Special Led Chip Array With Step Down Converters.

[MostHonorableSuzi]

So I'm having an issue understanding exactly what the issue was with my stepdown converters.

Before I go any further here are the details of the stepdown converters.
http://vi.raptor.ebaydesc.com/ws/eB...363&category=117000&pm=1&ds=0&t=1476497479916

I ran a 120mm fan x2 combo that I scavenged from a server, it also had a rail that had a triple fan assembly. Fans were rated for 12V 1.0A. They are Nidec BETA V TA350DC fans. I also have a TA450DC rated @ 12V (1.40A) And several Delta Electronics DC Brushless MODEL PFC1212DE's, @ DC12V (4.80A)

So I connected the pc fans to a laptop supply that was 12V (2.0A) And drove 2 fans and it seemed to work like I was hoping except that when I would adjust the voltage potentiometer I could only lower the voltage down to about 6 volts maybe it was slightly lower I can't recall exactly now.

I could'nt seem to adjust the current at all which would be obvious because the fan would slow dramatically and the current potentiometer adjusment knob did not appear to be working the way that my circuit was set up. They are 4 wire fans but I was just running the 12v and ground wire.

I believe they have a hall effect sensor and a transistor rpm controller. As far as I know the 2 wire fan circuit can't be overridden by the transistor because there's no power going to the transistor (rpm controller).

I wish that I could upload photos of the converters and fan together but its not really that important atm.

What's of main importance is the led chips I want to drive with a 32V-40V (9.75A) switching power supply. The chips require an input volttage that is lower than the lowest output voltage setting on the power supplies voltage potentiometer. Chips are rated for MAX Approx (91watts) 26V (3.5A).

Assuming the step down converter cannot increase current in exchange for voltage drop I am fine with the 1/4 Amp loss. However its worth knowing whether or not a Buck/Boost Converter would be able to manage the power in such a way to exchange voltage drop for current increase. What I'm definetly not aware of is how the power supply would react to changes one the other end.

When I tried to current test the laptop supply 12V (2.0A) The meter would blip read 1.9 maybe 2.0A and then drop to somewhere like 0.1A in a sec and stay there and when I would stop short circuit current testing the fan would go back to full speed. I'm not sure if I'm current testing correctly, I do know my meters fuse is a 10A fuse, and the output current of the led power supply is 9.75A so I don't want to blow a bunch of fuses over and over again.

I'm assuming the issue with the step down (buck converters) I have with that fan only adjusting voltage down to 50% had something to do with the switching laptop power supply. I really don't want a voltage surge to hit the led chips because they have diodes that have a lower power capacity and will burn out and ruin the chip because of the combination (Parallel single row 10 diodes down) & (10 rows of Series circuits across with 10 diodes each) burning out a row would increase load through others and the chip would eventually catastrophically fail.

Anyhow I'd like to figure out how to upload photos to the thread for the sake of better illustrating what I'm trying to achieve in terms of ensuring the chips get the proper input voltage and don't end up fried and no good to me.

I much appreciate any help I can receive on this matter so I can share the results with other people interested in knowing more about the particular led circuit, the chips themselves, the buck converters, and potentially the light assembly which I have big plans for *Spoiler* It's going to be a recirculating liquid led chip assembly with a fan-cooled radiator to chill the coolant before it flows through the CPU waterblocks bolted to the back of the chips.

Anyhow this is enough information for one post anyways.

 

[(*steve*)]

You don't actually tell us what happened?!?!

I recommend you adjust the output voltage of the buck regulator. reducing it will make the fans run slower (which may be what you're after -- again you don't really say).

Don't touch the current control on the regulator -- leave it turned up.

 

[(*steve*)]

Also don't short things out and what do you want to do with the LEDs?

 

[MostHonorableSuzi]

They are for urban agricultural applications.

There is going to be some difficult drilling and tapping in the next few days when I receive the chips and carefully bolt them to the water cooling blocks.

I'm not sure if teflon tape will be thin enough to wrap the threads of a
8/32 screw, I might have to resort to using threadsealant or some kind of other liquid gasket seal.

My COB knockoff led chips have very little room for error when I line up the holes to drill them all the way through front and back a 10/32 width might be too much.

Have to wait till the chips arrive before I can come to any conclusions.

Only other concern will be whether or not the converters have any memory to remember the last voltage setting, I know so little about the things, they seem very analog on the top surface of the units using turn screw potentiometers, but the mosfet chips on the bottom make them look deceivingly complicated.

I'm crossing my fingers that they work like I have previously imagined them working or I will have to buy a separate transformer, rectifier, capacitor(s) and a voltage regulator and make my own power supply.

I already sourced out a transformer @ digikey, they wanted $50 for the transformer. I haven't checked but I'm gonna guess maybe $25 low estimate total cost for the other components.

One way or another I'm going to have these leds driven right,

 

[MostHonorableSuzi]

The 100watt rated 7-band LED chips mounted to the heatsinks.
The heatsinks were cheap, $4 each, LEDs were expensive something like $92/pc including U.S. exchange rate, taxes, and canadian customs fee. Pink plastic is just the protective plastic insert.

100watt chips each, next time Im thinking 300watt chips because I have a nice 965watt power supply I want to use and because the LED chips are just so bloody cool.

 

[MostHonorableSuzi]

I am looking for a good price on waterblocks such as this with perfected internal channels for the utmost cooling potential. This is from CustomThermoelectric™. They're prices are expensive @ $100/pc, they better be totally custom @ that unit price.

 

[(*steve*)]

They are for urban agricultural applications

What sort of dope do you think I am?

 

[MostHonorableSuzi]

What sort of dope do you think I am?

Sorry but that seems like a bit of a malicious suspicious statement. From the information I've given as far as anybody here can prove this is a legitimate urban agricultural project.

Sorry your not satisfied with my complicit honesty I was not aware this too is a place where people come to over-moderate and police people.

I suppose people are going to complain now that I've posted a picture of some exclusive industry first LED technology.

I don't care if you don't like it, this is about people accessing information with regard to electronics.
You don't like it thats too bad I'll go somewhere else and discuss this.

There's been nothing illegal going on here you need to check yourself big time.

 

[(*steve*)]

I'm intrigued, what urban agricultural applications are there where the cost of power for growlights doesn't make it cost prohibitive?

Also, why do your LEDs cost so much?

 

[MostHonorableSuzi]

Alright, I may have been a bit hostile there, I've had a rough week, actually more like a yr but we'll call it a week.

They are expensive because of the microsoldering technique used to solder the diodes with gold thread. The diodes are actually epileds from taiwan which are making these nanometer specific wavelengths for plants. And the backing plate and plastic template of the chip is made by Bridgelux from America and it is made of solid copper and electroplated with aluminum it looks like, its twice as thick as most LED's.

Its a 7 band led, close red, far red, infrared, UV, close blue, far blue, and a yellowy green for chlorophyll production.

I also have a 965watt PSU from a IBM server.
I plan to combine the 12v rails to 36v which is what the cheaper LED chips are rated for. I actually have 3 cold white 100watt chips(most lumens) and they are rediculously bright. I could wire 9 of them together and get some pretty crazy lumens for overall growth potential but the grow chips are specifically purposed to address the active photosynthetic radiation strictly for quality purposes and for side by side comparison.

I think the chips are actually quite energy efficient, compared to HID's, Halogens, comparable to the new plasma HID's in terms of PAR scale (photosynthetic active radiation.)

Much better than CFL's certainly in everyway.

350watts running 18hrs a day for a month is like $29 canadian in utilities where I'm from.

 

[MostHonorableSuzi]

Sry I thought this was heading for a thread deletion. I spent a lot time preparing this thread in the first post so people who want to know more about DC converters, and leds, or who know something about it would be able to find this in a search.

Anyhow if you purchased this transformer
http://www.digikey.ca/product-detail/en/signal-transformer/DP-241-8-28/595-1097-ND/952967

You could then drive a led with just a rectifier, a small series of capacitors, and a voltage regulator wired to it. It would then be a constant current supply not a switched mode power supply.

The cost of the chip and shipping would bring the bill to around $150cad, the chip manufacturing company already offers a competitve 120usd 100watt fan cooled assembly, and a 300watt fan cooled assembly.

I wanted to know that the led chips were going to be cooled very well so the high entry price of the chips would be redeemed in longevity.

 

[MostHonorableSuzi]

Dimensions of the digikey transformer is quite beefy.

102.40mm L x 58.50mm W x 67.7mm H

It ends up being the size of a 20 pack of king size cigarettes but obviously way heavier. Does explain why those 100watt ac adapters are so expensive because they require an oversized plastic housing with a smaller production run and higher unit price.

 

[(*steve*)]

You could then drive a led with just a rectifier, a small series of capacitors, and a voltage regulator wired to it. It would then be a constant current supply not a switched mode power supply.

I home you're going to suggest a circuit where the regulator is set up to supply a constant current, not a constant voltage.

I'm also really interested in why you want to forgo the efficiency gains you would get with a switch mode regulator?