TPS61088 - Cannot get enough current

[Sphade]

I am trying to create a circuit that can take the input from a Li-Po battery and output a stable 5V with a maximum current draw of the load of 4A.

After checking the datasheet and using the design tool provided by the TI website, I managed to get this circuit:

And here is the layout just in case someone finds something wrong with it (left side is from something else):

After populating the components (only for the Boost converter, nothing else on the circuit) I noticed I can only draw about 1.2A before the voltage drops to around 4.3V when I hit the 1.3A threshold. Until that threshold is slowly dropping to about 4.98V at 1.2A.

I cannot understand, for any reason, why this is the case. I have to be honest here and mention that I have not actually checked the circuit myself and entirely relied on the TI design tool (a mistake on my side).

Inductors I have used with the exact same result:

• LMLP07A7MR68DTAS
• MWSA0603S-R68MT

To test this I am using an electronic load and a power supply which I have to mention that are NOT the problem. They are high end equipment capable to supply/withstand thousand of watts of power.

I have also not populated the LDO in the schematic and tried to draw current from the VDD5V pin.

At this point I have exhausted all ideas that came to my mind.

Thank you in advance for any help!

 

[dick_freebird]

Before you get all excited about placing blame,
look at the battery voltage at the converter input
(is it sagging?) and the ILIM pin (where does IOUT
(w/ its set-resistor) cross advertised threshold)?

 

[Easy peasy]

Dick has the good suggestions to follow first.

also the wires to the board - dropping volts ?

--- Updated Yesterday at 10:43 PM ---

OK - it's pretty easy to see the issue here - the freq is set for 267kHz, the curr limit is set for ~ 8A

say we are boosting from 3V to 5V, the steady state D will be ~ 0.4 giving 1.5uS ON time

V/L = di/dt, gives 3/680n = 6.6 amps in 1.5uS, once you get to 1.3 amps ave out the peak is now 7.9A - i.e. the peak current limit is cutting in because the inductor is too small - conversely the sw freq is too low,

fix one or other or both and you will get more Iout & Vout.

 

[Easy peasy]

What ... ? Oh - you're welcome . . .

 

[Sphade]

Hi guys,

Thank you for the reply!

@dick_freebird to address your question, I tested the circuit with a power supply and the voltage was 3.3V steady. No sagging at all. I didn't quite understand the second question. Can you please elaborate?

@Easy peasy The frequency is set for around 598kHz based on the calculation given in the datasheet:

As for the current limit it is set to around 11.9 A since in my actual circuit I have a 100kOhm resistor for Rilim:

If I am not mistaken, and didn't mess up the math, I should have about 3.44A of peak current through the inductor which should be just fine.
Even with your assumption of 267kHz sw frequency, the 7.9A is below the set limit of 11.9A which is set by the RIlim resistor.

So i guess that unfortunately does not solve my problem.

Additionally I have simulated the circuit with the model provided by TI themselves in LTSpice and with the exact same component values the circuit seems to work without any issue.

This leads me to believe that the issue is related to either a layout problem which I cannot spot, or with values of the components or parasitics, but I cannot really even think where to start checking.

 

[mtwieg]

Sounds like you should do some basic measurements on the actual circuit:
1. Verify the switching frequency is what you expect
2. Check the Vin pin voltage waveform. Even a few tens of milliohms of source resistance might be causing Vin to dip into UVLO territory.

 

[Easy peasy]

Yes - verify the Fsw, I simply read it from the graphs in the data sheet with the given R in the sch ! - and also that the inductor is not saturating ( build a square 4 of the L's and solder in temporarily on wires, this gives the same L, but twice the current capability )

I note that typical inductors used in similar apps are > = 1.2uH ( food for thought )

also we can't see if your compensation components are causing instability ( scope on V out )

you can try a small snubber from the sw node to local power Vin, 1 ohm and 220pF, this may help with noise on the current signal

and - as a last resort, a 10uF cap right on the IC from Vin to gnd, shortest possible legs