HI,
we have an application where we have one main OPA549 in an inverting gain configuration (ie. ~30x), and one or more OPA549's in 1x gain configuration as followers. The follower OPA549 outputs are connected in parallel with main OPA549 per the datasheet recommendation with 0.1 Ohm series ballasting resistors for even current sharing among amps. This all works fine, except we notice for example if one OPA549's goes into thermal shutdown, only it shuts down individually, it has no direct way to communicate it's thermal shutdown with the other amplifiers to shut them off as well. This happens sometimes when our overall amplifier network is initially powering some particularly nasty reactive loads, where we want the system to cough and ride-through the start up transient. In the configuration where we have one main amp and one follower amp, we can have the situation where the main amp coughs (thermal shutdown) yet the follower stays enabled. Some loads are so nasty that the load itself is the secondary of another transformer accidentally connected that is powered from the AC line! We want our amplifier to survive all kinds of abuse. In the lab we've never been able to blow up an amplifier with any torture test, yet in the field, when we occasionally have an OPA549 blow up, it is always the follower blown out, not the main. The heat sinking/air flow is identical for both follower and main and we want to keep it that way.
Currently our circuit has global OPA549 enable/disable capability via computer control, as well as computer monitoring which gets flagged if any OPA549 is in thermal shutdown. So we currently have the capability to shut down all OPA549s together if anybody is in thermal shutdown, via software. But, we don't want to rely on the computer control since it is slow, on a different board, prone to programming errors/micro freezes, etc. We want shutdown circuitry which achieves the following:
1) shuts down all OPA549s simultaneously (main+all followers within say 2usec) if anybody is in thermal shutdown.
2) allows all the OPA549s to resume operation together once the thermal shutdown is cleared (typically we see thermal overloads clear within 2msec in cough situations).
3) still allows global shutdown/global enable via a software loop with the understanding that if any amp is in thermal overload, the thermal overload over-rides the software controlled enable.
For 1,2 we want that circuit to be minimal amt. of marginal circuitry per OPA549 follower added to the system, and ideally with no programmable parts such as micro-controllers etc.
It turns out we have a circuit which should achieve the objectives 1,2,3, but it's a bit involved, using one comparator per OPA549 along with some diodes.
Now for the question: Is there any other known good way to achieve 1,2,3 in a TI application note or other document? I was hoping the E/S pins of multiple OPA549's could just be paralleled to achieve this, but I noted from a separate thread (and empirically) that that approach can't work (side note: too bad the E/S pin pull down capability isn't about 5x stronger than the pull-up capability).
Best,
Richard
