0
\$\begingroup\$

I have the following situation: I have a LiPo battery of 3.7V, 1200mAh. I want to use it to power a microcontroller that requires 3.3V.

To regulate the power, I want to use a TPS73633 LDO, in a 5-Pin SOT23 package. According to the datasheet, it should be hooked up like this:

Schematic: typical circuit for TPS736xx

So i connected the + of the battery to IN and EN, and the - of the battery to GND. I did not add the optional capacitors.

I was my first time soldering with soldering paste and hot air, at first I soldered at 400 degrees Celsius, which was probably not a good idea. When I hooked up the battery, the LDO was getting smoking hot, and not working the way it should.

I did a second attempt at 250 degrees Celsius, but still no success. The LDO is not getting hot anymore, but the output voltage that I'm getting out of it is the same as the input voltage from the battery (about 4.0V) Any idea what I'm doing wrong?

The datasheet of the LDO can be found here: http://www.ti.com/lit/ds/symlink/tps736.pdf

\$\endgroup\$
4
  • 1
    \$\begingroup\$ You probably did a short and damaged the LDO. \$\endgroup\$
    – Bradman175
    Commented Jan 7, 2017 at 10:30
  • 1
    \$\begingroup\$ "Smoking hot" - less probable that output was shorted to ground, as they are on opposite sides. if soldering is possible via manual soldering try (with a new chip). your connections are fine. \$\endgroup\$
    – User323693
    Commented Jan 7, 2017 at 10:35
  • \$\begingroup\$ Did you use a PCB like that shown in the datasheet on page 20 and 21, figure 37 and 38? The ground plane is important for power dissipation of the LDR. \$\endgroup\$
    – Uwe
    Commented Jan 16, 2017 at 17:10
  • \$\begingroup\$ No, because that's for the VSON package, I'm using a SOT-23 package. I did however have a ground pane in my original PCB \$\endgroup\$
    – ErikL
    Commented Jan 16, 2017 at 17:50

2 Answers 2

Reset to default
4
\$\begingroup\$

That LDO regulator requires a minimum output current to regulate properly. In the data sheeet it states that the output regulates with loads between 1 mA and 400 mA. It is likely that with no load resistor the output rises up above the stated regulated output voltage.

Elsewhere in the data sheet it "implies" 20 mA is the minimum (table 1, page 16). It's almost always a good idea to use an input capacitor so don't scrimp on this either (see 8.3.3.1). Output capacitors are usually needed to improve load transient responses and this is particularly an issue on digital circuits so please use one.

If all of this does not restor proper operation then re-check the pin-outs to make sure you have not wired it up incorrectly.

\$\endgroup\$
4
  • \$\begingroup\$ Thanks for your great explanation! Just soldered on a new chip, and connected a ESP8266 for the load (This is what I eventually want to use it for anyway). I now do get different behavior, so that's probably good, also the chip is not running hot. I do not get the battery input voltage as output anymore, but I get around 2.2v, not the expected 3.3v. The power usage of the ESP8266 should normally be above 20mA (see bbs.espressif.com/viewtopic.php?t=133) Any idea what else I could try? \$\endgroup\$
    – ErikL
    Commented Jan 7, 2017 at 16:11
  • \$\begingroup\$ Maybe the ESP is taking too much current and the LDO regulator is shutting down. Maybe the 3.7 volts from the LiPo is collapsing to a low voltage (try measuring it while you have the ESP conencted). Try measuring the current or test with a fixed resistor that would take ~20mA from the output. \$\endgroup\$
    – Andy aka
    Commented Jan 7, 2017 at 16:51
  • \$\begingroup\$ The ESP actually takes ~80 mA, so that should be OK. I did discover that there was a short in my circuit (the one I was testing and where I got 2.2v out of it) after all. Unfortunately, after resolving that short, I was back to my previous situation, where I got around 4v out of it again, even with the ESP8266 connected as a load. If I want to test with a resistor, can I just connect a resistor between gnd and the output of the LDO? (Sorry, I'm quite a noob in electronics) \$\endgroup\$
    – ErikL
    Commented Jan 8, 2017 at 12:05
  • \$\begingroup\$ Yes, just pick a resistor that should draw maybe 20 mA from a 3.3 volt supply - 150 ohms will take 22 mA. Put the capacitors on though - you can always remove them to test stability after confirming the 150 ohm test. Also, is your metre OK i.e. is the battery OK? \$\endgroup\$
    – Andy aka
    Commented Jan 8, 2017 at 12:28
0
\$\begingroup\$

After many tests, I'm starting to believe that my problems are in the PCB design (or maybe a combination with my soldering skills). Spitting through many data sheets, it still seems the TPS73633 is the right LDO for the job, so I decided to start over.

For that purpose I designed a test pcb where I can just test the LDO. This is what I cooked up:

enter image description here

I added the 3 recommended capacitors, although I'm not sure the values are ok for 3.3v. Any recommendations upon this design?

\$\endgroup\$
1
  • \$\begingroup\$ Where is the ground plane for power dissipation? \$\endgroup\$
    – Uwe
    Commented Jan 16, 2017 at 17:12

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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