[TCML] quench times again
Barton B. Anderson
bartb at classictesla.com
Sun Nov 25 15:07:52 MST 2007
Thanks for the link Chris. I'll read with interest. Yes, wip and the
ideas that culminate from the experience and the conclusions drawn are
always a good read. These things are what spawns new ideas and innovation.
Take care,
Bart
Chris Swinson wrote:
> Hi Bart,
>
> I just posted to another thread about SISG, no doubt you will read
> that one...
>
> I forgot to post the URL
> http://www.future-technologies.co.uk/IMPULSE/coil/coil.htm Half of it
> maybe total nonsense as this was some time ago now. Though you can see
> how things started out.
>
> the SISG fired on the zero crossing of the AC wave to increase
> efficiency, though I did not like SISG series fire operation, so I
> went for a parallel operation.. latter I started to add in resonant
> mains side charging.... and things just kept on going... the pages
> after that to be honest is "WIP" and "ideas" it should not be
> published really, though ideas are always good I suppose....
>
> I did plan to build my SISG design though I stopped work on it after a
> while, I moved onto much lower voltages using a single IGBT. much
> higher current and less voltages... and yes my High Q system does
> solve a lot of problems, but as you say opens up a whole new set of
> problems!
>
> All the talk on the list at the moment is my second idea to build a
> 20KV high Q system.... factors such as resistance are not much an
> issue above 1KV... though 10VDC at 1,000amps, even a few fractions of
> a ohm can seriously reduce efficiency..
>
> Also a small toroid should be used, or none at all. it increases the
> capacitance and lowers the Q of the coil. With such a high frequency
> anyway, a larger toroid is bad yet again.. Though at this point other
> than trying it one could speculate for years to come, and already
> spent the past year in speculation!
>
> I think some of the confusion with everyone is I have 2 systems in
> mind, what may work with a classic coil may not work with a high Q
> system... the spark gap seems the limiting factors in just about every
> aspect, so if you remove that problem and use solid state then you are
> onto a good thing and have much more freedom to experiment with "new"
> designs... however solid state is low voltage which opens up its own
> problems.. "win some, loose some"
>
> IMHO, Tesla coil's need a new direction to go in, as solid state seems
> to be the direction then building a "new" type of solid state coil
> should be interesting... If design can be made easier, using off the
> shelf parts, with much smaller toroids, lower voltages, 100turn
> secondary coils.. then the advantages speak for itself.
>
> No doubt if my design works, I will run into new problems in running..
> my secondary maybe under so much stress it could turn into a bright
> flash off light and vanish into another dimension for all I know!
> Simulations can only tell you so much!
>
> Waffle over...
> Chris
>
>
>
>
>
>
> ----- Original Message ----- From: "Barton B. Anderson"
> <bartb at classictesla.com>
> To: "Tesla Coil Mailing List" <tesla at pupman.com>
> Sent: Sunday, November 25, 2007 8:03 PM
> Subject: Re: [TCML] quench times again
>
>
>> Hi Chris,
>>
>> I agree that this can be done to a point and about the trade-off of
>> losses in this system. The fact is no one has done this that I'm
>> aware of and it "is" new territory. At first, I wasn't sure. Too
>> often a coiler thinks he is doing something new only to find out it
>> wasn't new. In this case, I think it is new. The only one I know that
>> has done something along these lines is Dave Sharpe with is IDR coil,
>> but that really is in the primary area. Your secondary is completely
>> different and sets up a whole new set of issues as well as
>> eliminating quite a few.
>>
>> You have the ability to use a high frequency and keep the Q very
>> high. It will be interesting to see if you can apply this method to a
>> high voltage coil (plan B). I think you've thought this through very
>> well and it's been an eye opener for me. We don't really know what
>> type of problems will be encountered.
>>
>> The ring up and how the energy is stored in the secondary and top
>> terminal will be interesting. Now I understand why you were looking
>> at dwell reduction. Unfortunately, that just doesn't help and I'm not
>> all that confident that a rotary is the best avenue for the coil.
>> Have you considered maybe an SISG approach?
>>
>> Take care,
>> Bart
>>
>> Chris Swinson wrote:
>>> Hi Bart,
>>>
>>> I think some confusion is that when I mean higher frequency, I do
>>> not mean smaller coils.. my Q factors always go up with the larger
>>> wire with larger secondary coils.. Though you can only do this to a
>>> point as if you make the coil length longer to keep the same turns
>>> but thicker wire, then the HD ratio is a poor figure and chances are
>>> it will not couple correctly to the primary..
>>>
>>> So you are right in what you state, but it is not exactly what I am
>>> doing...
>>>
>>> Higher frequency does increase losses in some parts, but in other
>>> parts losses can go down by a factor of 4... It is one huge epic
>>> counting up all these "new" pro's and con's, though overall the
>>> higher Q system comes out best..
>>>
>>> The other problem is higher Q( generally) is higher frequency, which
>>> needs a lower tank cap value, which reduces energy. So you cannot
>>> use large tank cap values.. I got around about 250khz before running
>>> into such problems.. So you have to rebuild and use a higher voltage
>>> to keep the same power input with a small tank cap. Joules input is
>>> a lot better at higher voltages and puts less current across the
>>> spark gap.
>>>
>>> IMHO, once you get to 100nF tank cap, the voltage should go up for a
>>> greater input power, not the tank cap value... In general 100-200nF
>>> is about the limits at say 10KV, but I think 100nF should be the
>>> largest tank cap which should be used. Rather than pumping more
>>> current, it should pump more voltage instead... once you start to
>>> pump more voltage input, you can use a lower value tank cap, so less
>>> current, and it opens up the doors to much higher Q coils.
>>>
>>> Chris
>>>
>>>
>>>
>>>
>>> ----- Original Message ----- From: "Barton B. Anderson"
>>> <bartb at classictesla.com>
>>> To: "Tesla Coil Mailing List" <tesla at pupman.com>
>>> Sent: Sunday, November 25, 2007 3:53 AM
>>> Subject: Re: [TCML] quench times again
>>>
>>>
>>>> Hi Chris,
>>>>
>>>> Only a little confusing, but not too bad. Anytime I start talking
>>>> decrease this, increase that, I personally have periodic
>>>> cross-thought errors (type just the opposite than I meant due to
>>>> wondering off on a different aspect).
>>>>
>>>> There are of course losses associated with higher frequency.
>>>> Usually, when coilers are talking a high frequency coil, it's
>>>> geometric size is small and Q is not high. Here again now I've got
>>>> to state, and yes log me down for this statement: "higher frequency
>>>> does not result in a higher Q coil".
>>>>
>>>> Increase frequency by taking any coil and reduce it in 1/2. Thus,
>>>> divide the radius by 2, the height by 2, the wire size by 2, and
>>>> keep the same number of turns. Your frequency will double and Q
>>>> will lower because the AC losses begin to increase. If it were not
>>>> for those losses, I would expect the Q to remain the same.
>>>>
>>>> In your case, there is a high Q due to the higher conductance. Eddy
>>>> and skin effects will not be hindered in your coil as it would in
>>>> one of the smaller high frequency coils. This should definitely not
>>>> be related to Q, but rather to the large wire size and it's low DC
>>>> resistance and unaffected AC resistances.
>>>>
>>>> It should be true that as we reduce the number of transfers, the
>>>> gap losses should decrease. I'm not sure that higher frequency
>>>> would help ionization at the gap except that it will help to
>>>> decrease the transfer rate (so more energy over a shorter period).
>>>>
>>>>>
>>>>> The idea really is that a higher frequency should allow a higher
>>>>> current pulse with upsetting the RSG too much. It was also my
>>>>> point about "making sure" by decreasing the RSG dwell time. As
>>>>> higher current will be harder to quench, then decrease the dwell
>>>>> time and it should help matters also.
>>>>>
>>>>> A lot of factors come into play, as pointed out by yourself, John,
>>>>> etc. Though this was really the overview of the "high Q" system
>>>>> which I had in mind. A lot of ideas and corrections brought up in
>>>>> all these posts thats for sure!
>>>>>
>>>>> Everyone has been down the classic road, wider coils, more
>>>>> inductance, larger toroids... So I am thinking of a "new"
>>>>> direction instead....
>>>> But when you see certain aspects like Q increasing, look at what is
>>>> different. In your case, it's really the few turns of large wire
>>>> over a large area. This is a huge difference. Just take your coil
>>>> and reduce the wire size by half and you'll see Q start to drop
>>>> without much affect on frequency.
>>>>>
>>>>> Another point which has not come into it yet, even though I
>>>>> mentioned it. Higher frequency should also increase efficiency in
>>>>> it its own right
>>>>>
>>>>> for example, running from a 12V test setup, at 15cm "range" ....
>>>>>
>>>>> 50hz =0
>>>>> 39khz =0.5mV
>>>>> 124khz=5mV
>>>>> 1mhz =50mV
>>>>> 1.2mhz =70mV
>>>>> 1.43mhz =120mV
>>>>> 1.87mhz =150mV
>>>>> 2mhz =200mV
>>>>>
>>>> But is that a result of the frequency or is that a result of the
>>>> coil geometry? Higher frequency is resulting in a shorter transfer
>>>> rate and as a result di/dt increases at the secondary which
>>>> increases the amplitude since the AC and DC losses are so low. But
>>>> the same cannot be said for a high frequency coil which is small.
>>>> The losses are huge then. For your particular geometry, I think
>>>> what you said is true, but not across the range of coils.
>>>>> I was wondering if this would also apply to coupling efficiency.
>>>>> In a way it looks like voltage is lost over the coupling. Tighter
>>>>> coupling would in effect reduce my "range" figure and double up on
>>>>> the voltage.
>>>> I don't normally look at coupling as an efficiency number. Coupling
>>>> will always be 100% regardless. There are of course losses over
>>>> time at the gap and over the transfer. But yes, tighter coupling
>>>> will increase di/dt.
>>>>>
>>>>> After a lot of testing I drew up that double the frequency gave x4
>>>>> the voltage output. As a relation, 10 times the frequency gave
>>>>> double the "range".
>>>> For your particular geometry.
>>>>>
>>>>> Going by these figures, if a normal tesla coil used 1,000 turns at
>>>>> 100khz, then it suggests a magnetic field which runs "out of
>>>>> steam" at 1,000 turns. So increasing turns does nothing at all
>>>>> other than to gain a few volts and increase resistance.... the
>>>>> point now that if we progressed to 1mhz then we should be able to
>>>>> use 2,000 turns and the magnetic field will run "out of steam" at
>>>>> the 2,000 turns mark.
>>>> In a normal coil, the losses in eddy and skin effects will come
>>>> into play and will be significant. But, if we go down the road of
>>>> increasing the wire size and coil size in order to achieve 10x the
>>>> frequency and double up on the turns, then yes, we can get reduce
>>>> those losses. However, in reality the coil would be physically to
>>>> big to build.
>>>>>
>>>>> Also as frequency goes up you get more voltage. take 124khz 0.5mV
>>>>> to 1mhz 50mV . This is all at 15cm "Range". When I say range, I
>>>>> mean the distance between the primary and secondary. Remember
>>>>> only the frequency changed and the voltage was constant at 12V.
>>>> You can only get more voltage if the di/dt is increased without
>>>> significant losses. For your particular coil which is really
>>>> extreme I can see that happening.
>>>>>
>>>>> It is one of those odd things which also confuses me about tank
>>>>> energy going from primary to secondary. My own tests show there is
>>>>> a voltage drop... if we take 124khz I input 12V and got 0.5mV output.
>>>> Sure, there's always a voltage drop for any given point in time. No
>>>> doubt about that.
>>>>>
>>>>> Another problem is that Q factor was not taken into account with
>>>>> the secondary. I used a variable capacitor to tune the secondary
>>>>> to the primary. So Q factor probably was going up.. Though in
>>>>> anycase frequency increase gave way to higher Q factor coils and
>>>>> gave greater efficiency.
>>>> The cap in the secondary is a terrific approach on your coil. I
>>>> agree, but due to the few turns, large wire size, and coil size to
>>>> accommodate the wire size I believe is why. Your coil is so far
>>>> outside the loss box that the main loss in your system will be the
>>>> gap. In a high voltage situation, it would be interesting to see
>>>> how the voltage stresses react.
>>>>
>>>> Take care,
>>>> Bart
>>>>>
>>>>> Even though I still have more tests to do. I got 16mhz as being
>>>>> the best solution. I made me first think that the secondary coil
>>>>> over the loose coupling would only obtain a fraction of the
>>>>> voltage. In which case energy would be lost over the distance
>>>>> between the primary and secondary coils.... always interesting
>>>>> none the less!
>>>>>
>>>>> Chris
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> ----- Original Message ----- From: "Barton B. Anderson"
>>>>> <bartb at classictesla.com>
>>>>> To: "Tesla Coil Mailing List" <tesla at pupman.com>
>>>>> Sent: Saturday, November 24, 2007 8:01 PM
>>>>> Subject: Re: [TCML] quench times again
>>>>>
>>>>>
>>>>>> Hi Chris,
>>>>>>
>>>>>> Another correction I need to make.
>>>>>>
>>>>>> As the number of cycles increases, the transfer rate will
>>>>>> "decrease".
>>>>>>
>>>>>> What you are doing is interesting and how you are going about
>>>>>> looking at how the frequency affects the transfer rate,
>>>>>> efficiency, and gap conduction. Very interesting subject to me.
>>>>>>
>>>>>> Take care,
>>>>>> Bart
>>>>>>
>>>>>>> As the number of cycles increases, the transfer rate will
>>>>>>> increase. Here is the relationship.
>>>>>>>
>>>>>>> Total Energy Transfer = (0.5/((1/(1-k)^.5)-(1/(1+k)^.5)))*(1/fr)
>>>>>> _______________________________________________
>>>>>> Tesla mailing list
>>>>>> Tesla at www.pupman.com
>>>>>> http://www.pupman.com/mailman/listinfo/tesla
>>>>>>
>>>>>>
>>>>>> --
>>>>>> No virus found in this incoming message.
>>>>>> Checked by AVG Free Edition. Version: 7.5.503 / Virus Database:
>>>>>> 269.16.4/1146 - Release Date: 22/11/2007 18:55
>>>>>>
>>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> Tesla mailing list
>>>>> Tesla at www.pupman.com
>>>>> http://www.pupman.com/mailman/listinfo/tesla
>>>>>
>>>>>
>>>>>
>>>> _______________________________________________
>>>> Tesla mailing list
>>>> Tesla at www.pupman.com
>>>> http://www.pupman.com/mailman/listinfo/tesla
>>>>
>>>>
>>>> --
>>>> No virus found in this incoming message.
>>>> Checked by AVG Free Edition. Version: 7.5.503 / Virus Database:
>>>> 269.16.6/1150 - Release Date: 24/11/2007 17:58
>>>>
>>>
>>> _______________________________________________
>>> Tesla mailing list
>>> Tesla at www.pupman.com
>>> http://www.pupman.com/mailman/listinfo/tesla
>>>
>>>
>>>
>> _______________________________________________
>> Tesla mailing list
>> Tesla at www.pupman.com
>> http://www.pupman.com/mailman/listinfo/tesla
>>
>>
>> --
>> No virus found in this incoming message.
>> Checked by AVG Free Edition. Version: 7.5.503 / Virus Database:
>> 269.16.6/1150 - Release Date: 24/11/2007 17:58
>>
>
> _______________________________________________
> Tesla mailing list
> Tesla at www.pupman.com
> http://www.pupman.com/mailman/listinfo/tesla
>
>
>
More information about the Tesla
mailing list