Ferrite Core Question. / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

Ferrite Core Question.

Move Thread

LAN_403

Proud Mary

Mon Sept 28 2009, 10:05PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

If we were to pack the lower 10-15% of a TC with MF/LF ferrite rods, so that they came just up to the height of the primary, how would this affect operation besides increasing L and Q, and reducing fo ?

GeordieBoy

Mon Sept 28 2009, 11:14PM

Registered Member #1232 Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881

I'd imagine it would increase the coupling coefficient (k) of the system, at least for a small-signal analysis. In a spark-gap TC the ferrite would probably saturate quickly though, given the many kilovolts typically applied across a handful of primary turns.

Typical power ferrites are also not particularly good insulators when presented with tens of kilovolts at hundreds of kiloherts. Therefore tracking over the surface of this ferrite slug is likely.

Myself and Alan Sharp did a lot of experiments with ferrite cored transformers base-feeding TC resonators back when Solid-state coils were in their infancy. After much testing we both came to the same conclusion that it is no the way to go.

I guess if you want to make a good CW power transformer with tight coupling, then using a ferrite core is the way to go. But for high voltage, loose coupling with resonance to overcome the losses is the way to go. In that respect I think the classic Tesla Coil is hard to beat.

-Richie,

Proud Mary

Mon Sept 28 2009, 11:20PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

Thanks Richie.

Extreme Electronics

Tue Sept 29 2009, 07:36AM

Registered Member #74 Joined: Thu Feb 09 2006, 09:17AM
Location: Nottingham UK
Posts: 99

Richie,

I remember a coil that was basically a magnifier with the primary and secondary wound on an ETD core, base feeding the tertiary coil, was that one of Alan's and yours ? If so what were the issues? I've always wanted to try that configuration. On a small scale of course !

Derek

Steve Conner

Tue Sept 29 2009, 09:12AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

I remember this being discussed years ago on the TCML. Somebody said that putting ferrite rods into the base of the secondary was an old trick to increase coupling. Of course Richie's warnings about the conductivity of ferrite apply.

GeordieBoy

Tue Sept 29 2009, 10:56AM

Registered Member #1232 Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881

Hi Derek.

Myself and Alan Sharp both "wasted" a lot of time trying to base feed "magnifier coils" using ferrite transformers.

The original train of thought went like this:

Theory says base of resonator is a low impedance at resonance, so it will draw lots of current from a relatively low voltage inverter! Great! Except... Once the resonator breaks out the voltage at the top is more or less clamped, and this causes the current at the bottom to be limited too. In short, the "input impedance" when base-feeding the resonator rises from the moment you get breakout from the top. This means that in practice it resists drawing much current from an inverter running from sensible voltages in the hundreds that common semiconductors can switch.

There's at least four ways round this problem:

1. Use an inverter running from thousands of volts (like Gary Johnson did)
2. Use a low voltage inverter running from 400VDC or so, followed by a ferrite transformer.
3. Drive the resonator via non-resonant link-coupling from a primary coil. (This gives step-up from primary turns coupling to the bottom few secondary turns depending on coupling.)
4. Use a series resonant primary. This gives step-up of the drive through series resonant in the primary.

The problem with the ferrite transformer approach is the design of the actual transformer. Specifically it is not an ideal transformer. It is the leakage inductance and inter-layer capacitance of the secondary that cause the problem.

In order to get several kilovolts out of the transformer at 300kHz you need to put a lot of effort into insulation, creapage and clearance distances. This has a detrimental effect on coupling. (By coupling, I mean just the coupling inside this ferrite power transformer, not overall system coupling.) The decrease in coupling causes a relatively high leakage inductance to appear in series with either of the transformer's windings. Lets refer it to the HV secondary winding for now.

Then we have the secondary winding itself with all of its many turns. This at least has the inter-winding capacitance between the secondary and the primary if it is wound in just one layer. If it is wound in more than one layer, then there are also layer-to-layer capacitances. These depend on whether it is wound in zig-zag fashion between the margins or wound from left-to-right with an abrupt return at the end of each layer.

Anyway, the problem is that the leakage inductance resulting from all the insulation requirements series resonates with the inter-winding and inter-layer capacitance of the secondary. The frequency that this happens at will likely be above the operating frequency of the TC if the transformer is large and carefully designed, but it is not much higher! Add to this the capacitance of the HV feed-line from the transformer to the base of the resonator and this lowers the frequency considerably.

When we drive our TC resonators with inverters we normally ignore the rich spectrum of harmonics in the drive voltage. These higher frequency harmonics have a habbit of exciting the stray resonances in this transformer caused by leakage inductance and internal capacitances with two fatal consequences. Firstly, it causes points of very high potential at positions along the HV secondary winding where series resonance at one of the drive harmonics occurs. And secondly, it results in a large current being drawn from the inverter at this harmonic frequency.

Myself and Alan have had many attempts at ferrite power transformers burn up, or blow up the inverter from series resonance long before they were ever connected to a TC resonator. And for those who don't know, i'm not new to the design of SMPS transformers, so do know a few tricks.

If you model what you get with a non-ideal transformer base-feeding a TC resonator, you end up with a high-order system like those ones that Antonio Carlos De Quieroz worked out for magnifiers. All of a sudden you have a massive complex system with resonances all over the place, and energy sloshing back and forth between the drive transformer's "non ideal properties" and the TC resonator. The square wave voltage out from a typical inverter is just about the worst thing to drive it with, because the fast edges excite all of these strays into resonance.

It turns out that its much easier to get your drive-voltage step-up by winding a handfull of primary turns around the base of the resonator. You still benefit from turns-ratio, but you can trade off between voltage gain from transformer action and voltage gain from resonance:

Tight coupling means lots of secondary turns see the flux from the primary so you have a large turns ratio and get lots of voltage gain from conventional transformer action. But tight coupling leaves less of the TC secondary available for resonant action.

Conversely, low coupling, means that less of the secondary turns see flux from the primary so the voltage gain from transformer action is lower. But it leaves more of the secondary free un-coupled and free to resonate.

If you take a pretty normal CW-SSTC with coupling around 0.35 you might be suprised to see how much voltage rise is due to transformer action in the coupled region at the bottom of the resonator, and how much is actually due to resonant rise.

-Richie,

Extreme Electronics

Tue Sept 29 2009, 11:29AM

Registered Member #74 Joined: Thu Feb 09 2006, 09:17AM
Location: Nottingham UK
Posts: 99

Thanks for the comprehensive answer Richie..

I'd thought of some of those issues, but obviously not the complete picture.

Ok, I'll scrap the idea of playing with that one then :)...

Derek

GeordieBoy

Tue Sept 29 2009, 11:54AM

Registered Member #1232 Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881

If you asked Alan about ferrite power transformer drive I think he'd probably get cold shivers, bad memories, and then you'd get the same answer as from me. It's not worth bothering with.

It's the deadly combination of leakage inductance _AND_ stray capacitance across the secondary that causes the problem. One or the other can be accomodated for at the design stage...

For instance you can design an H-bridge to be tolerant of some degree of capacitive loading across it's output. Those fancy Class-DE inverters actually make use of additional capacitors across the MOSFETs to achieve soft switching during a generous deadtime. If done properly stray capacitance across the inverter output (which is usually bad) can be turned to your advantage.

Conversely, if the transformer only had leakage inductance, then that wouldn't be a problem either. If you think about it, any leakage inductance referred to the secondary winding of the ferrite drive transformer would simply add to the inductance of the resonator being driven. So in that sense leakage inductance alone is not a problem. You just need to remember that the portion of the HV winding that is not coupled, and is acting like an additional part of the resonator (exhibiting resonant voltage rise) IS ACTUALLY INSIDE THE FERRITE TRANSFORMER. Therefore provide additional insulation to allow for this voltage rise!

The combination of leakage indutance and secondary capacitance is the problem. It's almost like it forms a tiny L-match circuit with a high resonant frequency right between your inverter and the base of the TC.

-Richie,

Kolas

Tue Jan 06 2015, 12:55PM

Registered Member #102 Joined: Thu Feb 09 2006, 08:15PM
Location: Raleigh, NC
Posts: 169

I hate to kick a dead horse but I found placing large pieces of ferrite rod right at the base of the secondary winding to increase output. now from what I've been reading, if you could somehow convert more of just the base of the TC into a transformer there is no reason why more initial voltage rise can be achieved, right? So while I haven't been able to research all of what you guys found out... were ferrite, former filling disks of only a few inches high tried in the resonators them selves? it seems like it could eliminate or reduce a lot of what you said you had issues with.

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob