TPS61022: Device get ruined, when used in connection with LM2662

Hi, 

The pin 5 OSC of the LM2662 should be floating, and the OUT pin should be connected to GND.

There should be a 10uF capacitor placed for the input of the TPS61022.

Please share the layout of the TPS61022.

Dear Jasper,

the vias at L1 are directly connected to +2.5V (i.e. a EDLC). GND of EDLC is the via in the left bottom corner. Because of the EDLC as an input of TPS61022 and beeing very close, I omitted the ceramic input capacitor. Sorry for the confusion in the schematic, but OUT and LV of LM2662 (pins 5 and 6) are connected to GND and OSC (Pin 7) is floating. Please see layout below. To me it's strange that both "boosters" work separately, but not in combination. I tested both out them in below's layout.

Best regards,

Matthias

asper,

Hi,

I am working with tps61022 for two months and spend a lot of time on it. I assembled about 20 PCB and I had 40pcs tps61022 and i burned about 21 of it.

during qualify control of PCBs, I saw strange behaviors from tps61022.

please take a look at my question here:

http://e2e.ti.com/support/power-management/f/196/t/824926

some of those was working fine when the battery was connected directly. but then as soon as I placed an amperemeter in its input, burned.

some of those burned immediately after I connected the battery.

one of those burned when my MCU hanged and repeatedly turn on and off an LDO which was after tps61022. the output of LDO was connected to a capacitor which drains a lot of current at the startup.

I work hard and tried different values of input and output capacitors. the problem seemed to be solved but still, some of them burned with no special reason. I think there is a problem with the start-up of tps61022. most of those burned immediately after I connected the battery.

when they burn SW become connected to GND. its low side switch burns I guess.

 you can see my layout in the link above. I think in some situations tps61022 became confused and burns.

I would be grateful if TI inspects tps61022. I manufactured my PCBs and I'm really worried about the quality of my final product. it should pass EMC tests.

Dear Jasper,

TPS61022 should be used in my application for boosting EDLC. So I cannot reduce input capacitor, as alireza did, because I need the EDLC...Is an additional small ceramic required then? ESR of the EDLC is quite low.

Do you think that LM2662 has such a big inrush? Capacitors are small 10uF. There was no load an the LM2662, because I had not yet connected it to the consumers, which should later be powered...

Really looking forward to your opinion? Or du you think that TPS61022 is not the right choice? I need to boost 2.5V (and also 1.2V, if EDLC is discharged) to 5V. I should be 200-400mA.

Regards,

Matthias

Hi Matthias,

I didn't suggest reduce the input capacitor, but add a 10uF ceramic capacitor because the EDLC would be high impedance in high frequency.

the behavior is unexpected behavior, as we are still trying to reduplicate the behavior in the lab.

could you try to disable the charge-pump firstly, after the TPS61022 finishes startup and VOUT reach 5V, then enable the charge pump? this may help us to find the root cause.

Hi Matthias,

I have sent friendship to you. Could you accept it? We can discuss this case deeply.

Dear Zack Lui,

sorry , I did not get any email. I'm really interested in a clarification.

In the meantime I checked several topics after adding the 10uF input capacitor:

Running just the TPS61022 without LM2662 doesn't burn the TPS. When I later on add the LM2622 it's still working. However, I realized that there is a really high in rush current, when I established this connection manually. There had been still no load on the LM2662. The whole system (i.e. TPS61022 + LM2662) is consuming approx. 25mA at 2.5V without any load. Is this ok? To me this seams to be quite high...

If I start-up the whole circuit, in approx. 50% of the cases the TPS gets blown.

Now I've left just one TPS61022.

Do you have any proposals.

Reagrds,

Matthias

PS: I'm a bit unsatisified that the TPS61022 seams not to work as expected per datasheet and that on top of it TI refuses to provide any samples of the device.... I've lost more that 20 of it.

Hi Matthias,

i had sent you a message with my address, please send me one board. we can't find the issue with the EVM. 

can't you apply the samples in TI.com. if not, let me know your address and company name. I can try to provide some samples.

Dear Jasper,

For reduplicate the behavior i suggest you to put some (e.g. 5pcs) 100uF low ESR tantalum capacitors and some 47uf ceramic capacitors in paralle and an additional resistive load. this setup will have a larg inrush current and you can inspect the behaviour of TPS61022 in handelling the start up.

Best Regards

Alireza

Hi Alireza,

are the capacitor and resistor placed in the output or input of the boost? we can try this.

could you also apply a TPS61022EVM and check if you can reproduce the issue in the EVM? this would help us a lot.

Dear Jasper,

Capacitor and resistors are placed in the output.

Dear all,

in the meantime I found out that it doesn't matter whether I put a 10uF input capacitor to the TPS61022. I think that the TPS61022 blows-up due to inrush of the LM2662, because:

- TPS61022 stays intact if I don't solder the LM2662

- TPS61022 stays intact of I connect the LM2662 after start-up

- TPS61022 stays intact, if I connect both ICs from the beginning, but on separate pcb, just connects with two wires (ground and output of TPS)

As I have blown now all my chips, except one, I can no longer do testing.

However, is the solution then to add an inductor in between LM2662 and TPS or any small resistor?

Regards,

Matthias

Dear Matthias,

Me too found out it was not directly related to the input capacitor.according to my experiments I think the point is small input capacitor doesn't regulate the voltage at start-up and cause the input voltage drop at inrush current and it limits the inrush current with the coast of larger start-up time. it's the possible reason why my tps61022s blowed-up when I was using large input current and worked correctly when I was using a small input capacitor. by the way, it is nonsense that a large input capacitor which is a mean of regulating input voltage and should be used for power stability , make a device blow-up. the problem is definitely somewhere else.

I look forward to Jasper to do an inspection on tps61022 start-up under high inrush current situation as I suggested in my previous post.

Regards,

Alireza

Hi Matthias, Alireza,

thanks for the information. we will try to find the root cause and solution.

Hi Mathias,

as discussing through email, we are still work on this issue and expect to have a update this week.

we still can't repeat the damage in TI's Lab with the customer board. Based on the  discussion, the issue seem to relate to the input super capacitor characteristics. we will try to find a super capacitor and do the test again.