Jim Mac's Figuera Thread

[lfarrand]

This is a kind of out of the box suggestion, but wouldn't a solid state approach work better than a mechanical device that is difficult to engineer, prone to wear & tear and requires maintaining?

The idea that popped into my head was to use a multiple 'cells' in parallel, which can be individually controlled over time using a microcontroller. Each cell would be a single MOSFET and resistor in series. Resistors in parallel result in an overall decrease in resistance, so when all the cells are on then the resistance would be minimum and max current would flow. Then when each cell is switched off in turn the resistance would gradually start to increase until all of them were off and min current would flow. Current would still continue to flow, so the magnetic field wouldn't collapse, similar to how an analogue rheostat works.

Resolution of the current change & magnetic flux would depend on the number of cells used.

Expanding on this idea further, instead of a number of cells in parallel you could probably just use a single MOSFET in triode mode and use a variable gate voltage to control the resistance and hence current. You would probably need active cooling of the MOSFET when used this way because the resistance will cause power to be dissipated in the MOSFET. In theory this should give you infinite resolution that will only be limited by the voltage steps you apply to the gate.

https://www.nexperia.com/applications/in...inear_mode
https://electronics.stackexchange.com/qu...ing-losses
https://electronics.stackexchange.com/qu...-in-mosfet

[Jim Mac]

Thank you both for the suggestions. 

Unimmortal-  I think your suggestion is indeed a possibility that could work. It may just be how figuera built his.  It is just a very daunting build because we are talking about hundreds of contacts still. When The brushes start rotating at high RPM, ensuring we may contact to each and every connection point at all times is going to be very difficult. 

Ifarrand-  if I'm understanding your proposal correctly, it seems either method we still need hundreds of mosfets.  Again, I am sure this idea will work, but the intricacies of building such a circuit and complexity of wiring and logic control again is a very daunting undertaking. 

Picture this. Say we have 6 in of nichrome wire secure to a glass platform.  Each end of the nichrome wire connects to an individual inductor.   Then finally we have a single copper rod that carries direct current positive. The rod slides or reciprocates back and forth the nichrome wire always making contact. 

If all the contacts and switches could be avoided, The output wave could be 100% smooth, exactly mimicking reality.  And it is easier maintaining a solid connection of a carbon brush on a slip ring or a solid smooth surface than it is jump in contacts. 

I am not opposed to using a solid state method instead of mechanical, but even wiring up that 12 h bridge circuit turned into a bird's nest real quick.  And that represents only a fraction of the switches I would need.

[Lasco]

Hi guys
I’m still scratching my head .
Perhaps donut from les conductive materials
That will have resistance in the diameter as per Jim’s picture
I don’t know if this can by used in casting https://micronmetals.com/product/nichrome-alloy-powder/

Or something like this
https://nickelinstitute.org/media/8daa61...eating.pdf

[Jim Mac]

Hi guys
I’m still scratching my head .
Perhaps donut from les conductive materials
That will have resistance in the diameter as per Jim’s picture
I don’t know if this can by used in casting https://micronmetals.com/product/nichrome-alloy-powder/

Or something like this
https://nickelinstitute.org/media/8daa61...eating.pdf

Nichrome powder should work fine for the conducting agent. I think many kinds of conductive metal powders will be suitable.  But the hard part is the binder.  

Epoxy is an insulator.  The epoxy is covering the metal powder preventing it from conducting.  Now that my rings are dried, they measure about 0.5 MegaOhm.  

What about Concrete?  Concrete can be sanded smooth as glass.  It is plenty hard enough to support a carbon brush.  Perhaps adding the right amount of conductive powder to concrete would make it mildly conductive..

[Jim Mac]

Well I tested a little anchoring cement with Graphite water.  Let it set overnight. I have mild conductivity but way too resistive.  And I have no clue if it will conduct evenly with a brush on the surface.

I shorted 22V into the copper leads and I drew 0.019 amps of current.  So I have somewhere around 1100 Ω of resistance between conductors.

   

I measured voltage across the copper leads and it is producing just under a volt.  Not no readable current.

Concrete may be a possibility but the chances of designing it reliably and correct are slim at best..

I might have to attempt to use Nichrome Wire with a sliding contact somehow.  The rheostat and electronic methods also remain a possibility.

[lfarrand]

Ifarrand-  if I'm understanding your proposal correctly, it seems either method we still need hundreds of mosfets.  Again, I am sure this idea will work, but the intricacies of building such a circuit and complexity of wiring and logic control again is a very daunting undertaking.

I am not opposed to using a solid state method instead of mechanical, but even wiring up that 12 h bridge circuit turned into a bird's nest real quick.  And that represents only a fraction of the switches I would need.

You wouldn't need more MOSFETs than was necessary to work within your voltage and current requirements. You might even be able to get away with using just one if it is a pretty beefy one.

For example, take the Infineon IMYH200R012M1H. It is rated to 2000V (actually 2600V in my testing) and can handle 94A @ 100°C. If you varied the gate voltage between 4-20V then you can control the resistance of the MOSFET. In effect it becomes a voltage controlled resistor between some high value ohms down to the RDS(on) resistance (20mΩ).

Guests cannot see images in the messages. Please register at the forum by clicking here to see images.

Do you need to control each coil individually? I thought the Figuera patent showed all of the coils sharing a common source.

Guests cannot see images in the messages. Please register at the forum by clicking here to see images.

[Jim Mac]

Ifarrand-  if I'm understanding your proposal correctly, it seems either method we still need hundreds of mosfets.  Again, I am sure this idea will work, but the intricacies of building such a circuit and complexity of wiring and logic control again is a very daunting undertaking.

I am not opposed to using a solid state method instead of mechanical, but even wiring up that 12 h bridge circuit turned into a bird's nest real quick.  And that represents only a fraction of the switches I would need.

You wouldn't need more MOSFETs than was necessary to work within your voltage and current requirements. You might even be able to get away with using just one if it is a pretty beefy one.

For example, take the Infineon IMYH200R012M1H. It is rated to 2000V (actually 2600V in my testing) and can handle 94A @ 100°C. If you varied the gate voltage between 4-20V then you can control the resistance of the MOSFET. In effect it becomes a voltage controlled resistor between some high value ohms down to the RDS(on) resistance (20mΩ).

Guests cannot see images in the messages. Please register at the forum by clicking here to see images.

Do you need to control each coil individually? I thought the Figuera patent showed all of the coils sharing a common source.

Guests cannot see images in the messages. Please register at the forum by clicking here to see images.

Okay I get it. So my next question is how to control the gate voltage?  

Would I need something like a digital potentiometer which is controlled by logic circuit?

[lfarrand]

These links might provide more insight about how to control a MOSFET in linear mode.

https://provideyourown.com/2011/analogwr...o-voltage/
https://www.powerelectronicsnews.com/pow...ar-regime/
https://apps.dtic.mil/sti/tr/pdf/ADA258384.pdf

I think what you will need to do is charge a capacitor using PWM to regulate the voltage.

The MOSFET gate is a capacitor, so maybe you can just use that. You will need to bias the MOSFET gate voltage to ensure it doesn’t go below, say, 4V so that the current never goes to zero and hence the magnetic field doesn’t collapse.

---

I came across this also which describes using a DAC to control the gate voltage of a MOSFET.

Replacing Digital Potentiometers with Precision DACs
https://www.ti.com/lit/ab/slaa906/slaa906.pdf

---

https://www.analog.com/media/en/news-mar...ngDACs.pdf

https://www.analog.com/en/product-catego...ory-detail

---

https://www.analog.com/media/en/referenc...dn538f.pdf

[Sandy]

Greetings. My first post here at mooker.
Sorry if I am missing the point but would an audio amplifier achieve what you are going for with the linear mode mosfet?
Kind regards, Sandy

[Jim Mac]

Greetings. My first post here at mooker.
Sorry if I am missing the point but would an audio amplifier achieve what you are going for with the linear mode mosfet?
Kind regards, Sandy

Hi Sandy and Welcome.

What I am looking to do is create a totally smooth sinewave (1 polarity or 2 or both)  using varying resistance.  I am not looking for a sinewave created with reactance from coils.  

But your idea is bordering on the next question I was going to ask Ifarrand..

Ifarrand-  What about firing such mosfet gate with a signal generator?  We create an AC wave and bias it with DC,  or just use the function generator bias settings.  

This way the function generator can easily control the frequency of the whole circuit easily.