I have been playing with power LEDs for a while. I bought a bunch of power LED chips and strips and used homemade LED dimmers to drive them which worked well. Although I already have a great collection of DC LED Dimmers, I was intrigued by an Amazon notification, and ended up buying a “DC12V-24V (8A) dimmer box!
Chinese DC LED Dimmers
There are an incredible number of picks when it comes to Chinese DC LED dimmers, but the selection of cheap devices thins out if you want comparatively good amenities like flicker-free dimming and ample load handling capacity. It is not often that things work smooth on the first try, so I thought I would share something about the ubiquitous Chinese DC LED Dimmers.
This is the LED dimmer box I got from the Amazon seller. According to them, the dimmer box will run from 12V to 24VDC flawlessly and has an 8A (maximum) load managing capacity.
An agile examination
Instantly after its arrival, I made a couple of quick tests (a rush job) and got reasonable outcomes. The first test was conducted with a 12VDC (8W) COB LED string, and the second was with one 12VDC/24W motor (not ideal, but what I had handy). Both tests were powered by a 12V sealed lead acid (SLA) battery.
As observed by me, the LEDs stays flicker free on both minimum and maximum levels, but the motor starts whining – quite natural, the pulse width modulation frequency and duty-cycle have a significant effect on the motor speed control.
On the inside
Since I trusted the seller’s description, I believed the inside electronics to be close to the image posted by the seller (see below). Yes, it seems like a pretty cute circuit board that holds two 8-pin ICs, one 3-pin IC, and a power MOSFET couplet. So far so good!
I commenced the teardown, and once opened I was surprised it looked just like a different device. After a quick Google search, I found out the very popular dimmer box comes with many different circuit designs and printed circuit board layouts.
I also tried to find ‘clone-of-clone’ related review posts online but could not find anything useful. Later I found by chance that mine is not the ‘original’ version but its ‘lite’ edition (ha-ha)! This picture is the inside view of my dimmer box.
As you can see, it is like the original, apart from the obtrusive difference in PCB size and layout. Look at its schematic captured by me. Sure enough, a cunning concept!
The schematic is exceedingly simple and satisfactory and hardly needs an explanation. However, note that the base PWM frequency is governed by the NE555 chip (IC2), and the capacitor (C2) wired between its pins 2 and 6 should be responsible for that frequency (<1kHz). Look below to get the waveform appeared at pin 2 of NE555 (TP1).
Also see the waveform at pin 1 of LM358 (TP2) with 50% duty cycle dialed by the 1K ‘dimmer’ potentiometer (P1).
My quick estimations
I only peeked, but datasheet of the workhorse in this design – the 09N03LA power MOSFET – assured that it can handle 8A load current at ease. Actually, the IPB09N03LA is a logic-level N-channel MOSFET with very low RDS(on), and ID of 50A (VDS=25V). I do not care about the concerned PCB tracks, but took the claim on 8A load handling capacity, and nothing more.
This is a snip from the official datasheet of IPB09N03LA released by Infineon Technologies:
Likewise, the recommended maximum input voltage of 78L05 linear +5V regulator chip is 30V (and VDS of the MOSFET is 25V). So far, at least, I could say that the online seller/circuit designer had proved his/her point.
Pulse width modulation views
Expending the prefaced dimmer box, I am now playing with around 1kHz PWM drive for the LED strips. Even with the comparatively high PWM, I can see them as stripes come apart when I move my eyes rapidly. I know different people are sensitive to flicker in other ways, however, it is a better idea to raise the PWM frequency as high as practical, to get the most consistent and pleasing light. If you want to try higher PWM frequencies to see what takes place, simply replace the 100nF capacitor (C2) with a different value capacitor as necessitated (10nF for ~6kHz, for example).
Admittedly, the dimmer box is primarily intended for dimming LEDs and/or small incandescent lamps. We can use it as a DC motor speed regulator, too. The motor makes some noise, because the PWM is at a lower (audible) frequency. For a few tryouts I ran my test motor in low frequency which produced a humming noise. I raised the frequency over audible frequency and the motor is quiet now. It is worth noting that in certain high frequency PWM circuits you may need devoted driver to rapidly charge and discharge the MOSFET gate capacitance to minimize switching losses (more on later).
MOSFETs in high frequency switching applications
While choosing a MOSFET, parameters that are focused on by most design engineers intuitively are VDS, RDS(on), ID. However, for high frequency switching applications, it is essential to pick up a suitable MOSFET as Gate charge plays prominently in such situations. Total gate charge Qg include Qgs and Qgd. Qgs represents the accumulation of Gate-Source capacitance while Qgd is the accumulation of Gate-Drain capacitance, also known as miller capacitor. In high frequency operations, Qg should be selected as small as possible. Another tip of Gate charge for picking up a MOSFET in power system design is that ratio of Qgd/Qgs be lower than 1 to preclude the circuit from shoot through.
When it comes to the next significant parameter – the dv/dt capability, peak diode recovery is defined as the maximum rate of rise of drain-source voltage allowed, ie, dv/dt capability. If this rate is exceeded then the voltage across the gate-source terminals may become higher than the threshold voltage of the device, forcing the device into current conduction mode, and under certain conditions a catastrophic failure may occur.
Feel free to post your comments below.
More on voltage-controlled pulse width modulation
In the inside electronics of the dimmer unboxed here, the voltage control PWM is generated by first using the triangle signal generator (IC2) which provides the base PWM pulse frequency and the necessary ramp voltage (rise and down) to produce the PWM signal. Next by continuously comparing this ramp voltage according to the voltage level produced by the potentiometer (dimmer knob) using the comparator circuit (IC3) the exact voltage control PWM is produced. By varying the threshold point voltage, we could also vary the on and off period of the comparator which is the exact behavior that we need to produce the required PWM signal to drive the output load (LEDs).
Simply, the ramp signal is provided by the NE555 astable that generates the triangle wave signal while the LM358 comparator that gets its input from the 1K potentiometer which provides the voltage threshold point and together with the triangle wave to produce the required PWM output. PWM basically is an on and off pulse signal with a constant period or frequency. By changing the PWM duty cycles we could change the average voltage across the output load terminals, this mean we could vary the LED brightness/Motor speed just by changing the PWM duty cycle – the lower duty cycle percentage produces less power than the higher percentage.
My next objective
As stated at the beginning of this article, I already have a great collection of PWM boards and modules, so now I want to hack this dimmer box to modify it’s as a simple linear ramp generator for use in some other projects. Fortunately, the beefy box has enough free space inside to accommodate one or more little circuit boards. Obviously, it is easy to take the ramp output from TP1 in the schematic so that it can be routed to the external world through an RCA jack. But the responsible capacitor in the 555 circuitry charges according to an exponential curve, and what I needed is a linear ramp. A constant current source for the charging circuit would be necessary.
Below you can see my basic idea which is in fact a piece of circuitry I borrowed from a textbook. In the circuit, transistor T1 in combination with LED1 and R1 forms a simple constant current source. The resultant ramp is reasonably linear, and, on average it ranges in voltage from 1.65V to 3.32V(at 5V Vcc) before being discharged. Pin 3 of 555 will deliver a square wave pulse as well.
The idea surely calls for some refinements as there is a need for one buffer (or something alike), because the charging circuit will be disrupted if I take the ramp signal directly. Further, I need a 0 to 5V sweep range, rather than the restrained 1/3 to 2/3Vcc range. In a future installment, you can read all about the proposed hack. Thanks for reading!
- Infineon Technologies AG (IPB09N03LA)
- Texas Instruments (LM78L05)
- Amazon (Dimmer Box)
- Banggood (COB LEDs)
- Taiwan Semiconductor (MOSFET Facts)
- International Rectifier (MOSFET Facts)