Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Discrete LED Driver Topology

Status
Not open for further replies.

Endymion

Junior Member level 3
Joined
Nov 18, 2011
Messages
29
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,281
Activity points
1,611
Screen Shot 2015-07-01 at 9.01.27 pm.png

I have drawn the above schematic by reverse engineering a locally available LED Bulb. I was expecting to find an IC in the driver board but instead found this discrete circuit.
Googling "Discrete LED Driver" does not return a lot of relevant hits.

From my understanding, I think C4 and L1 form a tanked circuit which drives the base at startup. Once the collector is pulled low, a voltage is induced on L1 via L2 and this cycle continues (?). I do not understand how this circuit is regulating it's current through the LEDs as it is a constant current driver.

I am hoping someone more experienced here could explain the workings of this circuit to me. The output voltage can go upto 94V and the current is a steady 133 mA.
 

  1. This is basically a continuous Buck regulator switching oscillator than regulates the pulse width for an average DC voltage across Vbe controlled by D1 zener voltage , Vbe and Ie*Re and duty cycle.
  2. Offline buck designs have improved over this and work in principle like active PFC to provide a sinusoidal current fluctuating around an average DC current of 133mA +/-10 % possible but more accurate with precision buck or hickup continuous converters.
  3. LC input filter reduces spike currents escaping and line transients injected.
  4. Transformer design is critical and difficult for newbies, so fixed inductor type may be better with PWM regulation of MOSFET, transistor or IGBT switch.
  5. LEDs are 1W type with heatsink and have 1 Ohm ESR (1/1W=ESR) so for a cool switch, RdsOn should be <10% of this rated for 2x line voltage.
  6. Diodes Inc make the best low Rce BJT, (IMHO) with an integrated solution for offline LED CC driver.
  7. Caps must be X1 rated for safety.
  8. TI and AD also have many solutions.

Tony Stewart
EE'75
 
R7 equals 1 Gig? Just asking.

I agree with Tony....the simplest circuits are the ones that hold the deepest secrets.
In this instance, the coupled inductor L2/L1 holds the key to a successful operation.
 

A good strategy on this case is simulate the circuit and see how it behaves during startup and load change.
 

Important;
the circuit, as drawn originally, has a mistake.

I've redrawn it correctly....see if you understand why the original was incorrect.

error.png
 

SunnySkyguy;

I've had to develop that critical-thinking capability, because I have noticed that there is a lot of information on the web with errors.
The unfortunate thing would be, that an experimenter like the OP decides to build the circuit as described, only to blow in his face.

The circuit as originally shown, would have lasted a dozen microseconds (the first switching cycle) before exploding.
 
  • Like
Reactions: lal883

    lal883

    Points: 2
    Helpful Answer Positive Rating
I had to do the same in an EE lab in '73 when we had to build a staircase generator following a device datasheet. I was the only student got it working, finding with an error on the published datasheet (GE) who sold out to Harris. My Prof wanted me to become a grad student because I kept pointing out little mistakes in class. But I didnt want any more theory, just start designing things. So I was lucky, only 1 job offer and it was Aerospace, Telemetry and Instrumentation for my 1st employment in '75.

I remember there was this Nuclear Power client (Ontario Hydro )who monitored every design change on this 1st SCADA robotic system we made, after we were late. My job was to make it all work, incl my own SCADA discrete stuff, including a discrete UART.

He would say ... are sure you can put a ceramic capacitor on the output of this TTL gate to snub a +ve glitch. ... race condition.. Considering I had a dozen ECN's a day on this 6800 system with 44 PCB's in 3 subsystems and we were late shipping 3 mos....
... i said, of course . TTL is a very reliable asymmetric switch with built in R values and this is simple and effective low pass filter for an unwanted error condition. .. my problem log was about 100 items at the time, on my own. So these forums are just a reminder for me... It's much quicker with expert help. We should have been so lucky....
 

Yeah I stared at it for a while and assumed there was a diode missing, nowhere for the energy in L2 to go. Good eye.

The schematic doesn't show the phasing of the transformer windings, but I assume it's set up for positive feedback. Current ramping in L2 will cause the BJT to turn on harder until the emitter current rises high enough to force the BJT out of saturation and into cutoff, at which point the positive feedback will cause rapid turn off. That will continue until the energy in L2 is fully delivered to the load, at which point the BJT will start to turn on by current trickling through R5+R6. So it's current mode transition mode buck converter.
 

everything good so far but how do you design/select L2/L1 turns ratio and how do you construct it , what core type , wire gage cet.
 

There are many hidden parameters derived from L dv/dt, Vac-pk, Vf.total ln LEDS, Diode ESR, ZPrimary, mutual coupling factor, self resonant frequency, impedance ratio, core permeability, saturation threshold, recovery time, etc etc, besides L ratio.

Eventually the oscillator must drive a very low ESR load of the LED string of 1 Ohm per LED @1W per device. This requires careful design of the magnetics and transistor parameters to work. It has no room for error or transient protection and might take months of effort initially unless you reverse engineer what you have to determine sensitivity to tolerances, thermal design and have a magnetics background.

The design also has poor PF and THD, but exempt due its low power, but may be a good noise generator.

Who's product was it?

I suggest you learn some well documented designs first, before you tackle the elegantly simple, yet complex ones.

https://www.fairchildsemi.com/application-notes/AN/AN-4159.pdf

https://www.ti.com/tool/pmp8004

closest match so far, but better specs
https://www.onsemi.com/pub_link/Collateral/NCL30081-D.PDF

this has one transformer design details
https://www.active-semi.com/sheets/PSG_LED_General-Lighting-Applications_Released.pdf
 

You are right SunnySkyguy thanks for the heads up and the app. note is good too, well what appears so elegant is really not so simple after all; the “devil” is always in the details isn’t it? The circuit behaves as a Quasi-Resonant tank with “inverted buck topology” using L2/L1 combination, I think?, what will the resonant frequency expression be in s.s operation any idea ? (in terms of L1,C4, and R3IIR3) . Also what do resistors R1//R2 and R4//R3 do exactly?
 

Re limits current in emitter, Rb limits current in base.
Both remove heat from transistor and spread out into a pair each to limit T rise.

I would expect this design to work over a limited AC voltage range only for 4 W out and possibly <70% efficient so it wont be duplicated much as most designs now are universal voltage and dimmable.

This one is very cheap and dirty.

I am hoping someone more experienced here could explain the workings of this circuit to me. The output voltage can go upto 94V and the current is a steady 133 mA
.

where did you measure these?

What is power spec? 133mA x 4 x 3V?
Or 133mA x 94V?

AC voltage range?
 

You are right SunnySkyguy thanks for the heads up and the app. note is good too, well what appears so elegant is really not so simple after all; the “devil” is always in the details isn’t it? The circuit behaves as a Quasi-Resonant tank with “inverted buck topology” using L2/L1 combination, I think?, what will the resonant frequency expression be in s.s operation any idea ? (in terms of L1,C4, and R3IIR3) . Also what do resistors R1//R2 and R4//R3 do exactly?
The circuit in the first post is not resonant or quasi resonant. Those control ICs work in a completely different manner.
 

Philips non dimming A19 (800 Lumen) use similar design.

Primary Inductor inductance define switching frequency (25-50KHz).
Inductance calculation can be used from app note from Buck LED Driver (NCL30002 for example)
Drive winding need to generate 3-5V. You can use EE16 core with gap.

Best Regards
Prasad
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top