I've been searching for examples of using Mosfets instead of BJTs for VI Limiters, but could not find any.
In some amp designs, the VI Limiters are having troubles robbing enough base drive current from the drivers, so they're unable to fully implement the calculated locus to keep things under control into the SOAR.
So despite having properly calculated values for a given locus, the limiter isn't capable of properly enforcing it.
I've tried using CFP combos instead of a single BJT, and that does bring some improvement, although not always quite enough to fully enforce a locus.
Using Schottky diodes in the BJT's collectors instead of regular ones also improves things slightly, but far from enough.
So I was thinking if possibly using MosFets instead of BJTs could make for a more forceful action and perhaps it would be able to better enforce the locus.
Has anyone thought of this and tried it?
In some amp designs, the VI Limiters are having troubles robbing enough base drive current from the drivers, so they're unable to fully implement the calculated locus to keep things under control into the SOAR.
So despite having properly calculated values for a given locus, the limiter isn't capable of properly enforcing it.
I've tried using CFP combos instead of a single BJT, and that does bring some improvement, although not always quite enough to fully enforce a locus.
Using Schottky diodes in the BJT's collectors instead of regular ones also improves things slightly, but far from enough.
So I was thinking if possibly using MosFets instead of BJTs could make for a more forceful action and perhaps it would be able to better enforce the locus.
Has anyone thought of this and tried it?
You can use a depletion MOSFET as a current limiter -- DN2540 is popular as a current source and would suit the same purpose.
I haven't seen any example anywhere. And there would be some slight adjustments to be made to a BJT based limiter to make it work with a MosFet.
I assume it would work better with a MosFet because of the much lower saturation voltage and that low Rdson. So it might be a more forceful enforcement. Is this right?
I guess we could make the base resistor used (sometimes) in a BJT based limiter much higher in value for the MosFet, due to the very low gate current needed to drive it.
And that, in turn, could allow higher values pretty much everywhere in the limiter to be even less "disturbing" to the amp's circuitry, to keep it as transparent as possible.
With a MosFet though, I'm wondering if it might be a little too "digital" in nature, by tending to operate more in an on/off mode.
The BJTs have some Vcesat that isn't quite negligible, added to the diode in series in its collector, when fully saturated it's not "squeezing" enough to rob enough base drive current from the drivers. So although per calculations, they're acting, the effect isn't quite as hard as it needs to be to follow what the calculations determined.
Plus the BJTs start acting far earlier than they should, with a knee bend much too round, which blurs that sharp line locus that we use for calculations.
I'd like to see examples with MosFets. I think simulations are in order. But I'm not familiar enough with FETs in general, to properly determine how it's going in simulations...
I assume it would work better with a MosFet because of the much lower saturation voltage and that low Rdson. So it might be a more forceful enforcement. Is this right?
I guess we could make the base resistor used (sometimes) in a BJT based limiter much higher in value for the MosFet, due to the very low gate current needed to drive it.
And that, in turn, could allow higher values pretty much everywhere in the limiter to be even less "disturbing" to the amp's circuitry, to keep it as transparent as possible.
With a MosFet though, I'm wondering if it might be a little too "digital" in nature, by tending to operate more in an on/off mode.
The BJTs have some Vcesat that isn't quite negligible, added to the diode in series in its collector, when fully saturated it's not "squeezing" enough to rob enough base drive current from the drivers. So although per calculations, they're acting, the effect isn't quite as hard as it needs to be to follow what the calculations determined.
Plus the BJTs start acting far earlier than they should, with a knee bend much too round, which blurs that sharp line locus that we use for calculations.
I'd like to see examples with MosFets. I think simulations are in order. But I'm not familiar enough with FETs in general, to properly determine how it's going in simulations...
MosFets need WAY more trigger voltage , say 3 to 4 V instead of 650mV and that´s WASTED peak voltage pulled from output, so nobody in his sane mind uses them.
Trigger voltage is derived across a resistor in series with the load, so the lower the better.
And Bipolars have ample current capability to clamp drive voltage/current; if not it´s because the circuit is poorly implemented.
Trigger voltage is derived across a resistor in series with the load, so the lower the better.
And Bipolars have ample current capability to clamp drive voltage/current; if not it´s because the circuit is poorly implemented.
The sense resistors are usually the emitter resistors, which can be 0.22 ohms or even lower, so a low threshold voltage is needed, in the region of 1--2V, and it needs to be stable.
That rules out MOSFETs with their large spread of gate-threshold voltages (which often drift with age too).
Comparators or BJTs are indicated.
That rules out MOSFETs with their large spread of gate-threshold voltages (which often drift with age too).
Comparators or BJTs are indicated.
Right, I see, I had not considered that trigger voltage issue.
And I'm not fond of using emitter resistors over 0.22 ohms or so.
It's true the BJTs may have the clamping capacity, but they also have their Vcesat, which added to the collector diodes, adds up a bit, and that limits their clamping abilities.
That's why I've been looking at potential alternatives where the clamping isn't energetic enough.
Using CFP did help somewhat in those case where the single BJT wasn't clamping hard enough, but not quite enough to enforce the locus fully...
Using Schottky instead of regular diodes in the collectors also can help a little bit..
And I've see some using the action from the limiters to trigger a shut down or something like that, instead of curbing the eagerness of the amp. But I don't like the shutdown concept for that. I think it's very obtrusive and the show must go on even when there is an overload. The amps should protect themselves against the overload, but keep going without shutting down.
And I'm not fond of using emitter resistors over 0.22 ohms or so.
It's true the BJTs may have the clamping capacity, but they also have their Vcesat, which added to the collector diodes, adds up a bit, and that limits their clamping abilities.
That's why I've been looking at potential alternatives where the clamping isn't energetic enough.
Using CFP did help somewhat in those case where the single BJT wasn't clamping hard enough, but not quite enough to enforce the locus fully...
Using Schottky instead of regular diodes in the collectors also can help a little bit..
And I've see some using the action from the limiters to trigger a shut down or something like that, instead of curbing the eagerness of the amp. But I don't like the shutdown concept for that. I think it's very obtrusive and the show must go on even when there is an overload. The amps should protect themselves against the overload, but keep going without shutting down.