blew.

Unstabilized power supplies have fluctuating output voltages based on line voltage and load, while unfiltered but stabilized supplies have noisy but mostly load independently stable outputs. Stabilized and filtered is the best for long led life and no flicker.

Stabilized supplies are a must for leds but filtered and stabilized ones are better.

 

So what would have happen if more LED are added to the string?