TPS51220A: TPS51220A incorrect frequency

Hi Adam,

For your questions:

1. No. The electrical performance if Auto version is as same as before.

2. Because this controller has two control modes: D-CAP and CM. Which one are you currently using? Can you provide a schematic diagram?

In addition, the maximum input voltage of our product for long-term operation is 28V, not 30V. Looking forward to your reply. Thanks.

Aurora

We use CM mode.

We use VIN max 24V.

We don't have schematic.

Hi Adam,

Sorry, maybe we still need your schematic or some waveforms to see why the frequency is wrong. In addition, I‘m not sure where the 30V come from? "By increasing voltage in ST-BY current doesnt decrease and by reaching 30V inductors are warm "

Aurora

RF=330k it means frequency circle 300kHz

SKIPSEL1=SKIPSEL2= GND  , continuous conduction mode

TRIP=5VREG  , discharge on

FUNC=5VREG , current mode OVP disable

It is standard aplication.

Hi,

I'm sorry, actually, I don't quite understand your current question. I know you think the frequency is wrong, but under what circumstances did you measure the frequency? Do you have a specific waveform? What are the voltage and current of the two channel outputs? Under what operating conditions did the problem occur?

And I see you have been using it since 2016, didn't you notice this problem before? Or did you make any changes to the design?

Can you see how different it is from our EVM? Because the frequency anomaly may be related to external parameter configuration and layout. If I can't see your actual circuit diagram and test waveform, it's hard to help you make more judgments. I hope you can understand.

outputs 3V3 and 5V0 , light load.

Wave forms square , frequency circle 150kHz on bot channels by VIN 10V, current consumtion 200mA, light load.

Increasing VIN up to 24V current consumption 220mA and inductors warm .

By increasing VIN duty changes to lower and current should be lower , it is incorrect.

Hi,

Can you see the waveform of the inductor current (IL)? Will there be negative current? And when you say the frequency is 150kHz, does it apply to all input voltage ranges or does it only apply to 24V? Will the frequency become the 300kHz you set under heavy or full load? Because this material is relatively old, we don't have its EVM, so we can't do more testing. If possible, please upload your test waveform, thank you.

Aurora

It is something bad in IC or project. If I touch surface of TPS with screwdriver current decrese. If connect screwdriver to GND current again a bit lower. It is creazy reaction of TPS51220A. We orderd Q1 version for checking.

Frequency on one channel is circle 300kHz , on second floating and unstable. Can not work with oposite phase what is advantage of this IC.

Hi Adam,

Got it. Pls change the IC to see if the circuit still have the same problem. If the problem just occur in this IC, you can contact the FAE to do a FA for this IC to double check. Thanks.

Aurora

Hi Aurora , After changing converter for Q1 version is one level better. Changing VIN in range from 5V to 25V we have 300kHz up to 15V. By 16V 3V3 channel switches himself for 150kHz and duty is double (normal) ripple double (normal), ripple on VIN (normal) and current consumption increase (incorrect). Before and after mathematical integrals from wave on mosftes are very similar for both frequency. Now the final question: what is construction idea TI engineer of switching frequency ? What conditions cause it ?

4150.TPS51220A FSW issue.xlsx

Please look on xlsx picture. I have smilar situation. after switching frequency all step-down parameters are worster.

I need to know mathematical reason of this switching done in chipset. Main idea of step-down converter gives constant power consumption from PS independent on VIN. It means: VIN up , current down. Mathematical VIN x  I=POWER= constant.

POWER= integral from wave on inductor. Of course POWER increase by lower frequency because converter needs compensate it with longer duty (it gives higher current).  Values of integrals for both cases should be similar and proportional to summary output consumptions LDO3+LDO5+VOUT1+VOUT2. when LDO are without load , will work only integral from VOUT-s.

Hi,

Does this problem occur when the power is first turned on, or after the output voltage is switched? Can you measure the change in RF pin voltage when the frequency is wrong? Also, this error can be reproduced repeatedly, right? I don't have an EVM now, but I have applied for several EVMs. If I can reproduce it repeatedly, I will test it after receiving the EVM. But this will take some time.

Aurora

The problem is in ideology of this IC, To many posts about frequency, You can not give exactly answer about mathematical idea .

This chip works in opposite direction. By changing voltage up (lower current) it switches frequency to half and causes double ripple on VIN and VOUT . It is abnormall. Our problem is temporary solved - using MB modification we have full range VIN and 300kHz on both channels. And VIN current of course decrease by increasing VIN .

Hi Adam,

Sorry for the late reply because of weekend. Yes, I can't give you a clear explanation at the moment. Because from your description alone, the chip is not working properly. But there are many possible reasons for the abnormality: application error, improper layout, or some bugs in the chip itself. In addition, because this product is relatively old, I don't have a suitable EVM for experimental verification, so I can't give you a conclusion. 

If you can currently find a way to make the chip work properly according to the settings, you can use this method first. Or you can also look at our other updated converter products, which will be easier to use. Thank you.

Aurora