Re: 420 Consumer Reports Competition - GrowLEDHydro 300w LED vs. 400w HID
I love watching things grow. GH nutes have always been good to me, also...
(
Warning: This turned out to be a longer than intended,
The Matrix, 'Down the Rabbit Hole' post. If you want to see how far down it goes, take the
Red pill and keep reading. Otherwise, take the
Blue pill and skip past it. Don't say I didn't warn you -
and don't shoot the messenger. Thanks!)
Haven't added to my
Swipe File in awhile; maybe it's about time. Anyway:
Ok I just got back from checking the wattage on my 300w GLH led unit.....
Amps 2.92
Volts 113.1
Watts 361 to 365
Power Factor 98
I also counted the rows of led's... My unit has 22 rows of led's wide and 12 rows of led's deep. My unit could hold 24 rows of led's wide. I'm not sure why their not in there..
Thanks dn
, I appreciate you checking. I'd love to know Power Factor (PF) for both units as well, but as mentioned above, different meters use different calcs to get their results. Unless you've got a purely resistive load (like an incandescent bulb), you're going to get imperfect estimates most of the time. That PF seems a little high to me, I'd expect a little more inefficiency, assuming we could get a truly accurate measuring device...
That said, I looked at the product pics earlier. (Both have 22 rows x 12 deep, or
264 LEDs total, as you two mentioned. Thanks for clarifying!). It looks like they took 5
IRs out of the old units, and added 2
reds and 3
blues in their place. At 500ma that should add ~3.5-4w to your total.
If I can trust those voltages, and the
actual draw current used by both was similar, then the change in voltage may account for the rest:
Watts = Amps x Volts
If amp use is constant, then as your volts increase, so does your power (watt) usage. Use a ratio to figure out the new watts used by the same system at a higher voltage:
Take dn's ~363w (avg) x (118/113.1) = 378w, + 4w (new LEDs), and you've got about SS's panel draw. Or thereabouts. Maybe an EE can give you a better explanation (anyone?), I'm probably forgetting something...
(The body is the second thing that goes. I forget what the first thing is--!)
---------------
Many power ratings for products are standardized at a particular voltage, and 110V is common. The power in
your area will generally be something different than the rating on the box, however.
We've got
dirty power in many areas of the country - with transient spikes, dropouts,
brownouts, under/over voltage, resonance, line noise - you name it. Rarely does what's coming into your house stay constant.
Many fluctuate within the 110-125v range more than you'd think, and outside of that even. One of these days, I'm definitely going
Solar, folks...
That's why computers and other sensitive electrical equipment usually have some sort of
Line Conditioner built into it - so it doesn't get fried by dirty power from the utility grid.
I can't vouch for the quality, though. Many of the units themselves or components that go into such equipment suffer from crap quality and poor design, esp. in the mass market. Every manufacturer seems like they just want to cut costs. I doubt the LED panels have much better in them. Many constant current (LED) circuits with a step-down transformer won't even have a decent
Filter Cap in the circuit in case of the load being momentarily disconnected.
(All transformers are also inductors and can store large amounts of energy. If you get a momentary power loss and then re-voltage situation, that could trick your system into wanting to continue providing current (that can't go anywhere), so your voltage continues to rise; then the current gets re-connected - and
POOF! -
kiss your $1000 LED unit goodbye.)
These things are basically
a bunch of computer chips stuck on a board that produce light instead of processing 0's and 1's. Take a $1000 laptop, stick it in a small box, and then subject it to tons of heat and humidity.
How long do you think it'll last like that? Exactly.
Cooling is paramount.
If you have one in the house,
treat it the same way as a laptop. Get a decent
line conditioner, (+surge protection and UPS), make sure it's got great airflow, and keep the humidity under ~55% or so. And for God's sake don't drop them!
---------------------
Differences that can result from manufacturing tolerances or other factors aside, a few points that the watt readings above reiterate to me are:
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1) Digital Ballasts on HIDs are
much more efficient than magnetic ones. (I normally see magnetic ballasts on a 400W pulling about ~460-480w on the line side.) Sure, we knew that.
2) Electric appliances almost
never draw what they're rated. Ok, so we knew that too. Let's see now:
HID w/Lumatek: About
7% higher (428w/400w) actual power usage than rated. Not bad.
GLH Spectra: Almost
27.3% higher (382w/300w) actual power usage than rated. Hmmm.
HID w/Magnetic: Forget about it--! (28.5%)
3) Overclocking your LED chips and running the board hot (over it's rated amperage - in this case, about 43%) mean you're
going to pay for a lot more juice than a typical flowering-spectrum board running at 350ma. No freebies there.
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
How much more $$$ ? Well...
This same board running at normal levels will pull about
250-260w with the chips it's actually designed for. Call it a
120w difference.
What's that save me $$$-wise with either a flowering (12h) or vegetative (18h) light cycle over time? Call it
$0.2 / kWh from PG&E.
Peak rates are - what? 46 cents now? More?
EDIT: Schedule E-6 'Time-Of-Use' is now 58 cents per kWh for Tier 4 and up! Holy crap!
Let's find out:
^^^^^^
(Column A is
Kilowatts, obviously.)
Hey, I could probably
buy another whole LED light for that, couldn't I? Maybe increase yield even more? Hmmm......
--------------
Here's the thing, folks (and folk-ettes!).
Regardless of what the chips are actually rated at, those
1w boards everyone uses are not designed to dissipate all that extra heat from running it at 500ma instead of 350ma (43% more!). Personally, I wouldn't fill a Metal Core Printed Circuit Board board using your typical 1W emitters more than ~70% full of LEDs-to-spaces (even at a nominal 350ma) anyway, and I'd try to keep them under 60% filled - and
preferably, closer to 50%. The heat sinks and cooling used are
barely adequate for the board when they're filled at their stated capacity, like most of the boards you see on the market now.
Which means my
safety factor is about a
200w difference from this unit, using this same (288 unit) panel. I'd rather buy two 140-150w units and get better, more consistent coverage, instead. And higher g/w overall. Yes, that'll cost a
little more - but not that much. And definitely not in the long run.
The larger heat sinks required simply won't
fit into the existing panel, you need to get a larger casing for that. Higher CFM fans (if you can get them) will only do so much. Past a certain point, they won't whisk heat away fast enough from the back of the LED. Junction temperature keeps rising...
Any more than keeping these ~70-75% filled (...let alone overdriving a panel with almost all the spaces filled--!
), and you will have to do at least four things:
A) De-rate the lumens produced per watt by an increasingly significant factor, TWICE - for increasing
a) CURRENT,
and b) HEAT, which go hand in hand. This is worse for red LEDs than blues and whites. Heat and Current (ma) = inefficient light production.
B) Shorten the lifespan of each LED and increase it's catastrophic failure rate. Heat destroys LEDs. That's why all those drop-in replacement LED bulbs from Lights of America (i.e. 'Lights of China') die after two weeks in a recessed lighting fixture, same as CFLs.
C) Reduce the residual lumens at the end of its lifespan, provided it's still working. Forget 70% remaining after 7 years, you'll be lucky to get 50%, driving a panel like this. Put this over mothers you no longer like!
D) You will get Color Shifting as well. Higher heat means the
chromaticity (emitted light frequency) changes. This is worse for red LEDs than blues and whites, also. So much for targeting specific PAR wavelengths and driving efficient photosynthesis. Too bad flowering lights use a majority of red...
There are several things I don't like about the
HGL units, also, but not filling the every space on the board with LEDs wasn't one of them. (Wasn't too happy about those 'light engine' clusters, though...) Leaving empty (dark) slots on the board was one thing they
did get right.
At least there's a 5 year warranty on these units. Buyers will probably need them.
--------------
Now, that's not to say they won't produce light (they will, and you'll get more from increasing the current like this - just not as efficiently, or long-lasting), or get results. Some folks are certainly already doing so. And if you're ok with all that, then
great!
If it works for you,
do it. I respect those who are early adopters of any new technology or idea.
And, having been in the product business, I'd be
really interested to see what the defect and failure rates are 4-5 years down the road. Maybe I'm overstating the case a bit - but I don't think so.
It's just that
people need to know what they're buying, and go into any purchase with eyes open. Which is what being an informed consumer is all about - and
the reason we're doing these tests. The proof is always in the pudding.
The better DIY guys at least have the sense to use huge heat sinks, and space the LEDs out more over a larger area to dissipate heat better if they're going to overclock them.
There's still a lot of crap out in the market now...manufacturers using low-bin LEDs in their units that are
30-60%+ less efficient even within the same product family--! ...than a good, high-bin, quality LED:
Wow, a
150% difference in lumen output between the lowest and highest bins for a white LED.
Wonder which ones most panels are using?
I'll take 30W in high-bin LEDs over 50W of your typical panel LEDs any day. Nights, too.
---------------------
As
Soniq420 mentioned earlier, Excel is the great equalizer. We'll just have to make sure to do the g/watt-hr yield numbers both on
rated as well as
actual power use.
Big difference - and results should be based on the total cost
you, the user, are actually paying to grow, right? PG&E rates aren't exactly cheap.
That said, I still
definitely want to see this test as much as anyone, to see what a LED panel with the
dials turned up to 11 can do against a worthy, known competitor.
That's why these are so much
fun--!
Cheers,
-TL
my brain is fried.. lol
for the info, +reps
