LEDBrick Project - DIY pendant w/ pucks
- Thread starter theatrus
- Start date
CH 1 is 750ma, ch2 and 4 are bridged and my driver will only send a max of 600ma for some reason and I'm running ch3 at 700 ma
theatrus
100-mile-commuter
You're probably just at the edge of the driver's output. As current increases (and temperature), the voltage drop in the wires and the LEDs goes up slightly.
theatrus
100-mile-commuter
Should be capable. Haven't built one yet
. I'm mostly interested due to the very low price of the needed components, so figured I would take a stab at it for the sake of science. I can compare it to the LM3414 design as well. I'm expecting lower efficiency from the RichTek ensemble.Very interesting side project 
So u3 is just a little cheap PIC that can take 0-10V analogue inputs (via voltage divider) or I2C signal and output PWM direct to the u1 drivers, rather than connecting an external PWM directly? Will check the data sheet for that PIC later, but I assume it does have hardware PWM.
What's u2? I don't remember you saying and just had a quick scan and didn't see reference to it.
Brilliant little board - thanks for sharing
Tim
So u3 is just a little cheap PIC that can take 0-10V analogue inputs (via voltage divider) or I2C signal and output PWM direct to the u1 drivers, rather than connecting an external PWM directly? Will check the data sheet for that PIC later, but I assume it does have hardware PWM.
What's u2? I don't remember you saying and just had a quick scan and didn't see reference to it.
Brilliant little board - thanks for sharing
Tim
theatrus
100-mile-commuter
Very interesting side project
So u3 is just a little cheap PIC that can take 0-10V analogue inputs (via voltage divider) or I2C signal and output PWM direct to the u1 drivers, rather than connecting an external PWM directly? Will check the data sheet for that PIC later, but I assume it does have hardware PWM.
What's u2? I don't remember you saying and just had a quick scan and didn't see reference to it.
Brilliant little board - thanks for sharing
Tim
Yup! It has two full featured 10-bit capture compare channels and two 10-bit PWM only channels, at a max frequency of ~ 32kHz. Also a 10-bit ADC on nearly all the pins, though the performance of it is not specified (Preliminary datasheet only) - I'm not expecting much, but even if 6-7 bits were "usable" thats plenty.
U2 is a high-voltage-input linear regulator (45V), Infineon IFX20002, the 5V output part: http://www.digikey.com/product-deta...002MBV50HTSA1/IFX20002MBV50HTSA1CT-ND/4692356
The PIC running on its internal oscillator at max frequency (32MHz) sucks about 3mA, meaning about 60mW lost in the linear regulator for a 24V input to the board - warm but totally acceptable. This can be improved by dozing the CPU core of course (leaving the oscillator running for the timers+PWM) - I think it will support wakeup on I2C activity as well.
Last edited:
jedimasterben
LED world domination!
Sexy LED array. Have you considered using the Luxeon C? They keep the high flux like the Luxeon T and are in an even smaller package - not quite as small as the Z, but still super tiny, with a dedicated thermal path. The C is also available in the new 'mint', which is lime with a thinner phosphor layer, so it has more blue.
For comparison:
For comparison:
Last edited:
theatrus
100-mile-commuter
Sexy LED array. Have you considered using the Luxeon C? They keep the high flux like the Luxeon T and are in an even smaller package - not quite as small as the Z, but still super tiny, with a dedicated thermal path. The C is also available in the new 'mint', which is lime with a thinner phosphor layer, so it has more blue.
For comparison:
I'm keeping my eye on them for sure. The undomed vs domed on the angular intensity charts are interesting, as there is less variance than I would have expected between the C and Z there.
My main driver for trying to Zs was the availability of a violet in the footprint. Its super expensive, but its there. The Luxeon C has a 720nm for plant applications.
jedimasterben
LED world domination!
Yeah, for cost it definitely still works better to use SemiLEDs chips. I used to run MEGA violet over my tanks (as in capable of 100 PAR on the bottom of the tank from violet alone), but I have been experimenting with less and less and haven't gotten any different results as far as pigment production, but the tank is significantly visually brighter given the same PAR value. Then again I don't have any SPS, but in the designs I've given to others that use little violet they've also noticed no real change in pigmentation, but since I can't personally verify for them I can't say one way or the other
The new 90CRI 5700K is probably the most interesting LED in the C line to me. I'm a big fan of the 90CRI Bridgelux Vero, but I would love to have a single board/array that had everything on it - a good high CRI cooler white, a little super high CRI warm white, mint, low wavelength royal blue, low wavelength blue, a touch of cyan, and a couple violets. Mmmmmmm!
LOL - I think 10 ADC should be more than up to it, as you say
Shame only 2 PWM channels tho 
At 10b they should be nice and smooth tho Should have realised that - sorry
I think you had mentioned onboard 5V regulator...Yeah, using a linear just for the PIC shouldn't be much of an issue
Thanks for the info
Tim
Ah - thought it was a bit strange using a PIC with just 2 PWMs on a 4 channel board, but when you said it only had 2 10b PWM only channels I read that as meaning it only had 2 PWMs!
I'm loving the sound of this little board. Might have to get some made up for a play. Will see what the total cost would be on this side of the pond - 4 LDDs would run at about $29, without a PCB or anything to mount them on! I know these are not directly comparable, but they do essentially the same job
Tim
theatrus
100-mile-commuter
Its a muRata NCP18XH103J03RB, B is 3380 and nominal 10k ohms.
http://www.mouser.com/ds/2/281/r44e-522712.pdf has a 5 degree chart for ease of checking resistance to temp - I should make a spreadsheet for this.
Maryland Guppy
New member
I am going to place two of these on my prototype board.
Looks like a great way to measure an untested setup.
So far I have added more copper to the "hot areas" and some homemade vias.
Hopefully within a week or two I can etch a board.
theatrus
100-mile-commuter
Maryland Guppy
New member
Is this a better choice than soldering small brass brads?
theatrus
100-mile-commuter
LEDBrick Project - DIY pendant w/ pucks
I've seen it done, but either method should be more than sufficient. I'll see if I can dig up any details.
I've seen it done, but either method should be more than sufficient. I'll see if I can dig up any details.
