DIY LED upgrade in an old PFO MH/VHO fixture led driver suggestions.

[geekengineer]

So I had this old 36 inch PFO 2x150W Metal Halide/ 2x VHO T12 fixture that at this point I was going to upgrade the VHOs to a LED strip. I happen to have a bunch of Royal Blue 10W COB modules that I never used, which I could use for replacing the blue light. I bought a couple 34 inch 2 inch heat sinks one each to replace each VHO tube. I was going to use 6 10W COBs on each strip. So I don't have the specs on these ones I bought but the specs currently for the on on ebay says they are rated at 9-11V with 1A forward current. I want to be have them dimmable with the 10V input so I can hopefully interface it to something like reefpi. So far I haven't had much luck finding a good constant current driver that is dimmable other than going over wattage which means more $$$ that fits. The only one I have found that seems to fit that isn't like overkill wattage wise is this 75 watt driver. LUB075M-108C This one would let me put 6 of them in series theoretically for each strip. I might have to tune it at the start to make it default to a bit less current etc. Would there be any other drivers I should be looking at or configurations I could do?

 

[oreo54]

So you need 54- 66v (2 strings of 6) or 108-132v if you run all 12 in one string.

HLG-120H-C1050
If you get the (ab) type you can turn the current down to about 550mA.
"B" is 1-10v dimming, "A" is current adj.
74 to 148v , 1.050A.

Your expected voltage is 108 -132v.

HLG-120H-C-MEAN WELL USA Switching Power Supply

MEAN WELL Switching Power Supply. MEAN WELL is a Switching Power Supply manufacturer in Taiwan. – SMPS, SPS manufacturer, AC DC , DC DC , DC AC, industrial power supply, Medical power supply, Din Rail power supply, LED power supply, LED driver, charger, inverter, adaptor, rack power, modular...

www.meanwellusa.com

Since ab drivers are hard to find I'd suggest not doing 1a. The ab above allowed you to tweak it based on how hot your heatsink was getting 1.040a to .55a (approx, see spec sheet)

https://www.digikey.com/en/products/detail/mean-well-usa-inc/HLG-120H-C1050AB/10659885

Going to 700mA and your specs are close.
HLG-120H-C700
107-215v.
If the "real" V(f) of 12 cobs in series falls below 107v (8.92v each) @ 700 mA you may have a issue.
Btw it is HIGHLY unlikely that all of them will have the same voltage at the same current.

You may want to test your cobs using say 700mA meanwell ldd he and a 24 v 1a or better power supply. That way you can get the exact voltage the ldd- outputs and have " the correct" measurement of each cob.
Make sure they are temporarily heat sinked.

Ton o driver's here. You know what to look for.

HLG-C Series Constant Current LED Driver from Mean Well

Ideally if you wanted one string of 12 cobs you'd look for one w/ 90 -144v spec. at your desired mA.
2.5v to 4v for each ind diode in a cob. Would be odd if the average was below or above those numbers.
108 individual diodes in 12 cobs.
Then you don't have to worry about the slop in the system.
Those cobs are 3 diodes in series 3 strings in parallel.
Their common assumption.
3x 3v = 9 or 3x 3.67v = 11

45-72 v @ whatever if you want to use 2 drivers or divide the current by 2 with 1 driver.

Again I'd avoid running them higher than 750mA.
Sooo 45-72 ( or better) @ 750mA
Or 45-72 @1500mA with a 1A fuse on each of the 2 parallel strings.
Hopefully this makes some sense and I didn't screw up any calcs

So now you can see why HLG-120H-C1050(AB) would, unfortunately, be the perfect choice. 1 string of 12, no worries that the driver is out of spec.
550-1040mA, 74 to 148v

Actually found one. Huh.
One driver to rule them all ....
Sorry.

https://www.trcelectronics.com/View/Mean-Well/HLG-120H-C1050AB.shtml?utm_term=&utm_campaign=Shopping+Display&utm_source=adwords&utm_medium=ppc&hsa_acc=9805199880&hsa_cam=164782104&hsa_grp=11827142184&hsa_ad=41002365624&hsa_src=g&hsa_tgt=aud-804527999442:pla-811491429052&hsa_kw=&hsa_mt=&hsa_net=adwords&hsa_ver=3&gclid=CjwKCAjw1ICZBhAzEiwAFfvFhK-s0_N9AMsvq4kNhjkX3oHGVO7nAs8PXJwCiIT7oqxVId7M7lgPvRoCddUQAvD_BwE

 

[geekengineer]

Yeah I was looking at that Meanwell one too initially and it looked like I couldn't find an AB on but I didn't think to check if that place actually had an AB. I didn't realize that 700mA would be the current that you would more likely run at. I was thinking maybe more around 900mA, but I couldn't find much in that space either. Anyhow I placed an order for one of the 120 ABs and I see if they have them in stock. I couldn't tell if they did for sure.

 

[oreo54]

Yeah I was looking at that Meanwell one too initially and it looked like I couldn't find an AB on but I didn't think to check if that place actually had an AB. I didn't realize that 700mA would be the current that you would more likely run at. I was thinking maybe more around 900mA, but I couldn't find much in that space either. Anyhow I placed an order for one of the 120 ABs and I see if they have them in stock. I couldn't tell if they did for sure.

Mouser lists having 1... and I mean"1".
In case the above falls through.
But more expensive though.

https://www.mouser.com/ProductDetail/MEAN-WELL/HLG-120H-C1050AB?qs=w%2Fv1CP2dgqrfNPnwkljNOg%3D%3D

What would you run a 1w class diode at?
Cobs 9 1w leds (3x3)
250mA (750/3) is about .825w

Anyways you now have a choice which is always good.

There is a tradeoff in efficiency vs drive current
It is not linear. Also ( and part of it) the hotter the cob the less efficiency
Sooo lower drive current, lower heating, better efficiency.
Longer it lasts.
You still lose gross photons but not as many as one would expect in a linear relationship.
The reason " black boxes" run "3w class" diodes at 550mA ( commonly).
2 watts +/-. Even then lots of fan cooling.
They'd probably quickly die without them. Hmmm wonder if someone has proved that?
Damage may not be visibly noticed.
Need a before/ after measurement.
Lux, par wouldn't matter for comparison.

 

[geekengineer]

yeah makes sense I was assuming they would be around 10W, but that is probably in an ideal world. The order says it expect to ship tomorrow so we will see. Nobody contacted or e-mailed me to say it wasn't available etc.

 

[geekengineer]

I got nervous and decided to buy the mouser one too. lol I was thinking that in the future I might want to have two anyway if I decide to replace the metal halides with white COBs in the center anyhow.

 

[oreo54]

I got nervous and decided to buy the mouser one too. lol I was thinking that in the future I might want to have two anyway if I decide to replace the metal halides with white COBs in the center anyhow.

LOL, now Mouser has none.. for now.

Though I don't use them (anymore) I was glad when Meanwell came out w/ the
"AB" models.
It was kind of a dumb choice between dimming or current/voltage adjustments.

I generally don't encourage high dc voltage drivers preferring less over a tank but they have their uses.
Less so in fw when you are looking at lower wattages overall.
Besides I kind of like the little Meanwell ldd's.. or custom driver boards

Last huge array I saw someone attempt to build with 3w emitters and high voltage dc strings was a nightmare for them.
Never got it working. Initial diode strings had issues and trying to troubleshoot literally about a 75 diodes (well really a LOT) was too much for him. Now just a funny story for me.
You have it MUCH easier w/ only 12.

That said a ldd-lw (low current) and a 24v power supply and 2 probes makes a nice troubleshooting tool.
With the right power supply one can test multiple diodes in series. Or just one.
VOM is good enough for 3V-ish diodes but doesn't work for bigger ones.

 

[geekengineer]

So on the one I ordered from TRC I ended up calling them and they called back and said they have a shipment of those coming in 2 to 4 weeks. I decided to go ahead with that one so I have a spare or if I want to replace the metal halides later. Heat sinks are coming this afternoon and the supply tomorrow. So hopefully I can make a bit of progress with it.

 

[oreo54]

So on the one I ordered from TRC I ended up calling them and they called back and said they have a shipment of those coming in 2 to 4 weeks. I decided to go ahead with that one so I have a spare or if I want to replace the metal halides later. Heat sinks are coming this afternoon and the supply tomorrow. So hopefully I can make a bit of progress with it.

Want to be "cutting edge" space the cobs like 2 not 1.
You will get a more even distribution with some version of 2.
Rough drawing.

Wish there was a mathematical formula (probably could be) for precise spacing.

Each "gap" should be progressively smaller.

 

[geekengineer]

That is a good idea I would have probably though to be sort of OCD with it and do it all the same. lol

I fixed some of the wiring so I could use some of it later if I need to and got the heatsinks put in I will have to work the COBs around the mounting bolts a bit so hopefully that isn't a big deal. I have an extra heat sink I could put down the middle if I decide to go all LED later. I put it in the middle so you could see its fins. Hope this works and it doesn't get too hot. the hood has push pull fans so it will be getting good amount of air through it.

 

[geekengineer]

Well that sucks the driver was suppose to come in but I never got it yet the status hasn't updated or anything. way to go UPS. So I have no way to really test it unless maybe I get brave and hook it up to the mains. lol

 

[geekengineer]

Here's what I got so far. I am worried they lost my package with the driver doesn't give a date or anything. I ran out of heat shrink tho. lol

 

[oreo54]

Well that sucks the driver was suppose to come in but I never got it yet the status hasn't updated or anything. way to go UPS. So I have no way to really test it unless maybe I get brave and hook it up to the mains. lol

Sometimes ups transfers it over to the usps for the "last mile"
Believe you can put the ups number in the usps tracking box.

UPS last mile delivery USPS

UPS and FedEx will often transport orders hundreds or even thousands of miles before passing the package on to USPS to complete the final mile of delivery.

 

[geekengineer]

Well this morning it says it's out for delivery now so hopefully it's on the way now.

 

[geekengineer]

OK so finally go the driver and got it working used the original ballast case for the driver hopefully there would be room for a second if needed.

 

[oreo54]

Yea... Got a point you can measure the voltage across the cobs?
What did you set the drive current at?

 

[geekengineer]

I set the current at 733mA. I though that would be a bit more but not too much hopefully. the voltage it was running at was about 113 Volts so around 82 watts.

 

[oreo54]

I set the current at 733mA. I though that would be a bit more but not too much hopefully. the voltage it was running at was about 113 Volts so around 82 watts.

Sounds good...

 

[oreo54]

BXRC-65S4001-C-7x
34.4v @ 1170mA. 4508 lumens.
$12.57 each, $50.28 for four, plus shipping.

https://www.digikey.com/en/products/detail/bridgelux/BXRE-65S4001-C-74/11203636

If you want to replace the mh's this would be a good start. These are 6500k Iwasaki " like"

Only 4 per the same driver as you have
18,032 total lumens. Yea lumens isn't the correct measurement but it is a proxy for photons.

Rough calc estimates about 550 par (Bridgelux only)
Used 6 sq feet ( tank water surface area, all lumens spread out over that) to convert lumens to lux. High cri 6500k white to convert lux to ppfd.

Now you need one more driver to add your blue/violet/uV,IR whatever channel so as to not be a daylight reef.

Btw now that you have a better estimate of your blue string you " may" be able to find a cheaper dimmable driver freeing up that one for a bigger work load.

Composite and info..
Bright reversed colors are the Iwasaki w/ no Bridgelux component.
Dull natural colors are Bridgelux exceeding Iwaski component..
To replace the 300w MH's takes a bit of work.

3rd channel added, other blues off..Blue and violets are 1W each (40/20 total)

True 12000k and a better match for the Iwasaki.

Note 9 "Bridgelux" Data was created using smaller cobs so I needed to compensate for the larger ones.
Dotted line is ocean at 5M.
KISS was part of the equation.

All leds..
60 + 80 + 160W = 300W

Sorry for the mess but was just an interesting puzzle.. No guarantees as to it is what "you" like though.
Feel free to ignore it all.
I use Iwasakis spectrum as to keep estimates in the theoretical ballpark. High growth poor color or so I am told..

Oh in case you were wondering.. ALL on est. spectrum
Used 80W (1w emitters) of RB. Since graphs are dynamic the little triangle coverage varys by peak height.
So usually can be ignored, at least in my opinion.
As designed you are only using 220W vs 300W w/ JUST the mh substitution.

 

[geekengineer]

So I was hoping for something a bit simpler so I would only need two drivers mainly cause thats all the room I have in the converter box and I have another one on order. I did already have a bunch of 10W 6.5K COBs but I was thinking of maybe getting some 10W 10-15K COBs to mix up with them so like 16 total of those maybe 8 6.5K and 8 10-15K. I assume that isn't going to work very well. I was also assuming I would have the acnitic blue ones I have already on or at least partially as well.