EEVblog Electronics Community Forum
Electronics => Beginners => Topic started by: paulca on January 11, 2018, 10:38:19 am
-
Given an example LED strip which is 5 meters long and claims maximum wattage of 18W per meter, that's 90W.
The strip is 5V, so that's 18A, call it 20A for grace.
5V @ 20A is an awful lot of amps to be routing around a living room. To be sensible, for a 24/7/365 installation that might be switched on 12 hours a day you would need something like 10 or 12AWG wire. Or 2.5mm twin + earth installation cable.
So I considered two things.
1. Raise the transmission voltage and drop it at the start of the LED strip. If I take the voltage up to 12V or 24V I can drop the amps to a much more reasonable 8,3A or 4.2A. Two issues then present themselves. 24V PSUs are not cheap. In fact 5V 20A PSUs are not cheap either! Nor 12V 8A. It would also require a fairly high current buck converter which is either going to be expensive or risky (Chinese ebay modules).
2. ATX Power supply. I can buy a fully UK safety rated main power ATX power supply which will happily deliver 20A on it's 5V rail. I can run the power over mains 240V IEC cable close to the LEDs, pull the 5V line from the ATX, bridge all the 5V bus wires together (having verified they are all the same bus in the PSU) and connect them to the LED strip.
3. The conductors on the LED strip do not seem all that thick. I realise that the 20A pull will only be if all 300 LEDs are on, full brightness, white. I expect the conductor size on the LED strip (around 18AWG) is assuming you won't do that. At 20A for longer than a minute I expect 18AWG wire will become quite warm. After an hour it could be pretty toasty.
I could of course buy a dedicated 5V, 20A DC block power supply, but again a lot of these come from China or are very expensive.
What do others do?
-
I'd get a 12V LED strip and add a power wire each x meters. High power LED strips for 5V are a bad idea.
-
I got a few 5m led strips lately for an installation (customers house) and also to test with my ongoing projects.
The first thing i noticed was a label on the back of each packing warning about proper power supply termination
The image shows three examples
a 2,5m strip, 3,5m strip and 7,5m strip.
For the 2,5m strip you connect the PSU to one side of the strip and you are done
For the 3,5m strip you connect the PSU both sides of the strip
For the 7,5m strip though you connect the PSU both sides and also at the middle of the strip.
No matter what cables you use, the strip is thin... for higher power strips (mine was 4,8W/m) i am sure the manufacturer will demand power connection at shorter ranges too besides having "heavier" strip copper
-
Hmmm. For my use case (running a 5m strip up the stairs) power in the middle would not be possible.
Also... I know this is a basic question, but does it need to be the same PSU at both ends? Running 20A DC power cables up the stairs just doesn't sound nice.
I might consider dropping the 300LED strip for the lesser 150LED strip which halves the power.
I did consider the 12V strips, but they were the lesser 150LED... which would again drop my amps I suppose. There is a risk they turn up in the post as 3x50 and not individually addressable, but addressed in 3s.
Then again, I could live with that. I just thought it would be cool to have a PIR sensor controller stair light, mostly white with an R, G, B tracer running up (or down) them.
-
Hmmm. For my use case (running a 5m strip up the stairs) power in the middle would not be possible.
Also... I know this is a basic question, but does it need to be the same PSU at both ends? Running 20A DC power cables up the stairs just doesn't sound nice.
I might consider dropping the 300LED strip for the lesser 150LED strip which halves the power.
I did consider the 12V strips, but they were the lesser 150LED... which would again drop my amps I suppose. There is a risk they turn up in the post as 3x50 and not individually addressable, but addressed in 3s.
Then again, I could live with that. I just thought it would be cool to have a PIR sensor controller stair light, mostly white with an R, G, B tracer running up (or down) them.
if you are going to use 5m of LED strip simply cut it in two and attach two smaller PSU's one on each side of the total length... problem solved
-
Running 20A DC power cables up the stairs just doesn't sound nice.
Also, if you connect power from a single PSU to two points of the same strip, you can safely use cables for half the amps on each side since they will distribute the power evenly... always add 10-20% more than required just to play safe of course, will not hurt!
PS if your led strip is say 20W/m @5V instead of using 20A cables you can connect power to the strip every 1m (6 cables from side to side) with cables rated for 4A...
-
PS if your led strip is say 20W/m @5V instead of using 20A cables you can connect power to the strip every 1m (6 cables from side to side) with cables rated for 4A...
Yes, but I would have to run 4 sets of 4A cables up the stairs.
Starting to sound like a mess. Or require a proper installation with the cables routed under the stairs. I was hoping to just stick the strip to the skirting board and power it at the bottom.
I'll look up the amp requirement on the 150LED 12V strip.
ATX power supplies are ground references aren't they? So I can't use two supplies, one at the top, one at the bottom or I would create a ground loop. How much does that matter?
Note that cutting the strip in two to use dual supplies won't work as I need to run a data line through all of them and that data line would need a common ground between the supplies.
-
I dont know what complications might occur by two PSU's connected to the same load......
But, i think your connection will be like the one for the 3,5m in the photo above... so, one pair of cables on the bottom and only one pair of cables running up halfway of the strip... as for proper installation my limited experience with these buggers tells me either use an extruded and diffused aluminum profile to house the strip (and cables if you do it right) or at least add silicon ties available for led strip installation like the one in the photo... dont count on the double side adhesive as it will NOT stick as you expect except if you reinforce or replace it altogether with something really sticky and reliable....
-
Aluminum strip with extra cables running up it sounds fair. Adds expense etc. So here how this has developed.
Run a 5m strip with 300 LEDs on 5V up the long side of the stairs, power it at the bottom.
Run a 5m strip of 150LEDs on 12V
Run a 5m strip of 150LEDs on 12V up the stairs with a enclosure and extra cables
- 12V LED addressable LEDs are hard to find, except from China and it's uncertain if they even are 12V
Run 5m strip of 150LEDs on 5V with 3 or 4 sets of power cables in the enclosure up the stairs
Run 3m strip of ~100LEDs up the short side of the stairs on 5V with enclosure and extra power cables
The cost jumps from about £40 to about £100.
Software throttling i an option with fuses to bad it up. Totally the brightness across the LEDs (or ensuring the total brightness never exceeds a preset value) is an option. eg. you can have a small number of full brightness LEDs or a large number of dim LEDs, thus keeping the amps low. Definitely need the fuses though, if the software limit fails.
By the way, this is fairly scary:
https://learn.adafruit.com/1500-neopixel-led-curtain-with-raspberry-pi-fadecandy/power-topology
86Amps!
-
Good luck however you finally approach this...
Scary? I was scared the first time i realized a brand new computer needed 800-1000W PSU!!!!!
-
That 90W strip is going to be pretty bright. Do you have lux or FC estimates for the strip? In the absence of that and making a somewhat educated guess, that 90W strip would produce light that is the equivalent of 6 or more 60W incandescent light bulbs. Pretty bright. Though, there are a lot of mitigating factors that probably lessen the actual effect. But for lighting a stairway I would think you need considerably less light. In my house, I have some tread lights and am using 9 watts (3 x 3 watt lights) per flight of stairs. Not the same as a strip and different LEDs but probably reasonably similar power needs. Certainly not an order of magnitude greater.
I've also used a number LED strips around my house. The biggest costs were for the aluminum channels + covers to mount the strips. I like the channels because they help with heat diffusion and, obviously, protect the strips. A 90W strip will definitely need help in that regard.
-
Yea, I'm coming to that conclusion.
Less LEDs, 12V and I would probably accept the older WS2811 grouped in 3s (3 LEDs in series) for each chip.
I can limit the brightness in software and do some calculations to bring the amps down. Then fuse it in case a software bug slams them all 100% brightness.
I gather the channels would be worth while. The downside for me is the place is rented, so I can't really screw them on, they will have to be put on in a way I can remove them without lifting the varnish off the skirting.
Maybe it's not such a good idea.
I have a 1 meter 12V strip to play with, which I'm hoping to make into a monitor background light which will keep me busy.
-
Mounting the extrusion could be easy depending on what your mounting to.
3M has done a great job with the damage free mounting strips under the Command Brand. They're two sided elastic adhesive strips that releases when you stretch it. Use as many as you need to support the extrusion and when it's time to put it out just pull on the strip tabs.
https://www.amazon.com/Command-Hanging-Value-Pack-48-Pairs-17024-48ES/dp/B001KYSAN4/ref=pd_sim_60_5?_encoding=UTF8&pd_rd_i=B001KYSAN4&pd_rd_r=G0BA1NMNY5EVC71HCES4&pd_rd_w=9A5w3&pd_rd_wg=IBmgs&psc=1&refRID=G0BA1NMNY5EVC71HCES4 (https://www.amazon.com/Command-Hanging-Value-Pack-48-Pairs-17024-48ES/dp/B001KYSAN4/ref=pd_sim_60_5?_encoding=UTF8&pd_rd_i=B001KYSAN4&pd_rd_r=G0BA1NMNY5EVC71HCES4&pd_rd_w=9A5w3&pd_rd_wg=IBmgs&psc=1&refRID=G0BA1NMNY5EVC71HCES4)
-
The LED strip power consumption is nothing like what they quote.
I use old XBox power supplies, which are pretty rugged and put out a good 12V rail for LED use.
There are also some neat aluminium mounting strips available now, with diffuse covers and snap-in mounting.
-
The LED strip power consumption is nothing like what they quote.
I gather this. I think they have taken the maximum theoretical power requirements of the LED and multiplied it by the number of LEDs per meter. Which is fine to give an "engineering grace".
The LEDs claim 20mA max current, there are 3 LEDs per 'chip' giving 60mA. 60 LEDs per meter = 3.6Amp per meter.
However my little test strip, 1m, 30 LEDs running the FastLED demo reel off my bench supply never shows more than 0.5Amp and that's a peak outlyer, it's usually around 100-300mA. I would test it with 100% white brightness, but I have soldered the strip into a perfboard and would need to pull the chip to reprogram. I ordered another strip with 60 LEDs and will test it.
The FastLED pattern is interesting. All LED colour brightness is calculated as a fraction of a single constant LED_BRIGHTNESS. If this was 50% (127) it might not necessarily mean half the amps, but it would be close. This would act as a current throttle. And as I said above, back that up with a fuse.
What I am struggling to find is the max current for the strip, I mean the actual conductors within the strips.
I guessing 2 amps is probably about the upper limit of the strip conductors. They are quite thin and the tracks look to only be about 2mm wide. Assuming they are PCB trace style copper tracks, does 2 amp sound about right?
Another factor is measured resistance across the length of the strip, the higher the current the higher the voltage drop by the end of course.
-
The LEDs claim 20mA max current, there are 3 LEDs per 'chip' giving 60mA. 60 LEDs per meter = 3.6Amp per meter.
a 150LED/m strip with 20mA per LED would take 3A max
60LED/m strip with 20mA per LED = 1.2A
I dont know what strips you got and there are a ton of them out there but what you got should be rated as W/m from that and their rated voltage you can make out the total A consumption at full brightness.
Those i use now are rated 4.8W/m they can be cut every 5cm and they got 3 LEDs in each 5cm block, so there are 60 LED's per meter for a total of 4.8W and they are 12V strips so they should draw roughly 400mA per meter which is about right according to my measurements. A 5m strip of this type would take 2A total current at full brightness.
The copper lines in the strip I am sure they take a lot more current than 2-3amps BUT they are heated both by the current through them AND by the heat from the LED's themselves.. since the construction is in fact a sandwich of copper and plastic i can only guess they are more concerned about a meltdown in some hotspot, esp when not properly cooled..
-
Maybe the confusion is that the strips I am using are RGB, so 3 LEDs in each little LED component.
I checked the datasheet, which doesn't really help.
If you check the application circuits, Fig. 2 (the 12V circuit) is what I have now, 1m x 30 LEDs. However they are also sold in 5V, 5 meter lengths with 60 LEDs per meter.
https://cdn-shop.adafruit.com/datasheets/WS2811.pdf
The current drive figures on the example applications do not say if that is 18.5mA per LED, red, green, blue or 18.5mA for all 3.
I'll wait till I get the new set and try and measure the current per LED, per colour, per brightness.
The very worrying thing about these LEDs is they are becoming very popular with the maker community, but, they do not wire up as you would think due to the high currents involved. There are also Adafruit et al articles about running them on Lithium cells without the slightest mention of handling high current. A 5M 90W 5V strip pulling 18 amps will explode most average lithium cells, such as laptop cells and would need high current 18650s from the vaping fraternity or at least a medium discharge LiPo pack from the RC fraternity.
Of course when you buy them off ebay they arrive in a plastic bag coiled up with absolutely no information or documentation. A lot of people are liable to just wire them to a PSU and not consider at full brightness white they will pull high amps and potentially start a fire in the wires.
I wonder how many people have set fire to stuff with them.
-
I dont know either... but the datasheet is from a siscreet IC not the one inside the led which probably has different power dissipation etc. But that is just my gut feeling, i am not sure at all....
I wouldn't worry of exploding lithium batteries though... they will deliver what they can deliver and your leds will not be as bright as you'd expect, unless the batteries get shorted, they will not even heat up!!!
I wouldn't worry also because at least personally, avoid to use lithium batteries unprotected from overdischarge/overcharge and short so yes, you can be pretty much sure you are 100% safe when doing so... I am also sure with all the 18650's craze over the past few years we'd know if they exploded so easy..
Actually, all the lithium batteries explosions i've seen on youtube or mentioned online are either faulty phone batteries (or faulty phone chargers?) and on purpose attempts to show what they look like exploding! LOL
The pack in the photo has 6 034500 cells in parallel, they are not only used but abused and the total capacity is around 8000mAh (in a compact size of 75x50x70mm)
The board on top charging and discharging the pack is my TP5100FLEXADV-v2.00 (i am almost done with v3.00 but added a lot to it so its going to be twice the size and with more functionality).
This is charged either via USB from or attaching a PSU in the screw terminal takes 5-18V charging current is set to 1.4V but i am going to change it to the full 2A once i get the respective resistors and you get power either directly from battery so, 4.2V protected or from the MT3608 booster onboard (4.2-24V)...
I used it during the Xmas period to power the RGB led controller you see in the photo and as i didnt have RGB strip of any kind i simply hooked 3m white strip 1m per color and the effect was amazing, i believe better than a single RGB strip... it run for 12-15hrs spanned in two days (two nights actually)... charging time is a bit long but that is expected.. i think 7-8hrs maybe more (but the batteries are beat up and take forever to terminate charging)
All this without ANY signs of overheating (the pack is temp protected via NTC) and the MT3608 runs cool at these loads)...
Do i sense a sort of "fear of lithium" into you my friend?
(missed to upload the photo - added one while starting to charge it happily at 1,4A draw)
-
Do i sense a sort of "fear of lithium" into you my friend?
Not at all. An understanding and a respect for lithium maybe. My fear is for others with lithium cells. Some times the warnings are overstated, some times they are understated. "You'll be fine", approach, "What could possibly go wrong?"
Most of my experience with them is in high discharge LiPos in 3 series. There are so many good documents online and also so many dumb mistakes that people have made with them. They can expand or even explode is overloaded as some can deliver flooded lead acid rates of discharge in 3 digit amps, 30C, 40C even 50C discharge are common.
If you want to talk battery abuse. My 450 RC Heli runs on a 2200mAh 3S pack. It draws about 200W sustained, 300W in bursts, 16-25 Amps for 7-8 minutes and the battery is dead. The battery will be slightly expanded, puffy and very warm to the touch on landing. It will be allowed to cool, balanced and charged at 2C 4.4amps to be flown again. All packs are balanced to storage voltage when I return home.
18650s are usually much lower output, they are also encased in a metal sealed container, so maybe you are right, they would be much safer, unless shorted.
The other thing that concerns me in the electronics hobby is that while 90% of people use them in 1SnP configurations, which is fine unless a cells fails and they are no fuses, the 10% of people needing 2S, 3S and 4S don't seem too bothered about proper balancing circuits. The number of youtube videos were people are trusting charging a multi-series pack to a single connection BMS and running several amps into it is quite high. Some believe they can occasionally check the cells manually and put them on a balancer once in a blue moon. This is NOT how it works. Series cells get out of balance fairly quickly, particularly at higher discharge and higher DODs.
The later is most likely due to the complete lack of available balancing BMS boards that actually work or have any guts/smarts about their balancing. 10R or 100R resistors and mosfets on a 10Ah+ pack is not going to cut it if it's in constant service. It just won't keep up and it will waste a whole load of energy trying. Some cells will end up undercharged before cut off, some overcharged during charging.
Most of better BMS multi-cell boards I have seen have tiny wee resistive balance circuits and offer barely usable and partially dodgy individual cell protection.
-
the 18.5 mA rating is per individual LED. So, 58.5 mA per RGB unit. By the way, if your's are anything like the loose ones I'm using, you'll need to cut back the green a LOT to get a decent white. The Blue is pretty weak. I'm using my camera to try to figure out the Black Body temperature but haven't gotten that far. So far, I'm trying 0xA0, 0x80, 0xFF RGB to get white and it's not there. Need to build up my test bed.
On your tiny85 board - build in an ISP header and you won't have that problem anymore. I have a pcb design around here somewhere. doesn't cost much to get made. I'll post it if I find it.
-
On your tiny85 board - build in an ISP header and you won't have that problem anymore. I have a pcb design around here somewhere. doesn't cost much to get made. I'll post it if I find it.
I was just connecting the Arduino UNO direct to the pins for ISP programming. The issue wasn't the ISP connections it was that I had the LED data line on one of the pins used for ISP programming and it would cause issues, including bricking an ATTiny corrupting the fuses.
In the next version I might try moving the LED data line that is not shared by the ISP programming connections.
-
So I got a new strip for bench testing. 12V 60 LEDs. (groups of 3 in series).
On max brightness, full white, 255, 255,255, all 60 pull about 0.6A according to my PSU. However if you read the datasheet it should be pulling (60*18.5mA*3) * 5 = 16.6 watts. 16.6 watts / 12V = 1.3 Amps.
So in test it's pulling half the max stated current. I don't know how to make them pull more. If it's pulling small bursts of higher current I'm not sure I care either as the PSU and capacitance/inductance will take care of that and it won't overheat stuff.
Actually currently the LiPo running them is down to 10.2V and they are pulling only about 0.4amps peak running the FastLED demo reel. Including the driver circuit.