I have tried using amps with inbuilt protection against excess PA over current, and tried adding it to my own. But they all seem to have a glitch in so far as the amp is powered on all the time and the signal from the driver just commences at the start of a TX. Two issues ensue, one is a high current surge on start up. set the over current level to admit this and it always seems too high to protect the FET's if anything happens to the output side like a shorted aerial.The damned FET's just seem to have such a low pain threshold that I gave up on electrickery other than a basic quick blow fuse. The other issues is the very high RF level seems to play havoc with complex control strategies.
Did you ever stop to think, hmm, maybe they're interrupting startup for a reason?
There are many levels of protection and control. If you add a latching AND gate immediately after the toggle f/f, you can have the toggle outputs turn the transistors on and off, except that, when the per-transistor protection fires, it resets that transistor. The transistor then stays latched off until the toggle turns it on again. A momentary upset (like transformer saturation, the most likely startup problem) doesn't crash the system. Now, this wouldn't catch a continuous fault, but then you could add a pulse detector, that watches for how frequent the fault triggers are; when they become too frequent, the whole thing is shut down.
Or you could drop the discrete logic altogether and harness a chip for control and protection. Even the basic TL494 has a valuable feature: it contains a free-running oscillator, that can be locked to your 272kHz reference just the same, and its output duty cycle is adjustable. You can use the dead time function to implement soft starting. You can route current protection through this as well (but it's not latching, so that still needs additional hardware). Others may be out there with even better options (and integrated gate drive, and..).
If you're getting noise and nuisance trips, that's a more basic problem. That's fundamentals: that's grounding, short switching loops, common mode filtering, that sort of thing. You need to fix that anyway, otherwise you'll be spewing who knows what in the HF bands!
When I designed an industrial inverter some years back (5kW per module, up to 400kHz), I put in desat protection (a comparator watches the drain voltage, and fires a fault signal if it exceeds what Rds(on) would give). At first, this gave us incessant nuisance trips: at the switching edge, a flurry of false pulses would be transmitted. (Or not even false, simply because sometimes the transistors must turn on with high Vds!) We added blanking around the switching edges, so that events within 1us of an edge were ignored. If a fault remained after that, the system was shut off.
This simple method saved
dozens of transistors during development. I don't think we accumulated more than a handful of transistors during the 2 1/2 years I was there. And that includes the big industrial IGBT modules we moved on to, later in development!
This system is so robust, you can short out the transistors, during startup, while it's running, and it simply ticks and turns off. No damage.
In contrast, our predecessors, on that project, had left a small box brimming with accidents and failed experiments. They didn't have a clue how to design these things, or even how to work with them safely. When you don't give yourself the tools to succeed, you can't get anywhere.
Tim
Home
Help
Search
Login
Register
