EEVblog 1617 - Architectural LED Lighting Build + Test
ฝัง
- เผยแพร่เมื่อ 13 มิ.ย. 2024
- Building and testing an architectural wall wash LED Lighting solution.
Part 2: • Wall Wash LED Strip Me...
Counting LED photons: • EEVblog #869 - Countin...
Forum: www.eevblog.com/forum/blog/ee...
Datasheets:
havit.com.au/products/hv9693-...
cdn.shopify.com/s/files/1/176...
cdn.shopify.com/s/files/1/176...
havit.com.au/products/hv9783-...
00:00 - Custom architectural LED wall wash lights
01:27 - Dual sided aluminimum profile
03:45 - 19W/m LED strip
06:08 - 32W/m LED strip
08:20 - Thermal adhesive backing
08:53 - EcoLamp LED Drivers
10:36 - Cable manage the design for replacement
12:30 - Testing
13:10 - LED strip analysis under the microscope
15:04 - Bench testing and verifiying the 19W/m spec
17:28 - Finding the SMD LED constant current driver chip
21:56 - Found it, the Crosschip DS-INK1101
24:47 - Bench testing and verifiying the 32W/m spec
26:52 - Bench testing the dimming
30:09 - Testing the driver adjustable dimming with external input
If you find my videos useful you may consider supporting the EEVblog on Patreon: / eevblog
Or with crypto:
BTC: 33BsprBQNBtHuVzVwDmqWkpDjYnCouwASM
ETH: 0x68114e40ff4dcdd384750500501e20acf3875f8c
BCH: 35n9KBPw9T7M3NGzpS3t4nUYEf9HbRmkm4
USDC: 0x68114e40ff4dcdd384750500501e20acf3875f8c
LTC: MJfK57ujxy55su4XicVGQc9wcEJf6mAoXF
Web Site: www.eevblog.com
Store: eevblog.store/
Other channels:
EEVblog2: / eevblog2
EEVdiscover: / eevdiscover
T-Shirts: teespring.com/stores/eevblog
#ElectronicsCreators #LED #Lighting - วิทยาศาสตร์และเทคโนโลยี
Nice! When you mentioned you were going down this route I was thinking you were just getting pre-made fixtures installed by someone, nice that you can just buy the extrusion and LED strips separately from local stores to get exactly what your application needs and DIY it. Nicely suited for future repair and/or upgrades.
Strange that the 32.6 W tape came out lower than claimed. Even if you use the 83 mm segment length from the datasheet (12 segments per meter), it still only comes out to ~29.5 W (~10% below spec). For it to meet the 32.6 W spec it would have to be driven at 113 mA. The INK1101 datasheet says the max drive current is 120 mA for 24 V applications with 7 LEDs in series which this appears to be (and 120 mA is typical max current for those 2538 LEDs), that would give 34.6 W, so the circuit could do it. Product picture in the datasheet also appears to have the same 6.2 ohm sense resistor, so it's not like they just sent you the wrong item, no idea where they got 32.6 W from.
Yeah, maybe some last minute change in production for some reason and no one notified marketing.
@@EEVblog It could be that probing just one segment doesn't give you the whole picture. There are always slight variations between segments and overall it can still average out to the actual spec.
Back in the late '90s I was asked to pull a part number for RAM chip of some type, and the designer gave me a preliminary datasheet that was dated about a month earlier. When I went to the mfr's website to get the current spec, it was flagged as LTB! I called to verify- yes, it was being replaced by a denser part with a slightly different pinout. The designer was not happy! That was the shortest product lifespan I ever came across, besides "vapor" parts.
I've done several of the DIY aluminum extrusion + LED tape lights. i've been using the high power density COB tape on my latest builds and skip the diffuser. To address voltage drop dimming at the far end of the tape, particularly on the really long ones , especially if 12V rather than 24V, I've taken to connecting the negative lead to the extrusion, then at the far end connecting the extrusion to the negative tape bus. This splits the tape feed to opposite ends so the voltage drop is the same anywhere along the tape and light output constant.
leading brainiac scientists are doing years of high tech pulsating brain research to increase the efficiency of leds by a few extra %.....
in order for manufacturers to cut everything down by more than 50% with thick diffusers overnight LOL
Dave! I think why you get 28 W/m out of the strip.
You measured 80 mm and 2,4 W of power consumption. That gives you 2,4 W / 0,08 m = 30 W/m
But if you measure the length not between the cutting lines, but between the lines with a pattern you get:
75 mm length (approximately) and 2,4 W. Then 2,4 W / 0,075 m = 32 W/m.
The manufacturer was a little tricky there haha.
In 1993 my father was in the hospital for a few days, and these LED fixtures you have remind me very much of the lights mounted above the patients' beds in the ward Dad was in. They were also fitted with lights that pointed up to the ceiling to provide room lighting, and then a light that pointed down towards the patent's bed. Of course those lights used fluorescent tubes, not LEDs.
Project time again! I missed this. That's an interesting profile and LED strip.
Took a while to find the right profile! Ideally we wanted a sloped upper side but couldn't find such a thing in combo with the lower one.
Light technician here: I hate wall washer lights. For a dinner room, you need 100lux to confirtably see other people faces and what you are eating. Most indoor LED lights today have an efficiency of 60 lumens per watt, which for the living room translates to ~1.7 watt per square meter. That's a blessing in terms of energy consumption. But they are to be installed perpendicular to the surface you want to illuminate. When lighting is indirect, the efficiency drops by ~5 lumens per watt, or less, which translates to ~20 watts per square meter for an illuminance of 100 lux.
I assume the additional energy goes into more evenly lighting the space and heating it? Doesn't sound so bad to me.
On second thought, I never lived with an AC so I forget the heat is not always good : )
Yeah, but the wife likes it, end of story. Will see what lux I get whne the install is complete.
@@EEVblog spousal acceptance factor is by far the most important factor.
Isn't 60lm/W a little bit low for recent models? The latest edision style replacement lamps I bought had ~130lm/W.
5.7W approx per meter just to heat the resistors? I've used 12v strips before with the resistors, and the efficiency is questionable. Due to the heat, you have to ensure they are stuck to a metal surface. If you try sticking them to plastic, they can soften the adhesive and fall off and the LED's overheat. A mistake I made when first using strips. For the 12v variety a big mistake is assuming they will be ok for a vehicle (boat) where the 12v can go up to 14.4v - they don't last very long if you do that.
That's what the huge aluminium profile is for.
@@EEVblog Maybe I should have phrased it as: every meter wastes enough energy to light a 60W equivalent LED bulb (I guess that's insignificant until 1000's of people all have the fancy lighting).
i question if that profile is enough to not cook everything
@@cannesahs The profile extrusion looks fine, if, and that's a big if, everything makes a good contact with it. Leds are quite efficient compared to traditional light bulb, but there is still probably be about 30-40W to dissipate in this 1.2m fixtures when full on. (6-7W in AC to DC conversion, 6W in lossy resistor, 20-25W of heath from the leds)
Aren’t these dimmers actually dimming by PWM? That would be typical for constant voltage LED drivers.
That was actually the same thing I know on the CV drivers it would be interesting to see some scope measurements on the output of the controllers in Dimming mode
can you take a dive into the efficiency differences of both types? comparing the lumen per watt. it would seem that the ones with the driver is much more efficient.
Hopefully the replacement for the 1101 lasts longer than the 1101s in my swimming pool lights!
That Aluminium housing seems to be a bit oversized, for the LED strips running at 32W or so, especially since the constant driver is built in a separate assembly.
For interrest sake, did you calculate of did they provide the Lifespan of these LEDs running at 19W, 32W 12hours/day? I assume the LEDs lifespan will dictate the warranty of the light.
whether its oversized or not depends on how often u like replacing LEDs
"Right up the clacker" you craic me up with your Ausie vernacular Dave
I think it would be relatively easy to make a pick and place machine for arbitrary length LED strip, you'd just need some tooling holes in the strip, and a couple cog wheels to advance it. Might even be able to make it in a continuous process, rather than being fed from wheels.
I like HavIt, they were the only ones selling high CRI downlights (that I could find). So far very happy $20 each IC4, aluminum, wide beam angle, etc. I'm curious about earthing on your project. As it's an aluminum body with mains inside (I know the power supply is earthed). Thoughts?
Nice test but I missed measuring the lumens. Perhaps the blue tape led strips emit the promised lumens at a lower power as specified and perform even better than the 3M type.
Coincidence! I'm DIY'ing a LED driver dim signal controller, these 0-10V / 1-10V dimmer inputs seem to be an industry standard.
Many thanks to the forum, they helped me come up with a dual OpAmp circuit to convert 3.3V PWM to 0-10V smooth DC. I had problems with the uC output not going full zero (0.3V when low), that was fed into an RC filter and amplified by a single OpAmp so my low output was 1V.
Luckily, I had ordered a dual rail-to-rail OpAmp, now I'm using the first one to convert 3.3V to 12V PWM as input for the RC, the 2nd one is used as a buffer, that goes near to absolute 0 (unloaded) output.
Now I'm hung up on implementing schedules on the ESP32-C3 micro. Never programmed anything serious in C, this would be so much easier in any memory-safe language...
I can already control the duty cycle via HTTPS and I'm thinking about just leaving the software as-is and implement scheduling on an always-on client. Ultimately, I want to spin a PCB that can plug onto the ESP DevKit boards (dirty cheap), 12V in, 0-12V out.
Check out Tasmota, ESPHome or WLED. Unless your Goal the learning you get from coding your own solution, these are much simpler to get a lighting device with a nice interface and standardized integrations.
Over spec-ing your Power Supply is also better for Efficiency.
(well, less heat longer life I guess) 😄
With LED lighting I go for high CRI like CRI>95 and take care that also R9 and R12 are given and have reasonable values. I accept CRI80 like I saw in your datasheet for storage rooms, cellar or staircase. The difference is easily visible especially when you light objects with saturated colours.
I'm surprised you didn't buy an extra to take apart. It dosen't appear to be UL listed. But, I may have missed the certification logo (which is sometimes counterfeit)
dave said its potted, so there is no need to take it apart to know precisely whats in there.
pro tip: its loaded with reject bin parts and toxic waste.
Why would a led strip with active drivers be able to handle morre current than a led strip with drop resistors?
How is flicker on these, especially dimmed. I’ve been finding nasty 60/50 Htx flicker on a lot of led stuff.
Appears to be zero. Of course the dropper resistor one has no switching.
I think probing just one segment doesn't give you the whole picture. There are always slight variations between segments and overall it can still average out to the actual spec.
Among the first humans. Wohooo!
Our species still has a chance.
All circuitry bits cool ...The big take away? The miracle of quality
Aluminium extrusion.
I'd love a diversion into HOW they force it through a die, and it comes out perfectly straight!! Know anyone with such a factory??
Yeah, rather amazing stuff, but I'm sure it's a deceptively simple process given the ubiquity of it.
One crowd funded Al prototyping extrusion project I backed described the process as after some length of extrusion comes out of the die a mechanism grabs it and pulls to straighten it
adjust the vfb resistor to get back the 32W!
Be careful with your eyes. I've worked alot with high lumen leds and I have really poor eyesight nowadays, don't know if it's from other sources though. But oh man alot of leds are soooo bright!
No one's used a shoe phone since that Agent Smart incident.
Not having > 90 CRI in this day and age suggests that these might be older type less efficient LEDs. 90 and even 95 CRI is starting to be standard for even generic consumer LED strips.
The IES measurement report is dated 2016, so yeah, pretty old.
I don't want to bloody know so stick that where the sun don't shine, So ner ner naa ner ner 😆......cheers.
Everything is voltage driven. You can't have current flow at 0V. Debate solved! ;)
That extrusion is pretty cool! I've never seen an LED strip like that higher powered one, though. Seems like far too many points of failure on an already failure-prone genre of product to me. I've dealt with a fair bit of architectural lighting, and all I can say is, the simpler, the better. LED strips, no matter the manufacturer, are self-cookers, and the only way to get any useful longevity out of them is to overspec the LED strip for the application, heatsink the snot out of them, and under-drive them. The 24W/m dual row 24V strips I like using need a 2" wide by 1/8" thick piece of aluminum with standoffs for adequate cooling, and I run them at 20V-21V. Even with heatsinks, I've had black-dot deaths and resister burnout at the "rated" 24V. LEDs are such tempermental snowflakes...
They make circa 50W/m strips.
For the longest time I had trouble wrapping my head around constant current power supplies. My question was "how do you drive something with a specific current at a fixed voltage?". And the answer is, you don't! You're still controlling the voltage, you're just doing it by monitoring the current. So you either decrease or increase the voltage to hit a current target. (well the controller is) Once I understood this, it's like a light went on and lot of electronics concepts made more sense to me.
Yeah I’m not a small time engineer, but I stay away from the Diodes brand because they do seem to have a long device life
wait on wago terminal hahah what a heresy haha
Surprisingly inefficient, 7.2V 20mA over that resistor means 0.14W out of 4.8W is dropped as heat or only 70% efficiency. I sometimes wonder if constant current drivers are worth the effort, in this case they definitely would be.
So dim that your hard disk jumped.
Don't go hacking off the dimming input wires! From your test, it looks like the analog control voltage method gives you a nice flicker free output with a wide dynamic range. Have you done the same tests using a triac dimmer on the AC input? From my experience, even drivers that are rated to handle triac dimming still struggle with visual flicker, poor performance at very low levels and even audible noise. Embrace modern lighting wankery.
Nope, I don't want any of that rubbish.
@@EEVblog Isn't the TRIAC control the rubbish solution? Wouldn't the driver have to basically use the triac produced duty cycle as a data signal to "estimate" what the requested light level is and using this to set the output current... and all this while the triac is modulating the input making things like effective PFC impossible.
I always though of dimming LED drivers with triacs as a kind of Rube Goldberg machine.
There must be a better way to transmit the desired light level over the power wires - without messing with the actual power feed.
@@TylerDurden-pk5km Triac is definitly the rubbish dimming option here.
@@gendronw It would also question whether linear regulators in the strips - instead of current control through the driver is a good solution for dimming - the driver can not know much voltage is "burned" by the linear regulators - so that there is possibly no perceivable brightness change for the upper 20% of the regulation range.
@ 1:40 You point out the heatsink fins on the upper section of the tube, and then @ 2:48 you point at the SAME set of heatsink fins on the now "bottom section", when they are in fact the same set you pointed at when they were previously mentioned.
To be clear, there are NOT heatsink fins for both rows of leds! 🤣🤣
The Australian economy must really have hit a rough spot: dumpster room empty and EEs have to go into low end manufacturing ... 😀
Trust but verify! Haha alrighty Mr. President Reagan 🤣
Eww only CRI >80, not even 90
Yeah, I wouldn't wanna go below 90 CRI either these days. Unfortunately high CRI lamps are painful to source. Most shops have thousands of lamps and no search filter for CRI value.
@@t0scanelli most people prefer a bulb that is better than a blown one.
there is a stage where they will pay for a glow worm. and u want a CRI search filter 🤣