Using bridge rectifier diodes for ESD protection

[matrixofdynamism]

Please have a look at the attached schematic. It is from Robert Feranec's iMX6 Development Baseboard.

The schematic excerpt shows USB connector J22 (5787617-1) with protection components D10, D11 (BGX50AE6327) and L5, L6 (ACM2012-900-2P-T002).

The primary question is that the components D10, D11 are bridge rectifier diodes. I would expect TVS diodes to be used here.

What is the meaning of using rectifier diode IC for ESD protection?

[T3sl4co1l]

Not much of a "rectifier". Did you check the datasheet?
https://www.infineon.com/dgdl/bgx50aseries.pdf?folderId=db3a30431400ef6801141cc1c4cd04e8&fileId=db3a30431400ef6801141cc423c404e9
They don't call it one.  Suppliers may by categorization (it's a FWB circuit only; that doesn't mean you want to use it to rectify real power), but datasheet is what counts.

Of importance, notice the low capacitance.

Tim

[schmitt trigger]

Nowadays, component miniaturization counts. Not only with smaller packages, but with higher integration.

A good candidate has been the series connected diodes to protect data lines. You could use a pair of BAT54 diodes, but there is a version, the BAT54S which incorporates both diodes in a single package of the same footprint as the BAT54.

This Infineon device takes the idea a little further: since many of the most popular serial buses are two-line, why not incorporate a pair of BAT54S in a single package?
BEFORE ANYONE ROASTS ME; I know that the Infineon device is not a schottky. I only compared it with the BAT54S as an illustration, because it is ubiquitous in its usage as a digital line protection.

[matrixofdynamism]

ok, lets see;

The datasheet for BGX50AE6327 says:

"Silicon Switching Diode Array • Bridge configuration"

Digikey calls this "Bridge Rectifier Single Phase Standard 70 V Surface Mount PG-SOT-143-3D" (Which I checked)
Mouser calls this "Diodes - General Purpose, Power, Switching AF DIODE 70V 0.14A "

Now lets look at a different place in this design we have D4 (BZA956A,115).
The datasheet calls it "Quadruple ESD transient voltage suppressor". These words are quite clear that this is for ESD protection. The datasheet also says "ESD rating >8 kV, according to IEC61000-4-2" and gives "non repetitive peak reverse power dissipation" of 16W. The mentioning of ESD in the name and the large reverse power dissipation figure shows that this component is a true ESD protection device.

On the other hand, the BGX50AE6327 datasheet that this question is about is missing both of these things. From its schematic in the datasheet it looks like something that would be used as bridge rectifier (i.e AC-DC conversion) so I am puzzled why it is able to work for ESD protection.

[T3sl4co1l]

Nowadays, component miniaturization counts. Not only with smaller packages, but with higher integration.

A good candidate has been the series connected diodes to protect data lines. You could use a pair of BAT54 diodes, but there is a version, the BAT54S which incorporates both diodes in a single package of the same footprint as the BAT54.

This Infineon device takes the idea a little further: since many of the most popular serial buses are two-line, why not incorporate a pair of BAT54S in a single package?
BEFORE ANYONE ROASTS ME; I know that the Infineon device is not a schottky. I only compared it with the BAT54S as an illustration, because it is ubiquitous in its usage as a digital line protection.

For that matter, it's an improvement [over BAT54S], not only being double (but there's BAT54XY if you need that!) but having much less capacitance too.  So the signal bandwidth is wide.

Schottky are kind of bad at signals, maybe somewhat ironically; the contrast is, if you have RF/microwave schottky, they're very dinky things indeed and have quite high internal resistance, low power ratings, and are ESD sensitive.  Indeed, for BAT54S, it's probably more the PN guard ring diode that does the protection, again, because of high internal resistance in the schottky.  This suggests it should have low ESD ratings, or at least so in comparison to its junction area; equally concerning is no datasheet gives an ESD rating anyway, but I can say from experience they're fine, have never seen one fail at 8/15kV levels.

Whereas your average switching diode, is fine switching signals up to some 10s of MHz or so, or surge clamping of signals much more than that.

They didn't provide a layout and frequency response, as it's still specified as switching diodes, not a TVS array or ESD clamp; but we can expect good bandwidth out to some 100s of MHz, given the low capacitance, and the compact SMD package.  TVS arrays are indeed available with comparable capacitance and Vf, and we can reasonably assume similar specs (junction area, doping level) in this device.

Tim

[fourfathom]

One issue with the bridge diode protection scheme is that any transient current is dumped into the Vdd bus.  This isn't really a problem, at least not for the transients you might otherwise use a TVS for since the supply bypass capacitors will soak up the current.  But for prolonged overvoltage clamping if the fault current exceeds the operating current of the circuitry you will get an elevated Vdd.  But that really calls for a different type of protection.

[bdunham7]

I would expect TVS diodes to be used here.

There's where you've gone astray--it is common to clamp various things by using a pair of diodes, one to each power rail, instead of clamping across them with a TVS or whatever.  This is just a package of 4 diodes arranged in such a way to provide a 1-chip solution to clamping 2 lines.

[floobydust]

BGX50A has no transient/surge current rating: "Non-repetitive peak surge forward current= BLANK". Otherwise 140mA I_F 
Is the design even tested for ESD? Why is there no protection on VBUS? You can't look at some schematic and assume it's all good.
It's common to plop down protection components but never actually test their effectiveness. They end up as a pacifier, theoretical.

NXP using SRV05-4 which has ESD ratngs/tests, in the i.MX6 Sabre automotive. As well ESD9B5.0ST5G on VBUS.

Also seen with i.MX6 Littelfuse PGB1010603MR interesting "... use polymer composite materials to suppress fast-rising ESD transients (as specified in IEC 61000-4-2), while adding virtually no capacitance to the circuit." Never seen that before, 0.06pF

IP4221CZ6-S

A few part suggestions: NXP Application guide: ESD protection

[T3sl4co1l]

Also seen with i.MX6 Littelfuse PGB1010603MR interesting "... use polymer composite materials to suppress fast-rising ESD transients (as specified in IEC 61000-4-2), while adding virtually no capacitance to the circuit." Never seen that before, 0.06pF

They're interesting, but also kinda crap..?  I would like if they characterized them a bit more.  My best guess is they're a latching device akin to ye olde telephone TVS, the thyristor (SIDAC) type, but acting fast enough to be useful for ESD; or put another way, a solid-state GDT.  Anyway, like these devices, the peak let-thru voltage is rather high (~200V?), which makes them not all that useful in some cases, but the latching means the energy is reduced considerably.  If you can handle the initial peak, then, it's fine.

Tim