[Amps] Direct rectification of AC mains to drive the amp, VDD Supply

[Manfred Mornhinweg]

Bill,

> Hi Manfred:  If this hotspotting problem occurs, say when doing RTTY at a
> high power level, how quickly does it happen? 

I frankly don't know for sure, as I haven't seen it happening yet. But 
since this is a process that involves very small structures, with very 
low thermal mass, I would imagine it to be really fast. The whole 
process of hot spots forming, running away and blowing up the MOSFET 
would probably take just milliseconds. On the other hand, maybe the 
devices need to heat up, before the hotspotting becomes really 
dangerous, so you would have the normal time constant from die to 
heatsink involved. In that case it could be several seconds.

But maybe the hotspotting will cause accumulative damage, rather than 
catastrophic. In that case it may take days or weeks of use, before the 
device blows up. Or it might happen only under certain load conditions, 
that cause higher dissipation.

> For example, if one is
> watching the drain current, would one see a gradual increase or a sudden
> catastrophic one? 

I think you wouldn't see any drain current change. Instead you would 
simply notice that suddenly the amp stops working.

> Also, could this be tested with only DC? If one ramps up the bias with DC
> only to where the push pull pair is dissipating 1500 watts without
> hotspotting, could one be confident they would be safe doing RF?

I think yes. In fact, such a DC test may even be harsher than the actual 
RF use, because over an RF cycle the transistors spend some time off, 
some time very close to saturation (which are both safe conditions), and 
only a smaller time in their linear range with high drain voltage 
applied, which is the dangerous condition. Instead with DC they are all 
the time in that high voltage, linear area.

So, if someone has enough free samples of these MOSFETs to risk 
sacrificing one to science, it could be a good idea to make a simple DC 
testing setup: Good heat sink, a power supply of the same voltage 
intended for use (300V for an ARF1505), and a gate bias that 
self-adjusts to give a certain drain current (via feedback). Then try 
the device at various currents, each for a relatively long time, always 
within its safe dissipation range (which depends on the heatsinking 
used). Ideally with a thermal sensor mounted to the device, could be 
through a little hole in the spreader plate. If the device blows up 
while dissipating a power level that should be safe at the current case 
temperature, we can probably attribute it to hotspotting. If instead it 
survives testing up to the power level that can be expected to heat the 
silicon to its max temperature, we could conclude that hotspotting  due 
to high voltage isn't a problem.

Manfred

========================
Visit my hobby homepage!
http://ludens.cl
========================