Sol-Skin, let the journey begin...

[StonerCol]

Got any more of what you're smokin'?

No. Fuck off..

 

[dbrn32]

You may not understand my perspective. I have been an accomplished electrical engineer for over 40 years and have been designing LED grow fixtures for the University of Florida Horticulture Research Department.for the past couple of years.

I understand your problem. I like that you are not using lethal voltages. I also prefer working in the 48V range. I like your white and red. I am using Olson SSL hyper red rather than XQE photo red.

A resistor is not the solution. A resistor will not help. A resistor will work when each and every LED has its own resistor. Incredibly inefficient as well. I struggle with the decision between inductors with DCR's of 110 mOhm, 120 mOhm, and 186 mOhm which represent efficiencies of 98%, 97%, and 96% respectively. I could NEVER add a resistor in the current path.
View attachment 3983438 View attachment 3983439 View attachment 3983440

An LDD-H (or custom buck driver) on each board is a good solution. An LDD-H for every 45V (16 x 2.8v) string is better.
A combination of individual CV powered CC buck drivers, CC drivers (e.g HLD), and load balancer chips is my idea of a solution.

This is a really cool circuit. The values or the resistors sets the ratio of current for each string. No limit to the number of strings. Just one chip, a resistor and cap is all the is needed to add another string. You can adjust the current on all strings using the dimmer on the current source.

View attachment 3983441

____________________________________________________________________________




I am working on explaining it to you guys. I will let you know when (and where ) it's ready. I will post in a new thread rather than hijack this thread. LEDs are so much more complex than they appear at first glance.

Now we're cooking!

I'm in agreement that there are more efficient ways. But I also think most are going to require more thought and knowledge than the average stoner is putting into their light build. Diy lighting has kind of transitioned into assembling some components and maybe needing a screwdriver and set of strippers. While I agree with you that Marks original description was not exactly on point, I think that his boards will fit the bill for the average guy on here wanting to take advantage of this tech. If in your case, or say an led nerd, wants to do better, I agree with you.

You typically don't have to leave the first page of the led thread to find at least 1 build by someone who doesn't have the slightest grasp of ohms law. And 10 people commenting that probably don't either. But you have to admit it's pretty cool that all of them will successfully build a working light. Maybe not as cool as one built by an electrical engineer with a big university research budget, but they'll grow.

I'll definitely keep an eye out for what you're working on, always cool to check out new stuff.

 

[OLD MOTHER SATIVA]

Nofucks

you have hit on something i have been hoping to solve

using 600 w cc/cv mean well hooked up parallel

after i hookup six of these panels i can visibly notice the last ones dimmer than the first three

so i take the other dc out cord from the driver and hook each one up to three panels

i cannot visibly see that any are dimmer now...

success?

....not totally

i bet that there is still less [current?..voltage?] making to the ends of the last panels even though its a cc/cv driver

is this correct ,why is this happening, and what can be dome about it..?

is it "just" the result of too long a path?

 

[dbrn32]

Shouldn't be difficult to do some Google work and calculate voltage drop for your conductor. Or take some voltage and current measurements. I read in a different thread you designed and spec'd your own boards for production right? Not being a dick, just assuming you would have the tools and already done this.

600 watts is a pretty good load. Probably wouldn't take much ti drag down a high current low voltage supply. But you should probably wait for the professionals opinion.

 

[welight]

This means that the resistors only take effect when the voltage is not identical in all strings. If all voltages are identical, the resistors do nothing. Basically, the voltage differences in modern LED's are very small, but in long strings the differences can already be a few millivolts. The resistors compensates that, right? Is not the same technic used in COB's?

correct

 

[welight]

You may not understand my perspective. I have been an accomplished electrical engineer for over 40 years and have been designing LED grow fixtures for the University of Florida Horticulture Research Department.for the past couple of years.

I understand your problem. I like that you are not using lethal voltages. I also prefer working in the 48V range. I like your white and red. I am using Olson SSL hyper red rather than XQE photo red.

A resistor is not the solution. A resistor will not help. A resistor will work when each and every LED has its own resistor. Incredibly inefficient as well. I struggle with the decision between inductors with DCR's of 110 mOhm, 120 mOhm, and 186 mOhm which represent efficiencies of 98%, 97%, and 96% respectively. I could NEVER add a resistor in the current path.
View attachment 3983438 View attachment 3983439 View attachment 3983440

An LDD-H (or custom buck driver) on each board is a good solution. An LDD-H for every 45V (16 x 2.8v) string is better.
A combination of individual CV powered CC buck drivers, CC drivers (e.g HLD), and load balancer chips is my idea of a solution.

This is a really cool circuit. The values or the resistors sets the ratio of current for each string. No limit to the number of strings. Just one chip, a resistor and cap is all the is needed to add another string. You can adjust the current on all strings using the dimmer on the current source.

View attachment 3983441

____________________________________________________________________________




I am working on explaining it to you guys. I will let you know when (and where ) it's ready. I will post in a new thread rather than hijack this thread. LEDs are so much more complex than they appear at first glance.

There are always better ways of circuit design, what you suggest is good but does not make sense for our board design, I can throw a bucket of money at a design to make it all it can be but commercially becomes unrealistic
Cheers
Mark

 

[welight]

They are never identical.


I use strings of 16 for white, green, and blue and strings of 21 for red, amber, and yellow. The strings of the same white LEDs (2.8v typ.), their Vf differed by volts, e.g. 42v-46v. They all came off the same reel and likely were manufactured on the same wafer. The closest two strings to have matching Vf I have seen, differed by 200mV. I ran them in parallel with an HLG-40-54 and the next day one string was dimly flickering and the other was very bright. Thermal runaway.

See the highlighted paragraph on page one of the attached PDF.

CoBs use parallel stings but they also use current balancing circuits.

Thats what our design avoids, unbalanced load and thermal runaway
Cheers
Mark

 

[PSUAGRO.]

Again, are you going to sphere test these boards in the future mark?

Don't concern yourself with the "electrical engineer", it's just Google kung fu..... his last account( Dr. Misunderstood), he was a crack smoking, prostitute loving , selling houses at a huge loss mofo....... at least he's upgrading his fake story every six months, that's something imo

 

[welight]

Again, are you going to sphere test these boards in the future mark?

Don't concern yourself with the "electrical engineer", it's just Google kung fu..... his last account( Dr. Misunderstood), he was a crack smoking, prostitute loving , selling houses at a huge loss mofo....... at least he's upgrading his fake story every six months, that's something imo

Its cool PSUAGRO, I dont agree with him, but everyone has a view.
Yes will be Sphere testing the boards but not until later next month
Cheers
Mark

 

[bakedPotatoe]

Thats what our design avoids, unbalanced load and thermal runaway
Cheers
Mark[/QUOTE
Im sure ive understanding this right so with them the loads they get are basic forced all the same and then dumping excess if there becomes excess out the other end?
Which leads to longer life when dealing with fluctuations in power and heat?
Electronics stuff me round at times..

And it seems i can get a trick today either with quotes....fml.....

 

[HenrikB]

Hi Mark. I been keeping a eye on the sol-skin page on your site and saw you have some new products up there now. Its a bit confusing when you have the same spectrum on all the boards even thou they have different diodes on them (kelvin wise) i guess the spectrum shown are the original 3000k/4000k/660nm spectrum you posted here? Are you gonna get spectrum measurements for all the different boards done? Information on recommended drivers would be nice and power guidelines like HLG have on their site for your boards would be excellent to see to.

will you stock heatsinks fro the boards?

Better documentation of the possibilities of your boards would make it easier for me to choose between the sol-skin boards and the QB boards. I think your boards are much better but since im missing information about your boards it feels like have to be a beta tester if i buy them........

Henrik

 

[OLD MOTHER SATIVA]

Shouldn't be difficult to do some Google work and calculate voltage drop for your conductor. Or take some voltage and current measurements. I read in a different thread you designed and spec'd your own boards for production right? Not being a dick, just assuming you would have the tools and already done this.

600 watts is a pretty good load. Probably wouldn't take much ti drag down a high current low voltage supply. But you should probably wait for the professionals opinion.

Thanks
I was more interested what he would say

I now only drive three panels before I give the next three a fresh feed direct from the driver
I can take voltage readings if I went fishing less
But
I admit I am no electrical pro
But the panels are doing darn good
Growing
As well the strips are giving me bigger seedings than I have ever got
Leds r fun

 

[welight]

Hi Mark. I been keeping a eye on the sol-skin page on your site and saw you have some new products up there now. Its a bit confusing when you have the same spectrum on all the boards even thou they have different diodes on them (kelvin wise) i guess the spectrum shown are the original 3000k/4000k/660nm spectrum you posted here? Are you gonna get spectrum measurements for all the different boards done? Information on recommended drivers would be nice and power guidelines like HLG have on their site for your boards would be excellent to see to.

will you stock heatsinks fro the boards?

Better documentation of the possibilities of your boards would make it easier for me to choose between the sol-skin boards and the QB boards. I think your boards are much better but since im missing information about your boards it feels like have to be a beta tester if i buy them........

Henrik

@HenrikB yes will update the spectrum charts on new boards and driver suggestions was suppose to have that done last week but ran out of days so will have Mapping, spectrum and driver recommendations done this week. 3x3 mapping test bed

Cheers
Mark

 

[HenrikB]

cool I was looking at the 3d rendering of a case for the sol-skin in the first post is that something that will come up too?

 

[NoFucks2Give]

Thats what our design avoids, unbalanced load and thermal runaway

You say that but it does not appear to be true. I see zero load balancing. Nothing at all. I'm not saying you boards do not work well, they do not NEED protection from thermal runaway. Load balancing would be nice. But saying you have load balancing when you do not is deceptive bullshit.

Fact: Resistors CANNOT do load balancing.

I just wonder why you said you use "Load balancing resistors", and then cannot not adequately explain why either. And this response is just not true. You do nothing for balancing. You board would have more value if you did but no, balancing is a legitimate corner to cut. Many customers would rather pay a little less less for their grow light and pay for the consequential inefficiencies over time over time. I have no problem with that. I think your boards with the deep red are possibly the best boards in this Samsung PCB market. I do not like someone baffling buyers with bullshit rather than dazzling with brilliance. I continue with my stand that resistors cannot balance the load as you said you do. What else does that leave but you were full of shit when you said that?

If you cannot answer the simple question of how the load balancing resistors you spoke of balance the load, so I conclude you are just full of shit.

 

[NoFucks2Give]

he was a crack smoking, prostitute loving , selling houses at a huge loss

If that were true, which it is not, why the fuck would it be relevant?

 

[NoFucks2Give]

you have hit on something i have been hoping to solve

is it "just" the result of too long a path?

No the imbalance comes from the dynamic forward resistance of LEDs. There are many factors such as current and temperature that affect the forward resistance. There is a range in which current and temperature affects each individual LED in a string and it affects each one differently. The more LEDs in a string the more the probability there will be serious imbalances.
The only way to fix it is to control the current in each string to be equal to the current in the other strings. The best approach depends upon the amount of current flowing in the strings.

A resistor in series with each string will not change the dynamics of the LED's forward resistance. There are some very inexpensive ways to balance the current. When the current is under 350mA, the best method, in my opinion, is a CCR such as the On-Semi NSI50150ADT4G (see attached). If I had more details about you situation I would be happy to help you further. Number of strings, LED part numbers, the current in each string (expected or measured), and thermal management.

Even though there is plenty of technical literature (see attached) to prove, I ran an experiment to test this. Here are the results:

 

[NoFucks2Give]

This means that the resistors only take effect when the voltage is not identical in all strings.

Correct, but the forward voltage is never identical. Adding resistance to the string will only increase the voltage but will do nothing to equalize them.

If you have 1 amp flowing through two strings and the Vf of one is 40v and the other is 39v, if you add a 5Ω resistor to each string the voltages become 45 and 44. Still not balanced. Just wasted watts.

 

[NoFucks2Give]

After a number of LEDs are added in series, the power supply may not be able to handle the higher output voltage required.

The solution in this case would be to use the correct power supply, not add voltage to the load with a resistor making it worse.

 

[NoFucks2Give]

y if the range of LEDs required is high enough to increase the range of the output voltage and auxiliary circuits.

WTF is an "auxiliary circuit"? You have a constant current power supply and strings of LEDs or a constant voltage supply and a separate CC driver. There are no "auxiliary circuits".