LED Driver unit with no output

[partyanimallighting]

Hi guys, I have an Acclaim Lighting LED driver unit that is not powering up at all. The unit is fairly straightforward, 120VAC to 24VAC power supply (26VAC read with tester) feeding a driver PCB which in turn outputs to a total of 8 LED units. So far I've bypassed a small ceramic capacitor on the live / neutral 120VAC line in and the voltage transformer is fine, outputting 24VAC. I then reconnected the power supply outputs to the driver PCB and I'm getting 24VAC at the bridge rectifier and 36VDC out but I'm stumped from there as there's not much more I can check based on my limited electronics knowledge (basically everything I've learned, I've learned right here on this forum). There is also a small transistor or triac or regulator with a heatsink that's warm / hot to the touch but there is nothing in the display and no output. I have clearer images if necessary so I'll wait for your feedback before I proceed.

 

[Harald Kapp]

The ICs are probably driven by a lower voltaeg than the 36 V from the rectifier. The capacitors and the IC in the upper right corner of the lower left image ( ) look suspiciously like a voltage regulator. Typically the center pin of the IC would be ground (although there are other pinouts, depending on IC type), one of the other pins would be input, the other output. Find the datasheet of the chip on the net or simply measure voltages from all pins to a know GND reference to find out whether this regulator is working.
But I may be wrong, tell us what's printed on the IC.

You don't have by chance a schematic diagram?

 

[partyanimallighting]

Nope. No schematic. I took my first initial look at the unit this morning around 3:30 AM so I just posted my preliminary findings with pictures. As you said, it's most probably a voltage regulator so I'll power it up and check pin to ground for voltage readings and post my findings with the part number. Thanks for the advice.

 

[partyanimallighting]

Back again. I checked and it's a voltage regulator, the part number seems to be LM2575, listed as a "simple step-down switching voltage regulator". I powered up the pcb and tested the voltage outputs at the pins and, according to the data sheet, I'm getting 35VDC at pin 1 (input, so that's correct) but 0VDC at all other pins, so this seems to be the problem. Should I go ahead and source the part and make the swap out or is there something lurking in the dark that I need to know? Also, if LM2575 isn't available, is there a substitute number I can use?

 

[(*steve*)]

The issues are:

1. There are several variants of the LM2575. It is likely that this is the 5V fixed version, but it would be best to verify that. (edit: it does say -5.0 on the third line, and I can't imagine this is anything other than the output voltage, EDIT2: that's "dash" 5.0, not "negative" 5.0 )
   
2. If it has failed, it may have failed for a reason that still exists, for example a shorted output. A replacement in this case will likely suffer the same fate.

With the power off, can you measure the resistance between pins 2 &3, and 3 & 4. Measure first with the black lead on pin 3, then with the red lead on pin 3. Please report all resistances clearly so we know exactly what each refers to.

 

[partyanimallighting]

Here they are. Resistance from Pin 3 (Black Lead) to Pin 2 - 1.1Ω. To Pin 4 - 1.4Ω. Resistance from Pin 3 (Red Lead) to Pin 2 - 1.1Ω. To Pin 4 - 1.4Ω. Hope this helps.

 

[(*steve*)]

Hmmm, a resistance of 1.1 ohms between the output and the ground is not expected. If you can remove this device and measure that again it would be good (measure the resistance on the board, not that of the 2575).

The resistance between pins 3 and 4 suggest that the inductor has a reasonably high resistance and thus this regulator is not supplying a lot of current.

 

[partyanimallighting]

I'm getting 2.6Ω and 2.4Ω across pins 2 and 4 from pin 3 with the LM2575 removed from the pcb. The Fluke tester is a little on the old side and may not be 100% accurate after years of use and I can't find my other tester at the moment so I can't take a second reading right now. Do these measurements sound correct?

 

[(*steve*)]

The readings between pins 2 and 4 are a lot lower than I would expect. These are taken on the board, correct?

 

[partyanimallighting]

Yes they were. I removed the regulator to try to source it tomorrow.

 

[(*steve*)]

I'm concerned that there is something shorting out the output of the regulator. I fear that if you simply replace it, the replacement will suffer the same fate.

 

[partyanimallighting]

I just found the other tester ( under the unit's housing on the workbench) and took readings with this tester and I'm getting 0.2Ω and 0.4Ω (2 to 3 and 3 to 4). Sounds better or worse? If the LM2575 is fairly cheap I'll buy 2 pieces, and if the first one blows after I install it, at least I'll have a spare one to troubleshoot with once the problem's found.

 

[(*steve*)]

I expect a low resistance like you're getting between output and feedback, but not between output and ground.

I would normally check the sous connected between the output and ground, but the resistance seems to be greater than the inductor, so it may be after that.

 

[partyanimallighting]

OK. You stumped me there. Sous??

 

[(*steve*)]

Would you believe "diode"?

Damn autocorrect.

 

[partyanimallighting]

LOL! I'll see if the part is available tomorrow, install it and see what happens from there. I believe the unit may have simply overheated because it was installed in a ceiling and was then smothered with sound proofing material so there was no ventilation whatsoever.

 

[(*steve*)]

It might be a good idea to connect a heatsink to the regulator of it doesn't have one.

 

[partyanimallighting]

It has a small heatsink. I'll put a larger one and maybe a small 12VDC fan if I can find a source on the circuit board.

 

[partyanimallighting]

OK. I got the part or something resembling it, LM2575T and cleaned up the solder points.

 

[partyanimallighting]

I also cleaned up the solder points as best as I could.