TPIC6A595: Chip Fails, when power applied (5v vcc and logic lines)

Expert is out for public Holiday days. please wait for her back. thanks.

Hi Mitch, 

Thanks for your patience.

For this device, the recommended operating range of Vcc is 4.5~5.5V.

Could you capture these voltage waveforms, including VCC, SERIN, SRCLK, RCK, SRCLR, DrainX..? 

It would be highly appreciated if you can double-check and show (or describe) the complete connection of these devices.

Many thanks!

BR,

Hailee.

yes, the vcc is 5v

and doing test, when the vcc is always applied, the chips dont fail

i can show you schematic, but everything works perfectly, unless the daughter board is removed/reinstalled with power active

and power active is that the outputs of the TPICs have 12v on them though the relay coils

the design has been working for many years now, its only when the circuit was split apart, and the TPICs put on the main board with the relays

and the rest of the circuit is on the daughter board.

Hi Mitch,

From your description, I think the issue may be caused by the connection between daughter board and mainboard, or the signals from the daughter board.

Let me reorganize the current situation and confirm some questions with you.

1) Did you mean that you did the test without the daughter board, when the Vcc is applied by 5V, the chip didn't fail.

How are other signals controlled? TPIC6A595 needs to be controlled by input data. 

2) Have you checked these signals from the daughter boards what are these control logics? how are these two boards connected?

BR,

Hailee.

as i explained, the daughter board contains the micro (Atmel atmega328p) using the SPI port to connect directly to the TPIC ics
nothing special here. the daughter board also contains the 5v vcc supply... when plugging in the daughter board... the chips fail

did a test, by making a small 5v power supply on the main board, just connected to the VCC pin of the TPIC, all seems ok,

when unplugging and plugging the daughter board, this time no failures, so it seems the VCC pin must always have power.

i was thinking maybe some damage was happening because when vcc pin has no voltage on it, the IC tries to  power its logic up via power through its ouputs... it shouldnt but something is killing the ICS

i will try to get you schematic... but its so simple.... 

here is a simplified schematic

micro controller and 5v supply on daughter board

mainboard has the TPIC ics with outputs connected to relays connected to 12vdc

8233.simplified schematic.pdf

thanks for sharing, it would be high appreciated if you could also provide the mainboard schematic, I'd like to see all the connections.

Hi Mitch, 

This device is controlled by Serial Input signals not by SPI interface.

you mentioned, when 5V is powered on, all seems ok, what do you mean this "ok"? the delay can work? especially my question is about your control signals. I hope you can provide some details and capture the waveforms....

"i was thinking maybe some damage was happening because when vcc pin has no voltage on it, the IC tries to power its logic up via power through its ouputs..."

About your concern, it is also related to the output connection, could you capture the voltage waveforms? or you can change to a new IC to see if the situation is the same.

BR,

Hailee.

we have already destroyed aprox $50 worth of ICS,

like i said all works fine, no issues... we control the relays (outputs) just fine.... 

the chips fail, when the daughter board is removed/inserted with power applied

if we keep 5v on the TPICs and do inser/removal of the daughter board, the TPICs never fail

nothing to do with the signals on the logic lines.

like i said the outputs connect to a relay, the other side of the relay connects to 12v (which is always active)

oh, i cant provide main board schematics, as i don't have access to them.... i do know the outputs of the TPICs are just simply connected to a 12vdc automotive relay, the other side of the coil is 12V, there is a protection diode across the coil... 

i had the customer disconnect all 8 outputs of 1 TPIC (only 1 in board now), so that only things connected are the grounds, the VCC pin, and logic pins

inserting the daughterboard, and the IC failed. when he reconnected the outputs, they no longer controlled connected relays

dropping a new IC in, worked fine...

question: should VCC on the TPIC IC have a bypass cap? (they didn't when the whole circuit was on a single board), now that the TPICS on main board, they are a little farther away from the 5v supply (that is on the daughter board) and could this cause the failure
when the daughter board plugged in (5v is from a LM2937IMP5.0NOPB, which is in the schematic)

this is very pricy testing, as these ICs > $3each!

Hi Mitch,

For Vcc, the recommended operating condition is 4.4~5.5V, can't exceed 7V of the absolute maximum ratings.

Some descriptions of datasheet are shown as below. That means the Vcc needs to be guaranteed to work below the rated voltage. 

If the customer doesn't take any protective measures and just plug/unplug the connector, it may cause overshoot on the device Vcc pin easily, especially with a long trace.

Please add enough caps near the Vcc pin to help filter overshoot and noise.

From my understanding, the Vcc of main board is from 5V of daughter board, right? and the customer can see the Vcc waveform to verify it.

you can see the 5v supply in the schematic...

its going to be hard for them to capture waveforms, the only thing they got it multimeter, and of course that shows 5v

everything works as designed, its just when the module is plugged in with power live... it doesnt make any sense to me

the daughter board would supply the 5v, and it should work, but the chips go bad

mitch

5V is the supply voltage under stable condition, but if you plug in the daughter board with power, it can easily introduce voltage overshoot and noise to the device pins during transient conditions, so for your application, we recommend the customer to add enough caps near the pins, especially for Vcc(the max rating is 7V, slightly above 5V). This device may be damaged when the pin voltage exceeds the maximum value, so the customer must guarantee the device can work within the specs not only in the steady state but also in the transient state.

i dont think thats happening. the power supply (you can look at it in schematic) isnt overshooting

besides when the TPIC ics where on same pcb with the power supply... and the whole thing was powered up... never a chip failure

now on the mainboard we added .1uf cap to vcc pin, and chip died when plugging in module

1. It is not recommended to plug and unplug while the power is on. The better sequence is to ensure that the connection between the mainboard and daughter board is stable before powering on. 

2. Avoid "Hot Swap", if the customer wants to achieve safe hot plugging and unplugging, they can consider adding specialized protection device or components (efuse and hot swap controller). eFuses & hot swap controllers | TI.com

3. 0.1uF cap is not enough, typical value of Vcc pin cap in our other EVM designs is 4.7uF//0.1uF in general applications, the customer can increase the value based on this. Besides filter caps should be placed on both mainboard and daughter board TPIC device Vcc pin.  

From my side, the key point is not to do plugging in or unplugging while the power is on or to take protective measures.