Jump to content
Cultivated Reef

Narrow Reflector for the NanoBox V3 array?


R_MC

Recommended Posts

Now you just need to put all the pucks on a servo controlled pan/tilt platform so you can emulate the angle of the sun as it rises and sets ;)

Right? ... But NO... I have to keep the scope down. This has already gotten out of hand.

Link to comment
  • Replies 173
  • Created
  • Last Reply

Right? ... But NO... I have to keep the scope down. This has already gotten out of hand.

I was gonna say.....

Link to comment

Unless you are using fully parabolic reflectors on the T5 you are going to have a lot of light loss putting the bulbs that high on the frame.

Link to comment

Hi Gus, I've run some simulations and the 5 degree reflector will recuperate a lot of the losses. It would work better if the fixture were narrower...but If it's not enough I will add an additional center reflector like /°| |°\ The extra 3" of space will give me this option. If this is still not enough I will angle the center reflectors so that it is more like /°\/°\. If that is STILL not enough I will create a few more bends - but my initial estimates have me thinking that a 5 degree, 7 inch extension of the existing parabolic reflectors will be more than enough. Once it's build up I'll show some working comparisons.

 

There are other ways I can modify the reflectors to get more throw, I think the depth of this fixture (~7 inches) will give me plenty to work with.

 

Also - I'll run some examples through a physical rendering engine tonight to show what can be expected of each.

Link to comment

Hi Gus, I've run some simulations and the 5 degree reflector will recuperate a lot of the losses. It would work better if the fixture were narrower...but If it's not enough I will add an additional center reflector like /°| |°\ The extra 3" of space will give me this option. If this is still not enough I will angle the center reflectors so that it is more like /°\/°\. If that is STILL not enough I will create a few more bends - but my initial estimates have me thinking that a 5 degree, 7 inch extension of the existing parabolic reflectors will be more than enough. Once it's build up I'll show some working comparisons.

 

There are other ways I can modify the reflectors to get more throw, I think the depth of this fixture (~7 inches) will give me plenty to work with.

 

Also - I'll run some examples through a physical rendering engine tonight to show what can be expected of each.

 

Why do all that though when you can just lower the bulbs? That way you have more room for ballasts and electronics above it plus some cooling.

Link to comment

If you read that before I edited it - ignore it.

 

The reason is that I need at least 7 inches of reflector to focus the light. Moving the light further down will result in more spill. At the current reflector depth (~7") almost no primary emission will spill out of the front or rear of the tank (sides are hopeless). The 5 degree 7" reflector can be considered to be an extension of the primary parabolic reflector design.

 

You don't typically see this because in industrial applications T5's / fluorescents are used for diffused lighting whereas point sources are used for high bay. It's true that I could go with a 150W halide instead, but that would restrict lighting options / draw more power / cause more heat. 75W is just too inefficient.

Link to comment

If you read that before I edited it - ignore it.

 

The reason is that I need at least 7 inches of reflector to focus the light. Moving the light further down will result in more spill. At the current reflector depth (~7") almost no primary emission will spill out of the front or rear of the tank (sides are hopeless). The 5 degree 7" reflector can be considered to be an extension of the primary parabolic reflector design.

 

You don't typically see this because in industrial applications T5's / fluorescents are used for diffused lighting whereas point sources are used for high bay. It's true that I could go with a 150W halide instead, but that would restrict lighting options / draw more power / cause more heat. 75W is just too inefficient.

 

That is not correct if the reflector is fully parabolic and the bulb sits super close to it. For example the reflector of an ATI Powemodule.

 

para2.jpg

 

The bulb sits almost touching the reflector on those units so the light spill is basically non existent as all photons are directed straight down no matter what.

 

The amount of side spill will be less with a reflector like that vs the amount of light loss going with what you propose.

Link to comment

Gus, that image only accounts for back radiation. The parabolic reflectors on the bulb will address back radiation, however in your diagram you have unfocused radiation coming off the remaining ~240 degrees. The front 40-50 degrees are ok, but that still leaves nearly 180 degrees / 50% wasted energy. Also, in my experience, the reflectors do not create perfect parallel projections as they're faceted / dealing with an area emitter rather than a point emitter.

Link to comment

vLtw6Wz.jpg

 

q5Gmjnq.jpg

 

Not a 100% perfect fit, but I think it's close enough. This cut-out will provide a mounting surface for the heatsink - allowing me to mount the entire assembly to the security camera bracket. The cut-out also provides a surface to secure the reflector.

Link to comment

Gus, that image only accounts for back radiation. The parabolic reflectors on the bulb will address back radiation, however in your diagram you have unfocused radiation coming off the remaining ~240 degrees. The front 40-50 degrees are ok, but that still leaves nearly 180 degrees / 50% wasted energy. Also, in my experience, the reflectors do not create perfect parallel projections as they're faceted / dealing with an area emitter rather than a point emitter.

I just used that picture as an example. The ATI full parabolic reflectors extend past the bulb by a big margin. That's why I asked if you were using full parabolic or quasi like the sunpower.

 

Also your housing with the 5 deg angle is basically a specular dome. Photons are going to be bounced all over the place.

 

My point was that if you want to maximize intensity and narrow spread a full parabolic reflector like the ATI would be best.

 

Also I don't see anything in your drawings about cooling. That dome is going to get hot without forced ventilation which is going to decrease bulb output significantly. T5 needs an ambient temp surrounding the bulb to be 35C to be most efficient. Hotter kills output and bulb life while cooler kills output. You are going to need a fan with controlled output pushing air out of the dome to keep it at optimal ambient temp.

Link to comment

Thanks gus - running some simulations now.

 

Here's an early preview:

 

EUYzDBi.png

 

 

 

In this example the right most image is run with reflectors only and the T5's are 7" lower. The center unit has reflectors at 5 degrees on the outside and 10 degrees in the center like / \/ \. The unit to the left has only the 5 degree side reflectors with no reflectors in the center.

 

4g3OoqI.jpg

 

Early on it seems that the center setup has an advantage in preventing light spill but not significantly increasing light levels in the tank over the T5's mounted 7" lower. In all examples the hood acts to minimize ambient illumination in the room. I think the reflector design can be further optimized - so I'm going to give that a shot. My guess is that if I used the middle options and angled both bulbs slightly in I would get even better results.

 

I'm using maxwell physical rendering engine and true photon simulation for this. I'd love to use a program that's more appropriate, but I feel that this particular physical rendering engine has given true to life results for me in the past. If only it had a par meter ;).

 

I'm going to let this simulation run over night and will post more accurate results in the morning.

 

 

Always learning here but what do you mean by specular dome?

Link to comment

Laser cut parts came in

 

PPRwkn9.jpg

 

 

Not gonna lie. First thing that popped into my head was piston connecting rods.

Link to comment

So, I let the simulation run for 8 hours or so, here is a more accurate result:

 

CLgrpf8.png

 

edit: woops - link included. Click to enlarge.

Link to comment

http://imgur.com/gallery/b3jDA/new

 

Here's a series of experiments showing different reflector layouts - best results were with the 10 degree center reflectors.

 

Setup:
d1Bj5g4.png

 

This image shows the 10 degree reflector with 3 perpendicular 3mm thick black blinds spaced evenly along the length of the fixture. The additional light shaping creates a projection with much less spill and a similar luminous intensity to the units mounted 7 inches lower.

 

g3JSojM.png

 

This last image is zoomed out to show how the light is distributed in a 20' x 30' space.

 

1uAD9eC.png

Link to comment

Specular means that the reflection angle of the photon is the same as the incoming source angle. By adjusting the angle slightly all it does it change the reflection angle a tiny bit. A full parabolic reflector because of the way it's designed has a slightly different reflection angle across the entire vertical plane and that is why all photons reflected by it will always go towards the same direction.

 

If you were to do a simulation with an actual parabolic dome you would notice that light spill is almost non-existant outside of the boundaries of the dome and that intensity would be greater.

 

With your specular dome photons are going to be reflected multiple times inside it before they exit and the angle of exit will be everywhere which is why you are getting quite a bit of light spill on some of your simulations. Not only that but all these multiple reflection points will decrease output by quite a bit.

 

A perfect example is the Giesemann Matrixx II T5 fixture. The reflectors are basically specular with one redirection angle which is still specular. Tests have shown the PAR output of unit to be lower than the Sunpower and considerably lower than the Powermodule/Diuna. It does have superior horizontal spread than the others though.

 

A simple test you could do is measure the LUX/PAR of the T5 reflectors all the way at the top of the dome and then again with them towards the bottom. You might be surprised at the difference in intensity for such a short distance.

 

The one thing you need ask yourself though is what you want to achieve with the T5. Do you want to maximize output or spread? If you care more about spread then your design takes full advantage of that.

 

Also ballast choice would help with output as well. There is no other ballast that you should be using on your build than the Phillips Centium Advanced ICN-2S39-T. Why? Because it's the only one out there with a ballast factor of 1.12 when using it with 24W bulbs. This means that a bulb will be 14% brighter than the HEP SI-218-40UNI (which is in ATI) and 24% brighter than Fulham Workhorse. Also the power factor stays at 1.0 so you will not be overdriving the bulb.

 

I also need to reiterate what I said earlier about cooling. That dome is going to get hot. If you want to maximize bulb output and life you need to take that into account.

Link to comment

I'm using T5's almost exclusively to supplement the LED spectrum. I think what I'll get out of the 5-10 degrees is about 30 percent recovered light that would have spilled just outside the edge of the tank. I'd agree with you completely if they were parallel, but 10 degrees makes a big difference.

 

Here's a very simple diagram which shows the difference a 6 degree reflector makes on the exiting beam angle. With a 0 degree reflector (or with the bulb positioned 7 inches lower) the exiting beam just grazes the edge of the tank. With a 6 degree reflector it strikes nearly at the middle. Consider this 5-10 degree reflector to be an extension of the last fin of the bulb's primary reflector. For all intents and purposes the 7" extension is a "parabolic" reflector.

 

ZzIoGTI.png?1

 

Could you point me in the right direction for retrofit reflectors with the best throw?

Link to comment

What are you lining the insides with to reflect the light? That will play a big part too. You going with something fancy like a sheet of Miro-Silver?

 

I will take out on my stashes of reflectors to show you one of the better ones I have found. It was super cheap too. Insanely similar to Sunpower reflectors. I have yet to find a DIY reflector like the Powermodule/Diuna.

Link to comment

Archived

This topic is now archived and is closed to further replies.

  • Recommended Discussions


×
×
  • Create New...