2
\$\begingroup\$

I recently designed a PCB for the TPS61169 LED driver. I used the sample layout from the datasheet, as it is a tiny SMD package and I couldn't breadboard it. I also followed the Typical Application in section 9.2.

However, it doesn't function properly.

Referring to 9.2 figure 4,

  • Supply voltage is 5.0V
  • CIN is 100uF tantalum + 0.1uF ceramic
  • L is 6.8uH inductor
  • D is an SS24 schottky diode
  • COUT is 2x 1uF ceramic
  • RSET is 2x 1.6R resistors in parallel (0.8R equivalent)

In testing, I have been trying to drive 3 green 3W LEDs in series.

Using the equation in 8.3.4, RSET = 204mV/250mA = 0.816R.

I calculated the inductor peak to peak current using equation 4 to be 3.84A, and the inductor peak current using equation 5 to be 2.17A. I don't think these are correct, however, as they seem high compared to the values in table 2 of the datasheet - Recommended Inductors.

Here is a schematic of the board:

Schematic A few notes: the two inductors and the two diodes are for different components - through hole and SMD - as I didn't know what components I had or could get at the time. Also, only two of the resistors R2 - R5 are populated.

Here is the board done in Eagle. Layout is per the datasheet.

Eagle board

Symtoms:

  • unexpectedly high current - my power supply reports a current of whatever its limit is set at, and will go up as high as I dare to set it. The current measured by my DMM in series with the LEDs is about 1/4th, and scales with the input current. I have no clue where the extra current is going.
  • voltage drop - along with the high current, my power supply goes into CC mode, and the voltage drops to 4.2V at 250mA. As I increase the current, the voltage decreases.
  • CTRL pin - the CTRL pin is used to dim the LEDs. However, feeding it a 20KHz square wave from my function generator has no effect on LED current. If I ground the CTRL pin, give it a voltage under 0.7V, or leave it floating, the LEDs will not light.

Here are some screenshots from my scope...

Waveform from between inductor and diode

Waveform from between inductor and diode

Waveform from between inductor and diode

LED voltage waveform

What I think is going wrong:

  • I had trouble soldering the TPS61169, as I stupidly didn't put thermals on the connected copper planes, therefore, I had to use excessive heat. I don't think I destroyed the TPS, as I made two of the boards, and they both display the same problems. It's less likely that I destroyed both.
  • more likely is that my inductor is saturating. Unfortunately, I don't have a different one yet, and the one I used was from aliexpress, and didn't have specs on current handling or saturation current.

As soon as I get a proper sized inductor, I will try that.

Am I correct in suspecting that the inductor is saturating? Are the symptoms consistent with a saturating inductor? Is there anything else that could be causing these problems that I am missing?

Thank you all for your help.

\$\endgroup\$
  • \$\begingroup\$ what current can your 5 volt supply produce? \$\endgroup\$ – Andy aka Dec 24 '16 at 21:31
  • \$\begingroup\$ It's a switching lab supply capable of 30V, 3A. \$\endgroup\$ – Hengy Dec 24 '16 at 21:32
  • \$\begingroup\$ Look at V across R2 with scope. This should ramp linearly across on period as current increases. If inductor saturates the current will change slope upwards - usually at an exponentially increasing rate as saturation increases. Less load should reduce this. \$\endgroup\$ – Russell McMahon Dec 25 '16 at 12:17
1
\$\begingroup\$

With a total load of 9 watts, given that the efficiency of the boost circuit might be 80%, the input power might be about 11 watts. So that would be a continuous current of 2.2 amps. However, the boost circuit takes energy in bursts; possibly 50% of the time it extracts energy from the supply and then hands this energy to the load for the other 50% of the time.

So the average power taken from the supply might be 11 watts but in reality it will be something like 22 watts for 50% of the time and zero watts for the other 50% of the time.

So, the peak demand is 22 watts at 5 volts. This is a current of over 4 amps and your power supply cannot hack it. Try just one LED to see if this works. Also double check that the peak current in the inductor is what you think it is. I have my doubts.

\$\endgroup\$
  • \$\begingroup\$ I should have mentioned that the LEDs are 3W RGB LEDs. I am only testing with one color - green - at a time, so only about 2.5W through the LEDs. Accounting for efficiency, about 3W. If it's only conducting for half the time, thats 6W. My supply is capble of handling that. \$\endgroup\$ – Hengy Dec 24 '16 at 22:22
  • \$\begingroup\$ I've been limiting the current from my supply current to about 400mA max. I now realize that this is too low. I was thinking about the 250mA limit through the LEDs, and not about the power conversion. D'oh! I was also, admittedly, nervous about putting so much current through such a small circuit. I will try increasing the limit to 1.3A or so and see what happens. I still think my inductor is underrated though... \$\endgroup\$ – Hengy Dec 24 '16 at 22:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.