Does anyone out there have any experience with this ic?
I have an idea to replicate a 10W automotive light that I have now, and it uses this ic. The circuit board on said light also has another ic, but there's no number. It's an 8P ic.
I've put a little circuit together (following the same principle as the datasheet), but struggling to get it over 4W, so it's not really bright enough. I have a couple of other things to try and will report back on those as and when I've tried them.
I can't really find any info for the LJY5200 online/YouTube etc.
There's plenty of info regarding PT4115 LED driver ic, but not sure how that will run at 10W as the voltage drop across the LED is 3.4V and the PT4115 ic max current output is 1.2A, so by my calcs, that's 4W.
I'd be most grateful for any links, info or whatever that you think may assist me.
I don't like the schematic in the datasheet. That is not how to draw a schematic. The 0.5Ω resistor is between the mosfet source and the GND.
It is a cost effective design. The type of led and the wires to the led might have influence as well.
With the 220k resistor, the frequency will be a little below 70kHz. I don't know what an inductor of 500µH does with 70kHz. It is just to reduce the sharp edges, or does it keep the led on all the time ?
If you want to build something that works, then you better stay away from these cheap dedicated chips. If you watch https://www.youtube.com/@bigclivedotcom then you see a lot of videos with such dedicated chips and minimal designs.
Do you have to use a single 10watt LED.
The setup would be more efficient if you could use three 3watt LEDs in series, or a 12volt COB LED (with three 3W LEDs in sseries), which you can drive with a PT4115, which doesn't need an external mosfet. The PT4115 also runs on a higher frequency, so a smaller size/value inductor can be used.
Leo..
I think I've watched pretty much all of Big Clive's videos. You might well be right about staying away from these chips, but I have to say that we've used a couple of hundred of these LED's using this chip, and haven't had a failure. But, I'm certainly open changing things if necessary.
I do also have a couple of PT4115's to play with, and they seem quite a popular solution to driving LEDs.
Due to how the light is assembled, I need just a single LED
Yes, the PT4115 looks like a good candidate, and I like your suggestion of a 12V COB. That's worth a shot. And a smaller inductor due to the higher frequency will definitely save space.
Hi guys.
I threw a circuit together with the PT4115, following the datasheet, with a 3W LED (forward voltage 3.5V) and things seemed OK.
Then tried 3x 3W LEDs in series and this lit up OK at first but then started to flash.
Tried messing with sense resistors and inductor values but could really get a reliable outcome.
Was hoping to play more today but other things got in the way.
Quick question, is it possible to use an external mosfet with the PT4115? So the switch pin on the PT4115 drives a mosfet (by way of a TC4420 driver) and then this drives the LED. As the LED I want to use is a single 10W LED with 3.5V forward voltage, so it'll need a little something to get it to full mojo as the PT4115 can only drive up to 1.2A.
Any ideas?
What was Vf of the LEDs. Supply must at least be 3xVf, plus 2volt for the PT4115, with an absolute minimum of 8volt. Maybe you should use two LEDs in series.
The LED Vf is 3V on the commercial unit (using LJY5200) I'm trying to replicate. And just stuck my current probe on with a mean current of 2.2A, so that's 6.6W if my math is correct (pre morning coffee).
I can get numbers for my circuit later. Other work getting in the way.
That board should also be able to drive two LEDs in series, for ~15W LED power.
Make sure you have enough heatsinking available. Hot LEDs won't last long.
Leo..
Hi
I've had mixed success so far.
Best is with LJY5200 ic.
Had 3x 0.33r sense resistors in parallel, so 0.11r.
13V input
147uH inductor (actually a 100uH and 47uH in series)
100k ROSC (using a trimpot as bored with swapping 0603 resistors!)
I ramped up the current on the power supply slowly. And ended up with input at royghly 13V and 750mA = 9.5W.
Very rough measurements on my LUX app on my phone showed about the same brightness as the light I'm trying to replicate.
However. I then turned the power supply off and back on again and the LED lit for a split second and then went off. And won't work again. Haven't found the problem yet but maybe something to do with using a 100pF cap on the PWM pin to ground when I think it should be a 100nF. This is to give a soft start, but thinking that the 100pF cap was anything but!
Need to tear it apart and find out what went wrong. Maybe tonight.
Going back to the schematic. There's 4x 2.2k resistors in series on the input. This drops the 13V to about 6.7V, but why do we need to use 4 resistors? Do you think I can use an LDO regulator and just set it to say 7V. I'm think an AMS1117-ADJ. I could then add one resistor to limit the current.
Thoughts?
Stubbornly I want to make this instead of buying a ready to go solution.
Appreciate your help.
Cheers
Matt
So for a fosc of 150kHz I should just ground the PWMD pin (pin 5), as per the datasheet?
You think that this is the way to go? And then a small value inductor?
For the sense resistor, I've been using 2512 SMD resistors and stacking them 2 or 3 high in parallel to get the value I want. Probably not the best method, but seems to work LOL!
I have all those calculations from the datasheet written down in an excel doc, so threw some numbers in and gave it a go.
Using 3x3W LEDs in parallel.
I have essentially a 200uH inductor fitted (2x 100uH inductors in series) with the following:
Vin = 13V
Rosc = 1M
Inductor = 200uH
Sense resistor = 0.1r (2x 0.2r in parallel)
Input power = 9W (according to my power supply)
Vf LED = 3.2V (measured on DMM)
Didn't measure current through the LED
I used a 100nF cap on the PWMD pin to ground and this gave a very nice soft start with no over shoot. Tempted to fit a higher value cap as taking a couple of hundred ms to fire into life is not a problem.
I have the 3x3W LEDs mounted on a big heat sink and that obviously got hot, but the PCB was particularly hot.
I left it there as I didn't want another depressing evening drive home wondering what went wrong lol.
Tomorrow I'll try a different approach following your suggestions . Higher frequency and lower value inductor. And go for bit more mojo. And I'll break out the scope and hook a few things up to get some meaningful measurements.
Thanks for your help.
Oh, one last question if I may. Concerning to 4x 2.2K resistors on the ic input. Why not just use a linear regulator for the ic supply?
Bad idea.
Small LEDs in parallel is not that bad, because they don't get hot, but power LEDs do, and then their Vf drops, which makes them 'steal' more current from the neighbouring LED, which makes them hotter. This is called "thermal runaway". Eventually one LED will fail, and then the two LEDs left will fail much faster.
Leo..
