CC1310: DC to DC converter damaged during contact ESD testing

Hello Simon,
Does the diode you suggest have any effect on the CC1310's harmonics, as I know that the CC1310 outputs very strong odd-ordered harmonics and I don't want to make them any worse?
The TVS diode will protect the RF input, but can an ESD event on the RF input leave the radio fully operational and only destroy the DC to DC converter?
Regards,
Tony.

The normal ESD methodology is to lead the ESD charge from the node receiving it and to ground before hitting the chip. I'm not sure how to think when the ESD charge hit a ground node. What I suspect in this case is that a large voltage/ current has been induced in the DCDC inductor destroying the DCDC_SW pin. You can try to remove the inductor and use the internal LDO and see if everything works afterwards.

How is the SMA connector placed compared to the DCDC part?

A ESD diode at the SMA connector should not cause any issues due to the very low capacitive load.

Hi TER,
I have added unidirectional TVS diodes between ground and pins VDDR/VDDR_RF, VDDS, DCOUPL and DCDC_SW on the CC1310.
The SMA connector is bulkhead-mounted and is on a flying lead that is connected to the PCB by a ufl connector.
I have changed the power supply that feeds VDDS, to an isolated (5.2 kV rated) DC to DC converter. My goal was to increase the impedance to ground to reduce the current pulse caused by the ESD and the isolated DC to DC converter achieves this. I also added an earth bonding strap from the SMA connector to ground. As I have raised the impedance to ground through the circuit's 0V rail, I hoped that the earth bonding strap would act as a low impedance path to ground and take the bulk of the current associated with the contact discharge.

Unfortunately this has not helped and the CC1310 was damaged when tested. The DCDC_SW pin survived and the CC1310 is still working, but the current drawn has gone up to 0.6 Amp and the CC1310 gets extremely hot.

I am awaiting delivery of the TPD1E0B04DPY diodes recommended by your colleague and I will fit one to the antenna feed as soon as they arrive.

I notice that an air discharge to the SMA ground is sufficient to lock up the radio, without damaging the CC1310. I don't know if this tells us anything useful.

Any assistance that you are able to offer would be gratefully received.

Regards,

Tony

Hi TER,

The TPD1E0B04DPY diodes did not help.  With the diode fitted, the CC1310 was damaged by the first 4kV discharge that was applied.  Normally it takes more than one discharge to damage the device.

I'm currently working with an EMC consultant on trying to solve the problem. I'll let you know the eventual solution. But if you have any suggestions in the meantime, they would be most welcome.

Regards,

Tony.

We got some feedback from the engineer that works some with the on chip ESD:

• If he runs contact discharge of a grounded equipment, the ground line may be not good enough? Or the ground is routed first to the CC1310 and then to the system ground? Which pin(s) is/are grounded? A PCB review might help here. Also more information on the total ground strategy could be helpful.
• If he runs contact discharge of an ungrounded equipment, how does he power his product? I have read that even isolated wall-mounted power supplies can impact ESD testing negatively.
• Also, that air discharge “locks the radio” might be a hint towards bad grounding or PCB loops. In this specific case, what does “lock” mean? Does it work again after reset?
• Unidirectional ESD diodes do not help, he claims. Maybe he can test +4kV and -4kV separately and tell us which polarity breaks his product?
• I did not really understand what he meant with “isolated DCDC”? Is he using the same ground on both sides? Anyway, the DCDC supplying VDDS can be higher rated, but if the injection point is “behind” the DCDC, this will not help with respect to ESD.

Hi TER,

Thanks for the feedback.  Taking the points one at a time:

The equipment is powered by a plug-in USB supply, so the grounding is reliant on the power supply's grounding.  I have seen different behaviour with different power supplies. When using one particular USB supply, the ESD gun indicated that it had not fully discharged after a 1 Hz sequence of 4 or 5 pulses, however another supply was identified that was able to discharge all 20 4kV pulses.  This is the supply I will use for further compliance testing.

  
The antenna connector's ground is directly connected to the 0V layer in the PCB. The USB's 0V pin is also connected to the 0V layer in the PCB, via a ferrite.  The layout has the antenna at the opposite end of the board to the USB , so the current pulse has to travel through the 0V layer, past the CC1310, on its way to ground.  In order to improve the grounding I have added a metal braid to the SMA connector and run this directly to the USB shield and USB 0V. As there is a ferrite between the USB 0V and the 0V PCB plane, the braid should act as a lower impedance path to ground.  My plan is to modify this scheme and replace the ferrites (there is also one on the USB 5V) with a common mode choke. I hope the common mode choke will increase the impedance of the supply lines to ESD, thus favouring the grounding offered by the SMA-braid-USB 0V path.

I note your point that isolated wall-mounted power supplies can impact ESD testing negatively.  I will test with a variety of supplies and select one that favours our equipment.

The radio spends nearly all its time in receive mode, however the reception has a timeout set.  When the radio locks up, the reception never returns, even though valid packets should be received or a timeout should occur.  The radio is totally unresponsive even a RF_Close/ RF_Open call cannot recover the radio back to an operating state. Even a watchdog reset with "Warm Reset Converted to System Reset"  cannot recover the radio to a working state.  The only way I have found to recover the radio is to disable "Warm Reset Converted to System Reset" and wait until two warm resets have been performed. After the second warm reset the radio is brought back into an operating state.

After adding the TVS diode to the antenna the CC1310 was destroyed by the first ESD pulse, which was positive.  Without the antenna ESD diode, the CC1310 inormally survives the 10 positive pulses (although the radio locks up on the first pulse) and is destroyed by the first negative pulse.  The equipment is fully discharged between the application of the positive and negative pulses in accordance with the test procedure.

'Isolated DC to DC converter': I used a Murata MEJ2S0503SC DC to DC converter.  This galvanicaly isolated the equipment's ground from the USB ground. The braid connecting the SMA connector to the USB shield/0V becomes the only path to ground for the ESD.  However I now understand that while this DC to DC converter provides 5.2 kV of DC blocking, it is the AC performance that is important for the ESD and there was still an AC path to ground through the converter.  Please don't worry about this as I have abandoned this approach to isolating the equipment and USB ground.

I will let you know if the common mode choke approach proves successful, but would still appreciate any insights you are able to offer.

Regards,

Tony.

   

Hello Tony,

Have you tried removing the bead between the 0V on the USB connector and the board ground layer?

How have you connected the shield of the USB connector to your board ground?

And how is the connection between your board gnd and ESD gun gnd?

-Simon

Hi Tony,

Thanks for sharing your solution!

Creating a low impedance path from the SMA to the USB shield seem to solve the problem.

Best regards,
Simon