LM3409HV: Output current/voltage not matching expected values.

Hi Mathew,

How many LEDs do you connect between L+ and L-? How many are in parallel? How many are in series? Could you provide the LED part number?

Best Regards,

Steven

Hi Steven, thanks for your help. It's a UV-A LED module with 25 individual LEDs, in a 5S5P configuration, typically used for SLA 3D printer applications. Here's a photo (the small passive in the bottom right is just a thermistor):

 

Unfortunately, since the LED came as a kit with heatsink, power supply and Fresnel lens, I do not have an exact model number for it. I've searched quite a bit to find one, but it seems to be custom. The datasheet for the entire kit specifies a 17V 5A supply to the LED.

The power supply that came with the LED is an LM25116MH-based design. It's much too large in size to fit my requirements, which is why I'm trying to build my own LED driver circuit using the LM3409. I can confirm the power supply they included does put out 17V at 5A when fed from the same 24V/10A bench source I am using to test my design, so I'm confident that's what I need. I'm just not sure how to have the LM3409 match that.

Hi Mathew,

Got it. I will carefully check your parameters and give you a feedback within 24 hours.

Best Regards,

Steven

Thank you very much Steven.

Hi Mathew,

Sorry but I am having an urgent biz travel these two days. 

I do apologize for my delay. I will reply to you once I am back in office. Thanks for your understanding.

BR, Steven

No worries, thanks Steven.

Hi Stephen, just checking in since it's been 5 days, would really appreciate your help if possible.

Checking in again. Still would appreciate knowing what I'm doing wrong.

Hi Mathew,

Sorry. I apologize for the delay. I am now looking at this issue and are working on it. I will give you the reply by tomorrow.

BR,

Steven

Hi Mathew,

• I have checked your calculation and I do not find any problems.
• I see you put too input capacitor Cin-1 and Cin-2 in your calculation details but I do not see them in your schematics and PCB layouts.
• The PCB layout can be improved a lot. But I am not sure at this time if the layout is the root cause of the "inaccurate current" problem.
• The below point is suspicious. Please try below items to help me identify:
  • Please try to measure the voltage of IADJ pin.
  • Please try to connect the IADJ pin to an external constant voltage source = 1.24V and see what the output is.

Mathew Carlick said:

The other thing I tried was a different sense resistor of 0.033Ω instead of 0.05Ω. This brought the current up to 3.6A, with the same 15.5V

Best Regards,

Steven

Hi Stephen, thank you for looking into my issue and verifying my calculations.

You are correct about the input capacitor, there is only one in my layout and schematic. I did not realize two input capacitors were recommended until after the PCBs were designed and ordered. During my testing, I tried both a single 22uF input capacitor, as well as a stack of two in parallel on the same footprint that I measured at 44.6uF. The extra input capacitance had no measurable effect on the output voltage and current to the LED. Regardless, my next board revision will have an extra footprint for a second input capacitor.

I measured the voltage of the IADJ pin during operation to be 1.241V. Despite being the correct voltage, I followed your suggestion and connected it to an external constant voltage source of 1.24V. This also had no measurable effect on the output voltage and current to the LED. I still measure 15.5V at 2.85A.

Hi Mathew,

• The second 22uF input capacitor may not be needed as per the theoretical calculation. But you can have that extra capacitor footprint for further decoupling intents if needed.
• I will test your parameters on our EVM today.
• Please measure the I_LED current waveform with a current probe and provide me with the oscilloscope capture. (If a current probe is not available, then you can probe the voltage of CSP pin and CSN pin and then do a mathematical calculation on the oscilloscope.)

Best Regards,

Steven