FIGUERA'S AETHER MAGNETIC FIELDS LINEAR PUMP, REVIVED

[lfarrand]

I'll elaborate a little bit more about my reasoning for suggesting abandoning the P channel MOSFETs and going with an all N channel MOSFET approach. P channel MOSFETs are not as common as N channel, and the choice of higher voltage P channel MOSFETs is severely limited.

There is a problem with using N channel MOSFETs on the high side though, which means you can't simply plug an N channel in place of a P channel without reworking some stuff.

• P-channel MOSFETs are typically used on the high side (between +V and the load).
  
  • They turn on when the gate is pulled low relative to the source.
• N-channel MOSFETs are usually used on the low side (between the load and ground).
  
  • They turn on when the gate is pulled high relative to the source.

Because of this, the control voltages needed are opposite, and the way current flows through the circuit changes.

If you simply drop an N-channel FET in place of a P-channel one:

• The source and drain voltages are reversed from what the N-channel expects.
• The gate drive circuit won't have enough voltage to turn it on properly.
• In high-side switching, an N-channel FET's gate must be driven above the source voltage - often requiring a charge pump or driver circuit that can go above the supply rail.

So, in a circuit designed for a P-channel high-side switch, replacing it directly with an N-channel usually means:

• It will never turn fully on, or
• You'll need to redesign the gate drive section.

You can use an N-channel MOSFET on the high side if:

• You have a dedicated high-side driver that can drive the gate above the source (like a bootstrap driver IC).
• The circuit is rearranged so that the N-channel is used as a low-side switch instead (between the load and ground).

This is often done in modern designs because N-channel MOSFETs:

• Have lower on-resistance (Rds(on))
• Are cheaper and more common

The good news is that my MOSFET driver board takes care of all the above, and allows you to use N channel MOSFETs on both high and low side. The board has a gate driver and an isolated DC/DC converter to make sure everything just works regardless of whereabouts you put it in the circuit.

[kampen]

@ Lee, 

This is a fantastic, solid, and very good suggestion. 

I agree that a back-to-back N-MOSFET "AC switch" will cleanly block in both directions when OFF and will conduct bi-directionally when ON, 
so the reverse/negative Voltage that was punching through body diodes in the P/N arrangement is then no longer a problem. 

The best way to drive an inductive load without a mechanical commutator.

If possible, send over the KiCad/Gerber or a block diagram. 
Thank You.

[lfarrand]

Glad you agree!

I've attached the KiCad files along with the Gerbers.

The isolated DC/DC converter that I've specified is manufactured by Murata and is about $15, so an expensive part. If you're using a MOSFET with low-ish gate charge then you can use a cheaper Recom or Murata isolated DC/DC converter. I've got another KiCad project for an adapter board (the power board attachment) to allow you to mount the other format converter and plug it into the MOSFET driver board.

The expensive Murata DC/DC converter is needed when driving MOSFETs with larger gate charge, such as those which allow larger amperage.

This was my lockdown project. I figured I'd need a good switching solution for the stuff I was working on, so invested a lot of time into it  I'm glad that it might be useful for you too!

[Ufopolitics]

Hello All,

First I want to WELCOME New Member @lfarrand to OUM!! (This is normally done at the WELCOME ALL NEW MEMBERS Topic)

So, if any Member wants to congratulate him, please go there and do it there...NOT HERE please!!, not to populate this Topic.

Then I want to thank L Farrand for his great contribution here about this new proposal on the Figuera Linear Generator Solid State Driver.

Yes, definitively looks like a great idea, and totally right that PFETs are limited as also looking for the right match to an NFET related to speed and response timings, plus having same Spec's.

Much more SOLID Switching System by only using same type of NFETs.

And those NFETs you are mentioning, those are really "Monsters" man!!...They can handle super powers!!

Regards

Ufopolitics 

[Ufopolitics]

Hello @lfarrand,

When we were designing this SS Driver, one main thing I requested to Member Kampen, is for the Sequencing Signals to be OVERLAPPED, (which He achieved on his second design) this is the way the Brush does between Two Commutator Elements whenever TRANSITIONING from one to the other.
In other words we also called them 'MBB' (Make Before Brake) NOT 'BBM' (Brake Before Make)
A more detailed explanation is at posts 343 & 344 on Page 69
Post 343: https://overunitymachines.com/index.php/topic,5.msg1910.html#msg1910
Post 344: https://overunitymachines.com/index.php/topic,5.msg1911.html#msg1911

Resuming that when we do MBB, the overlapped transition -between sequential coils- does NOT allow for Magnetic Field to collapse in any point of the sequencing.

These may also in a nano fraction of time, increase a bit the resistance, then lowering just a bit the currents, however, of regardless consideration comparing to allow Field to collapse.

It is VERY IMPORTANT we NEVER allow the MAIN FIELD to COLLAPSE!!

When Main Field Collapses, it will generate very high REVERSED VOLTAGES, BUT MAINLY, the OUTPUT will get momentarily interrupted...Then FIELD will have to RESTART from ZERO, DEMANDING MORE INPUT POWER.

We will obtain a 'FLICKERING' OUTPUT...as our INPUT will not be stable and constant, likes it needs to be.

Regards

Ufopolitics