LED driver

[Diver300]

I had to replace the power supply on an LED downlight after the old one got full of water. It was rated at 24 V, 150 mA, and I bought the 4-7 W version offered here:-

Pardon our interruption...

I thought it would be quite simple to cut down the current a bit.

The circuit of the LED driver was this:-

The datasheet for the IC is here:-https://www.lcsc.com/datasheet/lcsc...Silan-Microelectronics-SDH7611STR_C601125.pdf

I tried to increase the 1.6 Ohm resistor but I found that I couldn't increase it beyond 2.5 Ohms without the LEDs flickering and being very dim.

Adding a resistor load in parallel with the LED could also make the LEDs stop flickering. I ended up with 1500 Ohms in parallel with the LEDs, and the current setting resistor at 2.5 Ohms. The LED driver produced about 165 mA with the 2.5 Ohm current setting resistor and the 1500 Ohm resistor takes around 15 mA, leaving the 150 mA that I wanted in the LEDs.

Has anyone got any idea why the LEDs flickered when I tried to lower the current?

 

[rjenkinsgb]

The datasheet shows a snubber circuit across the transformer primary. I wonder if the flyback voltage gets too high without that and with a low load current, causing it to shut down?

With the transformer phasing shown, the snubber is effectively adding some parallel load to the output.

The datasheet also says the IC has some form of open load sensing based on transformer magnetisation, which could be triggering if the output current is too low??

 

[Diver300]

The datasheet shows a snubber circuit across the transformer primary. I wonder if the flyback voltage gets too high without that and with a low load current, causing it to shut down?

With the transformer phasing shown, the snubber is effectively adding some parallel load to the output.

The datasheet also says the IC has some form of open load sensing based on transformer magnetisation, which could be triggering if the output current is too low??

With those sensible suggestions in mind, I tried a snubber but it didn't end well.

I had run the lamp as I described before for 24 hours without problems, but there was quite a lot of heat from the power supply and the 1500 Ohm resistor that was in parallel with the LEDs.

I added a snubber and the lamp still worked. There was an average of 60 V across the resistor and capacitor. It was an 82 kOhm resistor so about 1mA on average. That seemed OK.

I removed the 1500 Ohm resistance that I had in parallel with the LEDs, and that worked, but I didn't want to run it for too long because it was likely to have been more current than the LEDs would normally take.

Then I adjusted the current sense resistor from 2.5 Ohms to 3.3 Ohms. When I turned on, the IC exploded.

There must be something that is happening with a higher value of the current sense resistor, which causes the IC to cut out, causing the flickering. It appears that the snubber circuit prevented the protection circuit from working.

 

[rjenkinsgb]

That is really weird!
Did you use the same diode-resistor-capacitor arrangement as in the datasheet?

 

[eTech]

What level is the AC supply voltage?
How many series LEDs?

 

[Diver300]

I did use the diode-resistor-capacitor arrangement as in the datasheet.
The supply voltage is nominally 240 V, probably actually 245 V, as normal in the UK
There are 8 series LEDs, in four packages in series, totalling about 24 V.

 

[Nigel Goodwin]

I did use the diode-resistor-capacitor arrangement as in the datasheet.
The supply voltage is nominally 240 V, probably actually 245 V, as normal in the UK
There are 8 series LEDs, in four packages in series, totalling about 24 V.

Actually in the UK the nominal supply voltage is 230V (unified across most of Europe), although in practice it's usually 240V as it always was, being regularly at 245V would be relatively unusual, though still within specification.

 

[eTech]

I did use the diode-resistor-capacitor arrangement as in the datasheet.
The supply voltage is nominally 240 V, probably actually 245 V, as normal in the UK
There are 8 series LEDs, in four packages in series, totalling about 24 V.

Are you sure the Vf is 3V?
The ones I've seen are 6V@150mA (Probably 2 or 3 internal LEDs in series).
Can you post a part number and/or photo of them?

 

[Diver300]

Are you sure the Vf is 3V?
The ones I've seen are 6V@150mA (Probably 2 or 3 internal LEDs in series).
Can you post a part number and/or photo of them?

There are 4 LED packages. The housing of the water-damaged power supply says 24 V / 150 mA. From what I have seen, there is no current at less than about 18 V.

For white LEDs, 6 V is going to be 2 LEDs in series. I think that is what I have four of in this assembly but it could be some different series/parallel arrangement, and it is on an aluminium PCB, so I'm not going to change anything.

 

[Diver300]

Well I bought what I thought was an LED driver of quite a different design, but when it arrived it was very similar. The eBay listing has a different price to what I paid, and shows what arrived, but I wouldn't have bought such a similar device knowingly, soI think that the listing has changed since I ordered it.

It uses the same high voltage driver IC. It has two outputs and this IC
https://en.publicxin.com/product/56.html
to switch between them somehow.

I have bypassed the output IC, and made the same changes as before, of a 2.5 Ohm current-setting resistor and around 1500 Ohms in parallel with the LEDs to stabilise the output.

The LEDs take 3.6 W, and the resistor takes 0.4 W. The input power is 5.2 W. I'm not going to try to improve it any further unless it goes wrong.

I'll run it continually for a day or so before installing it.