TPS2493: EMI/RFI susceptibility TPS2493

Hi,

Let me check with the team whether 400Mhz to 700Mhz frequency range falls in the charge pump frequency range of LM5050 or TPS2493. Please allow time till mid of next week.

Best Regards, Rakesh

Hi Rakesh,

Thank you I look forward to hear about the sensitivity of the charge pump next week.

In the meantime:

Please let me know if there is placed an internal EMI filter in the IC for the current sense pins?

Are there conducted some sort of EMI qualification (e.g. IEC 62132 and IEC 61967) of the part?

Please propose an EMI filter (or guidelines for an EMI filter) to be placed on the entrance pins of the IC.

FYI: By Short-circuiting Q2 the output voltage remain 24VDC (LM5050 cannot disconnect the output voltage due to the inherent body diode in the FET.)

Br Jørgen

Hi Jorgen,

Let me get back on the charge pump frequency information. That will help to answer the follow-up questions.

Best Regards, Rakesh

Hi Jorgen,

The internal charge pump frequency of our devices is in the order of 2MHz range. So, these devices should NOT be the reason for the frequency range 400Mhz to 700Mhz you are seeing in your board. That's also the reason for not carrying any EMI qualification tests on our device.

Best Regards, Rakesh

Hi Rakesh,

We are not "seeing" 400Mhz700Mhz on the board.

We apply an external radiated field of 10V/m, 80% AM modulated with 400Hz and and in the frequency range 400MHz to 700MHz the TPS2493 by malfunction disconnect the serial FET Q2.

• Please let me know if there is placed an internal EMI filter in the IC for the current sense pins?
• Are there conducted some sort of EMI qualification (e.g. IEC 62132 and IEC 61967) of the part?
• Please propose an EMI filter (or guidelines for an EMI filter) to be placed on the entrance pins of the IC.

Br, Jørgen

Hi Jorgen,

I got your question.. Is it a soft turn-off and fast turn-off?

Can you test again with below changes.

1. Add 47nF across R7 
2. Add 100nF between GATE to ground of TPS2493

Please let me know if there is placed an internal EMI filter in the IC for the current sense pins?

->here is no internal EMI filter inside the device.

Are there conducted some sort of EMI qualification (e.g. IEC 62132 and IEC 61967) of the part?

->No EMI qualification done on these parts.

Please propose an EMI filter (or guidelines for an EMI filter) to be placed on the entrance pins of the IC.

->Please check with the above changes. In parallel, I will discuss with the internal team whether there are any inherent limitations with the device.

Best Regards, Rakesh

Hi Rakesh,

Thank you for the fast reply.

 We do not know if it is fast or slow turn off, as it is impossible to have measurement equipment inside the EMI test room with 10V/m field strength.

 We know that with 10V/m without the 80% AM modulation 400Hz we have no mal function of the TPS2493 = it is the modulation that trigger the fault function in the TPS2493

We will try with your suggestion and please if any of your colleagues have other suggestion in order to make the TPS2493 more resilient to the AM modulation we would like to know this as this could save time for us.

Br, Jørgen

Yes Jorgen. I will check.

Keep us update on your findings with those changes.. those may give us some clue.

Best Regards, Rakesh

Hi Jorgen,

Does the below changes helped ?

1. Add 47nF across R7 
2. Add 100nF between GATE to ground of TPS2493

Hi Rakesh,

Sorry we do not have acces to the test room yet. we expect this to happen late this week.

In the mean time please if you or anny of your colleags have futher / other suggestions that we can try once we get the test facillity please let us know.

Once we have the test facility it is not a problem to test different "solutions" but it is difficult to get acces to the test room.

BR Jørgen

Hi Jorgen,

Unfortunately, we have not come across such issue (due to radiation) in the past. We would like to probe some critical nodes to debug but this particular environmental test does allow to do that. So, like to see whether our suggestions helped or not.

How much cap is used on +24VF rail ? In addition to the above tests, veify by increasing the cap on +24VF rail 

We have calculated the impact of your recommendation: 2. Add 100nF between GATE to ground of TPS2493

Our calculation of the turn off time in the situation with a hard short (0.1R) across the 24V output will be delayed about 1200us.

1200us delay is not acceptable as the current in this period will increase to 240A and the SOA of the FET is exceeded.

Please comment urgently.

BR Jørgen

Hi,

The soft start circuit in the front page of LM5069 datasheet is the solution to discharge the external GATE capacitance so that impact on short circuit response can be avoided.

BR, Rakesh

Hi,

Thank you.

Yes this will give a fast discharge.

But will there then bee a anny EMC benefit of the 100nF when it not are direct connected between Gate and GDD???

BR Jørgen

Hi Jorgen,

The device has GATE sinking current of 125mA, so the 100nF cap would discharge in ~20us. Can you let us know how you arrived at 1200us ?

Best Regards, Rakesh

HI Rakesh,

You are right, Toff ~20us. By mistake we calculated with the 2.4mA valid for Vuven, sorry.

In TPS2493 Datasheet page 24 a serial resistor of 1K is recommended in serial with the 100nF, we assume this should not be placed??

We expect to start test within 1 week.

BR Jørgen

Hi Jorgen,

Use 100 Ohm series resistance to limit the discharge current from 100nF into the IC. We suggested 1kOhm in the datasheet on very conservative approach.

Best Regards, Rakesh

Ok Rakesh,

BR Jørgen

OK. 

Hi Rakesh,
The 47nFp14-p13 does not change the current limit behavior of the circuit.

However with the 100nF+100R we far exceed the SOA of the FET when a SC is placed at the output. (50A/1ms@24V)
With 10nF+100R we measure 50A/100us@24V, that is within the SOA.

Please comment and do not hesitate to propose further improvements.
Tanks
Jørgen

Hi Jorgen,

The other option is to increase the cap on +24VF rail to support the load even if hot-swap resets. 

Best Regards, Rakesh

Hi Jorgen,

Have you increased the cap on +24VF rail and tested again.? Did that helped ?

Best Regards, Rakesh

Hi Rakesh,

Sorry with 330uF E-Lyt + 10uF ceramic does not help.

We will test 8th dec. and then informe you about the outcome.

Br Jørgen

Ok Jorgen