Is this a Vero 29 killer?

[nfhiggs]

No, I think they are stupid for selling a product that can kill their customer. 100VDC or more I will not touch. I'm too careless. You've got to know your limitations.

It's no better than the Quantum Board. If you can be patient, I'll hook you up with some PCBs. If necessary, I'll show you how to solder the SMD parts. It's easy.

I would be interested in some of your boards. I can solder SMD no problem. Been doing it for many years.

 

[nfhiggs]

Parallel strings driven with a shared CC driver will experience thermal runaway. When one string operates at a slightly different temperature than the other, the voltages will change, the current will change the temperature will change and over time one sting will be drawing most of the current. CoBs are matched pretty well in voltage and temperature, so it would take a long time, like days for it to happen if they were left on 24/7. If I were to fire up one string wait a minute then add the second string they would start out with a significant thermal imbalance, within 12 hours one would be burning up, the other flickering out. If I switched which one I fired up first then the other would flicker out. The more LEDs in the string, the higher the voltage mismatch, and the greater the difference in temperature the more likely it will happen.

"Parallel strings driven with a shared CC driver" - there's your problem. You don't use CC with parallel loads. You use CV, you let the current and temp stabilize, then adjust to your target. Recheck and readjust as necessary to keep it in range.

 

[nfhiggs]

I was kinda kidding about the ice maker, I did consider it though. The plan is to keep adding fixtures until the water starts getting warm so I know the thermal capacity. A friend of mines son is moving to Oregon next week to run a $Million indoor cannabis grow facility. So if things work out hopefully a lot of kW. Unfortunately I am not ready for them yet.

I have not had good luck using thermal dynamic formulas so would rather do an experiment than the math.

There are three kinds of people. Those that understand math and those that don't.

There 10 kinds of people. Those that understand binary and those that don't.

I don't try to do the math either, for the most part. I just experiment and figure out what works for my setup.

 

[DrMisunderstood]

You don't use CC with parallel loads. You use CV,

Okay, with a CV how do you limit your current? I've seen mostly Meanwell HLG being used here, those are CC. I use the Meanwell HEP which is a CV. Plus I have a shitload of HLG laying around all over the place. If I connected the HEP, the boards would fry. Generally with a CV a current limiting resistor is used. That would be too inefficient when you have multiple amps flowing. The CoBs I have, I power with Meanwell LDD-1500 modules which are powered by the HEP. Soon I will be using my own Driver using the TI LM3414HV with about 98% efficiency. Then when doing parallel loads I will use the TI LM3466 to balance the load. Those string will be driven with the LM3414HV.

This is my LM3414 Buck Step Down Driver with Thermal Foldback. I use the analog current adjust driven by a transistor biased with a thermistor on the LED strip. R3 set the max current. R4 sets the switching frequency. The L1 foot print will accommodate many different inductors. I can then adjust the efficiency for each board. After I populate the strips I can bin them by forward voltage and adjust the HEP to just over max forward voltage for excellent efficiency. Plus I can keep the current regulator in CCM without ever going into DCM with this flexibility.

I know you can get away with doing things that go against the rules, but nothing is free. There is a reason for the rules. My main rule is to know the rules so I know how to break them properly.

 

[nfhiggs]

Okay, with a CV how do you limit your current? I've seen mostly Meanwell HLG being used here, those are CC. I use the Meanwell HEP which is a CV. Plus I have a shitload of HLG laying around all over the place. If I connected the HEP, the boards would fry. Generally with a CV a current limiting resistor is used. That would be too inefficient when you have multiple amps flowing. The CoBs I have, I power with Meanwell LDD-1500 modules which are powered by the HEP. Soon I will be using my own Driver using the TI LM3414HV with about 98% efficiency. Then when doing parallel loads I will use the TI LM3466 to balance the load. Those string will be driven with the LM3414HV.

This is my LM3414 Buck Step Down Driver with Thermal Foldback. I use the analog current adjust driven by a transistor biased with a thermistor on the LED strip. R3 set the max current. R4 sets the switching frequency. The L1 foot print will accommodate many different inductors. I can then adjust the efficiency for each board. After I populate the strips I can bin them by forward voltage and adjust the HEP to just over max forward voltage for excellent efficiency. Plus I can keep the current regulator in CCM without ever going into DCM with this flexibility.

I know you can get away with doing things that go against the rules, but nothing is free. There is a reason for the rules. My main rule is to know the rules so I know how to break them properly.

View attachment 3929493

I use a 1 ohm, 5W current limiting resistor on the input of each COB. Each COB is getting 33V @~1.33 amps, six COBs currently on that power supply. Is it "efficient"? Dunno, don't care, it works.

Shortly I will be adding 4 more COBs to the fixture and reducing the current to each one to 800 mA. Total current will be the same but I should get more total light output since I will be reducing the power to each COB from 44W to 26W.

 

[DrMisunderstood]

I would be interested in some of your boards

It may be a month or more before I have more made fabed. I am working out the little details yet. I am first going to build a board that has a bunch of test circuits. Drivers, Voltage Regulators, PWM circuits, a micro with serial port and etc. I think I'm going to populate the strip PCBs with only LEDs, no on-board anything. Then I will have various daughter boards to control the strip.

The question d'jour is which foot print? I had a dual Luxeon Rebel and Luxeon Color C/SunPlus 20. Then OSRAM released a new Olsen SSL 150 aimed at horticulture and it blows Lumi out of the water. I just bought 250 Olsen SSL Royal Blue, 565mW for $0.12 from Digikey. Luxeon Color C and SunPlus Royal Blue is around 300mW and costs more. The twelve cents was likely a misplaced decimal point. They probably should have been $1.20. I have some on back order too. I really like Luxeon Color C 170° view angle. I ran the Olsen 150° through my app and the uniformity was phenomenal.

Dual Rebel / Color C



While the Color C is 170° and the Olsen only 150° Numbers are deceiving. The Color C is great.



Now Look at this 150°

60° @ 98% Flux. 95% at 70° Plus the flux is over twice Color C and cost is less.







The reason I went with the dual Rebel and Color C is because the Color C is too damn small and I was going to use the Rebel as a backup.

 

[nfhiggs]

And the nice thing about having a 1 ohm resistor in series with each cob is that you can measure the voltage across it and know how much current each cob is using.

Bingo! And you can get little digital voltage displays for about a buck each.

 

[DrMisunderstood]

I use a 1 ohm, 5W

Right there where you said "5W", that's inefficient. An entire ohm is out of the ballpark. A whole ohm might be used if the current is less than 100mA.

For me, on the LM3414 driver, efficiency is the difference between an inefficient 0.200 ohm inductor resistance vs. a 0.05 ohm inductor (all current to the LEDs goes through the inductor).

That's why we use drivers, they are A LOT more efficient than a resistor.

If I don't care about efficiency I'll use an on-board 0.01 ohm, 1% resistor.

To measure current with a resistor, this is the resistor I use:

 

[DrMisunderstood]

nice thing

I'm drunk. This is fuckin funny as shit!!! Sorry.

1 ohm is NOT nice!!! Fuck! Come on guys. Please.

Notice L1 is 0.170 ohm. This is inefficient.

Notice on the next image the resistance is 0.021 ohm. And that is for a voltage regulator for less than 500mA.

Okay, so when you are looking at a power supply and you check out the efficiency. A 1 ohm resistance anywhere in the supply path is going to drop the efficiency down to probably like 70%. I mean we look at the ESR of the output caps.

Previously I said you can break the rules but nothing is free. For whatever reason you chose to use a resistor and parallel strings you have no clue how badly you fucked up.

Are you guys high of something? Fuck I need to go get more beer.

 

[nfhiggs]

Right there where you said "5W", that's inefficient. An entire ohm is out of the ballpark. A whole ohm might be used if the current is less than 100mA.

For me, on the LM3414 driver, efficiency is the difference between an inefficient 0.200 ohm inductor resistance vs. a 0.05 ohm inductor (all current to the LEDs goes through the inductor).

That's why we use drivers, they are A LOT more efficient than a resistor.

If I don't care about efficiency I'll use an on-board 0.01 ohm, 1% resistor.

To measure current with a resistor, this is the resistor I use:


View attachment 3929544

5W is simply the size I chose to keep it from getting uncomfortably hot if I touch it. At 1 amp the resistor dissipates 1W of heat and drops one volt. As pointed out above, this makes it simple to monitor the current, by measuring the voltage drop with a voltage display - no conversion needed. Smaller resistors can certainly be used and will use up less power, but the drawback is they drop less voltage and may not be as effective at current control as the 1 ohm is - this is passive current limiting, not active. Again, I'm just not worried that much about losing 1/32 of the power going to the COB to a current limiting resistor. I'm not looking for every ounce of efficiency I can get, I'm looking for cheap and simple. Meanwell drivers are simple, but not cheap, which is why I don't use them.

 

[DrMisunderstood]

I see you're under a lot of stress

I am literally laughing my ass off. You are so fuckin clueless. I have zero fucks left to give. I have a few rats asses but I no longer give a rats ass.

I am fuckin serious. The shit that used to piss me off is now funny. Now you have to realize that I understand when someone come along and puts me down they are trying to elevate themselves.

News flash, that does not work. You also put women down because you have such low self esteem you think you have to put them down of they may discover they you are such a loser and they will leave you. And because you put them down, they leave you. Your limp dick performance anxiety did not help. So do not worry about me, take care of yoiurself.

3 out of 4 of my grandparents were dead at my age. The only thing that may cause me stress is I have so much shit in my head that I may dead before I can give it away to someone.

Example, I recently moved out of a $300,000 house (with $800,000 in code enforcement fines). I walked away. Did not give a rats ass for that either. Well a few day before it was going to be auctioned off by the foreclosure court I sold it for $15,000. I walked away with a smile.

I can't smoke pot. My lungs are shit. I can smoke crack it's like an anesthetic. And I love crack whores and they get me.

 

[DrMisunderstood]

I'm looking for cheap and simple.

You got cheap and simple. Still I am serious, even though I am drunk, one ohm is very inefficient. A 0.01 ohm does not cost much more. Still cheap, still simple and yet efficient. Wattage goes down too, it will not get hot. And the volt meter will still work, just change the decimal point.

 

[SaltyNuts]

This seems like a good opportunity. How about a thread dedicated to safety features and redundant thermal protections in DIY led setups. Something we can apply to cobs and discrete diodes.

 

[SaltyNuts]

Because it's not really addressed very thoroughly otherwise. So far hacking together a bunch of cobs seems ok as long as you use a decent driver correctly. And that seems to be working. But it's good to know the rules you are bending.

 

[nfhiggs]

You got cheap and simple. Still I am serious, even though I am drunk, one ohm is very inefficient. A 0.01 ohm does not cost much more. Still cheap, still simple and yet efficient. Wattage goes down too, it will not get hot. And the volt meter will still work, just change the decimal point.

Actually I did consider using a .01 ohm. But I do not believe it would be effective at passive current limiting as its voltage drop is much smaller. With a 1 ohm resistor a 100 mA increase in current from a temp rise will drop an additional .1 volt, and I know from experimentation that a .1 v drop on the voltage applied to the cob will decrease its current draw by a similar amount. I do not have any confidence that a .01 ohm resistor will do that.

 

[DrMisunderstood]

So far hacking together a bunch of cobs seems ok

Yes, everything is going great. But the World is still turning and will leave you behind if you do not keep up.
I have to admit the status quo is pretty fuckin' good.
I was happy with my results in 1975 using 4 growlux bulbs. The shit was kick ass. Tasted like shit.

The big question is, do you want to do better?

 

[DrMisunderstood]

its voltage drop is much smaller.

A smaller voltage drop is the goal. Smaller is better in this case. A CCR has a voltage drop of about 50mV and that is their appeal. The highest current CCR I know of is 350mA Oh wait a second, you are drawing an amp. You can put 3 of them in parallel. No, a buck driver with a load balancer would be cheaper.. A 350mA CCR is about $0.75 and you'd need 3 of them.


Do not think of it in terms of cock, think of it in terms of pussy, smaller is better.

Wait a second, I'm drunk. The resistor value is picked to select the amount of current.

No. I'll have to get back to you on that after I take an Adderall tomorrow.

 

[The Dawg]

I can't smoke pot. My lungs are shit. I can smoke crack it's like an anesthetic. And I love crack whores and they get me.

This Explains Everything

 

[DrMisunderstood]

This Explains Everything

Yes it does but not likely in the way you think. Your response says more about you. You have issues. You should work on that.

 

[sixstring2112]

This Explains Everything

take it easy on Dr nofucksgiven,he used the drunk disclaimer so he is allowed to be a jackwagon lmao.