Dear E2E,
we are evaluating the TPS61010 boost converter using SLVP157 evaluation board. The target use case is tested by simulating a constant load with 30 Ohms and simulating a single primary AA alkaline cell input voltage with a laboratory power supply unit (0.8V...1.6V). All measurements are taken by Fluke/Agilent multimeters. Unfortunately, the device seems to have some type of effciency fall-off at a battery voltage of approximatly 1.3V (see test data in second text box). What is the reason for this effect and how can it avoided?
Modifications on EVM:
L1 replaced with Wuerth 74477710, C2 replaced with 100pF, C0G.
Some attempts we made, trying to fix this efficiency problem:
1.) Compensation network; varying C2 (47pF...150pF) as mentioned in datasheet affects efficiency in all test points. 100pF brings the best results.
2.) Other inductors; Wuerth 744053100 -> less efficiency in all test points, Wuerth PD series with 22uH -> the efficiency notch seems to occur at a lower input voltage (~1.2V) => Is here some kind of resonance including the inductor that acivates the anti-ringing circuit of the TPS61010?
The current into L1 (at SW pin) was measured using a current probe/scope. The TPS61010 operates in continues mode all the time and the points of lower efficiency seems to occur if a duty-cycle of 50% is reached. Having some kind of hysteresis, the lower effiency (=higher input current) points can be found at fixed, repeatable voltages, when decreasing or increasing input voltage.
Best regards
Gunnar
