Hey Eldarion,

I also spoke to Earl about that circuit and I will be building and trying it out myself. If the transition time becomes a problem, you can just use the driver chips themselves to drive the coil. The mic4427 IC's are guaranteed to charge 1000pF in 20ns. That should be more than enough to charge the tiny bit of capacitance in our TPU coils if the Flip flop isn't beefy enough. But in either case, we still don't need the fets as long as the pulse is fast and short. I'm not sure how small the pulse width would be on the driver chips without the fet on them. When using the IRF840s, I was able to get the pulse width down to about 380ns so you should be able to do much much better just using the driver chip alone.

God Bless,
Jason O

Hi Bob,

I have modified the low-pass filter, adding a choke and a resistance (under the assumption that the HV bias current is very small).  The resistor would also make the choke a low-Q inductor and eliminate resonances in the choke.

Your comments?

Regards, Earl

Quote from: Bob Boyce on August 10, 2007, 11:50:34 PM

Quote from: Earl on August 10, 2007, 04:29:04 PM
Hi Eldarion,

In my opinion, the low-pass filter that I have suggested to keep RF out of the bias supply must have C3 at the TPU in order to keep the output power loop physcially as small as possible.  See my attached explanation.

Regards, Earl

@ Earl

You seem to be thinking transversely. Assuming that the primary load energy flow would be from induced phase differential from one end of the secondary to the other, which would be coupled to the load through both caps - as you show. Granted, there is energy in that mode, but that would only be a portion of the total energy available. Please try to entertain the notion that modulated longitudinal energy is induced into the secondary winding from outside, and into the secondary/earth dipole. Modulated longitudinal energy flows through C1, and through the load to ground. A capacitor at C3 would shunt a portion of that energy to ground, bypassing the load. Just ponder this alternative viewpoint.

Bob

Quote from: Jdo300 on August 11, 2007, 02:33:50 AM
Hey Eldarion,

I also spoke to Earl about that circuit and I will be building and trying it out myself. If the transition time becomes a problem, you can just use the driver chips themselves to drive the coil. The mic4427 IC's are guaranteed to charge 1000pF in 20ns. That should be more than enough to charge the tiny bit of capacitance in our TPU coils if the Flip flop isn't beefy enough. But in either case, we still don't need the fets as long as the pulse is fast and short. I'm not sure how small the pulse width would be on the driver chips without the fet on them. When using the IRF840s, I was able to get the pulse width down to about 380ns so you should be able to do much much better just using the driver chip alone.

God Bless,
Jason O

@ all
I have free samples coming from Texas Instruments. The UCC27321 / UCC27322 and the UCC37321 / UCC37322. I will test these to see just how clean and fast they really are.

Here is a link where you can order samples or download the datasheet.

http://focus.ti.com/docs/prod/folders/print/ucc27321.html
http://focus.ti.com/docs/prod/folders/print/ucc27322.html
http://focus.ti.com/docs/prod/folders/print/ucc37321.html
http://focus.ti.com/docs/prod/folders/print/ucc37322.html

Bob

Maybe this will help.

Bob

Hi Bob,

thanks for your reply.  What you are saying is that not only can three primaries serve in a catalytic action each primary with a wire hanging in the air, connected to nothing, but that the secondary can output appreciable power with one lead hanging in the air, connected to nothing.

Did I understand this correctly?

Regards, Earl