upgrade metal halide fixture to led

[Jadran]

.. and few shots more,

- temp swithc mounted on heatsink
- leds turned on

testing it over the friends tank

- T5s alone ( 2x54w actinic) that will do for mornings and evenings
- led pucks alone
- T5's & leds

 

[oreo57]

.. and few shots more,

- temp swithc mounted on heatsink
- leds turned on

testing it over the friends tank

- T5s alone ( 2x54w actinic) that will do for mornings and evenings
- led pucks alone
- T5's & leds

Nice job.. Curious as to how you are powering them though..
Channel arrangements, current ect..

This is the chip right?
http://www.topledlight.com/100w-cre...yan-490nm-red-660nm-white-7000k_p1856cec.html

your buck converter..
https://www.amazon.in/LM2596-DC-DC-Buck-Converter-Module/dp/B009P04YTO

diodes seem uncontrolled..except by sheer luck..well more like running out of voltage.
Your 2 chips are fed in parallel off the buck converter..

Those chips are meant to be driven w/ the likes of these AC/DC drivers like this:
ELG-100-54
https://www.onlinecomponents.com/me...VEeDICh2yjwc9EAkYASABEgLDjfD_BwE&gclsrc=aw.ds
One per chip...

Using your current power supply (assume 36V or greater) you could do an LDD-1500HW..
One per chip..
https://www.meanwell.com/Upload/PDF/LDD-H/LDD-H-SPEC.PDF

The "w" is just for err wired.. Will make things easier.
https://www.ledsupply.com/led-drive...MIxpPdp_ri5QIVjIvICh3VFQ5zEAkYASABEgJGWfD_BwE

only adds $16 to th cost.. plus shipping I suppose.. 5% off today..

 

[Jadran]

Thats the chip, right. But it is modded, UVs are replaced with blue and white, they would burn so they had to be taken out. Buck converter is there for fans, it converts 31,5 to 12v for them to run.

Yes chips are connected directly to psu, output is lowered to 31,5v. Im explained that coller the diodes less A they pull but giving the same power.

Well this is not my work, and I should not be credited for it. As I stated in posts earlier, Im from different background and electronics are not my thing, so I entrusted this job to an reefer friend who is in this sort of thing by trade. I did some asking arround to learn things for myself as well and been instructed to use drivers to provide constant current otherwise leds would burn. ... This was dismissed by my tech and explained to me that it would be dumb to go that way.. because it will lower the output (32v x 2A --> 64W) what we did here is more like 31,5v x 3A --> 94,5A per puck. UVs that were risk are replaced and that should do.

I could have bought it in set.

https://www.ebay.com/itm/660NM-490N...ltm5MbE4A3Y7JWa5HJdAszyB0-8fF6SxLG7ThR9ZrVLsM

but we decided to against that,, fans and heatsinks might be also of questionable quality..

 

[lingwendil]

Not bad

Not sure I understand the wiring as described, but hope it works well for you.

 

[Jadran]

Not bad

Not sure I understand the wiring as described, but hope it works well for you.

I guess the best would be to check in in a 6 moths and inform if it all still running fine

 

[oreo57]

This was dismissed by my tech and explained to me that it would be dumb to go that way.. because it will lower the output (32v x 2A --> 64W) .

That goes against probably 90% of those that DIY and

because it will lower the output

That is a choice based on COST alone. Plenty of AC/DC constant current drivers w/ enough amps set point to do that..
ELG-100-36 will output up to 36V (voltage ramps till Amp setpoint is reached) and 2.66 "controlled" amps..95.76W..


I've only known like one or 2 people who use "native" switching power supplies as the only "control circuit".


See given enough heat (say dead fan) and the diodes pull a lot more current at a lower voltage creating "thermal runaway" that will fry the diodes..

Basically you are just hoping the voltage runs out before the diodes go into self destruct current mode........

suppose, since I'm being picky, you barely have control in the 5 channels.. since current is shared over them w/ no accounting for differences in diode electrical characteristics..

 

[Jadran]

I've only known like one or 2 people who use "native" switching power supplies as the only "control circuit".


See given enough heat (say dead fan) and the diodes pull a lot more current at a lower voltage creating "thermal runaway" that will fry the diodes..

These other folks that use native ps.. what are their experiences...any long run success stories?

About dead fan and thermal runaway.. --> but weve got temp switch on each heatsink.. as soon as the temp reaches 47C it will shut down the plucks power feed...that should keep us in the clear wont it?



Poslano sa mog SM-A605FN koristeći Tapatalk

 

[oreo57]

Sorry never kept track and most were freshwater and frankly knew the diodes were disposable at like 10/$1...

There are a lot of things that work contrary to "best practices".. just not my cup of tea.

One thing though is I consider it an obligation to mention "alternate views" to people who have little to zero understanding of what they are doing..

https://www.powerelectronicsnews.com/technology/the-three-ps-of-led-driver-selection

Like I said, you can "properly" run those chips but adds about $100 to the cost.
In a sense you are putting a lot of faith in very cheap components..there is very little stability to begin with..

to end this let's just say that is not the way "I'D" do it..

OK key point if one could THEORETICALLY tailor the voltage output to the chip as to only allow a safe current flow (sadly ignoring each row will be electrically different to begin with) it is possible to run err "driverless" with success both short and long term.

Problem is components never stay the same and it doesn't take much of a change to have diodes "pull" more current than they are designed for especially if you are already running at the edge.
https://www.waveformlighting.com/pcb-designs/when-and-why-do-leds-need-current-limiting-resistors

On the upside, don't see you burning down your house...

Bottom Line:
LED devices are inherently current-controlled devices, and do not respond well to fluctuations in voltage.
If you are building an LED system using constant voltage power sources, you must absolutely be prepared to use current limiting resistors to ensure stable and safe operation of LED devices.

 

[Jadran]

Like I said, you can "properly" run those chips but adds about $100 to the cost.
In a sense you are putting a lot of faith in very cheap components..there is very little stability to begin with..

to end this let's just say that is not the way "I'D" do it..

OK key point if one could THEORETICALLY tailor the voltage output to the chip as to only allow a safe current flow (sadly ignoring each row will be electrically different to begin with) it is possible to run err "driverless" with success both short and long term.

Thank you for your explanations, highly appreciated. That is all very eloquently said and when layed down in such form I simply cant disagree that it is in my interest to have safe and reliable fixture. Of course it is my prespective to cut down on costs ( thats why I wont buy Giesemann or Maxspect in a first place) but not blindly and with intentional acceptance or risks. 100usd is surely gladly acceptable extra if its going to provide long life to my diodes.

it was just persented to me that drivers are not needed and the stuff that man makes speaks for itself. I dont see my friend gaining anything by ill advising me,,, thats all.

 

[oreo57]

He didn't "ill advise" you exactly. Just a matter of technique so to speak..