Yes, it should work nicely. I have three diy LED strips using Steve's square tubing and fan. They work great and are quiet. I have two double tube at 30" length with 15 XMLs (so one 30" section handles 8 XMLs) driven at 2100 mA...so about 6.6W ea...and it's just warm to the touch. Personally, I wonder why more people don't try it. It's relatively inexpensive and the fan is quality so it pushes a lot of air quietly.
Best LEDs today? Share info!
- Thread starter JMD
- Start date
captainmorgan
Well-Known Member
Thanks jc,do you know where I could find some quality 680-690 nm LEDs?
El Tiberon
Active Member
Anyone know about these? Someone jumped in on a thread of mine and dropped this link
http://plantphotonics.com/
http://plantphotonics.com/
JMD
Well-Known Member
If it boosts the photosynthesis, then why wouldn't it do it all day?
I have fried couple of motherboards and some other computer components when a cheap psu has failed. That, my very incomplete electronics knowledge and a general distrust towards cheap chinese electronics kind of drives my fear.I think it was around 40 RMB for 2 heatsinks, 2 large drivers and a smaller driver - but that was to China. Got someone to bring it to me
I will run a test of the drivers when I get them (should be at the end of the month). It's possible for everything to fail in a way that not only destroys your LEDs, but also burns your house down. But is it likely to happen? Luckily no!
All electrical can fail in a very destructive way, but it is highly unlikely that it ever will.
Your LEDs are quite safe, no worries.
I was hoping that you have an idea how the circuits are built and if there are some components that when failing, would produce a mode of failure I described earlier. I know that cheap capacitors going bad and the way those cheap computer PSU's are designed are the usual culprits for them taking other components with them.Have you thought about designing your own drivers? I looked at the flashlight scene, and saw some very efficient designs. Seems very logical to strive for best efficacy, why settle for 80% efficient drivers when using top of the line efficient leds? Of course that is out of my scope, seemed very complicated for an electronics idiot like myself but the idea of purpose built drivers was very interesting.
PSUAGRO, or someone else who owns a Hans-panel; could you tell us if the material on the backside of the panel is aluminium (looked like that in El Tiberons video) and if there are holes under the fan for cooling?
JMD
Well-Known Member
Yea, I know how they are buildI have fried couple of motherboards and some other computer components when a cheap psu has failed. That, my very incomplete electronics knowledge and a general distrust towards cheap chinese electronics kind of drives my fear.
There are a few main approaches on how to build the driver, some better than others. I won't expect a cheap driver to fail in a way that can damage your LEDs. Even if the current sense resistor fails, the circuit is most likely not able to deliver high enough current to damage your LEDs. This is based on that the duty cycle has a minimum or maximum, at which the highest current is achieved - the inductor is chosen to meet this requirement, which sets the limit.Yes, I have been thinking of designing a low cost constant current driver. But it takes time
Which I don't have much of currently.
PSUAGRO.
Well-Known Member
Yes backside is aluminium and front is plastic(?) coated with a reflective paint to match........no holes under fan other than the one for the fan power wire. the fan is mounted on 10mm(?) spacers and pushes the air down & over the the back panel (this is on the original panel, IDK about the other two revisions)I have fried couple of motherboards and some other computer components when a cheap psu has failed. That, my very incomplete electronics knowledge and a general distrust towards cheap chinese electronics kind of drives my fear.
Have you thought about designing your own drivers? I looked at the flashlight scene, and saw some very efficient designs. Seems very logical to strive for best efficacy, why settle for 80% efficient drivers when using top of the line efficient leds? Of course that is out of my scope, seemed very complicated for an electronics idiot like myself but the idea of purpose built drivers was very interesting.
PSUAGRO, or someone else who owns a Hans-panel; could you tell us if the material on the backside of the panel is aluminium (looked like that in El Tiberons video) and if there are holes under the fan for cooling?
edit: trying to get the bin #'s from hans on the new flower panel for you......
Big thanks JMD! Those cheap chinese drivers would save me quite a bit, so I'll trust your opinion about them for the first version at least. Another good point is that its easier to find single drivers with higher voltages than from meanwell for example. Could you explain effects of ripple in our use?
Thanks PSU! I had a very different idea about the structure, good you have corrected it. I think you could slap couple of nice heatsinks to the backside, use thermal glue so you don't even have to break the panel apart. Upgrade the fan to a 120mm silent one, space it further from the panel and voila, you should have a nice, cool running silent panel? Though you need to pay attention to power consumption of new fan, if its much bigger you should use an extra power source (if the schematic really is like guod described earlier.)
JMD
Well-Known Member
From an electrical point of view (I don't know how plants respond to flicker), there isn't really issue with ripple.
Lets assume that your LEDs needs to be driven at a constant current of 1000 mA. If your driver is rated at 1000 mA +/- 5%, it means that your LEDs will get 950-1050 mA. Is it a problem for LEDs rated at 1000 mA? Not at all. Many drivers run at 100+ kHz, which means that the human eye will never notice the flicker.
Essentially the drivers are build in this way:
AC --> DC --> Buck/Boost with the duty cycle varied according to the current measured with the current sense resistor (low voltage drop across).
Some (usually the better ones, like Phillips etc.) use IC's capable of converting straight from AC to constant current DC with very low loss.
stardustsailor
Well-Known Member
Hmm...
I really do not know to what you mean when you say "ripple"...
Usually as "ripple" in the electronics world is meant the SUM of inductive noise ,
(because of bad grounding path => 'ground loops' or excessive wiring "twirling" around)
which is 'multiplied' in the output of a switching power supply (can be in whatever mode : buck,boost,buck-boost aka inverse/negative boost ,etc...) .
Thus making a "dirty " output ...
The whole ripple thing starts from the pwm regulation of the switching IC ...
A small ripple filter ( coil-capacitor ,LC ) before the output (20μΗ & 100nF ceramic/MKT/tantalum) will decrease at least x10 the ripple currents.
If that's the ripple you're talking about ....
A very easy way to make a CC driver ,delivering up to 3A and 37 Volts output (57 V for the HV version ) is using the IC : LM 2576/ LM2576 HV Simple Switcher Series from
TI ( ~4$ each IC ) .Very few other components are needed ( a storage coil ,a " free wheel "diode ,a pot ,few caps ...)..
To make it dimmable or even pulsed at higher voltages/currents ,you'll need an Arduino and a transistor or mosfet .(i.e BD 711...Switching on-off up to 12 Amperes,with minimal heatsink requirements ... Up to 5V at base ..Directly from Arduino's output..Even without resistor in between ...Can't get simpler ....)
Up to 1 Mhz frequency ...
A pot can set the Duty cycle from 0% up to 100% ....
EDIT:
Or the simpler way ...
Two ( of the CMOS type ,preferably ) LM555 timers (or a single "dual timer " LM556 IC ...),and
timing caps,resistors,and pots .(Timer A at astable operation-setting the pwm frequency - ,timer B at monostable operation ,
triggered by timer A -setting the pulse width aka duty cycle .And a power transistor/mosfet playing the role of the "switch" ...
.Search "pwm circuits using LM 555 " for more info..
Seems complicated ,but ain't ....
......
.....
And a lot of great things start to happen ,then....
A total different world of driving the leds ....
A truly amazing world with lot's of things to discover ....
PS individual systems (both PS I & PS II ) ,at average ,photosynthesize approx at 4 Hz per system ...
But ,who says that the systems are "sychronized",in their total sum ...?
So ,probably ,plants photosynthesize all the time ,continuously ....
No frequency there...
But "pulsed" leds ? ....
Ah...A great story,indeed...
JMD
Well-Known Member
Ideally we have a pure nice DC supply with a constant current, but because this is a switched power supply (and not just a voltage drop across a resistor) we have something a bit different.
Average over just a few periods (which at 100+ kHz is a very short time) the current is as you want it to be. But zooming in on each switching, you will see that the transistors/MOSFETs are not ideal, and that it takes some time for them to turn on and off.
So when the IC measure a voltage across the current sense resistor that tells it "now 1000 mA is flowing", it will start to turn off the transistor/MOSFET - but it's not instant, so the current continues to rise.
How well the driver is able to maintain the steady state, is the +/- 5%.
There's some flaws, sorry to point them out.
- If I'm not mistaken, the LM2576 is a voltage source, not a current source.
- The LM2576 is a buck converter, meaning that you need a larger input voltage than you want out.
- You need a power supply for that IC.
The IC might be cheap, but once you're done, you've spend much more than $4 and you would have been better off just buying a driver.
What you are talking about, is just a simple PWM circuit - it doesn't really have any relation to driving LEDs at a constant current.
If you want to dim LEDs, then you will need a PWM circuit. 50% duty cycle directly relates to 50% light output.
Great, I didn't know. I just remember someone talking about plants being stressed if they only used CFL, due to the 100 Hz flicker (50 Hz system).
stardustsailor
Well-Known Member
The same thing with other words ..
Yeap,nothing more to add here ...
a)don't know if that helps a bit ...
...It works great though .....
b) yes....
c)yes...
A central psu can supply multiple ICs ...
If you go the way you're suggesting ( AC=> DC Constant Current ) ,be prepared to deal directly with the mains supply "qualities" ....
( EMI ,RFI filtering ,AC harmonics by-passing ,surge protecting,ripple,etc......)
All these sub-circuits for each individual of such drivers ,will raise definately the final cost ....
So,cost-wise,better a main PSU to deal with the mains AC and "offer" a DC voltage to sub-DC/DC CC drivers ..
...
50% duty cycle directly relates to 50% light output.
???
Of which output ?
Driven at suggested voltage/current ,of course ......
What about 35% duty ,at 500 KHz ( 0,7 microsecond pulse width- 1.3 microsecond off ) and 2500-3500 A ?
(For a led rated 800-1000mA max continuous operation ) ....
You know what ?
Except from increased radiometric output power ....
Cooling ' needs' -of leds- decrease a lot ....
Yeap ...Might be a clever idea also, to avoid 'harmonics ' of 4 Hz ,when dimming or pulse overdriving the leds ....
Who knows,really ?
JMD
Well-Known Member
Looks interesting, I will take a look. Have you calculated the BOM cost?
Driving a LED at a i.e. 1000 mA, switching it at 50% duty cycle, will yield 50% light output. Doing the same with a halogen bulb, won't. Just what I was trying to say
Considering that the LEDs become less efficient at higher currents, it doesn't make sense to pulse them with higher current at that high frequency.
Imagine dumping 3W (I use watt, because double current doesn't mean double power) into a LED at a constant current, and then dumping 6W but for half the time at a high frequency - you will get less light out at 6W than you did at 3W.
Pulsing LEDs with high currents is great for emergency lights like the ones on ambulances etc., where you need high output from a single source, but with a relatively low duty cycle.
When I think of it, I can't see why flicker would bother plants.. In nature there are clouds moving across the sky all the time - that causes much higher fluctuations than just a few %.Yeap ...Might be a clever idea also, to avoid 'harmonics ' of 4 Hz ,when dimming or pulse overdriving the leds ....
Who knows,really ?
captainmorgan
Well-Known Member
Very interesting driver talk guys! That overcurrent pulsed driving sounds extremely interesting, have you stardustsailor done any experiments if the leds last abuse like that and how long? I don't know if the current itself actually causes lowering efficacy, I think it is the temperature rise that higher current leads to. While pulsing, there would be more time for cooling, but would it produce any gains? Don't know, though I think that there would be some drivers like that already, if there would be significant gains to be made without sacrificing reliability or other important properties.
PetFlora
Well-Known Member
Interesting topic. Begs the question best at a given price or regardless of price?
Being pragmatic I felt good leds would offer a hell of a lot for those of us who cannot afford the most expensive/best
My results using 20w non-horticulture specific LED tubes speaks volumes. Unless you have money to burn for that extra 20% I am very happy using conventional led tubes and globes. Most important is what spectrums to use and when
Being pragmatic I felt good leds would offer a hell of a lot for those of us who cannot afford the most expensive/best
My results using 20w non-horticulture specific LED tubes speaks volumes. Unless you have money to burn for that extra 20% I am very happy using conventional led tubes and globes. Most important is what spectrums to use and when
I think performance/price ratio of DIY will beat the household bulbs and tubes, though it takes a lot of work to source parts and compare components. Of course a lot of labour too, no good to have a pile of components in your drawer, those won't grow any plants as opposed to a plug and play solution.
PetFlora
Well-Known Member
I think performance/price ratio of DIY will beat the household bulbs and tubes, though it takes a lot of work to source parts and compare components. Of course a lot of labour too, no good to have a pile of components in your drawer, those won't grow any plants as opposed to a plug and play solution.
Correct sir! People like me, don't have the skillz to trouble shoot a DIY fixture. I have read too many blogs where the builder ran into trouble and did not have the appropriate equipment to trouble shoot.
Thankfully I do have the skillz to retrofit a shop light, and enough knowledge about which spectrums are essential for a decent grow. A 6 bulb fixture with ~ 20w tubes should easily produce 2 dried ozs for many grows to come, with minimal degradation. Cost ~ $275 to cover a 2-3 x 4 area
~110w in action over 2 fat clones that are days from harvest
