fault IPS160H

[apenn.13]

hello,

I use the component in a board to control gas solenoid valves in a cooking plane.

I need to know for the regulations if in case of any fault or component breakage I always have a low or high impedance output value so as not to always excite the solenoid valve

That is to have a short between a VCC and OUT? or not?

[Domenico RAGONESE]

Hallo @apenn.13​ 

Both events (hiZ or short) can occur to the integrated power switch between Vcc and Out pins of IPS160H.

You can detect such event on DIAG pin (activated when IN pin is driven low by the MCU) thanks to the "open load detection feature".

However, if you need to guarantee the HiZ in case of any fault (and consequently protect the load), then you need to introduce some redundancy.

I mean, you need another IPS160H (or a P-channel) in series to your IPS160H (you'll have primary switch and secondary switch; the latter connected to the load)

MCU has to check the diagnostic signals and switch off the primary switch in case of short on the power integrated in the secondary switch.

Hope it helps.

Best regards,

Domenico.

[apenn.13]

HI,

SO DO YOU CONFIRM THAT THERE MAY BE THE CASE THAT FOLLOWING THE BREAKAGE OF THE IPS160H THERE IS A SHORT BETWEEN VCC AND OUT EVEN IF THE IN IS LOW?

BYE

[Domenico RAGONESE]

Hallo @apenn.13​ ,

the short between Vcc and Out can occur both outside IPS160H (at application level) and inside the IPS160H (power switch in low impedance whatever the IN pin status is).

However, the DIAG pin can report the event independently if internal or external to IPS160H.

The IPS160H is really robust, but internal failure can occur from the violation of the AMR (e.g. Voltage or EAS violations...).

Hope it helps.

Best regards.

[apenn.13]

hello,

the spark to ignite the gas flame causes high voltage pulses both on the component inlet and on the solenoid valves out and sometimes they break when smoking.

Aren't they protected by esd?

the component is powered at 48V, just smbj48ca as tvs both on in and on Vcc?

better to put Ccutoff = 1nF?

Regards

[Domenico RAGONESE]

hallo @apenn.13​ ,

the IPS160H is protected against ESD.

However, depending on the ESD level of your application, it could be necessary to further protect the inputs by a resistor in series to the IN pin and by a capacitor between IN pin and ground.

It's not clear to me if:

• the damage in your module are on IPS160H or some other component.
• all the modules are damaged or only a subset

The SMJ48CA (Vclamp = 77.4V) could not enough to protect the internal clamp of the IPS160HF (AMR = 65V).

Could I suggest to see the behavior replaing it by SMJ40CA (Vclamp = 64.5V)? Drawback would be ~1mA leakage at 48V (due to lower Vbr).

About Ccutoff = 1nF: it may helps when the power switch is activated and an overcurrent event is activated.

If the problem comes from some overvoltage I don't think it can help.

Best regards,

Domenico Ragonese.

[apenn.13]

Thank you.

I have another question.

Can i put 5 V to Vin to activate Out with Vcc = 48V?

[Domenico RAGONESE]

yes, you can.

The input pin of IPS160HF is compatible with 3v3 and 5v0 logic levels.

(see table 7 in the datasheet).

[apenn.13]

HI,

I STILL HAVE PROBLEMS.

I HAVE MODIFIED THE CIRCUIT AS IN THE ATTACHMENT, BUT IN THE PRESENCE OF SPARKS THE COMPONENT BREAKS.

IN PARTICULAR IN CORRESPONDENCE OF PIN 7 EVEN IF THERE IS NO LOAD OUTPUT (OPEN)

WHAT SHOULD I DO AGAIN?

[Domenico RAGONESE]

@apenn.13​ .

We need to investigate even deeper:

1) can you send me a picture of the damaged IC?

2) Do you have the possibility to monitor the voltage between Vcc and GND (e.g. pins 1 and 7) during the spike?

3) Could you re-do the same acquisition, but with a short in paralle to R275 (and D24)?

4) Could you share the layout of your pcb?

I want to check the Vcc path and the position of the TVS D16.

Thank you in advance.

Best regards,

Domenico.

[apenn.13]

PIN 7 REMOVED WITH THE BREAKAGE OF THE COMPONENT

BROKEN

I HAVE ALREADY DONE THE TEST TO SHORT CIRCUIT THE RESISTANCE AND THE DIODE, BUT I HAD THE SAME RESULT

I ATTACH VIDEO WITH SPARK . THE YELLOW TRACE IS VOUT.THE BLUE TRACE IS VIN. THE SCALE IS 20VOLTS.

TOP AND GND LAYER ( THIRD IPS160H ON SCHEMATIC )

TOP LAYERTOP AND BOTTOM LAYER

TOP AND VCC LAYER

[Domenico RAGONESE]

@apenn.13​ 

I've checked the layout and I can see many potential dangerous path from supply rail to Vcc pin of the IPS160H.

Similarly on the ground path, too.

Before to spend time (and money) on a new layout I'd try with some hand patches:

• maintain SMJ40CA as TVS
• maintain the short in parallel to reverse polarity protection diode
• connect by a short wire C207(+) directly to Vcc pin (#1 or #12) of IPS160H.
• connect by a short wire C207(-) directly to GND pin (#7) of IPS160H.
• do the same for C206(+), C206(-) and C208(+) and C208(-)

The actions above purposes to reduce the parasitic effects on Vcc and ground nets due to many vias on the route to IPS160H.

It would be appreciated If you can capture the waveform between Vcc and GND (shortest path as possible from IC pins) to see if any improvement is visible: we must exclude that the IC is destroyed for the activation of the internal clamp between Vcc and GND).

Hope it helps.

Best regards,

Domenico.

[apenn.13]

I HAVE MADE THE CHANGES YOU TOLD ME AS IN THE FIGURE RESISTANCE OF 10K AND 100nF INPUT CAPACITOR TVS IN SMAJ5.0CA INPUT SINCE I SEND A 5V IN INPUT INSTEAD OF THE 48 TVS ON VCC 40V

IN THE PRESENCE OF THE SPARK IN THE AIR I HAVE ON OSCILLOSCOPE THE VOLTAGES ON VIN (BLUE) AND ON VCC (YELLOW) HIGH EVEN IF FOR A FEW USEC.

THE VIN IS 5V AND IC IS ACTIVE WITH 48V ON VALVE.

CAN THESE VALUES BREAK FOR THE COMPONENT?

WHY DON'T TVS FLAT THESE VOLTAGES?

THE PROJECT IS VERY IMPORTANT AND THE PRODUCTION HAS HIGH NUMBERS (250,000 YEARS)

[apenn.13]

I HAVE MADE THE CHANGES YOU TOLD ME AS IN THE FIGURE BELOW

I HAVE MADE THE CHANGES YOU TOLD ME AS IN THE FIGURE BELOW

RESISTANCE OF 10K AND INPUT CAPACITOR 100nF

TVS IN SMAJ5.0CA INPUT SINCE I SEND A 5V IN INPUT INSTEAD OF THE 48

TVS ON VCC 40V

IN THE PRESENCE OF THE SPARK IN THE AIR I HAVE THE VOLTAGES ON VIN (BLUE) AND ON VCC (YELLOW) HIGH EVEN IF FOR A FEW USEC.

CAN THESE VALUES BREAK FOR THE COMPONENT?

WHY DON'T TVS FLAT THESE VOLTAGES?

THIS PROJECT IS VERY IMPORTANT AND THE PRODUCTION HAS HIGH NUMBERS (250,000 YEARS)

THANK YOU

A.PENNA

[Domenico RAGONESE]

@apenn.13​ 

The noise seems very fast (<1us) and it could be too fast for a TVS.

Could you send me a picture of the board (including the probe of the oscilloscope)?

It's not clear to me if you have implemented the following hand patches:

1. maintain the short in parallel to reverse polarity protection diode
2. connect by a short wire C207(+) directly to Vcc pin (#1 or #12) of IPS160H.
3. connect by a short wire C207(-) directly to GND pin (#7) of IPS160H.
4. do the same for C206(+), C206(-) and C208(+) and C208(-)

Alternatively to (2), (3) and (4) you could try by soldering a ceramic capacitor directly on pins 12 and 7 of IPS160H.

Thank you in advance.

Best regards,

Domenico.

[apenn.13]

HELLO,

WHICH CAPACITY VALUE TO INSERT BETWEEN PIN 12 AND 7?

I HAD TO MAKE A TEST BOARD FOR IC IPS160H WITH WIRES THAT ATTACH TO THE MASTER BOARD AS THE BREAKAGES HAVE AFFECTED THE TRACKS OF THE COMPONENT.

WE HAVE AN INPUT PWM OF A FEW HERTZ, CAN IT GOT ANYTHING?