Scrubber build

I'm really not sure where you got 9V from!!

I posted this

But that was for a driver that supplies a constant current and varies the voltage to maintain that voltage drop. A driver is designed to supply a constant current to whatever you connect to it, which is why there is a range of operation.

A battery is designed to supply a constant voltage and the current will vary based on the load resistance. An LED battery tester is meant to supply a roughly 3V constant for momentary testing of an LED only. 9V will blow any LED out completely, I suppose if you knew you had about 9V of drop across a string of LEDs you could test say 4 LEDs at once with a 9V battery. But not just one.

 

So were you able to determine what was causing the array to run dim? Was is one or both of the violets/blues?

 

I don't think that's the case honestly. I think if you are having issues with the violets, it's a quality issue. Violets are particularly finicky and are subject to delamination issues, at least that was the problem in the past with many brands. Steve's LEDs did a lot of research and testing and determined that the SemiLED hyper-violet performs the best in this regard, I started using that violet a few years back and have had very few issues.

The other aspect of it is how much you drive them, I would venture to guess that if you run them in parallel within the series string, you'll rarely have any issues, but if you run them full-current (700mA+) then you might run into some problems.

The other issue could be conduction to the heat sink. In general this would cause problems with any color LED running full rated current as the heat would not transfer to the heat sink efficiently. I've used the Berquist Thermal Tape from Rapid LED:

Berquist Thermal Pad (individual)
Berquist Thermal Pad (set of 10)

I've also used a one-part thermal adhesive from Steve's but it's kind or a pain to work with, you have to keep it refrigerated and then allow it to warm up to room temp before using without letting moisture get to it (which causes it to set up too fast) and it takes a few hours to cure, during which time, you can't move the LED or the thermal bond will be lost

Thermal Adhesive

You can use a 2-part adhesive that is available in many places, like this:

Arctic Alumina Thermal Adhesive

Or you can use screws and thermal grease, which many places sell, but I get mine from Steve's because they carry the best version of this IMO - I think that's the paste but I know it's not white like shown in this pic

Thermal Paste

Outside of that, your heat sink look more than adequate to me. I think what happened is that you had a partial delamination of one of your violets and this cause it to "choke" the current. The flashing when you power it on might be a symptom of this, it would seem that this is the issue because once you took those out of the series, the reds fired right up.

This is simple enough to test with 2 AA batteries and some tape and wire. Just take 2 batteries side-by-side and tape them together. Take a small piece of bare wire and lay it across one end connecting the + from one battery to the - of the other and tape that down. Then on the other end, take 2 wire leads about 5" long each (I would use small stranded wire for this, anything will do) and strip both ends of each wire about 1cm, then tape one end of each wire to a battery. Now you have 3V across the 2 leads. Touch these to the terminals of each LED for a second to test. + to + and - to -. But this might be difficult for yours since you sealed them in silicone...but this is something you will want to do as you unpackage the LEDs when you get them to make sure you don't have a bad one, then again after you solder them.

The soldering technique could be part of this also, you need to pre-tin the LEDs before you mount them to the heat sink. If you tin them while they are connected to the heat sink, you have to hold the soldering iron on longer to get the pad to temp so that it accepts solder, and this excess time can heat up the LED too much and shorten the life or damage it.





Then wiring them up on the heat sink is easier



These are older videos from when I made the Rev 1 but the soldering principles are the same

 

You want to heat up the pad with the iron (with the tip of the iron tinned with a bit of solder but just enough to coat it) and then push the solder on to the pad where the iron is touching - you need to heat the pad up so that the solder adheres to it. The technique you describe might work, but it could also result in a situation where the solder is only sitting on the pad and not making a full connection. This can result in what is referred to as a "cold" solder point.

That works OK in most cases, but it's better to put the wire on the pre-tin drop then put the iron on top of the wire to heat up the wire and transmit the heat through the wire into the solder so they reach the same temp, for the same reason (avoiding a cold joint)

Those are the hyper-violet, the lower bandwidth 415-420, 420-425, roughly the same

Not sure what mx4 is?

Are you trying to keep the LED in place with silicone? would not do this

Try LED Seal instead - it's a spray-on. The manufacturer has demonstrated this to run LEDs underwater actually, but I wouldn't.
LED Seal - Silicone Spray Sealant
You can get it on Amazon, ebay etc as well
https://www.amazon.com/LuxDrive-LED-Seal-Silicone-Sealant/dp/B00LWU1XVW
PS dries in 30 minutes

 

I use 20 AWG solid, for soldering IMO that works the best over stranded, unless you take the stranded wire and tin it first, and this is usually done best by using a solder pot but that just adds another layer of unnecessary complexity.

Wire Gauge Converter - AWG versus square mm

so 20 AWG ~= 0.75mm wire, can go larger like to 18 AWG / 1.0mm but any bigger is probably not necessary

 

If you're using the Arctic Alumna adhesive, you don't need anything else - that stuff is like glue, once it sets up, you're bonded to the heat sink so there is no reason to do anything else.

Thermal grease is another story, that requires something to hold the LED down and silicone is not appropriate, it allows movement and any air pockets that get in there will ruin the conductivity of heat. You need a tight fit, which is why there are cutouts on the star for a screw.

Also in that last pic you just posted I notice that the heat sink appears to be scratched up, did you do that? If so, why?

 

Nice!!

 

This is not necessarily accurate because he originally was using a Meanwell LPC-35-700 which is capable of powering a 48V drop, he just got the 20-700 recently

 

No, if you run any 2 LEDs in series, say 3.2V drop each, that's 6.4V drop. If you run this pair in parallel within a series string, it's only one 3.2V drop and the current splits between the two. So this does help reduce the overall voltage drop.

Correct, but if you take the 2 violets and put them in parallel, that's one 3.2V drop + the 25V = 28.2V total drop, and now you're withing operating range for the LPC-20-700.

This is a good point, but I've never seen this kind of warning from any of the LED suppliers that I've received a battery tester from. But I think the most important part is that you're not supposed to run LEDs on batteries like this for any length of time - it's meant to be a split second, momentary contact just to confirm that the LED has not been physically damaged and that it still operates. I literally hold one lead on one terminal of the LED and then brush the other lead across the other side and look for a "flash", which is all you need to confirm that the LED is good.