- I was looking here at the OLTC http://scopeboy.com/tesla/t3concept.html

(theres a few but this one is simple to explain my problem)

However, while building it is no problem I have a couple of problems with

the workings out there..

On that small OLTC he uses 1.88uF, other higher power designs he uses up to

30uF. It seems to run from anything from 100 to 1000 BPS so that would

assume the IGBT is running at the BPS figure (not in the khz range).

However, 1.88uF and 3uH primary is 67khz, which could match a secondary of

also 67khz, however, the problem I have is that what about the mutual

inductance from the secondary ? it would actually make the primary's

effective inductance more like 150uH.. and 1.88uF with 150uH = 9.5khz. Even

more so by the time he got to 30uH it would be more like 2khz.

his secondary coil calcs do not seem to tally good either. about 120khz is

10" dia and 30" long using 0.4mm wire. JavaTC states 130uH effective primary

inductance, so even with 1uF tank cap its only 12khz.

As he talks about notches it seems it *is* in tune, but I cant see how it

can be ?! I did wonder if it was just a single discharge pulse, but it is

not clear, in particular when he states 100uF discharge time and talks about

3rd notch quench, which suggests it is in tune, but it can't be.... ?

He has 200mH and 1.4uF which would give 300hz resonance for the DC charging,

so it should trigger the IGBT at 300BPS. no problems there.

his later spec http://scopeboy.com/tesla/t4spec.html up to 30uF.. but

again, just can't be used for resonance with the secondary, so a big "huh"

is on my mind ?

Chris - Hi Chris,

Chris Swinson wrote:>

You're right - it's not trivial. There's a balance between too low a k

>

> Hi Bert,

>

> I did read Richie's site a few times, be it a while back, during my more

> classic TC days! I will keep an eye out for the book, I actually like the

> older ones as they always seemed to explain it in ways which made sense.

>

> It would seem better to use a lower coupling to gain a higher secondary Q,

> but may not exactly be that easy.

(and getting excessive primary energy loss from primary resistance and

switching losses) and too high a "k" (exceeding the rate of energy

growth into the secondary, causing racing sparks or failure to quench in

a SGTC). Best performance usually occurs for "k" in a range of 0.12 -

0.18 for spark gap coils. This also seems to be a good range for high

power SSTC's, DRSSTC's and OLTC's. For a variety of reasons, VTTC's

perform best with a "k" of about 0.15 - 0.30 since energy growth is

already limited by the peak power available from the vacuum tube oscillator.

>

Operation is different for pulsed coils versus CW coils. Pulsed coils

> Though one other thing I still don't fully understand is the energy

> transfer

> between primary and secondary. For example, if you place 2 coils side by

> side and run one from 12volts and use a low coupling, the other coil will

> only see about 1volt if you are lucky. If it was a CW then that 1V would be

> in resonance with the coil and build up voltage. Though I do not see the

> entire primary energy actually being transferred to the secondary. Though

> with that idea, it would be better to use closer coupling. Can't really

> have

> it both ways!

depend on the "transient response" of the coupled tuned circuits, while

CW coils depend on both the transient and the "steady state" response of

the system. In a SGTC or OLTC, we have a fixed amount of initial energy

(the bang size) stored in the tank capacitor. This is ALL we have to

work with each time the primary switch closes (spark gap or switching

transistor is turned ON). Once we begin transferring energy, we also

begin losing a fraction of this energy to resistive losses and to the

switch. If we have a low "k", the energy transfer will take more time,

and we'll lose more energy before the transfer is complete.

Ideally, we'd like to transfer the energy very quickly (i.e., high "k")

but if the secondary can't absorb this energy quickly enough, it will

cause racing sparks on the secondary. This tends to limit maximum "k".

As you lower "k", more energy is lost before the transfer is complete,

and maximum amount of energy making it to the secondary declines. The

best "k" for a system is usually found experimentally.

>

There is a direct relationship between "k" and the number of cycles it

> Though the way its explained is if you have X amount of energy and 0.10K

> then you could assume maybe it will take 10cycles to transfer the energy

> from primary to secondary. But how can the energy not simply be lost over

> the distance between primary and secondary ? In general I just accept what

> is told is true, but I must confess I still don't follow it!

takes to fully transfer energy from P--->S or S---->P. For k=0.1, this

is about 10 half-cycles or 5 cycles. Let's assume you removed primary

and secondary resistive and switching losses. In this case, 100% of the

energy would transfer from P--->S over a number of cycles, and then

would transfer from S---->P over over the same number of cycles. Energy

would continue to transfer between the primary and secondary

indefinitely... as long as "k" is greater than zero.

Since real world Tesla coil primary and secondary circuits have losses,

the system energy continually declines as it transfers back and forth,

eventually declining to zero. By balancing losses, quenching, and "k", a

well designed SGTC can transfer over 85% of the initial bang energy to

the secondary over even though the "k" may only be 0.12 - 0.18. For k =

0.18, it will take about 3 RF cycles to completely transfer energy from

P to S or vice-versa.

>

You're looking at it incorrectly. There are no losses directly

> The way I look at it, if you had 10Joules and 0.10K and it took 10 cycles,

> then to me that would mean each cycle would be 1Joule (putting it simply),

> but would actually loose 90% of that energy due to low coupling. So on that

> note, the secondary would see 0.1J for 10 cycles.

attributable to "k". However, having lower "k" DOES require a longer

time (more RF cycles) to transfer energy between P and S (or vice

versa). As k decreases, more system energy will be lost through

resistive losses leaving less energy left at the end of the transfer.

You may wish to revisit Richie's site again with the above explanations

in mind...>

Hope the above helps..

> Probably not that simple since the capacitor discharges its energy will get

> less per cycle, or at least I assume so. Though I hope I explained the

> "problem" enough for you to follow.

Bert>

> Chris

>

> ----- Original Message -----

> From: "Bert Hickman" <bert.hickman@...

> <mailto:bert.hickman%40aquila.net>>

> To: <usa-tesla@yahoogroups.com <mailto:usa-tesla%40yahoogroups.com>>

> Sent: Sunday, January 03, 2010 9:47 PM

> Subject: Re: [usa-tesla] OLTC question

>

> > Hi Chris,

> >

> > Once the primary and secondary tuned circuits become coupled, all sorts

> > of complex interactions begin to occur. Your friend was indeed correct -

> > the tighter the coupling the greater the interaction between the primary

> > and secondary circuits, and the greater the effect the low Q primary has

> > on the high Q secondary. These interactions also impact the bandwidth of

> > the combined system, the resulting primary and secondary Q's, the rate

> > of energy transfer between the primary and secondary, and the formation

> > of two frequency peaks (one above and one below the uncoupled resonant

> > frequency) during high coupling. Some of these behaviors are discussed

> > in Richie Burnett's excellent site. For example in the section about

> > quenching, you can see the effect that various coupling coefficients

> > have on frequency splitting:

> >

> > http://www.richieburnett.co.uk/operatn2.html#quenching

> <http://www.richieburnett.co.uk/operatn2.html#quenching>

> >

> > Another very in-depth review of coupled tuned circuits is contained in

> > chapter 3 of Frederick Terman's 1019 page book, "Radio Engineer's

> > Handbook", 1943, McGraw-Hill. Although long out of print, used copies

> > can easily be found for under $10. It should be on the shelf of every

> > serious Tesla Coil researcher.

> >

> > See: http://used.addall.com/ <http://used.addall.com/> or

> http://www.biblio.com/usedbooksearch.bib

> <http://www.biblio.com/usedbooksearch.bib>

> > to locate a copy of your own.

> >

> > Bert

> > --

>

>