Snapback TVS diode behaviour

[ricko_uk]

Hi,

To protect a MIPI-DSI interface, I’m considering using this TVS diode: PESD4USB3UTBS-Q ( datasheet: https://assets.nexperia.com/documents/data-sheet/PESD4USB3UTBS-Q.pdf ).

It has a working voltage of 3.3V, but since it’s a snapback type, my understanding is as follows: once it reaches the breakdown voltage of 9V (typical), it snaps back to 3.5V (as indicated in Table 5, page 4). However, looking at Figure 5 (page 5), it seems that the voltage first drops to approximately 1.2V, and then, at the specified 5A current, it clamps to 3.3V.

What I am trying to figure out is:

1) Does the diode recover immediately after the ESD pulse ends, allowing the signal to resume normal operation? Or does it remain clamped, as I’ve experienced with a similar diode I used in a previous project (which I posted about here before)? What are the conditions to ensure it does not remain clamped?

2) How can I determine the specific conditions required for the diode to recover and allow the signal to return to its normal working voltage, without staying triggered? I am asking because for example, being used on a fast MIPI-DSI signal it might switch to high-level (1.2V to 1.8V) too fast hence not allowing enough time for the TVS to recover out of clamp mode).

3) Additionally, although the datasheet states that it is a unidirectional TVS diode, Table 5 (page 5) suggests it exhibits the same behavior for negative voltages, with identical voltage values. Could someone clarify this please?

Thank you as always

[PCB.Wiz]

You would need to measure it.
The data on these is usually very poor, and incomplete.
The AOZ8S205BLS data is slightly better with I(Hold) at least drawn on a graph, but alas, not specified anywhere. 

[Psi]

ah yes, the magical snapback diode.

I wish you luck figuring out the specs.

[selcuk]

I've not seen this snap-back diode before. Interestingly, it seems like a transistor circuit more than a diode. Is there a reason to use it instead of a regular TVS. Like this one having the same capacitance and leakage but not a defined dynamic resistance:

ULC3311CDN

[Marco]

Probably just a thyristor with a Zener on the trigger input. Aren't all the high speed interfaces mentioned in the datasheet AC coupled?

[ricko_uk]

Thank you all,

the high speed signals are mentioned in the datasheet but when I used it in the past with a 400KHz signal once triggered it remained triggered (I posted here back then).

I better stick tot he classic TVS even if it has a higher Vclamp.

Thank you

[PCB.Wiz]

Thank you all,

the high speed signals are mentioned in the datasheet but when I used it in the past with a 400KHz signal once triggered it remained triggered (I posted here back then).

Yes, such a thyristor needs to fall below a holding voltage, or current, to release.
Spice gives you a feel for what to measure, taking the Nexperia equivalent circuit and whacking in some low values gives the attached.

Trigger/release here  is ~ 150mA  for 10R VBE, as both Rbe need to conduct to trigger.
If your circuit is stiff enough to hold-up above the clamp voltage/current point, it will not release.

 

[ricko_uk]

Thank you @PCB.Wiz, I appreciate you doing that simulation.

[Benta]

It's basically a Silicon Bidirectional Switch (SBS), except it's unidirectional.
https://electronicsarea.com/sbs-silicon-bidirectional-switch/
When the input voltage reaches a certain point, it triggers and short-circuits the signal. It only resets when current flow is interrupted.
Also a lot like a DIAC.
Great for trigger circuits (SCRs etc.), but I fail to see the advantage in a USB interface.