Understanding electricity in the TPU.

[gyulasun]

...

Is it possible to use 5 to 10 of these JFETs in parallel to increase the handling capability because I would like to still use the IRF840 and the IRF9540 on the poles and use some of these in parallel in the center?
...

Hi Wattsup,

Yes it is possible. Unfortunately the jfet types you refer to are small signal types, this means a few mA (between 2-20mA) max drain-source current and 20-30V max drain-source voltage. By connecting several ones of the same type in parallel, (drains to drains, sources to sources, gates to gates) you can increase drain source current handling but not voltage of course.
These types have a rather high drain-source resistance, they are not designed directly for switching purposes. Of course you can use them as a switch in a circuit that considers this high channel resistance but in a resonant tank circuit this is not easy if possible at all. 
You can check with an Ohm meter the drain-source resistances of your jfets: choose any one of them and connect its gate to its own source directly and measure between this common connection and its drain.
What you will see on the meter it is the ON channel resistance at zero gate-source voltage and at about 2.5-3V drain-source voltage what your Ohm meter possess between its probes in that measuring range.

The Data sheet for instance for the PN4416A type says I_DSS ranges from 5 to 15mA measured at 15V drain-source voltage, this current defines the ON state channel resistance: if it is (suppose) 10mA for a particular device, then r_DS=15V/10mA=1.5kOhm!  not a friendly value in a high Q tank circuit...   and if you connect 20 of them in parallel, you still will have a switch with 1500/20=75 Ohm ON resistance (if all the 20 jfets had the same 1500 Ohm channel resistance which is the case only if you select them!).

For a hefty and good jfet switch you may have to resort to the type I gave earlier, probably expensive, but it has 1000V V_DSS and several Amper I_DSS and a channel resistance of  about  .07-.08 Ohm.  It is a normally closed switch (with zero gate-source voltage.

There is a solution to get a normally open switch from two power MOSFETs connected in series, see Figure 6 in Page 5 of this application note: http://www.irf.com/technical-info/appnotes/an-1017.pdf     
Because the PVI device (PhotoVoltaic Isolator) driving the common gates is rather slow in switching speed (only max. some hundred Hertz) you can use a small audio transformer or a pulse transformer to control the two MOSFETs by faster pulses. (The transformer's coil DC resistance would help discharge the gate-source nF capacitance, together with a few kOhm parallel resistor if needed.)
If you settle for  .1-1A current and 50-150V voltage, then you can get higher speed (some hundred usec) from solid state relay devices that include the series MOSFET pair and the PVI device in a single case, see here:
http://www.dionics-usa.com/product_index_2.htm  and there is a comparison chart here: http://www.dionics-usa.com/PDFs/relay_comparison.pdf  They have normally closed switch types too.

But in case you build your own series MOSFET switch with pulse transformer to drive, you can get the highest speed. And you can choose the switch parameters like ON channel resistance, max drain-source voltage and current, the speed is limited only by the pulse transformer.

rgds,  Gyula

[Mannix]

Or..use bottles as per the inventors advise

[giantkiller]

Device design topology #1
Tubes: Large polarized electron cloud.
Ferrous cores: Large polarized electron/flux reservoir.
Large copper loops: Large polarized magnetic/flux reservoir.
Eclipsing magnetic fields: Polarized magnetic/flux field manipulation.

Device design topology #2
Bipolar solid state: Minute nonpolarized electron flow.
Field effect solid state: Minute magnetic junction.

As one can see the solid state is not the environment to achieve the correct manipulation and has no volume for feedback. The design can not hope to achieve what the feedback is used for when dealing with such a small area. The configuration that works is a circuit that includes the topology for feedback of volume.

The device has characteristics of fast switching, emissions, and feedback, amplification. The amplification is inherent via the topology of the configuration. The feedback is achieved impressing upon the volume of usable electrons or flux. The fast switching is achieved by Bloch wall manipulation and no other way(The spark gap achieves this quite easily). The electron position or angular placement in space is the key factor and not at some junction or flow point where there is not enough volume for control.
Attempts at component control will only lead to extensive component research, purchasing and testing. The results happen in space and not at the component level.

This is what Spherics and Erfinder were driving at. I refer to the Atomic bomb tower setup by Spherics and the Telsa Ionizer patent and circuit description by Erfinder.

Or..use bottles as per the inventors advise

[giantkiller]

The Bloch wall would be moving around circumferrally at the same speed as the snake biting its tail. I believe it is moving faster than our current bench equipment can measure.
A charged iron ring is active as long as it is in a ring. Break the ring with an LED across it and you get a visual discharge. We know this. I was promoting all topologies as a mix. I was making the feedback the primary concern and stating where it lies with each of the topos. I am setting up a build right now with a gate and drain lines going through a ferrite ring. Drain line will then go to a larger horizontal loop.
The idea is to use solid state and to create a layer for feedback because SS doesn't have the cloud or flux field to communicate to.
I noticed that SMs tpus come in 2 types open and closed. The open ones don't show a horizontal run with a vertical run wrapped around it or have I fallen off the turnip truck? The closed ones have layers(maybe) but definately have an outer winding.
Wattsup has pictures of multiple tpus with the center toroid thing. I am going to put my drain and gate line magamp in that position and then in the center of a horizontal loop. This gives 2 layers of feedback of which I can alter the connection directions to play with.

Builds we all have done.
Builds we would like to try.
Builds that produce some effect.
Specifications, common and uncommon.

This is alot of information to grasp and utilize over the length of time for this involvement.

[giantkiller]

You have an iron horseshoe and a steel keeper on its end. A wire wrapped around a place in the iron. You tap a nine volt battery to the coil and the keeper stays forever. Have you ever seen anything read the activity that is keeping the keeper bound? If you have then there is OverUnity to a 'T'.

Back to current setup: