Modding a Vipraspectra, color opinions?

[dirtyd61]

In the process of modifying a VipraSpectra v165.

1. Welded up a stand mounted arm 0-16” is my height adjustment options with X/Y axis leveling.
2. Printed some shades.
3. Made a diffuser spaced 1” off the light housing.

I plan to remove eight 12k and two 6500 white LEDs from channel one and replace with four 430nm Violets and six 450nm Royal Blues. Any better color combo opinions?

Lastly going to relocate the Power supply’s to the cabinet to get the AC power way from water and possibly switch to quiet Noctua fans.

 

[Gumbies R Us]

@oreo54

 

[oreo54]

In the process of modifying a VipraSpectra v165.

1. Welded up a stand mounted arm 0-16” is my height adjustment options with X/Y axis leveling.
2. Printed some shades.
3. Made a diffuser spaced 1” off the light housing.

I plan to remove eight 12k and two 6500 white LEDs from channel one and replace with four 430nm Violets and six 450nm Royal Blues. Any better color combo opinions?

Lastly going to relocate the Power supply’s to the cabinet to get the AC power way from water and possibly switch to quiet Noctua fans.

Well first things first and probably something you already know.
1) The drivers have limits to the voltage it is VERY unlikely that your changes will exceed (you won't go under since you will be increasing voltage) but it should be kept in mind.
Royals = whits in voltage but the 430's will probably have a higher voltage. With just 4 it shouldn't make much of a difference and unlike the "blue" channel your reds have already dropped the cumulative voltage.
3.2-3.4 now is 3.5-4v. So you may add like max 3.2V. Like I said just something to be aware of.

Second moving all the line voltage out of the box involves a lot of wire modifications.
Not difficult just messy. It's a good idea. Do suggest using like solid copper bell wire in like the correct number of conductors. Usually 18ga. Max 55w on each driver to led board. Voltage can reach like 90volts @550mA.
18ga IS overkill. Calcs say like 25 -20 ga depending on acceptable voltage loss and wire length.
https://www.omnicalculator.com/physics/dc-wire-size
Ignore the copper board. That is an "add on". Anyway all the wiring from the driver will need to be extended. Se bundle on top of black brick. 2 fans 1 from each driver, so 4 wires there. 2 wires for each led board, another 4
Well I give up here.
Below image from here:

Thread 'Controlling Viparspectra Lights'

Mar 8, 2020

I was inspired by @AbjectMaelstroM's reef-pi thread controlling Viparspectra lights and @ph123uk's thread controlling Viparspectra. I picked up a couple of lights recently and I'm finally getting around to checking them out.

I spent a bit of time characterizing the lights today. Of course, I found the same things the others found as well, which is good news. The documentation on the board matches the pin out.

The light uses 200 Hz PWM at 3.3v.

It also works at other frequencies; I tested 100 Hz - 3 kHz. It flashes once you get...

• Michael Lane
• Replies: 148
• Forum: Do It Yourself (DIY)

• 

Now on to your ACTUAL question.
It's impossible to say without knowing what one likes and goals. AND what the Viparspectra contains "currently". Why currently ? They change these things on a whim.
I suspect you did the AI thing and came up with this:

• Channel 1 (Blues/Actinics): Contains Royal Blue (450 nm), Blue (470 nm), and Violet (420 nm) LEDs.
• Channel 2 (Daylight/Full Spectrum): Contains Cool White (12,000 K & 6,500 K), Green (520 nm), and Deep Red (660 nm) LED

Go to Viparspectra and you get this:

ViparSpectra® Updated V165 110W LED​

What does that mean? did they change the spectrum?
Point is you may not even know what you are starting with.


BTW there are plenty of "black box mods" that will offer guidance on swapping colors.

But lets play..
Currently that channel is probably like 8000k and bluish.
Decent color rendering for a "shallow" reef.
Your mods will push it to like this.
WHY did you pick what you did?
You just added more royal but also added a different violet.
Puts it into like the "burple" range.
Not saying it's wrong. It's a choice. Curious as to why.
If you want different violets it might be best off just modifying the blue channel to be a broader spectrum.
Like I said depends on "why" you want to change.

 

[oreo54]

Reference:

Thread 'The Modified Black Box Thread'

Jan 16, 2017

So, over the past year I've been changing LEDs and lenses to find a good spectrum, light dispersment, limited shimmer and most importantly less "disco" effect, on my 165w wifi black box. I think I've found a combination. I decided to keep the 120° lenses and mixed 90° beaded and not beaded lenses. Though I dont have coral growth progression shots, this will show you more of what I've done.

Before shot. Ignore the dead looking LED. I experimented with UV LED on a white channel. (failed)


AFTER.


Since this change, I've seen some very possitive changes in polyp...

• reeferfoxx
• 165w black box chinese led mars aqua modified
• Replies: 595
• Forum: Aquarium Lighting

• 

10 yrs ago this was a Viparspectra..
They apparently dropped the 410's
Which brings up another point, violets have a tendency to a much shorter lifespan. They dim faster.

 

[dirtyd61]

Reference:

Thread 'The Modified Black Box Thread'

Jan 16, 2017

So, over the past year I've been changing LEDs and lenses to find a good spectrum, light dispersment, limited shimmer and most importantly less "disco" effect, on my 165w wifi black box. I think I've found a combination. I decided to keep the 120° lenses and mixed 90° beaded and not beaded lenses. Though I dont have coral growth progression shots, this will show you more of what I've done.

Before shot. Ignore the dead looking LED. I experimented with UV LED on a white channel. (failed)


AFTER.


Since this change, I've seen some very possitive changes in polyp...

• reeferfoxx
• 165w black box chinese led mars aqua modified
• Replies: 595
• Forum: Aquarium Lighting

• 

10 yrs ago this was a Viparspectra..
They apparently dropped the 410's
Which brings up another point, violets have a tendency to a much shorter lifespan. They dim faster.

Validated the fV on the replacement diodes so should be acceptable. Not too worried about the wiring portion extension, more electrically inclined in that aspect than your average person. The LED replacement/reflow (and any futures if they burn out or dont look right) will be done at my friends electrical engineering company nearby. So I have access to board heaters, bench top power supplies, high-end equipment, and skilled labor there ha. Im sure if there is any issues with extending the power bricks, he can provide any guidance.

I have the "Updated" light which supposedly was primarily the timer control % increments, but if they have changed LEDs (whites are all I care about right now) Ill do the best with polarized glasses to try and determine what is Cool and what is Warm white. From everything I have read, people have requested they change the white balance, but it has not happened. Seems they intended to keep the fixture more general use, both fresh and salt as well as different worldwide market prefs in perceived color.

The choice for the royal blues and the violets (specifically 420-430nm) was based on two things; 1. Research on best spectrums for coral absorption and, 2. What other major manufactures are doing in terms of total percentage break down between colors/spectrum. In every case, comparing to Radion/Kessil/etc, the Vipra came in ~10+% higher in cool/warm whites than what anyone else is doing. My overall % change should bring the fixture closer to what seems to be the average standard. That, paired with the almost unanimous consensus across forums that the white channel is nearly unusable at anything beyond 5ish percent (most people staying at 1%) led me to want to make better use of that channel. Im working with a 32g cube (softies/some LPS), so I dont have the dimensions or the correct coral to be able to crank channel % up without some consequence.

With the amount of white that will still remain + the diffuser, I believed that the tank should still maintain a reasonable perceived color...but I could be wrong and we may end up making future changes.

 

[oreo54]

De soldering / removing the egg diodes isn't an issue. If the center slug is soldered normally you just twist them (and usually break them ) off. No need to get fancy and reflow the whole board or even hot air for that matter.

Lift the ears and twist.. :)

Don't forget the thermal compound with the new ones.

The board traces appear pretty hardy .. BUT.. YMMV

This is an extension of what you already have likely read.
There are minor differences in "philosophy" like the use of amber ( see 2-3 )
https://www.blueharbor.co.jp/en/Coral-Color-Management.html

However, until now, the importance of UV has not been well known, as only GFP and RFP were recognized by the public as fluorescent colors in corals. In addition, UV-LED emitters are very expensive and can melt the lens if not carefully product designed, so aquarium makers have not been very active in adopting UV-LED emitters. As a result, the general aquarium LED products, only blue light of 450nm is very strong and UV light is very weak.


Then, to ensure Wide-Band-Blue, you need to adjust your LED light's 5 channels (UV 400nm/Violet 420nm/DeepBlue 450nm/Blue 470nm/Cyan 500nm) to the same wavelength intensity. For most LED lights, you will need to weaken the 450nm channel considerably. Also, if you have an LED light that does not have a UV channel, you may want to add a UV spot lamp.

Problem is most plastics have a severe cutoff on transmission at nm below 400.

You know there is this catch that "average/standard" includes buyer perception and costs as well as effectiveness.;)

I'm in no way criticizing what you are doing nor suggestion you "need" to do it another way.. This is just options..
A quote from that 30pg "modify my light" thread above..

Honestly, without access to a spectrometer, this is all "Kentucky windage" at best. Meaning, there is no way to tell. All i know is that viparspectras grow SPS just fine and that the white channel is under-utilized; so I'm adding more of what seems to work on the blue channel, to the white channel.

My only personal take is getting rid of as many high K whites as you can.
R(orange, amber )G(cyan)B (violet) = white without the yellow.

BUT you did ask for "suggestions"..
So here your new 14000k "white" channel..
Cyan to feed the Perdinin
Amber per the Japanese (also see peak in t5 lamps like the blue plus)
Your violet and a bunch of R/G for blue to make white.
Limiting whites to 8000k or so and a low number to boot.
EDIT..just leave the 12000k's won't make much difference
IR to add the new shiny nm's. Peak may be distorted but it doesn't effect color ect.

Notice no real UV..that's your call.
Yes it may look "windexy" but that's what the other channels for.
Also visually cyan is very visible to he human eye. It is hard to predict the look
over the data.
Ignore power functions for these are 1W emitters.



both at 100%

 

[dirtyd61]

De soldering / removing the egg diodes isn't an issue. If the center slug is soldered normally you just twist them (and usually break them ) off. No need to get fancy and reflow the whole board or even hot air for that matter.

Lift the ears and twist.. :)

Don't forget the thermal compound with the new ones.

The board traces appear pretty hardy .. BUT.. YMMV

This is an extension of what you already have likely read.
There are minor differences in "philosophy" like the use of amber ( see 2-3 )
https://www.blueharbor.co.jp/en/Coral-Color-Management.html
Problem is most plastics have a severe cutoff on transmission at nm below 400.

You know there is this catch that "average/standard" includes buyer perception and costs as well as effectiveness.;)

I'm in no way criticizing what you are doing nor suggestion you "need" to do it another way.. This is just options..
A quote from that 30pg "modify my light" thread above..


My only personal take is getting rid of as many high K whites as you can.
R(orange, amber )G(cyan)B (violet) = white without the yellow.

BUT you did ask for "suggestions"..
So here your new 14000k "white" channel..
Cyan to feed the Perdinin
Amber per the Japanese (also see peak in t5 lamps like the blue plus)
Your violet and a bunch of R/G for blue to make white.
Limiting whites to 8000k or so and a low number to boot.
EDIT..just leave the 12000k's won't make much difference
IR to add the new shiny nm's. Peak may be distorted but it doesn't effect color ect.

Notice no real UV..that's your call.
Yes it may look "windexy" but that's what the other channels for.
Also visually cyan is very visible to he human eye. It is hard to predict the look
over the data.
Ignore power functions for these are 1W emitters.



both at 100%

We did some buried lead work on a old Nanobox array recently, these do look much easier. But since he is offering to do the work Im just letting him remove them how ever he pleases and not shoulder surf too much. Id take a crack at it but his soldering equipment is just so much better for this smaller work than what I have.

I did consider some UV's but hadn't sourced them yet. Looking over your color rec's, trying to clarify on that edit. Are you suggesting removing as many of the 12ks as possible first and leave the 6500s initially in my 10 diode swap? I had been planning to remove mostly 12ks and try and keep the 6500s, being more of a full spectrum diode, I could remove exclusively 12s.

For your complete channel replacement, I'll see what I can find for availability in those spectrums. It does look good, but that may have to be a v2 design for me. Only due to the fact I wasnt initially planning on an entire channel replacement.

 

[oreo54]

You will notice some similarity to the ati aqua blue special.
T5's don't add " white". It's all based on rgb.
The "gold standard " t5 is the blue plus.
Basically amber, broad blue and green .
Aqua blue special below.

Use whatever whites you feel comfortable with.
Actually the higher the k the less yellow but the more it is " just " a dirty royal blue.

Well you keep 8 out of the 28.
4 12000k, 2 green, 2 red.
If uncomfortable with 4 660nm reds lower nm reds are fine and maybe even preferred.
As long as you have green and red you have white.

 

[dirtyd61]

I will try the full no-white at some point or mod a second light. Having a little trouble finding the ambers but will keep digging. Figure if I stick to the retention of the original white, add some True UV and prevent any similar colors from sitting side-by-side id have a swap of:

Add:
2 - 395nm
4 - 430nm
2 - 450nm
2 - 495nm

Remove:
6 - 12k
4 - 6500k

Layout ending up something like this:

 

[oreo54]

There is a shortage in the states apparently..
Had to go to Aliexpress for a better selection. Cheaper too.. maybe .. Tariff stuff drives me crazy.

 

[oreo54]

There is something weird about all these chips though. Except for whites most of the "3w class" have a low
mA rating.. Like 3.4V @ 500mA (max). Most black box drivers are 550mA. Now not a great difference but technically they are listing 3w leds that are supposedly only good for 1.7 watts.

3.4 x .5 = 1.7 Watts.
I don't remember the tables being sooo low. Usually like 700mA max. Can't use "5w class" because the voltage doubles.

Anyways did this to give you the Emerson enhancement effect 4 IR leds (Amazon has) and cut down on the white a bit. You can change the placement.
Not sure how scientific the below is but.. can't hurt.. and you have too much white anyways..

 

[dirtyd61]

There is something weird about all these chips though. Except for whites most of the "3w class" have a low
mA rating.. Like 3.4V @ 500mA (max). Most black box drivers are 550mA. Now not a great difference but technically they are listing 3w leds that are supposedly only good for 1.7 watts.

3.4 x .5 = 1.7 Watts.
I don't remember the tables being sooo low. Usually like 700mA max. Can't use "5w class" because the voltage doubles.

Anyways did this to give you the Emerson enhancement effect 4 IR leds (Amazon has) and cut down on the white a bit. You can change the placement.
Not sure how scientific the below is but.. can't hurt.. and you have too much white anyways..

I’m very been able to find the 700ma rated modules on eBay. Straight outta China..

 

[oreo54]

I’m very been able to find the 700ma rated modules on eBay. Straight outta China..

And the voltage is correct?

 

[dirtyd61]

And the voltage is correct?

Looks to be. This shop has a decent selection but not everything. https://ebay.io/m/vqbItM

 

[oreo54]

Looks to be. This shop has a decent selection but not everything. https://ebay.io/m/vqbItM

Great, yea looks good. I actually didn't believe the Chanzon chart. DIY-ers been buying those for a long time.

 

[dirtyd61]

Great, yea looks good. I actually didn't believe the Chanzon chart. DIY-ers been buying those for a long time.

This might have something to do with it. The amazon listings are all over the place on the IR (700ma with fV 2.4 = 3w. What?). Looks like on their spec sheet they actually claim its a 500ma diode but theyve been tested at 700ma. So might be the reason some claim 500 and some claim 700. Interesting to note the Wavelength creeps as you increase voltage. Probably why these LEDs are usually rated in a range ex 420-430nm. Wish these were just available on Mouser or Digikey so you got legitimate datasheets lol.



Amazon IR LED Link

 

[oreo54]

This might have something to do with it. The amazon listings are all over the place on the IR (700ma with fV 2.4 = 3w. What?). Looks like on their spec sheet they actually claim its a 500ma diode but theyve been tested at 700ma. So might be the reason some claim 500 and some claim 700. Interesting to note the Wavelength creeps as you increase voltage. Probably why these LEDs are usually rated in a range ex 420-430nm. Wish these were just available on Mouser or Digikey so you got legitimate datasheets lol.



Amazon IR LED Link

I've gotten into the habit of just using "class" so as to not put a solid number to them.
Most do not ever even get to the top of the class.. pun intended.
"3W class" ect.

And of course all depends on bin.
Even Mouser wouldn't guarantee me a bin when buying luxeon emitters.

Only seller I ever really trusted as to real exact data was Steves LEDs
My first builds used a bunch of cheap leds off ebay except for a few "must haves" for me. Like 660nm red and cyan.
For a while China was selling a blue/yellow twin as "cyan". You could see the double peak and of course little real cyan.

 

[dirtyd61]

I've gotten into the habit of just using "class" so as to not put a solid number to them.
Most do not ever even get to the top of the class.. pun intended.
"3W class" ect.

And of course all depends on bin.
Even Mouser wouldn't guarantee me a bin when buying luxeon emitters.

Only seller I ever really trusted as to real exact data was Steves LEDs
My first builds used a bunch of cheap leds off ebay except for a few "must haves" for me. Like 660nm red and cyan.
For a while China was selling a blue/yellow twin as "cyan". You could see the double peak and of course little real cyan.

Wish they still had the Lux T/TXs. What’re you using now? Lux C’s?

 

[oreo54]

Wish they still had the Lux T/TXs. What’re you using now? Lux C’s?

Last thing I used were the z's. Tiny little things. 1.5mm sq I believe
Shot through glass so there is a ghost image. It's only 7 diodes, 5
channels.





Dimming the 660 red channel improves color "naturalness".

 

[dirtyd61]

This is my v3 revision currently running on the tank. V2 was still too white for me.