Author Topic: power rail protection beyond polyfuses  (Read 2013 times)

0 Members and 1 Guest are viewing this topic.

Offline helgeTopic starter

  • Regular Contributor
  • *
  • Posts: 73
power rail protection beyond polyfuses
« on: July 22, 2018, 10:39:42 am »
I've looking at input protection for battery powered circuits lately. I'd like to summarize the specs and what I have so far, concluding with two circuits I'd like to add to the discussion. To be honest right now I'm leaning towards Schottky diode / amplified Zener clamp + fuse (polyfuse or resistor), mostly due to cost and space constraints but we'll see where this leads below.

The specs are
  • +3.3 V / 1A to be supplied from an external buck / buck-boost converter
  • +2.85.. 4.2 / 12 V / 2A for battery or power supply operation respectively
I'm not exactly looking for autotomotive-style surge protection up to 60 V. I'm trying to cover the following 3.3V, 12V and GND permutation scenarios:
  • -3.3 V, +/-12 V on the 3.3 V rail
  • -12 V on the +12 V rail
  • battery voltage on I2C lines while 3.3 V sees reverse polarity (not much to be done about that I guess)

so in essence this is about:
  • 2.85 .. 12 V rail reverse polarity
  • 3.3 V reverse polarity and overvoltage
  • maintaining low quiescent current of the protection circuitry

What it's not about:
I doubt that at this point overcurrent protection will make a lot of sense on the PCBs since I cannot control cable diameters and lengths and inrush currents will occur often. Both supply rails are switched on to power external vaguely specified secondary boards, there's just an ultimate limit of 2A. I'll also skip analog / digital signal protection here, to revisit the subject this might be helpful:
https://www.digikey.com.au/en/articles/techzone/2012/apr/protecting-inputs-in-digital-electronics


Now there is a wide array of possible approaches to supply rail protection, from the humble Zener diode to full-featured protection ICs. Some have been investigated in detail on the forums and I'm quickly attempting to summarize them for reference:


TL431, LTC4365 OVP:
https://www.eevblog.com/forum/beginners/overvoltage-protection-using-tl431-and-mosfets/msg1361583/
general discussion, LT4356, "Zener with a crowbar thyristor with a polyfuse", "beefy Zener", automotive load dumping:
https://www.eevblog.com/forum/projects/over-voltage-protection-circuit/msg1129956/
MAX16969, automotive USB protection:
https://electronics.stackexchange.com/questions/289649/circuit-protection-incorporating-a-transient-voltage-supressor-and-polyfuse
Zener soft breakover, TLV431:
https://www.eevblog.com/forum/beginners/zenner-diode-selection-for-protection-not-regulation/msg1682126/#msg1682126

Technical documents:
LTC4365 UVP, OVP, reverse polarity protection IC:
http://www.analog.com/media/en/technical-documentation/data-sheets/4365fa.pdf
https://www.eevblog.com/forum/beginners/seeking-ltc4365-advice/msg414723/#msg414723

automotive OVP and load dump protection:
http://www.ti.com/lit/an/snva190b/snva190b.pdf
Reverse Polarity and Overvoltage Protection Reference Design:
http://www.ti.com/tool/PMP10737
discrete protection circuits and Ti eFuse:
http://www.ti.com/lit/an/slva862a/slva862a.pdf
automotive eFuse for higher voltages:
http://www.ti.com/lit/ds/symlink/tps2660.pdf
eFuse for PC hardware, more down my alley:
http://www.ti.com/lit/ds/symlink/tps25940.pdf

fusible resistors:
https://www.eevblog.com/forum/beginners/what-is-a-fuse-resistor/


Now let's see some circuits!


dual PMOS + Zener and resistors:
possibly not the greatest thing since sliced bread due to Vgs,th variations but ok to protect against 24V or worse on a 5V rail. R1 will blow up.


single PMOS + amplified Zener clamp:
refined version with a bipolar transistor (SOT89 or SOT223 style), more predictable in terms of parts variations and ageing but still has the resistor fireworks. On the upside a dual PMOS package can serve two power rails and there are dual NPNs available (PBSS4260PAN). Might also be used with polyfuses.


Non-destructive overvoltage protection and reverse polarity protection:
More along the lines of eFuse and LTC4365, this circuit uses a back-to-back PMOS arrangement. I'm thinking https://www.infineon.com/dgdl/irf9362pbf.pdf?fileId=5546d462533600a40153561158c41da7 due to the low Vgs,th to work with 3.3V and the lower price when compared to e.g. SI4925DDY. The only naughty thing here is that current limited reverse breakdown of the base-emitter junction occurs under reverse polarity. Hopefully transient 1 .. 20 µA will not degrade hFE.


Interestingly the BOM cost (~40-50ct) for the last solution is in the order of TPS25940(75-90ct). If not for the horrible specified Iq it would be attractive to just use this IC instead. LTC4365 also kinda sucks in that regard (it literally sucks 125 µA per rail).So if <10 µA at 3.3V and disconnect below 4V can be achieved, the discrete version will do.

Thanks for sticking around! Looking forward to replies.


ps. I have my issues with polyfuses because I've had them become unreliable and randomly cut power to a board below their rated current. That's something I really like to avoid. There's also some voltage drop associated with polyfuses that exceeds Rds,on.
PMEG4020ER + fuse resistor are starting to look pretty good right now :)
« Last Edit: July 22, 2018, 02:38:17 pm by helge »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf