Rosemary Ainslie circuit demonstration on Saturday March 12th 2011

[Rosemary Ainslie]

Which all brings me back to this answer to nul-points. 

The positive peak excursion across the CSR resistor, is a result of the sharply cut off Drain voltage of Q2 when it turns ON. Prior to this when Q2 is OFF, the CSR's inductance is energized by the up-swinging Drain voltage (via the 'Q1' and Q2 capacitance), and so this sudden cut-off causes the CSR's inductance to reverse it's voltage across itself (Lenz's law),...

- when I would have thought that the collapsing fields across the CSR result from Faraday's Inductive Laws.

And then

and this positive peak voltage is then "limited" by the forward-biased diode,...

when far from limiting this it would enable this.

And then

which is why the positive peaks appear "squashed" and widened. This latter part happens when Q2 is ON, so the diode is forward-biased into a 0-volt potential, which is why it can limit the positive peak, even though it may only be a couple volts in amplitude.

What 'forward-biased' diode is this?  The zener body diode?  Or the diode that you've now put across the cap to replace the Q1 MOSFET?  Because if you left the Q1 MOSFET as was earlier shown - then you actually don't get any 'squashing' or anything other than that clean sinewave shape. 

So.  Again.  I'm not sure that it's actually desirable to get rid of Q1 MOSFET.  I can't see how you'll be able to vary the input voltage from the source to up the power dissipated at the load and on the circuit generally.  So I'm not sure that it's a good idea to get rid of it.  And I'm not sure that it's correct to claim that any of these effects are due to anything short of Inductive Laws.  Unless of course Lenz Law allows for a complete breach of thermodynamic constraints.  And while your simulators can most certainly mimic our results - I'm also not sure that the simulations will be able to compensate for the input power at higher levels.  And I'm absolutely not sure what you actually think is going on here.  Perhaps you can advise us.

Rosemary

[Rosemary Ainslie]

Here's the kind of example that I'm trying to point to.  At this level we're dissipating in excess of 80 watts.  Just check out the math trace.   And I absolutely do not know how we could get here without the Q1.  I seriously propose that this is required.  But I'm open to correction.  Perhaps there's a way of doing it without that MOSFET.

Kindest regards,
Rosemary

Added

Sorry.  This is not the one I meant to post.  Here is only upwards of 40 watts.  But I'm hoping it's pointing to the issue at hand.

[Rosemary Ainslie]

Here's a better example where we're into wattage dissipation at something close to 80 watts.

Added.  I've been going through the hundreds of samples that we've got.  But the real 'runaway' numbers just happen too quickly to get a down load of the data.  I could however get a screen shot and shall try and do so this week.  The sample I've shown here is at a high frequency.  We can get this at slower frequencies - but I need to be nimble.  It often ignores the setting and then just keeps going north.  And I'm not that keen on spoiling any more of those FETs.

[poynt99]

All,

With reference to Q1's function in the circuit, I hope this helps:

How a diode can effectively convert a wave form with a mean value of zero, to one with a mean value that is non-zero. This is one of the two functions Q1 is performing in the circuit via it's internal body diode.

(The other function is to provide additional D-G feedback capacitance in parallel with Q2's own internal Cg-d).

This is a simple example that illustrates one of the effects Q1 has on the circuit. This ties in with my response to np's question about the positive portion of the "CSR" voltage. This hopefully gives you some insight into what can cause the MEAN negative voltage across the "CSR" resistor.

Let me know if anything does not seem clear and easy to understand.

.99

[poynt99]

I encourage anyone following along and wanting to get the most out of the information being presented, they save and print out this diagram, which is the simplified single-MOSFET equivalent of the original 5-MOSFET version.

Please pay careful attention to the labeling and the connection/location of the "CSR" resistor (Lcsr1 and Rcsr1), and what used to be the function generator, and is now replaced with the fixed DC voltage source (Vbat2).

There are still a couple changes that can be made to simplify it even more, and I am still hoping one of the readers here might spot these. 

.99