RGB LED Strip and Several Controllers

Some time ago, a friend asked me to write a review of one of his products. Yes, don't be surprised, it happens :)
And now I've finally gotten around to reviewing it. Unfortunately, links to some of the products are no longer active, but I think the review will still help you figure out what’s what.

This whole story with the controllers and the strip actually started last summer. As it happened, my friend thought one of the controllers worked over Wi-Fi. At least, that’s what the seller apparently claimed, as far as I understood. He also gave me several other controllers so I could compare them in a review, which is what I eventually decided to do.

As it happens, one of the controllers didn’t make it into the photo, but it will be covered in the review.


I’ll come back to the “smart” controller toward the end of the review. For now, let’s take a look at the strip.

The strip I ordered is RGB. That means it contains red, green, and blue LEDs.


More precisely, it uses 5050 three-color LEDs. Each LED contains three dies, one for each color.
I was careful to say LEDs of three colors above, because there are also strips with separate LEDs for each color. Those LEDs are usually smaller, but there are 3–4 times as many of them.

There are a lot of different kinds of LED strips. Here’s one way to group them:
1. LEDs per meter — 30 — 60 — 120 — 240
2. Supply voltage — 5 — 12 — 24 — 220
3. Color — red — green — blue — white (warm, cool, neutral) — RGB — RGBWW.
4. Protection — standard — sealed (silicone-coated).
5. Configuration — single-row — double-row
6. LED orientation — front-facing — side-facing.
7. LED type — through-hole — SMD
8. SMD LED package — 3014 — 3528 — 3825 — 5630 — 5730 — 5050.

Strictly speaking, these aren’t really different types, but variations in the components and construction. The strip under review is shown in bold.

There are also strips with “smart” LEDs that let you control each LED individually, but they require a compatible controller. Their low supply voltage also limits their use, since the current draw ends up being very high.

White LED strips are often used for task lighting. Here’s a tip: if you’re planning to install accent lighting, choose a high-density strip, such as 120 LEDs/m, and use a diffuser. For example, under-cabinet rails are popular in kitchens. If you use a low-density strip without a diffuser, you’ll see the LEDs reflected as bright dots, which is very unpleasant on the eyes.
For example, there are single-row strips with 240 LEDs per meter.




Also, silicone-coated strips aren’t always a good choice: silicone tends to darken over time and isn’t very easy to clean.
I’d recommend using aluminum channels with a diffuser instead. They cost more, but they’re more convenient and look better.

The strip consists of small sections, each with three LEDs and three resistors. The LEDs of each color are connected in series, and a resistor limits the current through them.
In this case, there’s one 330 Ω resistor and two 150 Ω resistors. The different values are due to the different forward voltages of the LEDs.


Let's first check the power. Here I also decided to show that LED strips have a nonlinear relationship between supply voltage and current consumption.
For example, I've come across questions like: “Will the strip work at 9 V?”
It will, but its power will drop dramatically.

So, let's test the strip in two modes, at 12 V and 10 V, and see how its power consumption changes.
You can also see that the power changes differently for LEDs of different colors.
1. Green: 13.8 and 6.75 W, a 2-fold difference.
2. Red: 15.3 and almost 9 W, a difference of about 1.7 times.


1. Blue: 12.2 and 5 W, a difference of almost 2.5 times.
2. All three colors together: 35.8 and 18.6 W, a difference of about 2 times.


The experiment showed that blue LEDs are more sensitive to a drop in voltage because they have the highest forward voltage. Red LEDs are the opposite, and show the smallest difference. With red LEDs, more voltage drops across the current-limiting resistor, leaving a small voltage margin.

What are the consequences of this voltage drop?
1. If you're trying to use a strip like this as a white light source (which is fundamentally wrong), the color spectrum will change toward the end of the strip because the voltage drops there: red will shine brighter and blue dimmer.
2. The overall brightness will simply drop toward the end of the strip.

I don't see any point in checking the first one, but I'll show the second. I've actually done this before in one of my reviews, but that was a regular white strip.
It's not very clear in the photo, but you can still see that the LEDs at the bottom are brighter than the ones at the top. It's not hard to guess that the LEDs at the top are at the end of the strip.


Here's another shot. The strip is shining so brightly that it makes photography difficult.


If you want to ensure uniform brightness along the entire strip, there's a very simple solution: connect the strip diagonally.
With this connection, the strip's overall brightness will remain roughly the same, but the unevenness will be gone.


Some of you might say, “How much voltage can really drop across a strip?” Quite a lot, as it turns out.
I applied 12 V to one end of the strip and measured the voltage at the other end.
1. Green: 3.1 V drop
2. Red: 2.5 V
3. Blue: 2.5 V
4. All four colors connected in parallel, with the strip in white-light mode: 2.7 V at the other end.
As you can see, even my test at 10 V doesn't tell the whole story. There, power dropped by about 1.7–2.5 times; here the voltage is even lower, so you can expect a 2–3-fold difference.


In some photos, you may notice that the strip's total power consumption varies, even though the power supply voltage is regulated. This is due to the LEDs warming up. The higher their temperature, the lower their forward voltage and the higher the strip's current consumption.
I didn't leave the strip on for long during the tests because I was testing it while still wound on the reel, and it gets noticeably hot in that configuration.
The thermal image shows the temperature rise over one minute.

By the way, people often write online that a cable wound on a reel heats up because of its inductance. Here's a clear example showing that it heats up simply because a large amount of energy is being dissipated in a very compact space. The same thing happens with an extension cord if you don't unwind it when it's carrying a heavy load.


But in reality, high-power strips can overheat even when unrolled, so special heatsinks are used for them.
These heatsinks can also usually be fitted with diffusers, mounting hardware, and end caps. So if you want your strip to last, get a heatsink for it, or at least stick it to a metal surface. After mounting it, I recommend checking the strip contacts and the heatsink for a short circuit.


Now let's move on to the controllers. In practice, only two of the four controllers I tested worked the same way, so I decided to put them through a few tests.

First, the simplest controller.
The manufacturer specifies a supply voltage of 12–24 V and a current of 18 A. But since there are 3 channels, that works out to 6 A per channel.
In most cases, this is more than enough current, since even at a 12 V supply voltage that's over 200 W.

The controller has three channels and comes in a neat little box.


The package includes:
1. Controller
2. Remote control
3. Double-sided tape
4. Manual.


The manual is in English, but you don't really need it. It says the controller has 20 operating modes.


I included this page from the manual only because it shows the wiring diagram.
It's simple: connect the strip's four contacts to the controller's four contacts.


My first thought when I saw the controller was, “It's a toy!” :)
It really does look tiny.

I'm not including links to the controllers shown in this review, since the links are already dead, and I don't think these controllers are any different from other similar ones.


The wires connect via screw terminals. You can supply power either through the terminal block or with a power supply that has a standard barrel plug.
Still, I have serious doubts that the terminal block can handle 18 A, let alone the connector. I think the real limit is 6–8 A with the terminal block and 4–5 A with the connector.

There was nothing interesting on the outside, so I opened it up. This was the first LED strip controller I'd ever had in my hands. I'd never come across one before, but there's a first time for everything.


The circuit board looks quite neat, and the terminal blocks are pretty good quality, so it might be fine up to 10 A.
The electrolytic capacitor on the board is a bit depressing, though. It reminded me of my first experience with a low-voltage PWM power controller, when I learned that capacitors can indeed get quite hot.


On the back of the board, you can see tinned sections of the traces to increase their cross-sectional area.
There are also lots of vias between the board layers. They don't help much, though, since most of the heat is conducted away from the transistor's two leads rather than its body.


The power stage is built around three MOSFETs NTD4963N.
These MOSFETs have an on-resistance of 9.6 mΩ. At a current of 6 A in an almost static operating mode, that means about 0.35 W of power dissipation. However, I didn't check the gate voltage (it's probably 4.5–5 V), so I'll also calculate for the worst-case scenario, with a 5 V supply. In that case, the datasheet specifies an on-resistance of 16 mΩ, or almost 0.6 W at a continuous current of 6 A.

For this package and this board, that's plenty of margin. I think the current could easily be increased to 8 A, although there's not much point; the MOSFETs have plenty of headroom.
A CD4050BM IC is used as the driver, and the 24C02 EEPROM is at the bottom right.

The whole thing is controlled by a microprocessor with its markings erased.
Remote control is handled by another IC, also with its markings erased, though I really don't understand the point of this kind of “encryption.”


The remote operates at 2.4 GHz and is powered by two AA batteries. It looks a bit like a bar of soap :)
The remote is entirely touch-operated, with no mechanical buttons at all, which I find very inconvenient.
The thing is, no matter how you hold it, you can still accidentally touch another sensor and switch to a different mode. Maybe it just takes practice, but I wasn't impressed.
There's a circular color touch sensor at the top. Run your finger around it to change the strip's color fairly smoothly.
At the bottom are six control sensors: brightness, transition speed, and effect selection.


I checked all the controllers for ripple. Actually, that's not quite right: all of them have ripple because they use PWM for dimming, so I checked two things:
1. The operating frequency, and therefore the ripple frequency.
2. Whether there was any ripple at 100% brightness.

They fail the first test: the PWM dimming frequency is only 125 Hz, which is low—very low. Fluorescent lamps with electromagnetic ballasts flicker at almost the same frequency. But lamps have phosphor afterglow; this strip doesn't, so I'd recommend this controller only for occasional use.


A short video about this controller. If you watch closely, you can see that the color transitions aren't very smooth; there aren't many color-mixing options.


The second controller is very similar to the first: a similar little box, just in a brighter color.
But this one is rated for four channels and a total current of 24 A.


The package is exactly the same as with the previous controller: controller, remote, instructions, and double-sided tape.


The instructions are also almost identical, but the effects are slightly different.


The device itself is almost identical too. The difference is the fourth channel, which controls a strip with a separate white channel, and the modified firmware.
With the first controller, selecting the lighting mode (white) turns on all three channels. Here, the three color channels turn off and only the white LEDs turn on.


The connections and construction are identical to those of the previous controller.


Although there are more changes on the board than just one extra transistor.
For example, the input capacitor is already a low-impedance type.


But the traces on the underside haven't been reinforced, even though the rated current is higher than on the previous version.


Overall, the board is assembled quite neatly.


Four 09N03 transistors are used. According to the datasheet I found, their maximum voltage is 25 V (so I don't recommend powering this controller from 24 V, as claimed), and their on-resistance is 9 or 12 mΩ, depending on the gate-drive voltage.
In terms of heat dissipation, it's much the same as with the previous controller, perhaps a little better, but not by much. So 6 A per output is quite realistic.
The same IC is used as the “driver.”

And, as before, there's a microcontroller with its markings erased, an EEPROM chip, and a radio receiver IC.


The remote is almost 100% identical, but the remotes aren't interchangeable: they presumably use different codes and don't interfere with each other.


The oscilloscope trace shows the same 125 Hz ripple and, again, no ripple at 100% brightness. This suggests that the controllers are identical, apart from a small change to the program for controlling the white-light channel, of course.


In this video, you can see that the strip goes dark when switching to lighting mode. That's normal, since the strip is RGB and the controller is RGBW.


This controller didn't make it into the group photo, and I actually forgot about it at first.
It's clearly different from the previous versions, at least on the outside.

It has a metal case, and the rated specs are the same as the first version: 18 A total, or up to 6 A per channel. There are three channels.


In my opinion, this version is a little better: you can screw the case onto something, and it uses more convenient, higher-quality terminal blocks. It also has a standard power connector.
/The strip and power connections are brought out to the terminal block.


As you can see in the photo, the terminal block has two parts. You connect the wires to one part, then plug that part into the controller. This makes wiring easier, especially in narrow recesses.
If you think the metal case is there for cooling, I'm afraid I have to disappoint you: the transistors don't just lack thermal contact with it—they're on the other side of the board altogether. Though judging by the previous versions, they don't need cooling anyway.


The board is neatly laid out. Since the case is metal and radio waves don't like passing through metal, the antenna is placed near the connector. In practice, this doesn't seem to affect the range much. Well, it does, but the range is still sufficient for use around the home.


As usual, the connectors were soldered on after the board itself was assembled, so you can see flux residue, and the traces haven't been reinforced.


The switching transistors are identical to those in the first controller. The board also has an unidentified microcontroller, EEPROM, and a radio receiver chip, but this time they have markings.
What it doesn't have is a gate driver for the MOSFETs, though at these low operating frequencies that hardly matters.


The remote, however, is completely different. I had to retake all the photos with this remote because it should be positioned with the buttons facing up. I only noticed when I realized the strip's brightness was adjusting in reverse :)
The manufacturer managed to get both things right and wrong here.
1. The good: the buttons aren't touch-sensitive. They're actually more convenient than touch controls because you can feel them before pressing them.
2. The bad: the color adjustment wheel is at the bottom, so it's easy to hit it with your hand while pressing the buttons. When that happens, the controller usually exits the last selected mode and switches to color adjustment mode. But it doesn't always respond this way; apparently it depends on the selected operating mode.

The remote is powered by 3 AAA batteries, which may be why its range is comparable to that of the controllers in plastic cases. The operating frequency is unknown. Judging by the antenna, I'd guess it's not 2.4 GHz like the previous ones, but around 433 MHz.


In terms of flicker, this controller is the worst of the lot. It has a low PWM frequency, and even at 100% brightness it can't provide continuous power, so you can see small dips in the right-hand oscilloscope trace (the trace is inverted).




A side-by-side photo of the remotes for the three controllers.


There was a reason I showed the remotes in the previous photo, even though I still had one more controller in reserve.
The next model doesn't come with a remote.

There was a mix-up when I bought this controller. A friend saw the 2.4 GHz operating frequency and the claim that it could be controlled from a smartphone, and assumed it used Wi-Fi. That was a perfectly understandable mistake, though I think if it supported Wi-Fi, the fact would be printed in big letters in the most prominent place.
On the plus side, the specs mention a microphone, programmable switching, and various other useful features.


The package is simple: the controller itself and an antenna. The controller is noticeably larger than the previous ones.


It became clear almost immediately while testing it that the controller works over Bluetooth: the first thing the app asked was, “Bluetooth is off. You should turn it on :)”
The range is surprisingly good; at least it worked everywhere in my apartment.


The strip and power connections use the same plug-in terminal blocks as the previous model.
On the other side are the power and antenna connectors, along with an LED (it blinks when there's no connection and stays on when the connection is established).


Assembled.


But I'm more interested in what's inside, which is why I decided to write this review in the first place.
The board is positioned in the case so that it can only be removed in one direction.


As you can see, the board is single-sided, with a microphone and several capacitors on top. The input capacitor is even smaller than the one in the first version of the controller. The board is made of phenolic paper laminate.


The power traces are generously coated with solder to increase their cross-section.
Overall build quality is mediocre.




Let's take a closer look inside.
1. The transistors appear to be ISL9N306AD3ST, if I've identified them correctly. Their specs are 30 V, 50 A, 6 mΩ. Not bad at all, except for one thing. The case is marked 30 A*3, which would formally mean three channels at 30 A each. Obviously that's complete nonsense; it should say 30 A/3, meaning three channels at 10 A each. But even a total current of 30 A would be more than the installed terminal blocks could handle, let alone the power connector.
The transistors themselves can handle 10 A without any problems and without additional cooling; at that current, they would dissipate up to 0.6 W.
The assembly and soldering are pretty dismal: the transistors are soldered haphazardly, and the rest doesn't look much better.

2. The transistors are driven by a ULN2003 IC, but it's not really suitable for this application. It provides the full gate voltage, but switches the transistors on slowly.

3. Microphone amplifier. I tested it, and it works, but its sensitivity isn't very high. Still, it should work if the controller is close to the sound source. It extracts the low frequencies from the audio signal, so the LEDs switch in time with the music. Overall, I'm not impressed.

4. Bluetooth module. At first, I didn't even notice that this controller doesn't actually have a microcontroller to manage its operating modes. It wasn't until I was preparing this review that I realized the smartphone handles not just the controls, but the entire operation. Essentially, they took a Bluetooth chip, connected the three LED channels and the microphone signal to its free I/O pins, and left the rest to the app. Not very convenient.


I also noticed fairly strong resonant interference at the device's output, caused by the transistors switching. This is partly because there are no diodes at the output to suppress these spikes—another cost-cutting measure.
Despite all its drawbacks, it does have some advantages:
1. The ripple frequency here is 1,000 times higher, at around 125 kHz.
2. There is no ripple at full brightness.
3. You can set the brightness very low; other controllers can't do that.

The high frequency is also a drawback: switching the transistors at this frequency is much harder, switching losses increase, and interference gets worse. A frequency of 1–10 kHz would be more optimal.


The app is very simple. I first tried downloading it from the app store, but it wouldn't even install. In the end, I went to the manufacturer's website and downloaded it there; after that, everything worked without a hitch.
The main menu lets you open the lighting settings, choose music (it simply plays music on your smartphone; nothing is sent to the controller), configure the timer, and open the connection menu.

You can connect to the controller while it's switched on.
I didn't really understand the timer at all. Having to keep your smartphone connected for it to work seems like a pretty clumsy idea.


The lighting control menu lets you turn on white (all three channels are on), and also emulates the color wheel found on standard controllers.
You can also adjust the brightness and LED switching frequency in effect mode.
The effects aren't very impressive, if you can put it that way. There are only four of them, technically; some respond to sound, but I didn't like them.

I didn't quite figure out the Lighting setting, though. As you adjust it up to halfway, it changes the strip's brightness from 0 to 100%; beyond that, it dims the light.




What can I say about all these controllers?
Personally, I wasn't too impressed with the coarse color-transition adjustment, and you can see it in the video.
The basic controllers have a low operating frequency, but they're completely standalone, unlike the Bluetooth version, which requires a smartphone to work.
All four controllers handle the rated current, but I have serious doubts that the power connectors can carry that much current.

In general, in my opinion, things like these are better suited to decorative lighting in shops, signs, and so on. Although my neighbors installed this kind of lighting at home, the point of it somewhat escapes me. It could work as festive home lighting, though—cheap and attractive.

The strip reviewed here is completely unsuitable for lighting, since white is essentially produced by three single-color LEDs. And combined with the low flicker frequency and their 100% duty cycle (when operating below 100% brightness), it's a complete mess.

A few tips:
1. If you plan to do more than decorate a room and want to use the strip for lighting, choose an RGBWW strip.
2. For accent lighting, choose a strip with a high LED density.
3. If the strip has a high power rating (roughly over 8–9 W/m), use a heatsink—there are plenty of different shapes available now.
4. A diffuser makes the light more even and individual LEDs less noticeable.
5. For uniform brightness, you can use diagonal wiring.
6. Not all controllers are worth using; it's better to choose one with a high PWM frequency. The simplest way to check is the “pencil test”: hold a pencil between two fingers and move it quickly. If you see distinct outlines of the pencil, that's bad.
7. In my experience, all the controllers I've tested are limited by the power connector, not by the transistors or their heating. You can easily increase the power by soldering wires from the power supply directly to the board.
8. For long strips, look for 24 V models; you'll have less voltage drop to deal with.
9. An inscription saying 2.4 GHz doesn't always mean Wi-Fi or Bluetooth; sometimes it's just the radio link's operating frequency, so pay attention.

That's all from me.

Happy New Year, everyone!
I hope you all make as many good and useful purchases as possible this year, and have as few requests for help or returns as possible. I also hope the only place you encounter the word “customs” is in the film “White Sun of the Desert,” and that you never have to deal with them.
And, of course, I wish the authors more readers, the readers more authors, and the administration more of both :)