Tips for DIY led lamp

joselo

joselo

Member
Greetings.
I have read several threads of the forum, about the leds, but the information is very dispersed and in some cases is not conclusive.

I have 3w leds,
Led Grow full spectrum 3.0-3.4V (10) (optional)

blue 450nm DC 3.0-3.4V (10)
Blue 470nm DC 3.2-3.8V (16)
Green 520nm DC 3.2-3.8V (3)
Orange 610nm DC 2.0-2.2V (3)
red 625nm DC 2.2-2.6V (16)
far red 730nm DC 2.2-2.6V (10)
UV 420nm DC 3.8-4.6V (2)

I am currently solving the problem of regulated voltage, but what I really need to know is the following:

Pulse, PWM, and appealing to the experience of some of you, percentage of deep blue/blue red/far red.

I need to know, the ideal frecuency for the leds, a.k.a "pulsating light" (if is different for red/far red, or deep blue/blue). And pwm if applicable. The pwm is easy for me, but, the pulsating light, and pwm worth it?

I have microcontrollers, mosfets for this goals and technical capacity to do so... I would like to share experiences.

I want something like this (I am only interested in the type of light, not the rest of the functions):
This lamp, for example, alternates between red and blue. IF you set 10 seconds, in growth, will be 7 seconds of blue and 3 of red, in the flowering is the reverse, 3 seconds blue and 7 seconds red. You can set the percentage of red / blue.

Suggestions?
 
iHearAll

iHearAll

Well-Known Member
Pwm is not to pulse the light it is to pulse a mosfet and gets resisted by an RF choke inductor with a smoothing capacitor bank. You privabky knew that but imjust making sure were on the same page. Typically thesee pwmm integrated circuits function on feedback to control current or voltage. Check out the tl494cn datasheet to see equations of designing a buck converter from the component level.

If you really wanted to determine the average power you would integrate 1/T x integral from 0-T of (Vp(t)IP(t) )dt you add your separate integrals together over the course of period decay if that makes any sense. But Id say just pic an IC and follow the datasheet. Frequency only really matters to determine your choke inductor and ripple smoothing caps.

To get a separation of spectrum I would make two buck converters and have a single square wave driver that has a variable duty cycle timed to a 10 second frequency. I would imagine you will need a few hundred farads as your timing capacitor to alternate spectrum's based on signal polarity but super slowly. if you have an N channel gate controlling one spectrum and a P channel gate controlling the latter you flip flop spectrum with the duty cycle. Probably just an lm4558 would be efficient enough for your square wave or your microcontroller.

Make sense?
 
joselo

joselo

Member
iHearAll said:
Pwm is not to pulse the light it is to pulse a mosfet and gets resisted by an RF choke inductor with a smoothing capacitor bank. You privabky knew that but imjust making sure were on the same page. Typically thesee pwmm integrated circuits function on feedback to control current or voltage. Check out the tl494cn datasheet to see equations of designing a buck converter from the component level.

If you really wanted to determine the average power you would integrate 1/T x integral from 0-T of (Vp(t)IP(t) )dt you add your separate integrals together over the course of period decay if that makes any sense. But Id say just pic an IC and follow the datasheet. Frequency only really matters to determine your choke inductor and ripple smoothing caps.

To get a separation of spectrum I would make two buck converters and have a single square wave driver that has a variable duty cycle timed to a 10 second frequency. I would imagine you will need a few hundred farads as your timing capacitor to alternate spectrum's based on signal polarity but super slowly. if you have an N channel gate controlling one spectrum and a P channel gate controlling the latter you flip flop spectrum with the duty cycle. Probably just an lm4558 would be efficient enough for your square wave or your microcontroller.

Make sense?
Click to expand...
Yes, nice, had not thought of that. But my other problem is the type of light and the amount of each.
 
iHearAll

iHearAll

Well-Known Member
joselo said:
Yes, nice, had not thought of that. But my other problem is the type of light and the amount of each.
Click to expand...
That's preference. I would suggest white spectrum's over blue and red. I tried growing with blue red for a year and although it does work and fairly well at that, you best bang for your buck is lm561c diode boards or city clu048-1212 or vero 29. The lm561c have a better heat profile than the others but you could just got more 1212 or 29s and power at lower voltage. Did I read correctly you already own the red blue? You could have blues and UV together as one series circuit and the red orange and green as another series circit. Then have a potentiometer to adjust the duty cycle to change seasonal spectrum ratio on the fly. You could easily have an arduino read a trim pot to control dc.
 
Last edited:
iHearAll

iHearAll

Well-Known Member
But what to do with the full spectrum's? Maybe divide five to one channel and five to he other.

Also heat won't be an issue for you come to think of it. Plenty of cool down time it appears
 
joselo

joselo

Member
iHearAll said:
But what to do with the full spectrum's? Maybe divide five to one channel and five to he other.

Also heat won't be an issue for you come to think of it. Plenty of cool down time it appears
Click to expand...
Yes I have some ~4000k leds, and 10 grow leds.

Now, the blue light, in theory, Makes the plants grow more leafy and with little inter-nodal distance. The red light, makes the plant spike more, the inter nodal distance is greater. This is simple, by observation, we will know which spectrum of light we need more...

You can simulate sunrises and sunsets with far red(730).

Yes, I have leds, Deep blue/blue red/far red, UV, orange and green. Orange complements a little, and green light has effects on carotenoids.

Here my design. The heat is not problem. There are the blues and reds, uv, orange, and greens. Suppose, we will work blue, red and the DC is 7/3, the white leds whom do they accompany? :wall:... I have decided to work without glasses, only with "diffused light".

Captura de pantalla de 2017-07-15 18-08-39.png​
 
iHearAll

iHearAll

Well-Known Member
joselo said:
Yes I have some ~4000k leds, and 10 grow leds.

Now, the blue light, in theory, Makes the plants grow more leafy and with little inter-nodal distance. The red light, makes the plant spike more, the inter nodal distance is greater. This is simple, by observation, we will know which spectrum of light we need more...

You can simulate sunrises and sunsets with far red(730).

Yes, I have leds, Deep blue/blue red/far red, UV, orange and green. Orange complements a little, and green light has effects on carotenoids.

Here my design. The heat is not problem. There are the blues and reds, uv, orange, and greens. Suppose, we will work blue, red and the DC is 7/3, the white leds whom do they accompany? :wall:... I have decided to work without glasses, only with "diffused light".

Click to expand...
I would think the white and full sprectrun is constantly on during daylight hours or half are in the 7 and half are in the 3. ..
 
You must log in or register to reply here.
Top