karimwassef
Active member
Agreed! That's why I'm offsetting that with a liquid cooled DIY fixture. I'm also making the LED chips easily replaceable and expandable.
Looking for a specific DC CC driver for LEDs
- Thread starter karimwassef
- Start date
karimwassef
Active member
Wow... so here's what I've found out so far.
I decided to use an old 320W 36V Meanwell HLG-320H-36A with a pot to adjust Vo and Io.
It has a foldback feature that determines if it runs in CC or CV mode.
<a href="http://s1062.photobucket.com/user/karimwassef/media/CBF73D67-35A1-4E83-8C2A-FA495765FC1D_zpsekev1tku.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/CBF73D67-35A1-4E83-8C2A-FA495765FC1D_zpsekev1tku.jpg" border="0" alt=" photo CBF73D67-35A1-4E83-8C2A-FA495765FC1D_zpsekev1tku.jpg"/></a>
It should br adjustable up to 39Vo, and runs in CC mode in 18-36Vo. I pushed it up to 42V and 1.8A ~76W. There was no current adjustment (Io pot change didn't effect output), so it was still in voltage mode. This was the maximum I could get out of the Meanwell.
I decided to use an old 320W 36V Meanwell HLG-320H-36A with a pot to adjust Vo and Io.
It has a foldback feature that determines if it runs in CC or CV mode.
<a href="http://s1062.photobucket.com/user/karimwassef/media/CBF73D67-35A1-4E83-8C2A-FA495765FC1D_zpsekev1tku.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/CBF73D67-35A1-4E83-8C2A-FA495765FC1D_zpsekev1tku.jpg" border="0" alt=" photo CBF73D67-35A1-4E83-8C2A-FA495765FC1D_zpsekev1tku.jpg"/></a>
It should br adjustable up to 39Vo, and runs in CC mode in 18-36Vo. I pushed it up to 42V and 1.8A ~76W. There was no current adjustment (Io pot change didn't effect output), so it was still in voltage mode. This was the maximum I could get out of the Meanwell.
karimwassef
Active member
Then I went with a standard 48Vo power supply with an adjustable pot from 43Vo - 54Vo.
I hooked this up to the Sure Electronics CC Buck converter.
<a href="http://s1062.photobucket.com/user/karimwassef/media/8B837BDA-051D-4A74-8828-6741D9046D93_zpsnjallcc9.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/8B837BDA-051D-4A74-8828-6741D9046D93_zpsnjallcc9.jpg" border="0" alt=" photo 8B837BDA-051D-4A74-8828-6741D9046D93_zpsnjallcc9.jpg"/></a>
The Buck's maximum input voltage is 45Vin. I started at 43V and slowly trimmed the 48V PS up.. The Buck output was ~1V lower than input (43Vin = 42Vo).
The Sure Buck continued to run up to 46V, but the current seemed to reach a peak of ~3A around 45.4Vin and 44.4Vo! At higher input voltages, the current dipped down. At lower voltages, it dipped again. I'm assuming that a current limit of 3A is kicking in at higher input and output voltages. But that does mean that it's possible the LED can run at an even higher output voltage??
For an LED that's supposed to work at 34Vo, this is a significant departure (10V!)
The sweet spot is what I have in the image. 44.4V and 2.96A going to the LED. That's 130W going into the 100W LED. The CPU heatsink and fan are now putting out some serious heat!
I'm going to let it run for 24hrs to see if this is sustainable.
Next I'm going to take a PAR reading.
Then I'm going to set a lower current setpoint and see what happens.
I hooked this up to the Sure Electronics CC Buck converter.
<a href="http://s1062.photobucket.com/user/karimwassef/media/8B837BDA-051D-4A74-8828-6741D9046D93_zpsnjallcc9.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/8B837BDA-051D-4A74-8828-6741D9046D93_zpsnjallcc9.jpg" border="0" alt=" photo 8B837BDA-051D-4A74-8828-6741D9046D93_zpsnjallcc9.jpg"/></a>
The Buck's maximum input voltage is 45Vin. I started at 43V and slowly trimmed the 48V PS up.. The Buck output was ~1V lower than input (43Vin = 42Vo).
The Sure Buck continued to run up to 46V, but the current seemed to reach a peak of ~3A around 45.4Vin and 44.4Vo! At higher input voltages, the current dipped down. At lower voltages, it dipped again. I'm assuming that a current limit of 3A is kicking in at higher input and output voltages. But that does mean that it's possible the LED can run at an even higher output voltage??
For an LED that's supposed to work at 34Vo, this is a significant departure (10V!)
The sweet spot is what I have in the image. 44.4V and 2.96A going to the LED. That's 130W going into the 100W LED. The CPU heatsink and fan are now putting out some serious heat!
I'm going to let it run for 24hrs to see if this is sustainable.
Next I'm going to take a PAR reading.
Then I'm going to set a lower current setpoint and see what happens.
karimwassef
Active member
This simple setup is actually pretty cost effective for 130W (assuming that's turning 130W into light).
12V PS ................ $8 ............... 6A For the Fan and anything else
LED .................... $5 ............... 100W-130W
HS & FAN ........... $10 ............... per LED
Sure Buck ........... $10 ............... per LED
48V PS ............... $35 ............... 350W
So, it's possible to run 3 of these LED pendants at 350W total for ~ $35 + 3 x ($5+$10+$10) + $8 = $118
There's some wire and screws, so ~$7?
So 350W of LED for $125 ~ 2.8 W/$
The real question is reliability and PAR...
The light is also very harsh. It really needs a lens to make it more effective. I have one so I'll dig that up and see what it does.
I also have a few other LEDs to try.. whites and blues and one UV (I think). I got them all several months ago.
It's not a Kessil
, but compared to cheap Chinese light fixtures at $110 for 165W... this is half the price and a pretty easy DIY.
I haven't implemented the digital dimming in the Sure Buck, but that can be bridged with a basic Arduino Uno ($20) and some wiring to connect to an Apex ($5). This will run a lot more than just 3 LED pendants, but that's $150 for 350W with Apex dimming.
12V PS ................ $8 ............... 6A For the Fan and anything else
LED .................... $5 ............... 100W-130W
HS & FAN ........... $10 ............... per LED
Sure Buck ........... $10 ............... per LED
48V PS ............... $35 ............... 350W
So, it's possible to run 3 of these LED pendants at 350W total for ~ $35 + 3 x ($5+$10+$10) + $8 = $118
There's some wire and screws, so ~$7?
So 350W of LED for $125 ~ 2.8 W/$
The real question is reliability and PAR...
The light is also very harsh. It really needs a lens to make it more effective. I have one so I'll dig that up and see what it does.
I also have a few other LEDs to try.. whites and blues and one UV (I think). I got them all several months ago.
It's not a Kessil
, but compared to cheap Chinese light fixtures at $110 for 165W... this is half the price and a pretty easy DIY.I haven't implemented the digital dimming in the Sure Buck, but that can be bridged with a basic Arduino Uno ($20) and some wiring to connect to an Apex ($5). This will run a lot more than just 3 LED pendants, but that's $150 for 350W with Apex dimming.
Last edited:
karimwassef
Active member
so theatrus - can your PS handle even higher output? I think this LED could be even higher than 44.4Vo
Could your buck run 48Vin, 47Vo at 3.5A?
Could your buck run 48Vin, 47Vo at 3.5A?
theatrus
100-mile-commuter
The voltage limit right now are the input/output capacitors. Thermally is should be capable of up to about 5A, though I haven't run it from a bench supply.
A 1V voltage difference on the regulator isn't likely to engage the actual regulation (runs at max inductor current, not average LED current). I can run some more tests to confirm the behavior.
A 1V voltage difference on the regulator isn't likely to engage the actual regulation (runs at max inductor current, not average LED current). I can run some more tests to confirm the behavior.
karimwassef
Active member
Yeah. I was surprised that the Sure Buck can run 45.4Vin and 44.4Vo. That's a very short off time (~97.8% duty cycle). I don't know that it's even running at PWM.
theatrus
100-mile-commuter
From my tests with the LM3409, it's not. If you sweep the voltage up, due to the variable frequency of the converter, you can usually hear when the switcher runs in normal regulation mode (but then I miraculously can still hear lots of switch mode supplies when they're at low frequencies)
theatrus
100-mile-commuter
To be specific, in dropout (near equal voltages), the switch is kept on and any voltage drop and regulation would be across the shunt sense resistor and RdsON of the PFET. After this point the controller enters discontinuous mode (DCM) which effectively applies the peak current through the LED. Only after about 2.5V does the regulator enter full CCM mode (constant switch mode), and the average current is applied.
karimwassef
Active member
That's what I thought. Maybe I'll try to run my 48V PS directly.. It is 7.3A capable though... I can start at 44.4V and slowly increase the voltage and track my current draw.
Little dangerous, but at most, I would have lost a $5 LED.
Thoughts?
Little dangerous, but at most, I would have lost a $5 LED.
Thoughts?
theatrus
100-mile-commuter
The driver seems to be rated for only 45V:
http://store3.sure-electronics.com/ps-sp12153
which is curious as the LM3409 comes in 42V or 75V (LM3409HV) versions:
http://www.ti.com/product/LM3409
Can't tell if this is limited by the FET, capacitors, or the controller.
My controller has a 60V rated FET which with margin means I wouldn't want to go above 50V, which is coincidentally the limit of the MLCC ceramic caps
http://store3.sure-electronics.com/ps-sp12153
which is curious as the LM3409 comes in 42V or 75V (LM3409HV) versions:
http://www.ti.com/product/LM3409
Can't tell if this is limited by the FET, capacitors, or the controller.
My controller has a 60V rated FET which with margin means I wouldn't want to go above 50V, which is coincidentally the limit of the MLCC ceramic caps
karimwassef
Active member
I think it would be fine up to 50V. But pushing a higher voltage limits output in my case. The current goes down from 3.0 to 2.8, 2.7... as I increase input (& so also output) voltage.
I expect there's a current/power limit?
Thoughts on running 48V straight off the voltage supply with a resistor and fuse?
I expect there's a current/power limit?
Thoughts on running 48V straight off the voltage supply with a resistor and fuse?
karimwassef
Active member
Just a tangent here... Has anyone else run a multichip with a CPU HS & fan?
It seems so simple and cheap that I expect someone's already done this?
It seems so simple and cheap that I expect someone's already done this?
theatrus
100-mile-commuter
There is no OVP, so anything could happen. 48V is out of spec
If you have a slow blow fuse that should be plenty to prevent runaway explosions.
karimwassef
Active member
Cool. 2012 start!
But it seems like overkill? Those fans and heatsinks are rated for 3x the LED power?
In terms of OVP, that's why I can push it above the datasheet max... Question is: how do I measure the true drop across the LED without a driver that can deliver 48V CC?
Maybe I should look for a boost CC DC-DC driver...
But it seems like overkill? Those fans and heatsinks are rated for 3x the LED power?
In terms of OVP, that's why I can push it above the datasheet max... Question is: how do I measure the true drop across the LED without a driver that can deliver 48V CC?
Maybe I should look for a boost CC DC-DC driver...
theatrus
100-mile-commuter
Keeping temps low will keep output higher and longevity up. Might as well if cost or space isn't an obstacle. Also, fanless options.
It's hard to measure at load Vf over temp without actually running it. I'd personally just strap it to my 65V bench supply.
On the topic of boost, I'm mucking about with the A6265 controller which can run buck-boost, but would also have a voltage limit at 50. TI has some new controllers which were boosting to 75.
It's hard to measure at load Vf over temp without actually running it. I'd personally just strap it to my 65V bench supply.
On the topic of boost, I'm mucking about with the A6265 controller which can run buck-boost, but would also have a voltage limit at 50. TI has some new controllers which were boosting to 75.
karimwassef
Active member
Nothing off the shelf?
michaelrc51
Member
Nice discussion!
It may be overkill but it would be much better to design the system this way for output and longevity and also heat issues. Plus, you have some margin built in in case you ever decide you want to upgrade the LEDs.
Similar threads
