In older versions of Braids schematics/boards, there used to be resettable fuses between the connector and rails. On the recent version these fuses are going. I wonder why is that? Seems like a downgrade to me.
Another difference is the way the diode protection is done. On the old version the diodes were connected backwards between each rail and ground, so in normal operation the current wouldn’t flow through diode, but in case of reversed polarity, the rail would short to ground and fuse would disconnect the rail shortly after. In the current version, there are diodes between connector and rail, so the current is going through them all the time during normal operation. Also causing a small voltage drop, which probably doesn’t matter here…and a bit of heat…
I wonder why would Olivier ?upgrade? the schematic this way. The original version was protecting not only against accidental polarity reversal, but also against anything that would cause lot of current going through the module damaging it. Seems better to me.
1/ Series polyfuse + shunt diode. When the current is applied in the wrong direction, it flows through the diode which acts like a short, sucking a lot of current from the PSU. Then the polyfuse “blows” and disconnect everything. Note that the polyfuse is absolutely necessary for this to work.
2/ Series diode, preferably Schottky (low voltage drop). When the current is applied in the wrong direction, the diode blocks it. This protection scheme doesn’t need a polyfuse.
So why did I stop using 1?
With 1/, the small resistance of the polyfuse causes a voltage drop that varies with the current consumption of the module. The result is that the +12V supply for the op-amps slightly follows the current consumption of the digital section. Not good.
What happens when a module is plugged backwards? With 2/, nothing happens, it’s like an open circuit. With 1/, for a short time, the PSU will try to generate as much current as it can. I thought some PSUs might not like this at all. And there’s the question of very beefy PSUs - some PSUs are now so beefy that they’d happily dump 2A of current - in other words, the shunt diode should tolerate very large currents.
If you’re aware of other protection schemes let me know!
Note that I continue using the polyfuse + shunt diode for circuits which uses the V2164 - since this IC absolutely needs to be protected against a missing V- - and this requires the shunt diode.
I think I just tested this theory. I built a couple L-1 PSUs and got the ± backwards on the 12v legs. The version of Braids I have is v0.3. It doesn’t boot up. Is there a diagnostic I can do to see if this Braids module is recoverable?