TPS564247: Short circuit protection for TPS564247 and TPS2117

Hi Hossam,

The device has over current limit protection, please see section 7.3.6 in datasheet. Thanks.

Regards,

Shipeng

hello Shipeng,

so the device when short circuit happens will immediately shut down with no harm to it or anything else can occur? i read that section but i guess it is for over current but not short circuit. they are both the same initially but short circuit is a lot more current than just overload. 

can you verify the behavior using your dev board? i mean run it regularly and short the output voltage pin to ground and check the behavior. i highly appreciate this.

what about tps2117? since this one is very important as i put it after the buck regulators to choose the desired output. i need this ic too to have some sort of protection for short circuit condition. please link the expert of this ic here to answer.

but overall, since buck regulator comes before this tps2117 ic.. when short circuit happens, will the buck shuts down before any high current reaches the tps2117 ic or not? still i really like the tps2117 to have protection of its own.

thanks !

Hi Hossam,

Short circuit is part of overload at our side, so we always do short circuit test when develop the device. It's OK for device to protect when short circuit happens. please see Figure 8-19 to Figure 8-21 in datasheet. I don't responsible for TPS2117, you need to create a new thread.

Thanks.

Regards,

Shipeng

hello,

i saw the figures but i have some questions about them.

in figure 8-20 i see the purple switching waves are still happening during the high current (short circuit) situation, then it stops. what i understood is that when short happens for roughly 300uS according to figure, the switching changed its behavior then after this brief period of 300uS the switching stopped and output voltage stopped to 0v as well. but during the period when switching existed, output voltage was at 0v.

is my understanding correct? please explain the situation for me as an extra understanding always better.

also, i don't understand figure 8-21 and what it does. please explain it to me. 

when short circuit finished, does voltage and current return to normal operation or i need to to power off then power on?

thanks !

Figure 8-21 is shows when keep short circuit, the device will try restart every 16ms. So if short circuit is removed, the device will successful restart, you don't need to repower on.

hello,

in fig 8.21, if it keeps trying to start while in short, will it cause damage if stayed for a while? i see it bumps itself to around 0.25v and 7-8 amps for a very short period (2ms). how much can it tolerate in this status? 

you didn't comment on my words for fig 8.20.

Hi,

It's safe when work in hiccup mode, no damage will happen. Figure 8-20 shows when short circuit happens, the device will continue to work for 256us then shutdown and go into hiccup mode show in figure 8-21. During work time, the current will be limited at 7A. It's also safe.

hello,

nice to know it is safe. 

however, when exactly does protection kicks in? meaning at which current? the rated current is about 4 amps, does protection happen at 5amps for example?

 please see section 7.3.6 in datasheet. Thanks.

hello,

i have read it but i think it does not answer my question. my question is about which output current is the max the device can offer before short circuit protection happens?

ultimately i only want a confirmation that the circuit will be safe when short happens, but it is good to know the max current allowed.

thanks

hello,

so yes 6 amps is the limit. and it takes roughly 30 uS to sense the short and shuts down as i see from the figure 8-20.

in my design there is a buck for 5v connected directly to output pin, this one i need it to shut down after the 6 amps or so.

the same 5v rail is going into tps2117 then from that to the output.. here was a problem... since tps2117 does not tolerate more than 4 amps thus the current must be stopped.

tps2117 support here suggested that I use an eFuse just before tps2117 and after buck output caps.

the efuse i got is TPS259533 calibrated to shutdown at 3.7 amps, so I guess it will shut the current going to tps2117 before it even reaches 4 amps. while at the same time does not affect the current for the 5v going to output pin directly.

do you think there is any more danger i need to worry about?

hello,

i have modified my design to use a big physical switch which can tolerate up to 16 amps normal operation.

i do not need accurate protection in this case, just a quick and safe shutdown when short circuit happens. i guess now the circuit should be ok and your device can handle such requirement. correct?

Our device TPS564247 can quick and safe shutdown when short circuit happens.

hello,

yes thanks, i have verified that.

however, what happens when you short 2 rails together using this regulator.. like 3.3v and 5v? what if i short another rail like 9v or 12v but with a linear regulator to this switcher output... does it also go into hiccup mode or what protection?

another thing i have: right now on the new design i have total capacitance after dc bias at 3.3v to be around 116uF which is composed of:

3x22u 1206 X5R

6x10u 0805 X5R

1x4.7u elec. cap for damping, low capability and high esr.

1x22uF Panasonic polymer 35SVPF22M

1x100nF 0402

I have 150pF FF capacitor and 1.5uH main inductor.

I tried calculating the Fp in equation 4 to be 1 / [2*pi* sqrt(L*C)] = 1 / [2*pi*sqrt(1.5x10^-6  *  116x10^-6]> it will be Fp=914. so this in khz? is this ok?

previously the circuit worked fine without the panasonic capacitor but i'd like to add it to see if it can help even more, not required but nice to have.

will this amount of capacitance be ok?

thanks

I will reply after weedkend.

hello,

will be waiting for this, as i am about to send a prototype pcb order and would need to make sure it will work.

i hooked a crappy 33uF elec. capacitor to 5v output and it didn't show any bad thing but that cap was kinda high esr bad quality. the polymer is 22uF 35mOhm good quality.

i can wait for monday but at least kindly tell me the equations of which to know if it is stable or not. as i listed my attempt at eq. 4 above but didn't reach the same result you showed in a previous thread (using different caps). \

you will respond monday?

thanks

Hi,

There's Pspice model for TPS564247, you can run simulation for your design to see if it's stable.

hello,

unfortunately, i am not really good at simulation, only did basic stuff and never stability stuff like phase and margin. 

you previously posted some calculated values but when i tried to use equation 4 i didn't reach the same result. you may guide me to how you did it please. 

or how to do the simulation easily to obtain the stability results. i tried webench for this part and always reverts back to 1v output and does not accept my modifications. does not have stability results too.

 Hi,

1 / [2*pi*sqrt(1.5x10^-6  *  116x10^-6)]=12065Hz, please check.

And simulation is a good way to verify the stability, please try to learn that. Thanks.

hello,

looks like your equation is correct. so the result is 12khz, datasheet suggested this pole to be at roughly 30 khz. so for our 12khz result, how does it do? i mean is it on the good or bad side according to datasheet info? notice that i am also using 150pF Cff which will help.

i will install TINA tonight and try to do the simulation, but kindly give me what type of simulation i must choose and how to see the results. i think i need to put an ac source at the feedback pin of the device, correct? 

if you do have a simulation file for this device please send it here and i will do necessary adjustments.

thanks 

Hi,

As you put more capacitors than we recommend, the pole becomes lower, so you need to use Cff to improve transient performance. Another way is apply for/buy a TPS564247EVM board and do experiment on it. 

hello,

your recommendation of 100uF maximum, I put 16uF more or so. how much of a difference would that be?

i have a working design, I added a 33uF and worked fine. however, this 33uF is a bad electrolytic capacitor not low esr polymer.

i could add more ceramic capacitors equal to the additional polymer and see the result. but I only have an oscilloscope (no bode plot) and my regular application + dc load. previously i tested it with these to be very nice clean rail even at 4 amps.

so if i got the same end result with this newly added capacitance, can it be considered ok? i mean without doing bode plots.

you mentioned using Cff, and i actually use 150pF within the recommendation. if i exceed the 100uF limit by a small amount or so, will the Cff be enough? like, what bad thing can happen if i put a larger Cff like 220pF for all cases?

sorry to keep bothering you with questions but I really like this to be my final working prototype.

thanks!

hello,

Previously TINA TI didn't accept the model, now I am trying PSpice for TI.

i think it will do ac analysis with 1v ac injected into feedback pin right? if not, please tell me the correct testing method so i can work it.

Hi Hossam,

Pspice is difficult to get bode plot by a transient model, so a better way to check loop stability is run load transient and check Vout waveform.

Hello Shipeng,

you mean i run transient in pspice for ti and program a load transient and see the effect?

still i'd like to do bode plot

in my run i used about 122uF capacitance (no dc bias settings so i guess it is full 122uF) at the output with 150p Cff. i saw the 3.3v at the output but there is ripple which is not small as the one i actually measured in practical circuit. 

thanks

Hi,

Yes, run load transient and see if there's oscillation on output voltage.

hello,

i will try that today, but i noticed the simulation is so slow despite having a powerful computer. i had to make time to 100m seconds to be able to see the final 3.3v result as it spent about 50ms to reach it. is that normal?

any tip to make simulation settings better?

also as i told you, i noticed ripple is not small, did you get the same result? i would like to get your simulation file if possible.

at the same time, i will solder more capacitors on my currently working board and see what happens... i can solder 2x22u 1206 extra caps.

however, my application does not have any load transients... just regular load forever. if so, then you think nothing wrong can happen in terms of stability? 

Hi,

You can setup initial voltage of all your output caps, then the circuit can enter steady state quickly. Transient simulation can help see margin of stability, help prevent potential instability considering variation.

Hello,

I am trying simulating it again to see if it got better this time with small tweaking. i don't know how to get it faster or put initial charge in caps anyway.

I will send you my simulation file too once i finish simulating this session.

Anyway, I tried adding 2 of this capacitor: https://product.samsungsem.com/mlcc/CL32B226KAJNNN.do

to my already working design... nothing was bad at all, i tried running it on DC load which i moved the load fast but still did no problem. I hooked it up to my final target device and ran perfectly fine.

this cap is 1210 22uF X5R, at 5v DC bias it is about 17.5uF meaning 2 of it are gonna be 35uF which is well above my added 22uF panasonic polymer cap.

so I guess it will be ok, right?

____

more info:
my capacitors REAL data are as follows:

CL21A106KAYNNNE
10uF 0805 X5R
7uF   @ 3.3v
5.1uF @ 5v
1.5uF @ 15v



CL31A226KAHNNNE
22uF 1206 X5R
17.5uF  @ 3.3v
13.65uF @ 5v
4.4uF   @ 15v

for 3.3v rail i have 6x10u + 3x22u which means in dc bias: 6x7 + 3x17.5 = 94.5u + 4.7u elec for damping + 22uF polymer = 122.1u total capacitance.

for 5v rail i have 6x10u + 3x22u which means in dc bias: 6x5.1 + 3x13.65 = 71.55u + 4.7u elec for damping + 22uF polymer = 98.25u total capacitance.

Today's test was on 5v rail which has 71.55u ceramics + 10u elec for daming (will be replaced with 4.7) = 81.55u --> I added 35u now -> = 116.55u and still stable and showed no problem.

do you think it is fine this way after all these tests? like almost 117u was stable today so an extra 5u won't be bad... not to mention that there is 150p Cff there.

thanks

hello

sim file here: https://filebin.net/trmnuck9ly3skjvh

after many hours it is complete and the result as you see:

same high ripple.

i noticed that simulation result file size is 25 GB which is insane!!