Off the shelf retail store screw-in LED and CFL bulb comparisons

[indagroove]

I just bought another 4-pack of 10 Watt Noma SILs today and all the LED segments are in series and run off ~120 V DC.

If there is an Osram SIL (or other commonly available brands) that are designed the same way but with a driver separate from the LEDs, then it would be easy to replace the driver with something like I posted and run a whole wack of SILs with one external driver. (probably about 20 for a 200 Watt driver with 10 Watt SILs)

Plus, then you can crank your SILs.
You just have to keep an eye on the temperature and put them on a heat sink.

Your total lumens go way up but your lm/watt goes down.

The total lumens output and lm/watt are kinda wishy washy numbers.

My LED strips are rated 185 lm/watt but they measure it at 0.35 Amps to get the best numbers.
They can run at 1.4 Amps if you don't burn them out.

SILs are the same. They want lots of lumens output but they also want the bulbs to last 10,000-20,000 hours.

In my experience Osram makes great lamps in general, although I haven't seen a lot of their LEDs around. I was at my local hardware store the other day, and I couldn't find any over the counter SILs that were above 100 lm/w. I did see some LED screw in floods which I remember being higher lumens , but I didn't really consider lm/w at that time.

 

[indagroove]

These are a bit expensive for SIL's, but they do look kind of badass. You get 4000 lumens out of 23watts of LED: https://www.sansiled.com/27w-led-bulb-3000k-2-pack.html

EDIT: wow, I didn't realize that they have a whole line of SIL grow specific bulbs and even a cheap 70w panel: https://www.sansiled.com/led-grow-light.html

 

[PCBuds]

Those would be good if you want a "diffused" type of lighting. They spread the light around by their design.

A normal SIL uses the globe to diffuse the light and we concentrate it by removing the globe.

 

[Hookahhead]

I opened the 9.5w Osram 6500k with the chipped diode.

It was pretty easy to disassemble. There is a metal heat sink in the base. Check it out, no wires on this design. Looks much more robust than the generic bulb.

I wired it up to the 9w LED plate from last night. The LEDs flicker and there is an audible clicking noise associated with the flicker. A bit of a disappointment, I was hoping to Frankenstein this thing a bit.

 

[PCBuds]

Those bulbs look to be perfect for a separate offboard LED driver.

If the bulbs are cheap, you could buy a bunch of them and remove the LED "plates", connect up to 20 of them to heat sinks (like maybe pop cans. Lol) then wire them all to a single Meanwell driver.

You've got 30 LED segments using 9.5 Watts and I've got 15 LED segments using 10 Watts, so each segment on your board is putting out about 1/2 the lumens and Watts as my bulb.

That means you've got more room to Crank it up.

The only issue would be the heat that can overheat and burn out the LEDs.

I'm almost positive you would only burn out the LEDs and not burn up the LEDs and start a fire.

Using a separate LED driver gives you control of the output of the driver and the output of your SILs.

If you don't mind reducing the lifespan of your bulbs and your grow cab can deal with the extra heat from being cranked, then I say go supercharged!!



Start with measuring the output of the driver.

I'm guessing it's ~ 120 V DC. (probably a bit less to give it headroom for adjustment)

 

[PCBuds]

You might be able to "double-up" the driver to two identical LED boards but the driver is designed to be in the "heat zone" inside the base so it can self adjust it's output as it gets hotter.

If it were mounted in a cooler, separate place, it might overdrive the LEDs.

That may not be a serious thing?
It could just mean they get brighter but just don't last as long?



EDIT: That could be a more serious thing if the LEDs burn out during a grow and your girl gets all freaked out and decides to start growing boy parts.

You could have a cascade effect where if one LED burns out it allows more current to the other segments and they all burn out.

It the whole "constant current" vs. "constant voltage" thing and whether the LED boards are connected in series or parallel.

 

[Hookahhead]

The only issue would be the heat that can overheat and burn out the LEDs.

I'm almost positive you would only burn out the LEDs and not burn up the LEDs and start a fire.

This was my experience today. Just for fun I cleaned up the chipped diode and ran a pencil over the plate. This is an old trick from over clocking computers back in the day, but it was just lazyness. I probably should have put a glob of solder. Anyhow when I tested it, it lit up bright for a few seconds then a pop and a puff of magic smoke. The LED to the right burnt out, upon further testing the same thing happened to the LED on the left. The light stopped working when the LEDs popped.

I’ve thought of the SIL boards as mini quantum boards, they seem very similar as far as I know. So the idea of running multiple seems logical. My only other LED experience is with the 12v LED tape. These often come in rolls of 300 to 600 LEDs. I’ve ran multiple of these off drivers like the ones indagroove posted. However, as we’ve pointed out these are designed to run on 12v DC systems. In this case you just need to match the amperage with the draw you want to pull.

I don’t really understand how or why these boards are running at 107V DC. Is it somehow related to the fact we’re driving 15-30 LEDs instead of 300-600? I’m not sure if they’re different wattage LEDs in these applications.

 

[PCBuds]

I don’t really understand how or why these boards are running at 107V DC. Is it somehow related to the fact we’re driving 15-30 LEDs instead of 300-600? I’m not sure if they’re different wattage LEDs in these applications.

I think it's because of the voltage drop across that fuse/resitor.



I bet the missing voltage is across that resistor.

 

[PCBuds]

... So the idea of running multiple seems logical...

It does but the built-in driver is designed to run 9.5 Watts worth of LEDs and circuits.

Running two in parallel puts the Wattage up to 19.0 Watts.

The driver is designed with some headroom so it isn't maxed out making it more reliable but doubling its output might be too much for it.

 

[PCBuds]

My 50 cobs came today.



I thought they would be bigger.
It must be a tiny little Chinese hand in the picture. Lol





I also found an 8' piece of aluminum channel in the garage that I forgot I had.





I cut off a piece and brought it inside so I can get stoned and stare at it and dream up a science/art project. Lol

 

[Hookahhead]

I’m not surprised that there is a small voltage loss converting AC to DC, what I don’t get is why these boards require +100V to run. The LED tape is designed to run 12V, your LED bars are designed for 19.5V and the longer bars are 48V. I Looked up info on quantum boards very quickly. It seems like they run 45-48V too. All of these configurations have many more LED than the SIL. Therefore I would expect the converter to be pushing something like 12 or 24V?

I would not recommend driving multiple SIL boards from a single SIL driver, but I think if we figure out the actual voltage we can drive multiple SIL boards from an upgraded driver (Meanwell or big box style)

 

[PCBuds]

I’m not surprised that there is a small voltage loss converting AC to DC, what I don’t get is why these boards require +100V to run.

It's because of that resistor/fuse dropping voltage across it.
It reduces the voltage getting to the circuit, so the circuit needs to run off 100 V.

That resistor needs to be there to allow the circuit to regulate the current/voltage according to how hot it gets and to "blow" like a fuse if it gets too hot or there is a short circuit in the bulb.

But it's wasteful.
That's ~10 V multiplied by the current of the bulb turned directly into heat.

I would not recommend driving multiple SIL boards from a single SIL driver, but I think if we figure out the actual voltage we can drive multiple SIL boards from an upgraded driver (Meanwell or big box style)

Exactly.
A separate adjustable driver can be dialed back to the exact voltage/current needed by the LED segments.

 

[Hookahhead]

I don’t think you’re following me correctly... I don’t understand why a small SIL LED board with 15-45 diodes would be designed to run on +100V, when all other applications (involving hundreds of diodes) I know of are 48V or less?

 

[PCBuds]

I don’t think you’re following me correctly... I don’t understand why a small SIL LED board with 15-45 diodes would be designed to run on +100V, when all other applications (involving hundreds of diodes) I know of are 48V or less?

Because I've got 15 LED segments in series and it takes 100 volts to get them to lite.

You've got two parallel 15 LED series runs still requiring 100 volts.

A 12 V LED string has only 1 LED seeing the voltage and each LED is in parallel, so you just add more and more strings.

Every LED sees 12 V but the total current goes up as you add more and more, requiring a thicker wire to carry the current and a more powerful driver to supply the current.


A 45 LED bulb (at 120 V), has 3 parallel 15 segment runs.

 

[PCBuds]

If you've got two 10 Watt 800-lumen bulbs with the same size disk of LED segments but one has 15 segments and the other has 45 segments, the 15 segment bulb would have a brighter output per segment making each segment hotter.

The 45 segment bulb would share that heat with 3 segments making them run cooler, but, you still have 10 Watts of heat getting into the metal disk that needs to be dissipated.

The 45 segment bulb most likely has a much more dependable life span.


I still think liquid-cooled LEDs and crank them!

The water will absorb and remove the heat right at the source of heat. The LED segments themselves.

 

[PCBuds]

Then there's this early edition dimmable Great Value 10 Watt bulb.



It's got 8 segments in series and 5, 3 and 2 in series?

I have no idea what's going on with that?

 

[PCBuds]

I just fired up one of my little cob lights.

It was lit at 9.17 V but it was flickering (the lowest I could turn down the power supply)



It was nice and bright at 12 V.



Then it blowed up at ~18 V.

That's one down, 49 more to go.

It's fun to blow shit up! Lol


I thought I'd see if I could run them in my car which reaches about 14.5 V max. but I only got to 13.2 before it blew.

2 down. 48 to go.

I realized it was getting pretty hot in the few seconds I had it powered so I decided to add a heat sink.

I used my lucky 1947 square nickel as a heat sink with some heat conducting paste.
(it had some dude on the back, .. the Queen's dad I think?)



Still couldn't get past 13.4 volts.

That's three down.



These cobs suck!
They won't work in my car (without a control circuit. (aka a resistor)) and they're no damn fun to blow up.

No smoke, no noise, no spark, no fire, .. just a dead light.

 

[Hookahhead]

I understand what you’re saying about your string of 15 LEDs in series. However, If you do the math that rational doesn’t add up. I THINK most white SMD LED run around 3 - 3.4V. 15x3.3 = 49.5V

The board that I tested at 107.4V has 36 diodes. If you do 107.4/36 = 2.98V... which seems reasonable if they are ALL in a series. I still don’t really understand why you would want a +100V series of LEDs. Since it’s such high voltage, it must be really low AMP? W/V = A... 9.5/107.4 = .088A per diode?

Is 100V DC any more dangerous than 100V AC?

 

[PCBuds]

When I tested an led segment it wouldn't light at 6 V but would at 7.5 V.

I'll try a segment on that board again with my new fully adjustable voltage supply and see how much voltage it takes.

Page # 325 of this thread.

 

[PCBuds]

The segment was lit at 9.4 V and that was as far as I could turn down my power supply.

So hooked up two segments in series and they just started go glow at 14.0 V, so 7 V each.



Perhaps that "forward voltage" spec. is for when the LED is hot and conducts better??