Mixing colors in a string blade using four strings.

[Sincenatic]

It is difficult to know if I really had “a new idea”, but I will at least share some details on an experiment. Space Windu has made a lot of interesting LED-string blades and I found the color mixing blades particularly interesting. (See for example his viridian, ice-blue and platinum). Clearly these blades are the source of inspiration for my experiment. Since the procedure I used was a little bit different, I think you may find it interesting. I have made a white-blue mixing, where I use four 3mm LEDs in a square, where the LEDs in diagonal (seen from above) are of the same color. The idea is that the blue color always will have a white beside and behind and that this will blend the colors.

Building technique
I show the procedure in the photo and I add some explanations below. First I consistently start to bend out the longer leg, and for the white I bend the shorter upwards, while for the blue it is downwards. To fit the LEDs in a square I sanded by hand the inner part. I drilled a hole in a wood board to make it easier to drop a little super-glue in the middle. This was an important improvement as I started out using the helping hands – which was really slow, and also the LEDs sometimes where not perfectly on the same height. Notice that the legs that are closer are always negative leads.




The next step is to join these legs, and I did this by bending one below (and upwards) and thereby using it to bend down the other. I soldered these connections and cut away the ones pointing upwards. I call this square of four LEDs a set.  Next I joined these by bending them by hand. Basically I introduced one set into another and bended out the legs of the set in the position above. Once all the four legs were more or less in position I used a tool to press them together. When I started I had the tendency to get the sets too close to each other, but I later improved my control of the distance.
 
For the blade I made 6 segments, each consisting of two sections of 9 sets. Accordingly, this means 18*4=72 LEDs per segment and a total of 432 LEDs for the blade. Since the sections of 9 sets had different length I simply ordered them and used the shortest closer to the hilt. I also used super glue to make the string blade rigid, even between segments.
 
Specification of LEDs
The LEDs that I have used are actually very cheap, but I have no complaints about them. They are flat top with the following specifications; Blue, 3.2-3.4V, 5000-6000 mcd, White, 3.2-3.4V, 16000-20000 mcd. It says "Power Angle" 120-140 degrees, but I am not sure if they actually refer to viewing angle. Price with shipping is less than 6 Euros for 100. (I am not sure about rules on the forum, but if I am allowed, I will be happy to share the source).

I had bought a lot of blue LEDs, but only two packages of white. The packages actually included more than 100 in each, which made it possible to build the blade including 216 white LEDs. I lost a few on the way, but almost all was due to bending and gluing mistakes. 

Evaluating the result
I am overall very satisfied with the work. The technique gives a fairly good blending of the colors, but I work with a sanded polycarbonate tube, and the diffusion is not perfect. I have noticed that with lower power it becomes worse. I used a foam-wrap for diffusion, but I didn’t include that many layers, because I wanted to be able to insert the blade in the polycarbonate tube without creating wrinkles. This means that I have less diffusion in the blade compared to a try-out (of only five sets) that I initially did. In the try-out the blending was perfect. I include some photos below. I make some comparisons with a green blade that consists of 80 5mm LEDs; 540PG2C, (3.0-3.6V, mcd 16000, 40 degrees viewing angle.) The photos of the blade were taken during daytime (but in shadow), except the last photo, which obviously is during the night.









Difficulties to keep in mind
When I had finished, I had used all of my white LEDs, which meant that I had to use some sets that are a little inconsistent in height (for example a LED turn out to be slightly above the others). This is something that should be avoided, because the idea is to have the LEDS in a set exactly on the same level.

In the process I made some poor solder joints, and I had to spend some time testing and re-soldering. This is, of course, due to lack of experience and I improved on the way. On the good side, with more sources of light in each set, I do not find any problems of shadows on the blade, despite using quite a lot of solder.

I have some heat problems at the first section, and some of the foam seems to have melted. I have reduced the brightness a little bit. This problem could actually be related to the previous. I still have a poor contact in this position and when it is lost, the rest of the LEDs will get too much current.

I was also a little careless when wiring the common positive and the negative to each segment. I could have made these connections smaller (and closer to the core). This would, possibly, have allowed some more foam for diffusion.

Initially I wanted to keep all legs to be able to fine-tune a test version with resistors to obtain a particular mixing, but I never tried this because it was too many legs and the bending was close to impossible. I also had an idea to make the segments of the different LED-strings inconsistent in length to obtain a smoother light-up effect. This was also too messy, and I dropped the idea.

Conclusions
The blade is very bright, and the blue is clearly present on the surroundings, while the blade itself is very pale (close to being white). Some blue parts are still visible to the eyes, but with minor improvements I think this building technique can provide a good “white core” blade. The improvements that should be done are: being more careful in soldering the positive and negative wires to allow a little more foam for diffusion and using a trans white blade or some other kind of “professional” blade tube, instead of the sanded polycarbonate tube. Maybe this would also allow for more foam to be included in the wrap, if wrinkles would be less visible. (I actually don’t have any experience of these options, so please let me know what you think!)
 
It is also possible that using white and green would be even better as the green LEDs that I have are brighter (12-14000 mcd) and closer to the white LEDs (16-20000mcd). I will make more blades mixing colors in this way. Next up is white and green, but I may also try blue and green.

I use the blade with the DIYino board and LSOS. Hopefully I will provide a video to show you how it looks like in “action”.

[Obi_1]

Hey William, wonderful work on that string, they way you assembled your string is certainly unique, I've read every source of information about strings a while back and never encountered anything like this, so you can rest assured that you made a new design which is worth to explore to those people fixed on string sabers. I have a fellow here in Germany, he is also on the forum, I will recommend your technique to him, he explained he wanted to make the brightest LED string blade, he had quite a success so far, with your design he might even come closer to his ultimate goal. 400+ LEDs, that is crazy!!! On a video we will never see the nuance differences like the core being white and getting bluish towards the balde periphery, but if it really looks like how you described it is closest to movie feeling. If you can control the white segments independent of the blue, xxx you can even pulse the white core and leave the blues at constant brightness, that would look awesome!

But you need to make a better sketch of the LED bending technique, from the few pics you've posted I could not decipher the method. Maybe a walkthrough with more pics, I think people might want that pretty soon! Don't keep us waiting! Most interesting LED linking method since I've first read about the arm-buddies.

[Sincenatic]

Controlling the white and blue separately would need an alternative way to connect the LEDs, because as it is now a positive for a white is connected to a positive for a blue (and the same for the negative). The only reason to maintain four leads is to make it sturdier, but I could actually cut away two more from each set.
 
The reason that the blade is slightly bluer on the edge is because that part only is the polycarbonate tube, while the core has its foam (which is more lit up). This effect could be lost with a trans white blade, but I still would like to try it, because I want a better blending in the core. Even though I used a fairly good camera, imperfections are simply not captured on the photos.
 
Concerning the bending technique when joining one set to another; unfortunately I didn’t take any photos. The arm-buddies technique is usually explained with some particular tool to bend at the correct distance. In this case it is not possible. Instead I found that what is important is to hold the sets fixed at a particular distance, and only loosely bend the legs. Once all are more or less in position I press them into a more fixed position. It’s difficult to explain, and I will document this step better the next time. I actually have more LEDs waiting in the postal office, so I am ready to start the project as soon as I can pick them up! What is missing is an alternative blade tube, but I will start to build and make sure to add more photos.

[Sincenatic]

I was a little bit quick to answer earlier, and I need to add a clarification; While it would be possible to cut away another two leads from each set, this would require joining the positive to the positive, and the negative to the negative below each set. This would be a little bit messy to do. Hence, the reason to keep four leads is both to make it more rigid and for convenience. On the other hand, if someone really would like to control blue and white separately, I do think it would be possible, without keeping all 8 leads. However, it would require some creativity below each set.

[EXAR KUN]

Looks great! Awesome experiment and advancement here.

Did you email me with this design and we discussed? I think that must be you because the same cross pattern. Great work bringing this into reality I know that is a lot of soldering and all the while hoping it will work. Very cool. Can't wait to see the next colors! And also more pics of this one in future.

[Sincenatic]

Yes, I commented about this idea in an e-mail directly to you some time ago. I only introduced it to the forum now, because I wanted to see how it worked in the saber with the DIYino board first. The next blade will be quicker to actually test, because now I already have a functional lightsaber. There are quite a lot of details that I just didn’t bother to document, partly because I was not sure how it would work. Now I know that the concept works, but there are of course things that can be improved. I will provide more details the next time, and all suggestions and recommendations are of course welcomed.

[Sincenatic]

I have still not started working on the next blade because I feel that I have to do some refinement of the current blade. I’ll provide you with some more information. The brightness parameter on the LSOS is now set to 255, which is its maximum. Despite this the voltage drop for each segment is about 6.2V, which implies 3.1V for each LED. 3.1V is actually the minimum FV for the LEDs according the information sheet. The typical FV is 3.4 and unfortunately I do not know its maximum. This is one explanation why my try-out example actually blended the colors better. I use two Keeppower 18650, 3000mAh, which are fairly good. There are better batteries, but I do not know if those would be able to provide more power, or if it is the DIYino board and LSOS that are providing the restriction.

I add some more photos, where I compare the experiment to the green blade mentioned above and also a red blade.







When using fully charged batteries I measure a voltage drop of 3.7 for the green blade, and I am actually overdriving the LEDs (which have typical FV as 3.0 and 3.6 as maximum FV). (We’ll see how long they survive). The red blade consists of 5mm (560PR2C), 4000mcd@20mA, 60 degrees viewing angle. With fully charged batteries the voltage drop is 2.5V (while the typical is 2.0V and the maximum FV is 2.6 for the LEDs). The daytime photos were taken at about 11.30 with blue sky but in shadow. The camera captures the green, but somewhat washed out due to being at daytime. For the blue-white mixed blade the camera still provides a white core effect, but there are imperfections that are left unnoticed in the photos. The blending is better when I am in configuration mode of LSOS (and only one segment lights up at each time). In that case the voltage drop can be as high as 7.6 for a segment, meaning 3.8 for the LEDs. The blending then comes as a cost of overdriving the LEDs which I prefer not to do for this blade. The alternative way to obtain a better blending is to improve the diffusion (but at a cost of reduced brightness). I am not that concern about reduced brightness because it is very bright as it is now. Accordingly, I will try a transparent white polycarbonate tube.

[Sethski]

This is awesome1 Still digesting it, thanks for sharing, loving the various innovations with string blades at the moment 

[Obi_1]

That blue blade looks great! Certainly the brightest among all of them.

As to the reduced voltage drop over the LEDs, with my blue string blade I recently found a new house for, I also only measure ~something between 6V and 7V (Serial II setup). The MOSFETs used by the DIYino has an Rdson (channel resistance) of max 0.2Ohm (typ 117mOhms), and a segment draws maybe a few hunderds of mA at most, so the voltage drop an max drive should be worst case 100mV. I think the main contributor is all the solderings and connectors used in the path over the LEDs.

To rule out any effect from the software I will write an interactive LED characterisation sketch, so that people can experiment with whatever setup they have. Because it's one thing what battery and therefore voltage you use and what really arrived at the load (LEDs).

[Sincenatic]

Today I got a shipment of an inner diffuser tube and a white transparent polycarbonate blade. I include a photo to show how they look like.



I replaced the sanded polycarbonate tube for this set-up, but I was still not satisfied with the blending of the colors. The next step was to remove some layers of my 0.4mm foam and replace it with baking paper. This improved the blending, but at a cost of brightness. I add some more photos.





In action mode the blending is now fairly good, but not perfect. The cyan tendency (on the photo) is an effect of the camera. In config mode the blending is almost perfect. I tried to make a similar photo (as I did before, in darkeness) to see if the change reduced brightness in a visible way. The difference is very small. In both cases the effect in person is less blue and it is the camera that exaggerates the blue tone. 

In the building process I used quite a lot of soldering, and I can definitely reduce the amount for the next experiment. Learning-by-doing is important here, and the first segment was worse. If anyone knows a way to allow LSOS to provide more power in action mode, please let me know. The blade itself is now finished, and I am ready to start the next; a green-white mix. I will document the steps...

[Wolf69]

Nice!
How did you roll your foam since it is 50cm only?
Whats your preference between baking paper and foam after your "experience"?

[Sincenatic]

The foam was 10m, 50 cm wide, and 0.4mm thick. It was not perforated at any place, so I just cut the length of the blade (out of the 10m). The string (square) is about 9,5mm, so when I start to role the string the wrap never uses all of the 50cm. I initially cut away about 15 cm (more or less), but later I reduced it so that the baking paper would fit. I prefer the foam, since it protects the blade in addition to diffuse it. The baking paper diffuses more, but only in extreme cases it is necessary. For example, without mixing colors I would not use it. I have just started the building of the green-white mix, and once the string is ready, I will start with the foam, and only if the blending is not perfect, I will replace some of it with baking paper.

[Wolf69]

Right!! How silly of me!

[Sincenatic]

Mixing colors in a string blade using four strings – green experiment.
A few weeks ago I finished another blade using four strings. I have delayed the report, simply because I’m not entirely satisfied. This time it is green – white mix. As you can see on the photo I made one important change.



A slightly changed idea…
I decided to make 60 sets (type A) using green on the diagonal and white on the off-diagonal and 48 sets (type B) with the opposite structure. The reason is that I wanted to have a white LED below and above of each green LED. I use 9 sets per section, and two sections becomes a segment. All of the sections start with set of type A, then a type B, and so on, and since there are 9 sets, the last is also a type A. When I connect two sections I rotate the upper section 90 degrees because I want to connect its two negative legs to two positive legs on the next section.  Accordingly, all the green LEDs will have white LEDs below, above, behind and also on the side. The idea is of course that the blending will improve.

Building details
Before I explain the results, I will explain some of the details of the building process. I am not going to repeat too much of what I already have explained, but I am trying to fill in on some details that I previously didn’t explain. First, when I have made a set according to 1), I take the upper positive leg, and bend it below the lower positive leg and then upwards. The lower positive leg is bent downwards. This concerns both the left side and the right side. I do the same for the negative legs. Notice that the negative legs always are close to each other, while the positive legs are further apart. This is also visible on 2) on the photo, where I have soldered and also cut way the legs that I had bent upwards. In 3) I grab the set (type a) in the negative leg that is slightly further apart from the other legs. I introduce another set (type b) from below. The legs from the upper set are twisted around the corresponding legs for the set below. I work on all of the 12 sections at the same time, so I spend some time bending, and then soldering all of them. Then I go on to add the third set in the same way.

As I explained above, to join two completed sections I rotate the upper section 90 degrees. Before joining the two negative to the two positive (with the standard LED string technique), I bend the last positive leg below its set to the other side, joining it with the other positive leg. This can be seen in 4). It should of course be done with a safe distance to negative legs, and if it becomes too close, some hot glue or transparent plastic can be used to avoid future problems. (If you notice the ugly solder joint in 4); the reason was that I forgot to bend the positive leg to the other side, before joining the two negative to two positive legs… and I had to cut it to do it afterwards! Note for the future: pay attention!). The positive leg, which comes out in the middle of all segments, will later be connected to become the common positive lead. In 5) you can see how I connect negative to negative below the set. All of the three remaining legs are cut away, but I actually use them to “grab” the first LED on the next segment. I use super glue to keep them in position. Even though I use these to “grab”, they should be cut on a safe distance. These negative leads all have their individual wire.

Evaluating the result
I include a photo of the results; first in shadow during daytime and then in darkness.
 


The blending of the colors is almost perfect. It looks good both in action mode and in configuration mode – although it is brighter (with a whiter tone) in the latter. While the technique provides a very good blending, I couldn´t help to be a little bit disappointed that the blade is very light green – not the white core that I was hoping to achieve. I guess this is a consequence of that the green LEDs are very bright and the brightness of the white LEDs was not enough to wash out the green in the core. The day-time photo represents fairly well the actual color.
 
While testing the string (without the foam) I had an incidence; a chip (W2S) got completely burned out. I am not sure about the reason for it; probably a short circuit. Afterwards I couldn’t find any problems in the blade itself. The board still works, but only with five segments. I could, of course, change the blade to only have five segments, and program the board to only use these. For the moment it is just left aside.

Conclusions
The idea to use four strings to mix colors performs very well to achieve a good blending. If I make another one, I would probably use less LEDs to keep down the power cost. The white and green mix does not provide a white core, but it is still possible to improve the technique. A natural next step is to work with different viewing angles for the white and the green LEDs. Usually very narrow viewing angle is avoided in led-strings because much of the light will be absorbed in the next LED, instead of improving the brightness of the blade. In this case it could actually be an advantage when it comes to the white LEDs. Basically I want the white light to remain in the blade, but for the green or blue LEDs I am more interested to use a wide viewing angle.
 
The idea with building four strings in sets is that each green would have a white behind it. The idea to use different sets of type A and type B is to improve the wash out effect from the LED below. With a narrow viewing angle this effect would improve, while the blending due to the first idea might become slightly worse. Basically it is a trade-off and another experiment is necessary... This would, however, require buying different LEDs so I will postpone this idea some time. (Feel free to try it!). Instead I have ordered 50 white and 50 blue 8mm, 100ma LEDs. I am actually not sure about the viewing angle, because the information says 40-50, while on another place it says 20-25. It will be another white-blue mix, but with one string. I might sand or drill out the blue LEDs to make the viewing angle wider, while the white LEDs will be used as they are. I have some 1-3 weeks of waiting of this shipment, so it will take some time until I can report on potential progress...   

[Obi_1]

Hi William, good job again, you have a real pioneer Spirit, ready to experiment and I totally admire the effort you are willing to spend in order to reach your goal. A ligtsaber which looks white inside and have the characteristic color only on the edge fascinates me too, that would be for sure the most screen accurate!
I understood from your first experiment that you used 3mm LEDs, so I'm just curious whz now you want to go for 8mm/100mA? I'm not sure if they are not too big to put 4 in a square formation and still be able to pull a blade tube over it? But the idea to Experiment with 2 different colors with different viewing agles is defintely unique, I'm looking Forward to the results.
Do you also have a Video showing your blades in Action?