Question about hijacking other devices gate transistors.

[msuffidy]

So I was thinking of getting some cheap used RGB strips that appear to have connector leads, that I am guessing are some sort of straight up power with no logic and are coloured RBG,black. Some of these have boxes, which I am guessing have some sort of gate transistors as a last thing in the chain of devices.

Say the control pin of the transistor was severed from the trace, and a new lead was introduced. If you have an arduino, could you somehow control the transistor directly, or do you have to like use an optocoupler and power it on the strip side off of it's own circuitry? The main problem here is a risk of backwash to the arduino, and also different interpretations of like ground?

Here are some suspect transistors.

[Zero999]

It would be much easier to buy some transistors and use them to control the LED strip, with the Arduino driving them.

[SteveyG]

It's not clear what you're hoping to achieve by modifying the controller. Are you hoping to use the integrated power supply as well?

[Zero999]

It's not clear what you're hoping to achieve by modifying the controller. Are you hoping to use the integrated power supply as well?

The board doesn't appear to have any mains on it and if it does, it shouldn't be used because the creepage and clearances are insufficient.

[msuffidy]

The idea was to take over the switching of the led lights with the arduino, and use PWM for dimming. I have not got the strips yet, so this is not the exact board, but they seem to follow this general pattern of having these 3 transistors. There is an external brick so there should not be any power regulator stuff on the board.

Thanks for the comments though.

[msuffidy]

I have been thinking about it and I still think it would basically work to cut the bias trace  of the transistors, and solder a wire to a like 150 ohm resistor and then to the arduino. You probably don't even have to connect ground as the mains returns the circuit.

[thm_w]

I have been thinking about it and I still think it would basically work to cut the bias trace  of the transistors, and solder a wire to a like 150 ohm resistor and then to the arduino. You probably don't even have to connect ground as the mains returns the circuit.

Should work yes.
I would connect a ground lead as well (the bottom trace on the PCB).

[Zero999]

I have been thinking about it and I still think it would basically work to cut the bias trace  of the transistors, and solder a wire to a like 150 ohm resistor and then to the arduino. You probably don't even have to connect ground as the mains returns the circuit.

Why do you want to do that?

It would be much easier to solder three transistors on a piece of stripboard, than hack that PCB.

How long is the LED strip? What voltage is it? And how much current/power does it use per metre?

EDIT:
This is what I'm talking about. The MOSFETs can be logic level devices, such as the IRL3705, or IRL540.

[msuffidy]

It is only at a hypothetical stage right now, but there is some ability to use the supposedly existing power brick and basically make a small device as shown in the schematic.  I don't generally order electronic parts, so the easy way for me to get a load of rgb leds and even the wiring to the connectors would be to buy a LED strip and replace a portion of it with something else.

[msuffidy]

I want to scale up basically what is going on in this video. If it works as I expect I can get basically whatever RGB values to a large sets of LEDS as I want as controlled by whatever effects I want to. I am thinking of making something like a cylinder post with the strip to wrap around it so it is a stand up light.

[Zero999]

I want to scale up basically what is going on in this video. If it works as I expect I can get basically whatever RGB values to a large sets of LEDS as I want as controlled by whatever effects I want to. I am thinking of making something like a cylinder post with the strip to wrap around it so it is a stand up light.

I don't have time to watch a 10 minute video and doubt many others do. Briefly explain what you're trying to do.

If you need to control a large number of LEDs, then addressable LEDs, such as NeoPixel.
https://learn.adafruit.com/adafruit-neopixel-uberguide

If you just want to control one or two RGB LED strips, use the simple circuit I posted above. I could build it much more quickly and easily than I could hack a pre-built board.

[msuffidy]

I think it is pretty much answered, but what I was doing was merely controlling a RGB led strip from my computer using PWM dimming, as to get a certain RGB approximation. I was not building a multi LED cube display. I was mentioning a cube because I was describing color morphing but they were going to be all the same.  The scheme uses the arduino as a pwm signal generator.

[msuffidy]

I got a new idea about some stuff I have around here. Another solution was to take over this $4 light puck I got from the dollar store.

I have edited this post to reflect some things I have learned since writing it originally:

I was thinking I would power if from a separate usb line to a usb hub tied to where the battery used to go.  (yes but the usb is 5v and the batteries are 1.2 x 3 = 3.6V, so I would ideally have to convert them. )

Also the arduino I think should connect to the points I have drawn on here. I think maybe I could even use the existing resistor, so I just have to connect a few things to the arduino carefully. (yes but the control signal was 3.6 previously and would be 5v so I am not sure I can directly do that )

 

[msuffidy]

Hey I did it and it seems to pretty much do what I wanted it to do. I left it running on batteries instead of making a wire jungle. Bad news is I hit 2 of the smd rgb leds with the soldering iron sort of messing up the packages. Good news is it seems to still work anyway. Well anyway proof on concept. I will be putting up a youtube video soon demonstrating it. This is the setup though. Oh, BTW I broke off those 3 gpios from the micro. Only real problem is pwm setting 1 of 1000 is still pretty bright, I mean it is darker, but it should be totally off. So there is a low end gamut problem. I don't know if it is like switching time issue or something maybe on their transistor. I put in the gamma formula at gamma 2.5 and it works a lot better than in the video below.

[msuffidy]

[msuffidy]

OK that's it for a while

[msuffidy]

Ok I could not resist trying the method out on a USB led strip that I got for like $10. Looking at it, it seemed like it was pretty much the usual problem. Everything was on a really tiny scale. I tried soldering wires to the edge of smt resistors, but the entire resistor came off the board. I just continued to fumble around with the 2mm resistor on my soldering tip until I managed to sort of get it back on the board. I measured it and it was still the same value as the other 2. Then I scraped the board trace and soldered wire to that instead, which can not be a good thing. Anyway I encased the whole thing in glue gun glue so I am hoping the bonds will hold up if no forces are acting on the solder joints. I'd like to say a total success, but I have caught it doing weird things and there was a second chip. Maybe there was a via to the other chip. I am hoping it dies out after a while and goes away. I don't know if the remote still works or not for the chips. It does not look like it is doing anything.

**OK I noticed today that when I touched around the wires the green went bezerk. I un potted it a bit and the green wire is not the green wire, but another color so I went to the wrong spot. Go figure. Looks like a bad solder joint. Did it again. Working great.

[Zero999]

I don't have time to watch a 10 minute video and doubt many others do. Briefly explain what you're trying to do.

If you just want to control one or two RGB LED strips, use the simple circuit I posted above. I could build it much more quickly and easily than I could hack a pre-built board.

OP might be like many of us were - starting out with little access to the raw ingredients, not knowing the full theory, but being able to hack on compiled hardware. and hack away OP should

I could've built the circuit I posted earlier, when I was 12. The original poster doesn't appear to be open to suggestion.

[Zero999]

I could've built the circuit I posted earlier, when I was 12. The original poster doesn't appear to be open to suggestion.

i'm going to be charitable and assume you're having a bad day.

How am I being unreasonable? One of the first circuits I built, beyond the basic LED & resistor, was an astable multivibrator (you know the one with two BJTs, cross-coupled with RC networks) when I was about 11 years old. It definitely is easier to build simple circuits from scratch, than it is to reverse engineer and modify them.

If the original poster was open to suggestion, he/she would have asked more questions, rather than just posting their videos. Don't get me wrong. There's nothing wrong with creating a thread with videos and schematics about a project, but it's little frustrating when someone asks questions, then lacks interest in the answers and doesn't respond, when someone asks them reasonable questions such as: what? how? and why?

[msuffidy]

The way I see it is like this:
1) I want to control a reasonable brightness and size led strip under PWM intensity control from an Arduino
2) The cheapest way to get the entire led array done for me is to buy a LED strip from Walmart
3) I don't have to build anything at all if I mess with what they sold me
4) If I did want to build the stuff I would been some level of power supply pre fabricated or all myself: I would have to know the current and voltage specs of the existing LED strip. I have some parts lying around but I would need 3 matching transistors for switching. I would need some means of safely(like in a enclosure) assembling the supply and transistors to something like a bread board or DIY pcb.

The way I see it, it is mostly done for me and I just had to hotwire a part and did not have to do anything else.
I purchased 2 Arduino Nanos for I can recall like 2 for $25 or something. This was like a clearance strip from Wallmart from Monster co and it was like $12 after tax. That is good because I think the ones they are selling now are mixed white/color.  So this project cost me I'd say $25 Canadian abouts.

I think why I really asked here was I was worried about killing my PC with any electrical issue from what was connected by USB. The strip runs off of a USB apple charger at 5V, so I gather pretty much anything that could go wrong would be ok as USB is generally protected from low voltages in the first place? I still expect optocouplers would have been the totally safe way to avoid backwash into my USB but I am not that worried.

[Zero999]

I would not advise building the power supply from scratch. I recommend buying a pre-build power supply.

The only part I suggested you make, is the board with three transistors, switching the LEDs, because it's much easier to do, than hacking a board. The transistors don't have to be match. They just need to have sufficient voltage and current ratings for the LED strip. Heck do whatever you feel more comfortable with. If you prefer to hack, then go ahead.

The voltage rating of the LED strip can be determined by looking at it an counting the number of LEDs vs resistors. The blue LEDs have a voltage drop of about 3V, with the resistor dropping the remaining voltage. If there's one resistor per LED, then the strip is designed for 5V operation. 12V and 24V LED strip has, 3 LEDs and 6 LEDs, per resistor

The current of the LEDs on the strip can be determined using Ohm's law.
I = (V_S-V_F)/R
Where R is the value of the resistor. This can be obtained by looking at the resistors on the LED strip. It's normally the standard surface mount resistor code, with the last digit representing the number of zeros and the preceding digits the value, so 101 is 100 Ohms.
V_S is the supply voltage
V_F the LED forward voltage. Normally about 3 for green, blue and white and 2 for red. Of course this is multiplied by the number of LEDs in series, so if your strip has thee LEDs, per resistor, that means there are three LEDs in series.
That will give you the current through each series string of LEDs.

Define reasonable brightness and size?

Note that higher voltage LED strip is generally more efficient. 5V strip is poor as 3V goes to the LEDs and 2V is dropped in the resistor, so 40% is wasted. 12V and 24V LED strop, drop 3V, or 4V in the resistors, which is 25% wastage. 24V is generally a bit better because the current is lower, so there's less voltage drop along the strip.

If you're using 5V LED strip, don't worry about killing your PC. It's virtually impossible.

If you're using 12V, or 24V LED strip, it's unlikely you'll kill your PC. The driver transistors are connected to the Arduino, rather than directly to your PC, so the Arduinio is more likely to blow up, if the driver transistors fail. Using resistors in series with the gates of the driver transistors (refer to my schematic a few posts ago) it will provide some protection for the Arduinio, against the driver transistors faling.

[msuffidy]

Well in the end thanks for the information. I am moving on to other things now granted the strip continues to work properly. I did assume a sort of attitude that the transistors would be in a steady ready to go state, some designs may have had that rail on a voltage control and I did not at all look out for that. That was not the case here.

The code examples are sort of hard wired for me and not really commented where you would need to know about it.
The main things are the 3 data lines have to be re arranged to the color and line, or the program does when it sends the request. Long story.
Also I like to do this: Find the weakest led of RGB. Make that 100% then find the other 2 that seem to create a white color.
Then a value of 0-255 needs to be multiplied by some number to make it a number out of 1000, where the strong led is 1000. So it is like that.

Ardunio code
http://web.ncf.ca/fs864/pickup/rgb-pwm-v3-puck.ino
Linux calling model
http://web.ncf.ca/fs864/pickup/tether-puck-no7.c
Windows calling model
http://web.ncf.ca/fs864/pickup/tether-7-windows.cpp