TPS54340: Overshoot in output 5.0V of TPS54340

Hi Jagdish,

Try to put a aluminum capacitor maybe 22-47uH with maybe 50-100mohm ESR?

Thanks
-Arief

It's possible that when the boost converter is enabled, the board is more noisy and noise affected the operation of this buck.

I would suggest reducing the switching frequency of the TPS54340 to 500kHz or 400kHz. The 15uH inductor works for 400kHz or 500kHz. Actually it works better at lower frequency. It is too large for 1MHz 3A buck, making the control loop to have low signal to noise ratio, which makes it sensitive to noise. You will need about twice the amount of Cout. Compensation may need to be adjusted as well.

Hi Arief,

I already tried adding 100uF electrolytic over 44uF MLCC but this didn't helped. Also the overshoot is not constant, sometimes it is 5.25V & sometimes it is 5.52V. This is irrespective of adding electrolytic capacitor. However, it is never more than 5.52V.

Regards,
jagdish

Hi YangZhang,

Please find my comments as below,

1. Made freq=484KHz & as per datasheet equations modified  various components as RT=200K, Cout=94uF, R4=5K (as per datasheet), C5=54nF, C8=120pF, L=15uH but still issue didn't resolved.

2. Observed that if VIN is discharged completely & we wait 60s or so we didn't get overshoot, please find attached waveforms.

3. If VIN is not discharged completely we may get overshoot. I do checked other voltage rails but all return to zero instantly after power off except VIN. I am not sure why VIN requires some time in between after discharge for no overshoot.

4. Output also is sometimes non monotonic. Attached waveforms.

5. Also captured inductor current at power on if that helps.

6. Doubting particular sample, checked performance in other sample also with 963K freq but is same.

The above behavior is irrespective of freq. change to 500Khz, meaning the same also exists in 963KHz.

Any further inputs or clue to resolve this issue.

Regards,

jagdish

Hi YangZhang,

Above observations seems to be absurd but that is what I have seen. I confirmed the things in second non modified sample with 963K freq. However, one thing is sure that when VIN is brought to 0V after 30-40 seconds probability of getting overshoot gets reduced a lot but it still happens some times.

=> One thing I was thinking about of current slew rate during power on & various capacitor charging in circuit. I believe TPS54340 has fixed soft start, not sure if any slew rate change in current can still cause overshoot. Better you can comment & advice.

Regards,
jagdish

First observation from the attached scope shots is that the timing when TPS54340 is enabled is not consistant in Figure 1 and Figure 2/3. The VIN waveforms always stuck at 4V for some time before rising all the way to the final value.

In Figure 2 and 3, Vout rises pretty fast (2.1ms soft start time typical) to 5V, comparing to Vin slew rate. But when Vout tries to reach 5V, Vin is stuck at 4V. Vout continues to rise after Vin is high enough. That's why Vout waveform is non-monotonic. The difference between the two, I believe it is because Vout has some load in Figure 2 and no load in Figure 3.

In Figure 1, you can see Vout started to rise after the 4V flat piece of Vin. When Vout reach 5V, Vin is barely 5V, so a small dip there is understandable.

During the start up of the TPS54340, because of the 3.5A peak inductor current (needed to charge the output capacitors), Vin will slow down rising or even have a dip, as seen in all three figures.

Please check how the TPS54340 is enabled. I'd suggest adjust the EN threshold such that the part will be enabled after Vin>5V. Then the non-monotonic issue shouldn't exist anymore. I still recommend change the design to a lower frequency around 500kHz.

I can review your schematic and layout if you prefer.

-Yang

Thanks for your continuous support. Please find response as below,

=> EN pin was open in my application circuit. However, adding voltage divider to EN pin of GND to 10K to 22K to VIN didn't resolved the issue. I still see non monotonic output & this is perhaps due to loading of output. Frequency is already 500K. Pls find attached different waveforms. 

=> Regarding your queries of having some load in Fig:2 & no load in Fig:3, please be informed that load is same in both scenarios.

=> Also took waveform after removing MIC2288 boost converter IC & later I see no overshoot in output. I checked this 15-20 times. Attached are scope shots after MIC2288 removal. FYI, I don't see any soft start feature in MIC2288 & hence doubt more that.

Please share your email ID, so I can share schematic & layout with you.

Awaiting your further response,

Regards,

jagdish

Please send your files to simpleswitcherapps@ti.com

-Yang